Solvent effects on the barrier to C-N bond rotation in N,N-dimethylaminoacrylonitrile: Modeling by reaction field theory and by Monte Carlo simulations

Solvents are known experimentally to influence strongly the barrier to rota tion about the conjugated C-N bond of N,N-dimethylaminoacrylonitrile (DMAAN ). The barrier increases with overall solvent polarity, but solvent hydroge n-bond donor ability does not have a measurable effect. Two solvation model s were explored in an attempt to reproduce the experimental data and obtain insight into the causes of the observed solvent effects. Calculations base d on the isodensity polarizable continuum model (IPCM) encoded in Gaussian 94, a representative dielectric continuum-based procedure, yielded fair agr eement for aprotic, nonhalogenated, nonaromatic solvents. The model predict s a linear correlation with the Onsager dielectric function, (epsilon - 1)/ (2 epsilon + 1), which was observed experimentally for this set of solvents . However, the model underestimated the magnitude of the solvent dependence by approximately 30%. As a representative example of an approach based on the use of explicit solvent molecules, Monte Carlo simulations were carried out with Jorgensen's BOSS package. The simulations strongly underestimated the influence of cyclohexane, consistent with earlier Monte Carlo studies of amides in nonpolar solvents. The simulations also underestimated the sol vent effects in acetonitrile and methanol, but reproduced the experimental data in water quite closely. Radial distribution functions from the water s imulations showed that the lack of an explicit hydrogen-bonding contributio n to the solvent effect resulted from a generally weak set of interactions between the cyano nitrogen and the nearest neighbor water molecule. Further more, these interactions changed very little as rotation about the amino C- N bond took place. The simulations suggested that hydrogen bonding to DMAAN is far more pronounced and variable in methanol, but the experimental data did not support this conclusion. None of the simulations showed significan t hydrogen bonding to the amino lone pair. The possibility is raised that s ome of the apparent inconsistencies in the calculations might result from t he inappropriate treatment of the transition state as a species for which t he solution environment is equilibrated.