The mapping of large-amplitude translational or intramolecular degrees of freedom onto effective harmonic bath models is often deemed to be a reasonable approximation which can correctly capture qualitative behaviors. This mapping, which forms the basis of Marcus’ electron transfer theory, is understood to be valid only for the purpose of extracting the dynamics of observables pertaining to the subsystem of interest and implies Gaussian behavior of the medium. It is not clear a priori how good this approximation is for a given system, although this question can in principle be addressed within the classical mechanical framework. Most importantly, no methodology is available for estimating corrections to the linear response approximation in a systematic fashion.
Related articles:
- The linear response approximation and its lowest order corrections: an influence functional approach
- Dissipative tunneling in a bath of two-level systems