In recent years, significant research efforts on fundamental catalytic research have been undertaken, driven by the need to control and improve heterogeneous catalysis processes at a molecular level. Model systems have been designed and studied to investigate catalytic processes, while most of the work so far has involved metal nanoparticles deposited on oxide supports. Recently, another type of model catalyst system has been suggested which consists of a metal single-crystal surface decorated with submonolayer quantities of an oxide phase. This "inverse-catalyst model system" offers interesting possibilities to explore the effects of the metal-oxide interface on the reactivity.

In this thesis, we use density functional theory, with the GGA-PBE functional, to investigate the ability of vanadium oxide clusters, supported on Ag or Au, to break the C--H bond in methane. We perform a thermodynamic analysis to show that the VO4 cluster is the most likely oxidant and then proceed to calculate the energy of the dissociative adsorption of methane and its activation energy with the Nudged Elastic Band (NEB) method. It appears that isolated VO4 clusters supported on Au (111) are promising catalysts for the first step in methane activation, the breaking of the C--H bond. We explain some peculiar features of the reaction path and propose that they are general for alkane activation on oxides.

We also observe that the support makes a substantial difference and that Au is a much better support than Ag. This probably happens because Au makes weaker bonds with the oxygen in VO4, or, equivalently, because VO4 binds less strongly to Au than to Ag. We propose that if one compares the activity of the same cluster on a variety of supports, then the reaction energy for the dissociative adsorption is higher when the bond of the cluster to the support is weaker. We emphasize that these "rules" are, at this point, based on very few examples and need to be tested further.

An important role in catalysis is the surface mobility of active species in the catalytic process and also during activation or regeneration treatments of the catalysts. To quantify the degree of difficulty for VO4 clusters to move on Au (111) and Ag (111) surfaces, we calculate, by using the NEB method, the energy barrier for the motion of the VO4 cluster along the surface. The energy barrier for VO4 moving on a Ag surface is higher than that on a Au surface, which is not surprising considering the bonding character between the cluster and metal surfaces. The high mobility verifies our observation that the bonding connection between the cluster and the surface is much weaker after the dissociation and less sensitive on the positions of the VO4 cluster. We observe that the adsorption and the dissociation of methane on the cluster are equally probable for all intermediate cluster geometries during the motion from one lattice site to another.

As another model system for catalysis, MgO oxide surfaces have been an active research topic in the last decade and it is believed that oxygen vacancies on oxide surfaces can participate in various chemical reactions. The formation of oxygen vacancies on MgO surfaces has been studied extensively. However, previous works are largely limited to the neutral case and the studies of charged defects are mainly focus on their geometric and thermodynamic properties. Furthermore, the doping effect is barely included in the previous theoretical works due to its complexity. In this work, we study the geometric and electronic properties of the oxygen vacancies with different charge states on the MgO (001) surface. We have developed a methodology for calculating charged defect formation energies at surfaces. From first-principles calculations we obtain the formation energies of the oxygen vacancies on MgO surface with 0, +, 2+ and -1 charge states under both oxygen-rich and oxygen-poor conditions. We investigate the effects of doping on the formation energy and concentration of oxygen vacancies. The doping effect is considered in our calculations through the corresponding change of the Fermi level. We further study the absorption of oxygen molecules at these vacancies, calculate their relative energetic stability, and analyze its implication on their surface catalytic properties. This study will shed light on the catalytic activity of charged oxygen vacancies on oxide surfaces.