Wolfgang Ranke

Bulk and surface phases of iron oxides in an oxygen and water atmosphere at low pressure

Thermodynamic stability ranges of different iron oxides were calculated as a function of the ambient oxygen or water gas phase pressure (p⩽1 bar) and temperature by use of the computer program EquiTherm. The phase diagram for Fe–H2O is almost completely determined by the O2 pressure due to the H2O dissociation equilibrium. The formation of epitaxially grown iron oxide films on platinum and ruthenium substrates agrees very well with the calculated phase diagrams. Thin films exhibit the advantage over single crystals that bulk diffusion has only limited influence on the establishment of equilibrium phases. Near the phase boundary Fe3O4–Fe2O3, surface structures are observed consisting of biphase ordered domains of FeO(111) on both oxides. They are formed due to kinetic effects in the course of the oxidation to hematite or reduction to magnetite, respectively. Annealing a Fe3O4(111) film in 5 × 10−5 mbar oxygen at 920–1000 K results in a new γ-Fe2O3(111)-like intermediate surface phase during the oxidation to α-Fe2O3(0001). A model is suggested for the growth of iron oxides and for redox processes involving iron oxides. The formation of several equilibrium surface phases is discussed.

Adsorption of water on FeO(111) and Fe3O4(111): identification of active sites for dissociation

The adsorption of water on epitaxial FeO(111) and Fe3O4(111) films structurally well characterized by STM and LEED was investigated with photoelectron and thermal desorption spectroscopy. On the FeO(111) surface terminated by a close-packed oxygen layer monomeric water spe- cies get physisorbed. On the Fe3O4(111) surface terminated by ¼ monolayer of Fe atoms located over a close-packed oxygen layer underneath water dissociates resulting in adsorbed OH groups. The OH saturation coverage corresponds to the number of surface Fe atoms, which is much larger than the surface defect concentrations. Therefore, the dissociation takes place at Fe sites exposed on the regular Fe3O4(111) surface, and the FeO(111) surface is chemically inert because no Fe sites exist thereon.

Structure and reactivity of iron oxide surfaces

Epitaxial films of different iron oxide phases and of potassium iron oxide were grown onto Pt(111) substrates and used for studying structure–reactivity correlations. The film morphologies and their atomic surface structures were characterized by scanning tunneling microscopy and low energy electron diffraction including multiple scattering calculations. The adsorption of water, ethylbenzene, and styrene was investigated by temperature programmed desorption and photoelectron spectroscopy. A dissociative chemisorption of water and a molecular chemisorption of ethylbenzene and styrene is observed on all oxides that expose metal cations in their topmost layers, whereas purely oxygen-terminated FeO(111) monolayer films are chemically inert and only physisorption occurs. Regarding the technical styrene synthesis reaction, which is performed over iron oxide based catalysts, we find a decreasing chemisorption strength of the reaction product molecule styrene, if compared to ethylbenzene, when going from Fe3O4(111) over α-Fe2O3(0001) to KFexOy(111). Extrapolation of the adsorbate coverages to the technical styrene synthesis reaction conditions using the Langmuir isotherm for coadsorption suggests an increasing catalytic activity along the same direction. This result agrees with previous kinetic experiments performed at elevated gas pressures over the model systems studied here and over polycrystalline iron oxide catalyst samples. It indicates that the iron oxide surface chemistry does not change across the pressure gap and that the model systems simulate technical styrene synthesis catalysts in a realistic way.

Water adsorption and growth of ice on epitaxial Fe3O4(111), FeO(111) and Fe2O3(biphase)

Deuterated water adsorption on epitaxially grown FeO(111), Fe3O4(111) and Fe2O3 (biphase) films was investigated in the range 110–320 K by infrared reflection–absorption spectroscopy (IRAS) and temperature programmed desorption (TPD) spectroscopy. At 110 K, a first water layer forms on Fe3O4(111) and Fe2O3 (biphase) before the second and higher layers develop. The first half layer on Fe3O4 adsorbs dissociatively. The second half layer develops features characteristic for hydrogen bonding and the formation of dimers is concluded. Also on Fe2O3(biphase), initial water adsorption is dissociative. A strongly bound minority species is observed. Heating to 169 K causes formation of ice clusters. On FeO(111) adsorption is molecular and weak. On all studied surfaces, thick ice layers grown at 110 K are amorphous. On Fe3O4(111) they transform at 170 K into hexagonal ice (IceH) while up to 10 L on FeO(111) remain amorphous. The mechanisms for adsorption and ice formation correlate with structure and termination of the different oxide surfaces.

Interaction of ethylbenzene and styrene with iron oxide model catalyst films at low coverages: A NEXAFS study

The adsorption of ethylbenzene and styrene on well ordered epitaxial iron oxide model catalyst films with different stoichiometries was investigated using near-edge X-ray absorption fine structure spectroscopy (NEXAFS). On the iron-terminated Fe3O4(111) and α-Fe2O3(0001) surfaces chemisorption of ethylbenzene and styrene is observed occurring initially on the iron sites ia the π-electron system of the phenyl ring. This forces the molecules into an almost flat configuration (η6-like ring adsorption geometry). In the case of ethylbenzene this adsorption complex is supposed to lead to an activation of the C–H bonds, thus facilitating the dehydrogenation to styrene. The tilt angle of the aromatic ring systems increases to about 40° when approaching monolayer saturation. In contrast, the interaction with the oxygen-terminated FeO(111) surface is weak and of the physisorption type. The adsorbate–adsorbate interaction dominates and causes a tilted adsorption of the molecules from the beginning.

