POWER SPECTRA AS A DIAGNOSTIC-TOOL IN PROBING STATISTICAL NONSTATISTICAL BEHAVIOR IN UNIMOLECULAR REACTIONS

The possibility of utilizing different types of power spectra obtained from classical trajectories as a diagnostic tool to identify the presence of nonstatistical dynamics is explored by using the unimolecular bond-fission reactions of 1,2-difluoroethane and the 2-chloroethyl radical as test cases. In previous studies, the reaction rates for these systems were calculated by using a variational transition-state theory and classical trajectory methods. A comparison of the results showed that 1,2-difluoroethane is a nonstatistical system, while the 2-chloroethyl radical behaves statistically. Power spectra for these two systems have been generated under various conditions. The characteristics of these spectra are as follows: (1) The spectra for the 2-chloroethyl radical are always broader and more coupled to other modes than is the case for 1,2-difluoroethane. This is true even at very low levels of excitation. (2) When an internal energy near or above the dissociation threshold is initially partitioned into a local C-H stretching mode, the power spectra for 1,2-difluoroethane broaden somewhat, but discrete and somewhat isolated bands are still clearly evident. In contrast, the analogous power spectra for the 2-chloroethyl radical exhibit a near complete absence of isolated bands. The general appearance of the spectrum suggests a very high level of mode-to-mode coupling, large intramolecular vibrational energy redistribution (IVR) rates, and global statistical behavior. (3) The appearance of the power spectrum for the 2-chloroethyl radical is unaltered regardless of whether the initial C-H excitation is in the CH2 or the CH2Cl group. This result also suggests statistical behavior. These results are interpreted to mean that power spectra may be used as a diagnostic tool to assess the statistical character of a system. The presence of a diffuse spectrum exhibiting a nearly complete loss of isolated structures indicates that the dissociation dynamics of the molecule will be well described by statistical theories. If, however, the power spectrum maintains its discrete, isolated character, as is the case for 1,2-difluoroethane, the opposite conclusion is suggested. Since power spectra are very easily computed, this diagnostic method may prove to be useful.