510 zyxwvutsrqponmlk S. L. Craig and J. I. Braurnan: Unimolecular Dynamics in Bimolecular Ion-Molecule Reactions zyxw Unimolecular Dynamics in Bimolecular Ion-Molecule Reactions Stephen L. Craig and John I. Braurnan*) Department of Chemistry, Stanford University, Stanford, California 94305-5080, USA zyxw Key Words: Chemical Kinetics zyxwvu / Energy Transfer / Mass Spectrometry We have measured the translational energy dependence of a series of chloride exchange reactions in the gas phase. The translational energy dependence varies with the changes in the potential energy surfaces across the series of reactions in a manner that is consistent with the predictions of statistical reaction rate theory. These results in- dicate that despite reports of non-statistical behavior in simple SN2 reactions, the assumptions of unimolecular reaction rate theory (strong coupling, energy randomization) are valid for the complexes formed in many bimolecular ion-molecule collisions. Introduction Unimolecular reaction rate theory has been applied exten- sively to the complexes formed in bimolecular collisions [ t, 21. The lifetimes of neutral-neutral and ion-molecule colli- sion complexes have been calculated quite accurately with statistical theories such as Rice-Ramsperger-Kassel-Marcus (RRKM) theory [l - 31, and this has facilitated quantitative- ly accurate descriptions of ion-molecule phenomena such as three-body association rates [4] across a range of pressures including the fall-off region. Extended to isomerizations within bimolecular complexes, unimolecular reaction rate theory has been used successfully to interpret the kinetics of many bimolecular reactions. Bimolecular nucleophilic sub- stitutions, or sN2 reactions, are one class of ion-molecule reaction to which statistical theories have been extensively applied. Despite the apparent success in modeling the ex- perimental kinetics and predicting energies of activation for these reactions zyxwvutsrq [5 - 131, recent experimental and theoretical results suggest that certain SN2 reactions behave “non-sta- tistically” [14 - 291. These results raise important concerns regarding the use of unimolecular theories to model SN2 reaction kinetics in particular and ion-molecule reactions in general. A central concern is whether or not all of the degrees of freedom in the system are sufficiently coupled so as to allow the available energy to sample the entire potential energy surface statistically. In particular, the coupling between relative translational and vibrational modes (T - V cou- pling) seems to be generally poor for a series of halide ex- change sN2 reactions. Three independent investigations report persuasive evidence to this effect. Hase and co-workers have performed trajectory calcula- tions and RRKM analyses on the sN2 reactions of chloride ion with methyl chloride and methyl bromide (Eq. 1 a and b, respectively) [18-21, 24-28]. Their calculations show that the probability of reaction depends on the initial trajectory of the system. For example, *) Author to whom correspondence should be addressed. excitation of the carbon-halide bond in the neutral reactant leads to a much greater reaction probability than does in- creased relative translational energy. In the limit of statisti- cal energy randomization, the probability of reaction should be the same for a fixed total energy and angular momentum regardless of the initial state of the system. In other words, the intermediate complex should have no “memory” of how it was formed. The complex, however, does remember the process by which it was formed, and this indicates a bottleneck to energy redistribution between what Hase terms “intermolecular” and “intramolecular” modes. These modes are closely coupled to relative translational and vibrational degrees of freedom, respectively, and the bottleneck between them reflects what is referred to in this paper as inefficient T - V coupling. Soon after Hase’s initial work, Graul and Bowers report- ed the kinetic energy release distribution (KERD), or rela- tive translational energy of the bromide and methyl chloride products, from the decomposition of metastable chloride- methyl bromide reactant complexes in Eq. (1 b) [16]. The experimentally observed KERD was significantly lower than that predicted by statistical phase space theory (PST), revealing that the energy generated in the exothermic reac- tion was trapped in the vibrational modes of the neutral molecule rather than distributed statistically between vibra- tional and translational modes. Further work by the same authors demonstrated that this result is general for a series of halide exchange reactions [29]. These observations are consistent with the dynamics observed in the trajectory calculations in that the non-statistical energy partitioning implies a bottleneck to T - V energy transfer. Concurrently, Viggiano and co-workers [30 - 321 devel- oped a technique to measure directly both the relative energy dependence and the overall temperature dependence of ion-molecule reactions in a selected ion flow tube (SIFT). They reported that, for the SNz reaction in Eq. (lb), the overall reaction rate exhibits a negative dependence on kinetic energy, but no dependence on the internal energy of the neutral methyl bromide [32]. The differential response to the same amount of translational and vibrational energy is explained neither by the change in the rate of complex formation nor by the increased potential barrier due to higher angular momentum at large translational energies [27]. The dynamics within the collision complex, therefore, Ber. Bunsenges. Phys. Chem. 101, SIO-515 (1997) No. 3 zyxwvutsr f VCH Yerlagsgesellschaft zyxwvut mbH, 0 4 9 4 5 1 Weinheim, 1997 wO5-9021/97/0303-0510 zyx 6 lS.W+.2S/O