On the mathematical treatment of the Born-Oppenheimer approximation

TL;DR

This paper surveys the mathematical treatment of the Born-Oppenheimer approximation, reframing the problem as a semiclassical, energy-fixed reduction built from the electronic Hamiltonian $Q(x)$ and a finite-rank projection $\Pi(x)$. It develops the adiabatic operator $\Pi H \Pi$ (and refinements) as the core tool to approximate molecular dynamics, and surveys rigorous results for bound states, resonances, scattering, and time evolution, leveraging spectral resolution, direct integrals, Grushin problems, and complex distortion techniques. A central theme is the handling of continuous spectrum and thresholds, which the mathematical framework addresses more robustly than traditional heuristic derivations. The work clarifies the connection between electronic structure and nuclear dynamics, highlights the limitations of naive approaches, and points to future roles for semiclassical methods in extracting chemically relevant predictions from rigorous BO analysis.

Abstract

Motivated by a paper by B.T. Sutcliffe and R.G. Woolley, we present the main ideas used by mathematicians to show the accuracy of the Born-Oppenheimer approximation for molecules. Based on mathematical works on this approximation for molecular bound states, in scattering theory, in resonance theory, and for short time evolution, we give an overview of some rigourous results obtained up to now. We also point out the main difficulties mathematicians are trying to overcome and speculate on further developments. The mathematical approach does not fit exactly to the common use of the approximation in Physics and Chemistry. We criticize the latter and comment on the differences, contributing in this way to the discussion on the Born-Oppenheimer approximation initiated by B.T. Sutcliffe and R.G. Woolley. The paper neither contains mathematical statements nor proofs. Instead we try to make accessible mathematically rigourous results on the subject to researchers in Quantum Chemistry or Physics.