Measuring nonadiabaticity of molecular quantum dynamics with quantum fidelity and with its efficient semiclassical approximation

TL;DR

This paper introduces a rigorous measure of nonadiabaticity in molecular quantum dynamics using quantum fidelity, and proposes an efficient semiclassical approximation called MSDR, applicable to complex systems and various couplings.

Contribution

It develops a novel quantum fidelity-based measure for nonadiabaticity and introduces the multiple-surface dephasing representation (MSDR) for efficient semiclassical estimation.

Findings

MSDR accurately measures nonadiabaticity in model problems

The method evaluates the importance of various Hamiltonian couplings

Second-order nonadiabatic couplings are shown to be significant

Abstract

We propose to measure nonadiabaticity of molecular quantum dynamics rigorously with the quantum fidelity between the Born-Oppenheimer and fully nonadiabatic dynamics. It is shown that this measure of nonadiabaticity applies in situations where other criteria, such as the energy gap criterion or the extent of population transfer, fail. We further propose to estimate this quantum fidelity efficiently with a generalization of the dephasing representation to multiple surfaces. Two variants of the multiple-surface dephasing representation (MSDR) are introduced, in which the nuclei are propagated either with the fewest-switches surface hopping (FSSH) or with the locally mean field dynamics (LMFD). The LMFD can be interpreted as the Ehrenfest dynamics of an ensemble of nuclear trajectories, and has been used previously in the nonadiabatic semiclassical initial value representation. In addition to propagating an ensemble of classical trajectories, the MSDR requires evaluating nonadiabatic couplings and solving the Schr\"{o}dinger (or more generally, the quantum Liouville-von Neumann) equation for a single discrete degree of freedom. The MSDR can be also used to measure the importance of other terms present in the molecular Hamiltonian, such as diabatic couplings, spin-orbit couplings, or couplings to external fields, and to evaluate the accuracy of quantum dynamics with an approximate nonadiabatic Hamiltonian. The method is tested on three model problems introduced by Tully, on a two-surface model of dissociation of NaI, and a three-surface model including spin-orbit interactions. An example is presented that demonstrates the importance of often-neglected second-order nonadiabatic couplings.