The kinetics of heterogeneously catalysed reactions is often described by highly simplified models. For example, the reacting adsorbates occupy one kind of site, surfaces do not reconstruct and lateral interactions between adsorbates are often neglected. Particularly the latter is only allowed for surfaces that are nearly empty. When adsorbate concentrations are high, interactions between adsorbed species become important and they affect both the bonding configuration of the adsorbates and the rates of the reactions between them. The research described in this thesis focuses on the strength and the nature of these interactions and the role they play in catalytic reactions. Throughout this thesis, a combination of several spectroscopic techniques and computational tools is used to visualize and quantify these interactions. The surface chosen for these studies is the (100) face of rhodium, which has the advantages that it is relatively reactive but still relatively easy to prepare and clean in ultrahigh vacuum. Moreover, the structure is highly symmetric, which is an advantage for computational methods. On the experimental side, the most important techniques used are Reflection Absorption Infrared Spectroscopy (RAIRS) to probe the bonding geometry by vibration analysis, Temperature Programmed Desorption (TPD) to monitor the evolution of gas phase products, surface concentrations and adsorbate bond strength, and Low Energy Electron Diffraction (LEED) to reveal the structure of ordered adlayers. Two computational tools are used: Density Functional Theory (DFT) to calculate relative stabilities, bonding geometry, and reactivity of adsorbate structures and kinetic Monte Carlo (kMC) to investigate the role of lateral interactions in adlayer ordering. Chapter three is an introduction to the behaviour of CO on the Rh(100) surface without any co-adsorbates. CO can order into three different adlayer structures with the structures at higher coverages being less stable. At low coverages, CO predominantly occupies the top site; whereas in more condense adlayer structures also the bridge site becomes occupied by CO. Under conditions with equilibrium between CO adsorption and desorption, the surface concentration of CO is determined with X-ray Photoelectron Spectroscopy (XPS), Secondary Ion Mass Spectrometry (SSIMS) and RAIRS. This can be done accurately with all three techniques, but only in a certain coverage range for SSIMS and especially RAIRS. Both techniques suffer from non-linearities in signal intensity when the adsorbate coverages increase and in RAIRS lateral interactions may even quench some of the IR absorption bands for the CO molecule. In Chapter four, the lateral interaction parameters between adsorbed CO molecules are obtained by fitting measured TPD traces with Monte-Carlo simulations using differential evolution methods. CO coverages from 0 to 0.83 ML and configurations including both top and bridge sites are considered. The parameters of the fit are the preexponential factor of desorption, the adsorption energy of CO on top and bridge sites, and four types of lateral interactions between CO molecules. The challenge is to obtain a good fit of the experimental desorption traces with simulations in which both the relevant adlayer structures are reproduced, and the kinetic parameters are physically realistic. For this purpose, dependencies between interaction parameters have been included. There is no exact fit between experiment and simulation and the magnitude of the obtained kinetic parameters can only qualitatively be determined. This lack of accuracy is mainly attributed to the fact that CO is not allowed to relax out of the ideal sites in the simulation. The role of interactions between CO and atomic H is discussed in Chapter five. Experiments and DFT calculations are combined to create a model of a co-adsorption structure hitherto unknown in literature. At 150 K, a c(3v2×v2)R45° structure is formed with CO occupying bridged sites and hydrogen occupying both bridge sites on the surface and octahedral sites below the surface. CO is influenced by atomic hydrogen by changing adsorption site from its favored top position on the empty surface to the bridge position in the presence of hydrogen. Hydrogen, on its turn, is affected in three ways by CO: it is destabilized, partly displaced and moved into sub-surface sites. CO strongly facilitates recombinative desorption of hydrogen, as the activation energy for H2 formation decreases from 85 kJ/mol on the clean surface to 24 kJ/mol in the presence of CO in the rather dense c(3v2×v2)R45° structure. Chapter six deals with the influence of atomic N on CO, as is seen in its desorption. When the Rh(100) surface is saturated with atomic