Surface science studies of iron oxides as model catalyst supports

Abstract

The performance of heterogeneous catalysts often depends on the properties of the supporting substrate. The surface structure, its stability, its ability to support the active species and its structural evolution upon interaction with adsorbates all crucially affect the catalytic activity of the whole system. This thesis deals with surface science studies of two iron oxide surfaces: hematite α-Fe2O3 and magnetite Fe3O4, both of which are commonly used as catalyst supports or catalysts themselves. The first part of the thesis deals with the interaction of three dominant iron oxide surfaces with water. The (11 ̅02) surface of hematite (α-Fe2O3) was studied both as a pristine surface and after exposure to water vapor and liquid water. Two surface reconstructions with (1 × 1) and (2 × 1) symmetries were observed and studied by scanning tunneling microscopy (STM), noncontact atomic force microscopy (ncAFM), x-ray photoemission spectroscopy (XPS) and low-energy electron diffraction (LEED) and the results confirm the bulk-truncated model of the (1 × 1) surface. Water adsorption on this surface was studied using temperature programmed desorption (TPD), STM, ncAFM, XPS and LEED, which in combination with density functional theory computations clearly show that water adsorbs in the form of partially dissociated water dimers at all coverages up to multilayer formation. The very same spectroscopy and diffraction results acquired in UHV after exposure of the surface to liquid water indicate that the surface survives immersion and the same partially dissociated species are still stabilized on the surface. Based on the UHV results it cannot be said whether such species are present even during the actual immersion in liquid water, but such scenario is consistent with previously published in-situ diffraction studies. Water adsorption in UHV was also studied on the (001) and (111) surface of magnetite (Fe3O4) using the same approach. NcAFM imaging with a CO-terminated tip was used as the main method to image the individual water species. On both of these surfaces, water first adsorbs as small partially dissociated water agglomerates, but at slightly higher water coverages complex hydrogen-bonded networks are observed. While on both surfaces the water TPD spectra show desorption features at similar temperatures, the initial water agglomerate formation mechanisms are quite different. On the (001) surface, partially-dissociated water dimers and trimers were clearly identified, which are formed at specific sites of the surface and at higher water coverage these anchor the hydrogen-bonded network. On the (111) surface the smallest water agglomerates were clearly experimentally identified as trimers, but their observed shape in ncAFM images was inconsistent with the supporting computational results. This discrepancy can be explained if the water adsorption modifies the surface by pulling out an additional Fe cation to from the bulk to the surface, and the same mechanism could stabilize highly specific water hexamers which were experimentally observed at higher water coverage. These results suggest, that the Fe3O4(111) surface might be surprisingly dynamic even at cryogenic temperatures. The last chapter of this thesis deals with Ni, Rh and Ir adatoms and small clusters on the Fe3O4 (001) surface. Depending on sample preparation, all these adatoms can occupy multiple sites on Fe3O4(001) and it was studied how the different local environment affects the adsorption properties and reactivity. The interaction of the Ni/ Fe3O4 (001) system with water was studied with a view towards the enhanced performance of NiFe compounds for the water-gas shift and oxygen-evolution reactions. It was found that Ni adatoms can coordinate multiple water molecules above room temperature, but the water adsorbed on Ni incorporated in the octahedral sites of the support has similar characteristics as water adsorbed on intrinsic surface defects. The Ir and Rh/ Fe3O4(001) systems were studied to understand the high activity of Ir and Rh-based catalysts for CO oxidation. It was found that both Rh1 and Ir1 adatoms can coordinate either a single CO molecule or two CO molecules, and in both cases a square-planar coordination geometry is formed. In the monocarbonyl case, this is achieved by forming an additional bond to a subsurface oxygen atom. On both Ir and Rh, CO adsorption significantly stabilizes the single adatoms on the surface, which is rationalized by the analogy to rules from coordination chemistry and the availability of the preferred coordination geometry on the surface. Both Rh and Ir systems were found to catalyze CO oxidation via a Mars-van-Krevelen mechanism at elevated temperatures, but are deactivated by adatom incorporation once the adsorbed molecules desorb. Additional experiments on Rh/Fe3O4 were carried out to find whether the Rh1 species could catalyze CO oxidation via a Langmuir-Hinshelwood mechanism. TPD experiments with isotopically-labeled gases have shown that the system is active for such CO oxidation mechanism at low temperature if O2 is dosed before the CO, but it is not the single Rh1 species which are responsible for this activity. Instead, it was found that the O2 adsorption mobilizes the Rh which then agglomerates to form small oxidized RhxOy with weakly bound oxygen which acts as the oxidizing agent. If CO is dosed prior to O2, the highly stable carbonyls are formed and the system is blocked for O2 adsorption. The mechanism of CO oxidation on this system is therefore determined by the order in which the system is exposed to the two reactants. Overall, the results presented in this thesis provides significant fundamental insights into the surface structures, stability and reactivity of clean and modified iron oxide surfaces. On all the studied systems, previously unforeseen phenomena were observed, which highlights the importance of detailed surface science studies carried out on a case-by-case basis.