A robust, automated nudged elastic band method in internal coordinates
by Björn Hein-Janke
Date of Examination:2025-02-28
Date of issue:2025-12-18
Advisor:Prof. Dr. Ricardo A. Mata
Referee:Prof. Dr. Jörg Behler
Referee:Prof. Dr. Burkhard Geil
Referee:Prof. Dr. Konrad Koszinowski
Referee:Prof. Dr. Martin A. Suhm
Referee:PD Dr. Tim F. Schäfer
Files in this item
Name:Thesis_corrected2_noaff.pdf
Size:5.39Mb
Format:PDF
Description:Final version of the PhD thesis
Abstract
English
In this work, a robust and reliable implementation of the Nudged Elastic Band (NEB) method is developed and tested. It is a method for finding reaction paths and TS, which can be used to theoretically compute reaction barriers, allowing a theoretical understanding of kinetics experiments. It can also be used in mechanistic studies, for finding reaction mechanisms, or explore entire networks of interconnected reactions. A short overview and comparison of methods for modeling reaction paths and finding transition states is provided, including the NEB method. The challenges of reaction path modeling and transition state finding are discussed, both in general and in the particular case of the NEB method. The NEB method and its implementation are discussed in detail. Various issues and pitfalls when using the NEB method are addressed, and solutions proposed and implemented. The NEB implementation thus developed is tested on a reaction set for benchmarking, and used in kinetics studies on real-world systems. One of these systems is the FeMAN challenge, a series of reactions where theoretical chemistry groups are challenged to accurately predict barrier heights and experimentally measured rate constants. The other set of syste ms are mixtures of substituted benzyl alcohols with methyl glycolate in the gas phase, which form a number of possible dimers, which can interconvert through a network of reaction paths. The results for these systems are discussed in detail. The NEB implementation was able to find sensible reaction paths and approximate transition state geometries for many of these reactions with comparatively little human oversight, allowing a ’semiautomatic’ workflow. This work’s aim to make NEB more robust and reliable was a success, but there is still room for improvement. Remaining issues, and possible avenues for future research are discussed.
Keywords: nudged elastic band; reaction mechanism; reaction kinetics; computational chemistry; chemical reactions; transition states; reaction path modeling; potential energy surface; chemical structure; structure optimization