N, a clear c(2×2) pattern is observed with LEED, which should correspond to a coverage of 0.50 ML of N-atoms. XPS experiments have shown instead that only 0.35 ML N is present under these conditions and hence the perfect c(2×2) structure contains many defects where CO can adsorb. Kinetic Monte-Carlo simulations have been used to visualize the distribution of nitrogen atoms in such structures. Two frequently occurring defects are identified consisting of small groups of nitrogen atoms in triangular configurations. CO within these groups occupies the centre position of these triangles, due to repulsive interactions between N and CO. As a consequence, the activation energy for CO desorption has decreased from 135 (clean surface) to 70 kJ/mol when surrounded by N-atoms. By relating this destabilization to the number and distance of the nitrogen neighbours a repulsive CO/N pair-wise lateral interaction energy of 23 kJ/mol is inferred. The influence of lateral interactions on the CO oxidation reaction is the subject of Chapter seven, where we report a strong correlation between CO oxidation kinetics and the proximity of the reactants on the surface. Three adlayer structures are found for the oxygen adlayer: p(2×2), c(2×2) and (2×2)-p4g. In the latter, the high surface coverage forces the surface into the well known clock-wise rotation reconstruction. Based on LEED structures and RAIRS spectra, co-adsorption structures of CO with O could be identified. These correspond to different regimes of surface reactivity, which show up in the temperature programmed reduction experiment. The reactivity increases with the density of the CO+O configurations. Near zero coverage, isolated CO and O species react relatively slowly, as evidenced by the high reaction temperatures of 400 to 500 K. In the second regime at ¿O = 0.25 ML, oxygen occupies four-fold hollow sites in a p(2×2) ordered structure, with CO on bridge positions in between. CO oxidation now occurs around 360 K, which is faster than at low coverage. In the third regime at 290 K, reactivity is determined by top CO in defects of the c(2×2) oxygen layer. The most reactive situation occurs in the densely packed structure on top of the reconstructed surface. Here oxygen occupies the three-fold hollow site and reacts with CO on bridge sites incorporated in the oxygen layer. The experiment shows how CO oxidation kinetics depends on the structures and the proximity of the reacting species. Such effects are generally overlooked in studies on the kinetics of catalytic reactions. CO oxidation is among the simplest catalytic reactions and is therefore often used to investigate fundamental phenomena in surface reactivity. A slight more complicated reaction is that between CO and NO, described in Chapter eight. It proceeds via decomposition of NO to form N and O, followed by the reaction of CO with O to CO2 and the recombinative desorption of N2. Dissociation of NO is the crucial step as it determines the availability of atomic O and thereby the formation of CO2. NO decomposition is highly affected by neighbouring adsorbate species and the concomitant lateral interactions. Undisturbed, NO dissociates on Rh(100) easily around a temperature of 230 K. However, when surrounded by an excess of CO, NO decomposition is blocked until the temperature reaches values of around 350 K. At this temperature, an explosive surface reaction occurs which is triggered by the desorption of a fraction of CO. Free sites are created enabling NO to decompose. The evolving oxygen atoms react immediately with CO, thereby creating more free sites, which make this cascade of reactions autocatalytic, and in fact explosive. NO can also inhibit itself in the dissociation. If the surface contains an excess of NO, its decomposition becomes blocked till 425 K, where desorption of NO becomes feasible. The reason that the reaction starts at higher temperature than when CO is a neighbour reflects the stronger chemisorption bond between NO and the rhodium surface. The decomposition reaction mechanism of ethylene glycol is described in Chapter nine. Ethylene glycol is a model molecule for sugars and hence, understanding how ethylene glycol reacts may provide some insight in how biomass or sugars derived from it react on metal surfaces. At low temperatures, ethylene glycol adsorbs through its hydroxyl group to the surface, upon which both hydroxyl bonds break simultaneously, creating an (-OCH2CH2O-) intermediate and H-atoms. This intermediate decomposes further via an unstable (-COCH2O-) intermediate to form CO an H2 or synthesis gas, around room temperature.
|Qualification||Doctor of Philosophy|
|Award date||28 Apr 2010|
|Place of Publication||Eindhoven|
|Publication status||Published - 2010|