In times of an environmental and energy crisis, heterogeneous catalysis is more than ever a key technology. It enables the economical synthesis of bulk and specialty chemicals, as well as efficient and environmentally friendly energy conversion (“green chemistry”). The main challenge, also within this thesis, is to establish structure-function relationships of materials down to the atomic scale to understand and design suitable (i.e., cheap, efficient, and stable) catalysts, overcoming the still wide-spread empirical approach. A crucial role is played by the often very complex catalyst surface, whose function can be mimicked by well-defined and simpler surfaces of so-called “model catalysts” (single crystals, foils, thin films, supported nanoparticles or nanoclusters).The present thesis is structured according to contributions to three different research strategies and developments carried out in the field of heterogeneous model catalysis: (1) construction of a new flow reactor for kinetic studies of model catalysts, (2) modification of metal and oxide surfaces by Li and (3) composition analysis of pure and doped oxide-supported Au nanoclusters.First, a new flow reactor dedicated to measure catalytic reactions on single crystal and metal foil catalysts with a typical size of roughly 1 cm2 was designed, constructed, and optimized. The goal was to combine catalyst surface preparation under (ultra-)high vacuum (i.e., Ar+ sputtering, annealing, O2/H2 treatment and physical vapor deposition (PVD) at temperatures up to 650 °C) and investigation of catalytic parameters and performance under atmospheric pressure flow conditions (at temperatures up to 250 °C), employing both mass spectrometry (MS) for qualitative and gas chromatography (GC) for quantitative product/reactant analysis. The setup’s centerpiece is a stainless-steel reaction cell consisting of three components (i.e., (1) a microreactor that also functions as a sample stage and includes a K-type thermocouple connection for temperature read-out, (2) a heating anvil, and (3) a gas supply nozzle), which can be assembled and sealed under vacuum and separated again upon evacuation. In closed configuration, the small dimensions of the cell result in a reaction volume of only 4 ml, which enables product formation sufficient for flow operation, despite the small active surface area of the inserted catalysts. This new setup was successfully tested using precleaned Pd and Pt metal foils as catalysts for the ethylene hydrogenation reaction. The activation energy obtained for the Pt foil (41.2 0.8 kJ∙mol-1) as well as the much higher activity of the Pd foil (77 % conversion at 20 °C) are well in line with results reported in the literature.Second, as a first step towards the preparation of LixCoO2(-like) thin films as well-defined model systems for research on thermal and electrocatalytic heterogeneous reactions as well as Li-based batteries, the surface of a Co3O4(111) thin film grown on an Ir(100)-(1×1) substrate was modified with Li. Beforehand, a polycrystalline Ni foil and Ni(100) single crystal were used as test substrates, utilizing a self-built Li evaporator for Li-PVD. Surface-sensitive characterization of the chemical composition and structure of deposited thin films was performed via X-ray photoelectron spectroscopy (XPS), low energy ion scattering (LEIS) and low energy electron diffraction (LEED). Li, Li2O, and LiOH species, accommodated at substrate surfaces upon stepwise and extensive Li-PVD experiments, were identified by XPS. Based on their quantification and the model of a uniform thin film deposited on a semi-infinite substrate, Li-deposition rates and film thicknesses were calculated. This allowed the calibration of the evaporator to deposit controllable amounts of Li on Co3O4(111). Deposition of Li for 3 min led to the formation of Li2O with higher accommodation (4 monolayers (ML)) than expected from results acquired previously with clean Ni substrates, which was attributed to “reactive deposition” (i.e., consumption of abundant surface lattice oxygen by Li).Third, three types of oxide-supported Au nanocluster catalysts, namely (1) Au38 on CeO2,(2) Pd-doped Au25 on TiO2, and (3) pure Au25 as well as Cu- and/or Pd-doped Au25 on ZnO, were characterized via XPS in different stages (i.e., as prepared, after oxidative and/or reductive pretreatment and reaction). Therefore, effects of pretreatment and reaction could be studied, including the influence of the oxide support. Removal of the protective thiol ligand shell required for subsequent catalytic CO oxidation was achieved by oxidation at 250 °C in case of Au38 on CeO2. However, during the same pretreatment at 150 °C, the catalyst was poisoned due to sulfuric deposits of ligand fragments, which migrated to the CeO2 support. It was observed that Au25 nanoclusters on TiO2, doped with several Pd atoms, most likely formed a PdxAuy alloy (x = 0.25-0.33, y = 1) during oxidative and reductive pretreatment at 250 °C. After CO oxidation, the sulfidic sulfur of the ligands was found to be fully oxidized to sulfate and remained at the catalyst surface. At higher temperatures and due to the reductive environment of the water-gas shift reaction, pure and doped Au25 nanoclusters were partially encapsulated by the ZnO support according to transmission electron microscopy (TEM) images due to strong metal-support interaction (SMSI). This was confirmed by unusually large downward shifts in binding energies of all cluster species present, which is explained by an electron transfer from the support to the clusters. The amount of sulfur decreased significantly throughout, while remaining in the sulfidic state, suggesting either the formation of ZnS or encapsulated ligand fragments. Quantitative evaluation of the doped clusters used yielded the average stoichiometries PdAu24, Cu2Au23, and Cu3PdAu21.Overall, significant progress was made in all three areas, pushing forward the boundaries of sensitivity and complexity. In the case of the newly developed flow reactor, proof of concept was provided, enabling sensitive acquisition of kinetic data from model catalysts. The preparation and characterization strategies demonstrated in Li-modification of Ni and Co3O4 surfaces lay the foundation for upcoming studies on mixed-oxide model catalysts. Preparation, pretreatment, reaction conditions, doping (down to a single atom per cluster), and interactions with the oxide support all affect the electronic structure of Au nanoclusters, which has been successfully detected with XPS, complementary to other highly surface-sensitive methods, from which further studies will benefit considerably.
