Statistical approximations, potential energy surfaces

In the chapter on reaction rates, it was pointed out that the perfect description of a reaction would be a statistical average of all possible paths rather than just the minimum energy path. Furthermore, femtosecond spectroscopy experiments show that molecules vibrate in many dilferent directions until an energetically accessible reaction path is found. In order to examine these ideas computationally, the entire potential energy surface (PES) or an approximation to it must be computed. A PES is either a table of data or an analytic function, which gives the energy for any location of the nuclei comprising a chemical system. 

Statistical theories, such as those just described, are currently the only practical approach for many ion-neutral reactions because the fine details of the collision process are unknown all the information concerning the dynamics of collision processes is, in principle, contained in the pertinent potential-energy surfaces. Although a number of theoretical groups are engaged in accurate ab initio calculations of potential surfaces (J. J. Kaufman, M. Krauss, R. N. Porter, H. F. Schaefer, I. Shavitt, A. C. Wahl, and others), this is an expensive and tedious task, and various approximate methods are also being applied. Some of these methods are listed in Table VI, for example, the diatomics-in-molecules method (DIM). 

A wide variety of dynamical approximations have been applied to cluster dynamics and kinetics. Most calculations to date are based on simplified potentials and classical mechanics or statistical methods. In the near future, we can expect to see more work with detailed potential energy surfaces (both analytic and implicitly defined by electronic structure calculations) and progress in sorting out quantum effects and treating them more accurately. 

The determination of the microcanonical rate coefficient k E) is the subject of active research. A number of techniques have been proposed, and include RRKM theory (discussed in more detail in Section 2.4.4) and the derivatives of this such as Flexible Transition State theory. Phase Space Theory and the Statistical Adiabatic Channel Model. All of these techniques require a detailed knowledge of the potential energy surface (PES) on which the reaction takes place, which for most reactions is not known. As a consequence much effort has been devoted to more approximate techniques which depend only on specific PES features such as reaction threshold energies. These techniques often have a number of parameters whose values are determined by calibration with experimental data. Thus the analysis of the experimental data then becomes an exercise in the optimization of these parameters so as to reproduce the experimental data as closely as possible. One such technique is based on Inverse Laplace Transforms (ILT). 

Thus, by using a judicious combination of the LQA method and approximate evaluation of higher order terms in the Taylor-series expansion of the path, the potential energy surface information that is already available for performing statistical or dynamical calculations of the chemistry can be used to more accurately follow the path. 

Expression (67.Ill) can be considered as a "statistical formulation of the rate constant in that it represents a formal generalization of activated complex theory which is the usual form of the statistical theory of reaction rates. Actually, this expression is an exact collision theory rate equation, since it was derived from the basic equations (32.Ill) and (41. HI) without any approximations. Indeed, the notion of the activated complex has been introduced here only in a quite formal way, using equations (60.Ill) and (61.Ill) as a definition, which has permitted a change of variables only in order to make a pure mathematical transformation. Therefore, in all cases in which the activated complex could be defined as a virtual transition state in terms of a potential energy surface, the formula (67.HI) may be used as a rate equation equivalent to the collision theory expression (51.III). 

The situation is quite different in bimolecular reactions with an activation energy (E >0). In particular, the "diatomic" model is certainly a bad approximation for radical-radical rebinding along a double bond in which the maximum of the effective potential (35 IV) lies near the saddle-point of the potential energy surface /141/, In this case no central forces govern the nuclear motion hence, the total angular momentum is not a constant, which means that the reaction cannot be rotationally adiabatic. Therefore, in this situation the statistical theory cannot correctly reproduce the results of the simple collision theory. 

---