From low dimensions to full configuration space: Generalising models for nonadiabatic molecular dynamics

Abstract

This thesis aims to bridge the development of nonadiabatic dynamics methods and their application for studies of real molecular systems.

First, this work explores fundamental concepts of photochemistry by investigating two different pictures, arising from the Born-Oppenheimer and the exact factorisation representation. Based on a simplistic model, a photochemical experiment from the excitation up to the formation of photoproducts is simulated. This study then compares the Born-Oppenheimer and exact factorisation representations of the processes. Subsequently, the influence of the Born-Oppenheimer picture for approximate nonadiabatic dynamics is investigated on two-dimensional model systems around conical intersections. The effects of neglected couplings and geometric phase are evaluated for ab initio multiple spawning (AIMS), a method for nonadiabatic molecular dynamics based on classically moving Gaussians.

Afterwards, this work introduces a standardised test set of molecules to connect between tests of newly developed nonadiabatic dynamics methods on one-dimensional model systems and their intended application to full-dimensional molecules. Inspired by the widely used one-dimensional Tully models, three molecules are selected to form the molecular Tully models, which undergo similar photophysical processes, but in a high-dimensional space. In addition, the recently proposed stochastic-selection AIMS framework is also tested on two molecules undergoing ring-opening reactions to explore the strengths and limitations of the method.

Finally, a direct comparison between experimental and computational results is presented. The photochemistry of 2(5H)-thiophenone is probed during and after the initial ring opening using time-resolved photoelectron spectroscopy. Static and dynamic calculations unravel the photoprocesses and identify a variety of photoproducts. Using the computational results, the experimental signal can be translated to insights into the ongoing photochemistry.

Overall, this thesis aims to bring models in nonadiabatic dynamics in a real-world context. This work contributes to facilitating the transfer of new nonadiabatic dynamics methods towards the study of molecules in their full dimensionality.