Understanding heterogeneous catalysis on an atomic scale: a combined surface science and reactivity investigation for the dehydrogenation of ethylbenzene over iron oxide catalysts

In order to study the dehydrogenation of ethylbenzene to styrene, epitaxial iron oxide model catalyst films with Fe3O4(111), α-Fe2O3(0001) and KFexOy(111) stoichiometry were prepared in single crystal quality on Pt(111). They were investigated using surface science techniques before and after atmospheric pressure reaction experiments in a newly designed single crystal flow reactor. As expected from low-pressure measurements, Fe3O4(111) is catalytically inactive. The catalytic activity of α-Fe2O3(0001) starts after an activation period of about 45 min. After that, the surface is essentially clean but shows a high concentration of defects. On the potassium-promoted films, however, the activation period is much longer, the activity then is higher and the surface gets covered completely with carbon and oxygen during reaction. This indicates a different reaction pathway on the promoted films with a carbon–oxygen species as catalytically active species.

Surface structures of alpha-Fe2O3(0001) phases determined by LEED crystallography

We present a dynamical tensor low energy electron diffraction (LEED) study of α-Fe2O3(0001) surface structures that form in an oxygen pressure range from 10-5 to 1 mbar. Epitaxial α-Fe2O3(0001) films were prepared on Pt(111) in defined oxygen partial pressures at temperatures of around 1100 K. In 1 mbar O2 strongly relaxed oxygen-terminated surface structures are formed, while in 10-5 mbar O2 three different surface structures yield rather good Pendry R factors. Further experimental evidence from scanning tunneling spectroscopy (STM) and ion scattering spectroscopy (ISS), in combination with a critical review of the literature, is only consistent with a hydroxyl termination forming in 10-5 mbar O2. The stabilization of both structures is discussed on the basis of electrostatic arguments considering the boundary conditions at the oxide–gas as well as the oxide–substrate interface (autocompensation). For oxygen pressures between 10-4 and 10-1 mbar O2, the two domains coexist as analyzed using a new, modified version of the symmetrized automated tensor LEED program package. The system investigated in this study turns out to be very complex and the LEED analysis alone is not capable of identifying the involved surface structures unambiguously. Only in combination with results from other surface-sensitive methods was it possible to deduce models for the most likely surface structures.

On the preparation and composition of potassium promoted iron oxide model catalyst films

Potassium promoted iron oxide model catalyst films were prepared by deposition of potassium onto epitaxial Fe3O4(111) films at 200 K, followed by annealing in the range 200 to 970 K. Their formation and composition were investigated by X-ray photoelectron spectroscopy (XPS) in combination with thermal desorption spectroscopy (TDS) and thermodynamic considerations. Already at 300 K a solid-state reaction occurred and the iron oxide was partly reduced. Around 700 K a KFeO2 phase was identified which transformed at higher temperatures into KxFe22O34(0.67<x<4). This transformation started from the bulk of the film so that initially a potassium-rich KFeO2 layer was formed on top of KxFe22O34. The formation of a single-crystalline KxFe22O34 (x = 0.67) layer, which is terminated by a submonolayer of potassium, is assumed to occur at 970 K. For a certain potassium content, this surface develops a well ordered phase with a (2 × 2) superstructure. The potassium containing phases are not stable in water atmosphere: In 10−8 mbar H2O, potassium hydroxide forms, then decomposes and desorbs beyond 400–500 K resulting in a potassium-depleted near-surface layer.

Adsorption of water on Fe3O4(111) studied by photoelectron and thermal desorption spectroscopy

The adsorption of water on Fe3O4(111) films grown epitaxially onto Pt(111) was investigated by ultraviolet photoelectron spectroscopy (UPS) in adsorption-desorption equilibrium and by thermal desorption spectroscopy (TDS). With increasing coverage both methods reveal the existence of three species on the surface: Dissociatively chemisorbed water (γ), physisorbed monomeric water (β) and hydrogen bonded condensed ice (α). The corresponding isosteric heats of adsorption qst and desorption energies Edes were determined by UPS and TDS, respectively. For the α- and βspecies they compare well, for the γ-species Edes is higher indicating an activation barrier for the dissociative adsorption, which is also supported by the observed slow adsorption kinetics. The dissociation is assumed to occur at iron cations with neighboring oxygen anions acting as proton acceptors.

Determination of adsorption energies and kinetic parameters by isosteric methods

If adsorption is reversible and sufficiently fast, isotherms or isobars can be measured in adsorption–desorption equilibrium at low pressures on single crystal surfaces. The isosteric heat of adsorption can easily be derived from them using the Clausius–Clapeyron equation. Reaction orders and frequency factors can, in principle, be deduced from a fit of the isobars (or isotherms) using the kinetic equations for adsorption and desorption. Immobile and mobile precursor kinetics can be included in the analysis but the fit fails when structural phase transitions in the substrate or the adlayer cause the kinetics to become complex. We review the methods, strong points and limitations of isobar (isotherm) measurements and of their kinetic fits by discussing the adsorption of water, ethylbenzene and styrene on FeO(111), Fe3O4(111) and Pt(111) and the adsorption of ammonia on germanium surfaces. Where the kinetic fit was successful, mobile precursor kinetics is quite common and frequency factors for desorption deviate considerably from the often assumed value of 1013 s−1.