Theoretical investigations into single-atom catalysts on ceriumdioxide: deposition and activation

Identifier:  TDX:3005

Handle: https://hdl.handle.net/20.500.11797/TDX3005

Authors:  Daelman, Nathan

Abstract:
The main topic of this doctoral thesis is the investigation of single-atom platinum depositions on ceria and their potential for catalysis, all from a computational perspective. The US Department of Energy is looking for new car-exhaust catalysts that operate at lower temperatures of 150°C. These are crucial in the development of more environment-friendly, fuel-efficient combustion engines. It is demonstrated that single-atom platinum (a) is a candidate once activated and (b) outperforms commercial grade alternatives in oxidation reactions. Both results are the first-ever ab initio verification of experimental observations. The gain is traced back to single-atom platinum’s exotic oxidation states and their dynamic nature. An underlying mechanism is presented that couples the Strong Metal-Support Interactions with lattice phonons. The scope of this new Phonon-assisted Metal-Support Interaction is shown to be much broader, extending for example to single-atom nickel on ceria. In the last part, the challenges of state-of-the-art methods are addressed in two different areas. Firstly, a new, low-scaling implementation of second-order Møller-Plesset perturbation theory is benchmarked for the reduction of bulk phase CeO2 to Ce2O3. This is the first fully-converged wavefunction study of this reaction. As such, it is free from the debilitating self-interaction error of density functional theorem. Second are the implications of machine learning for the design of more inclusive, multi-scale models. A strategy is proposed to construct a high-dimensional potential energy surface for ceria using neural networks, similar to force fields in Biochemistry.