Influence of Nonbonded Interactions on Molecular Geometry and Energy: Calculations for Hydrocarbons Based on Urey—Bradley Field

Abstract

A modified Urey—Bradley potential energy function comprised of quadratic terms for bond stretches, bond-angle bends, and torsional displacements together with analytical expressions for pairwise nonbonded interactions was chosen to represent the force field for hydrocarbon molecules. Quadratic constants were taken from the spectroscopic U–B analyses of Schachtschneider and Snyder [Spectrochim. Acta 19, 117 (1963)], while the nonbonded functions adopted were those proposed by Bartell [J. Chem. Phys. 32, 827 (1960)]. Reference bond angles for the quadratic terms were taken to be 109.5° or 120° for tetrahedral or trigonal coordination, respectively. Reference single-bond lengths and the torsional constant were adjusted to fit the experimental data for CH4 and C2H6. Double bonds and ring bonds in cyclopropyl compounds were considered to be rigid. The above selections served to establish a universal model force field for hydrocarbons with no remaining adjustable parameters. The potential energy functions for a variety of saturated hydrocarbons and several olefins and cyclopropyl derivatives were minimized with respect to independent structure parameters (i.e., bond stretches, bends, and internal rotations). Even though all C–H (and C–C) bonds were input to be identical to those in CH4 (and C2H6) except for nonbonded environment, the bond lengths and angles corresponding to the minimum potential energy exhibited an appreciable variation from molecule to molecule, as did also the strain energies of geometric and rotational isomers. Calculated trends in structures, isomerization energies, and barriers to rotation agreed quite well with experimentally observed trends, provided that experimental isomerization energies were corrected to 0°K and zero-point energies were taken into account. Some novel features of the results and applications of the model for predicting deformations in strained systems are discussed. The present study differs from previous work in the area of ``molecular mechanics'' in the use of a more general force field, in allowing the strained molecules to relax in all degrees of freedom (except for unsaturated groups and cyclopropyl rings), in the selection of molecular systems, and in a detailed comparison with experiment.