Theoretical Insights into Catalysis on Nanoporous Gold from Ab Initio Molecular Dynamics

Abstract

Nanoporous gold (np-Au) has recently emerged as a highly selective catalyst, potentially suited to environmentally friendly and low-temperature applications. In contrast to the more extensively studied gold nanoparticle catalysts, the mechanistic understanding of catalytic processes on pristine and oxide-coated np-Au is far less developed. Quantum chemical methods, in particular, those based on density functional theory (DFT) can be used successfully to achieve a mechanistic understanding at the microscopic level, which is needed to optimize materials based on np-Au with respect to their catalytic properties. In this thesis, I summarize my computational work devoted to a deeper insight into the chemistry and physics of np-Au as a catalyst. Modern surface science has revealed that a catalyst is not a rigid body but undergoes rapid (sometimes irreversible) dynamic changes during chemical processes occurring on its surface. Whereas many theoretical studies still use an oversimplified model of a metal catalyst as a rigid, clean, and perfect surface, my PhD work examines dynamic processes occurring on the surface of np-Au in response to changes of the chemical environment, such as oxygen-induced surface restructuring, the formation of surface oxygen chain structures, and elementary catalytic reactions on np-Au, by using ab initio molecular dynamics (AIMD) simulations. In addition to AIMD simulations, traditional “static” DFT computations have been performed to verify minima and transition states in the reaction energy landscape and to construct reaction energy diagrams. The main results of the publications comprising the foundation of this thesis can be summarized as follows. Although np-Au consists of almost pure gold, silver atoms are also present on the surface as residues of the preparation process. Therefore, its surface chemistry turns out to be more complex than anticipated. Interactions between Au atoms and O atoms pre-adsorbed and/or generated during catalysis and the involvement of Ag impurities result in complex surface dynamics. First of all, with regard to pristine np-Au, the theoretical studies reveal that surface O atoms dynamically form one- and two-dimensional –(Au–O)– chains on a stepped Au(321) surface and lead to surface restructuring. In contrast, no chain formation has been found on Au(111), pointing to higher structural flexibility and propensity for restructuring of the stepped surface. Furthermore, our study predicts migration of subsurface Ag atoms to the surface, i.e. adsorbate-induced Ag surface segregation, in the presence of adsorbed atomic oxygen. Second, my thesis addresses the physics and chemistry of np-Au material functionalized with cerium oxide. To probe the reactivity of these systems and the involvement of inherent particle-support interactions towards CO oxidation, AIMD simulations and static DFT computations were carried out. As models, Ce10O20/19 NPs supported on thermodynamically stable Au(111) and on the stepped (rough) Au(321) surface were employed. For the ceria/Au(111) system, the simulations revealed the preference of a Mars-van-Krevelen type of reaction mechanism, in which a CO molecule first reacts with a lattice O atom of ceria rather than with an activated O22- surface species, forming CO2 and leaving an O vacancy behind. This vacancy becomes subsequently refilled by an O atom which diffuses from the site of the reaction of O2 with another CO molecule (at the gold-ceria perimeter). My studies also revealed that, in contrast, CO adsorption on the stepped Au(321) surface (in proximity to the ceria nanoparticle) may lead to the dynamic extraction of Au atoms from the surface resulting in Au-CO carbonyl species, which may subsequently diffuse on the Au surface and react with lattice O of ceria to CO2 with very low activation energy. After the reaction step, the extracted bare Au atom attaches to a step on the Au surface. As this is not the original site, this part of the catalytic cycle leads to a rearrangement of the surface structure. The second part of the cycle is likely to be the same as found for the ceria/Au(111) system. Finally, this thesis discusses the correlation between two types of catalysts that are “inverse” with respect to each other, namely, Au nanoparticles deposited on ceria and ceria nanoparticles deposited on Au (being the focus of this thesis), with respect to all aspects relevant for the surface reactivity, such as surface dynamics, charge transfer between the gold and the oxide phases, and the mechanism of CO oxidation