Abstract
dc:description.abstractDesigning next-generation materials for e.g., heterogeneous catalysis or gas storage requires a deep understanding of their surface chemistry. Simulations provide the needed atomic-scale insights, allowing for fundamental quantities like the adsorption energy and oxygen vacancy formation energy to be calculated. Consistently achieving accurate predictions will require going beyond the standard density functional theory (DFT) to methods like coupled cluster theory with single, double, and perturbative triple excitations [CCSD(T)]. However, the high computational cost of CCSD(T) typically necessitates the use of embedded cluster models, where the surface is modelled as a finite cluster [treated with CCSD(T)] that is coupled to the environment. Unfortunately, designing converged yet computationally tractable clusters is challenging, which has limited the routine application of CCSD(T) to surface chemistry. This thesis introduces the SKZCAM (pronounced "scam") protocol to overcome the limitations of embedded cluster models for metal-oxide surfaces. It provides a general, automated framework applicable to a wide range of ionic crystal structures, surface terminations, and adsorption/defect sites. In particular, it leads to small cluster sizes that are amenable for CCSD(T), enabling us to resolve long-standing debates across surface chemistry. For example, we reach consensus with experiments on the adsorption energy for the notorious CO on MgO(001) and 18 other molecule-surface systems, including molecules on the TiO₂ rutile(110) and anatase(101) surfaces. At the same time, we have uncovered new insights into complex surface systems, such as the structure of Au₂₀ and the binding behaviour of H₂O and CH₃OH on MgO(001). In addition, this set of adsorption energy and vacancy formation energy CCSD(T) references -- the largest in surface chemistry -- has also offered new insights into the performance of common density functional approximations and dispersion corrections in DFT. By integrating the SKZCAM protocol into an open-source workflow, we have now set the stage for accurate, low-cost predictions of surface reactions for reliable catalyst design.
Degree
thesis:*- 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 and committee
dc:creator, dc:contributor.*- Author dc:creator
-
- Shi, Benjamin Xu
- Advisor dc:contributor.advisor
-
- Michaelides, Angelos
Subjects
dc:subject × 7Rights
dc:rightsIdentifiers
dc:identifier.*- DOI dc:identifier.doi
- https://doi.org/10.17863/CAM.116944
- OAI identifier oai:identifier
- oai:www.repository.cam.ac.uk:1810/381924