From Quantum Mechanics to Experimental Observables: Computational Investigations of Energy-Related Heterogeneous Catalysts

Citation

Qian, Jin (2019) From Quantum Mechanics to Experimental Observables: Computational Investigations of Energy-Related Heterogeneous Catalysts. Dissertation (Ph.D.), California Institute of Technology. doi:10.7907/SPEJ-5X35. https://resolver.caltech.edu/CaltechTHESIS:12072018-215548214

Abstract

One of the most severe challenges in this decade is assuring more secure, more efficient, cleaner, and more sustainable energy to power our world. This work takes a catalytic approach to help overcome this challenge.

The Haber Bosch process is one of the towering achievements of industrial chemistry. It consumes a huge amount of energy due to the high temperature and high pressure reaction condition, and in turn, has enabled us to produce enough nitrogen fertilizer to feed the current world population. An essential goal of present research is therefore to dramatically reduce Haber Bosch energy cost by improving the catalytic performance of the presently used Fe-based heterogeneous catalysts. We use quantum mechanics (QM) and kinetic Monte Carlo (kMC) to predict reaction mechanisms and kinetics for NH₃ synthesis on Fe(111) – the best Fe single crystal surface for NH₃ synthesis. We find excellent agreement with a predicted turnover frequency (TOF) of 17.7 sec^-1 per 2x2 site (5.3 x 10^-9 moles/cm²/sec) compared to TOF=10 sec^-1 per site from experiment, and we further predict that top-layer Co doping leads to an acceleration by a factor of 2.3 in reaction rates of ammonia synthesis.

Compared to the industrialized Haber Bosch reaction, renewable energy technologies are still in their infancy with a great deal of questions unanswered, as well as a lot of barriers to overcome. Here we report our atomistic understanding of how CO₂ and H₂O molecules adsorb on the catalyst surface and interact to initiate CO₂ dissociation and subsequent product formation. Using synergistic experimental and theoretical analyses, we show that Cu and Ag operate entirely differently for the first step of activating CO₂. We develop a method of predicting the ambient pressure XPS spectrum in an ab-initio multiscale fashion: from electronic structure, to atomic picture, to chemical reaction network (CRN), and eventually to the experimental observable. We bridge both the qualitative and quantitative gap from quantum mechanics to XPS, and demonstrate our approach by decoding the initial H₂O adsorption and complex formation on Ag(111) surface, which we encourage to be the new standard protocol in this field.

Thesis Files