Off-lattice kinetic Monte Carlo methods and the Fe-H system: saddle-points, symmetry, and coroutines

In this thesis we apply, enhance, and extend the off-lattice kinetic Monte Carlo (OLKMC) method. The motivation for our research was building a general simulation framework, capable of modelling the complex interactions between crystal defects and hydrogen (H) in iron (Fe), into the timescales required to study the mechanisms of hydrogen embrittlement (HE). The underlying mechanisms of HE act on the atomic scale but manifest at the macro scale (particularly in the time dimension). We needed a method that could bridge the gap between these two universes and link them causally. Our primary contribution on the path to this goal include: \begin{itemize} \item An error-tolerant replacement for topological analysis of atomic local-environments which enables comparison of atomistic local-environments with a single, physically-meaningful parameter. \item Fully and self-consistently incorporating symmetry into saddle-point searching and local environment cataloguing, this can reduces the computational work by a factor of the number of symmetries in a local environment (as large as a factor of \num{24} in the case of BCC systems). \item Building a fully-portable continuation-stealing fork-join framework to efficiently parallelise heterogeneous work across heterogeneous processors, which was necessary to practically enable the simulations above. \end{itemize} We apply our OLKMC implementation to study the diffusion, disassociation and recombination of vacancy clusters in the presence of hydrogen. We demonstrate OLKMC is capable of reaching embrittlement timescales, of-the-order-of seconds, while simultaneously resolving the atomic motion of H-atoms. Through OLKMC, we are able to study the atomic mechanisms through which H impedes the diffusion of vacancy clusters and elucidate the effect H has on the disassociation barriers of small vacancy clusters. Crucially, we are able to link these atomic scale quantities to the macro scale and deliver experimentally testable hypotheses to explain HE.

Name dc:type.qualificationname

Doctor of Philosophy (PhD)

Level dc:type.qualificationlevel

Doctoral

Grantor dc:publisher.institution

University of Cambridge

Year dc:date.issued

2024

Author dc:creator

• Williams, Conor

Advisor dc:contributor.advisor

• Elliott, James

• hydrogen
• monte-carlo

Licence

• https://www.repository.cam.ac.uk/bitstreams/dfb5d8e8-159d-43d5-93ff-19f48bea19dc/download
• http://purl.org/NET/rdflicense/allrightsreserved

Language dc:language

eng

DOI dc:identifier.doi

https://doi.org/10.17863/CAM.120472

OAI identifier oai:identifier

oai:www.repository.cam.ac.uk:1810/387848