Coadsorbate Layers Research Articles

Lateral interactions and non-equilibrium in surface kinetics

Work modelling reactions between surface species frequently use Langmuir kinetics, assuming that the layer is in internal equilibrium, and that the chemical potential of adsorbates corresponds to that of an ideal gas. Coverage dependences of reacting species and of site blocking are usually treated with simple power law coverage dependences (linear in the simplest case), neglecting that lateral interactions are strong in adsorbate and co-adsorbate layers which may influence kinetics considerably.My research group has in the past investigated many co-adsorbate systems and simple reactions in them. We have collected a number of examples where strong deviations from simple coverage dependences exist, in blocking, promoting, and selecting reactions. Interactions can range from those between next neighbors to larger distances, and can be quite complex. In addition, internal equilibrium in the layer as well as equilibrium distributions over product degrees of freedom can be violated. The latter effect leads to non-equipartition of energy over molecular degrees of freedom (for products) or non-equal response to those of reactants. While such behavior can usually be described by dynamic or kinetic models, the deeper reasons require detailed theoretical analysis. Here, a selection of such cases is reviewed to exemplify these points.

Experiment-Based Kinetic Monte Carlo Simulations: CO Oxidation over RuO2(110)

Kinetic Monte Carlo (kMC) simulations of the CO oxidation over RuO2(110) have been performed for a variety of different reaction conditions ranging from 10–7 to 10 mbar and temperatures from 300 to 600 K. The kMC simulations are based on reaction rates of elementary steps including diffusion, adsorption/desorption, and recombination of surface CO and O. The activation barriers for these elementary processes are mostly taken from temperature programmed reaction and desorption experiments of well-defined coadsorbate layers on RuO2(110) under ultrahigh vacuum conditions. We show that the experimental kinetic reaction data under steady state reaction conditions both in the 10–7 mbar range and in a batch reactor up to 10 mbar are reconciled within this experiment-based kMC approach. Experimental in situ reflection absorption infrared (RAIR) spectra in the frequency range of the CO stretch vibration depend sensitively on both the adsorption site and the local environment of the CO molecules, encoding thus the s...

Jürgen Behm: Building Bridges from Surface Science to Catalysis and Fuel Cell Research

Jürgen Behm, to whom this special issue of ChemPhysChem is dedicated on the occasion of his 60th birthday, 1 started his career at one of the best places imaginable for a young surface scientist in the late seventies—the laboratory of 2007 Nobel Laureate Gerhard Ertl at the Ludwig Maximilians University in Munich. Jürgen Behm′s early research focused on the structure of H and CO adsorbate layers on transition metals and remains a classic work in surface science. However, early on he also developed a major interest in surface reactions, which he pursued during his post-doc sabbatical with Dick Brundle at the IBM Almaden Research Center. Although having worked successfully with a wide variety of techniques, he is arguably best known for scanning tunneling microscopy, notably building up one of the two first instruments of this type in Germany after returning to the Ertl group in 1984. After Jürgen Behm had accepted a professor position in Munich in 1987, the STM technique became his major work horse. In the following five years, a stream of high-impact publications on atomic-resolution studies of clean and adsorbate-covered metal surfaces as well as on chemical processes at semiconductor surfaces and electrochemical interfaces emerged. These publications have been cumulatively cited more than 5000 times by now. In 1992 Jürgen Behm was ready for the next big leap. Moving to the newly created chair of surface science and catalysis at the University Ulm, he set out to build a group that covered all aspects of heterogeneous reactions, from atomic-scale studies of the surface reactivity of planar model systems up to real supported catalysts. In particular, he focused on electrochemical and gas-phase reactions relevant for low-temperature fuel cells—a topic for which he is nowadays as well known as for STM. Throughout his career, Jürgen Behm’s work has been characterized by utmost scientific rigor as well as by a great willingness to cross boundaries between different approaches, such as fundamental studies and applied research, and different disciplines, such as UHV surface science and electrochemistry, resulting in more than 300 publications. The contributions in this issue, coming from the large international community of his former and current collaborators, reflect the breadth of his interests.

Coadsorption of Hydrogen and CO on Hydrogen Pre‐Covered PtRu/Ru(0001) Surface Alloys

The influence of pre-adsorbed hydrogen on the adsorption of CO on well-defined PtRu/Ru(0001) surface alloys and the modification of the adsorption properties of both adsorbate species due to coadsorption was studied by a combination of temperature-programmed desorption and infrared spectroscopic measurements. The hydrogen binding strength is weakened both by surface alloy formation and by CO coadsorption. In the densely packed coadsorbate layers, which can be formed after CO post-adsorption, the adsorption properties of the adsorbed CO molecules are also affected significantly. Depending on the Pt surface concentration, the CO adsorption process on hydrogen pre-covered surfaces is blocked to a different extent with a decreasing blocking tendency for Pt-rich alloys.

O- and H-induced surface core level shifts on Ru(0001): prevalence of the additivity rule

In previous work on adsorbate-induced surface core level shifts (SCLSs), the effectscaused by O atom adsorption on Rh(111) and Ru(0001) were found to be additive:the measured shifts for first-layer Ru atoms depended linearly on the number ofdirectly coordinated O atoms. Density-functional theory calculations quantitativelyreproduced this effect, allowed separation of initial- and final-state contributions, andprovided an explanation in terms of a roughly constant charge transfer per Oatom. We have now conducted similar measurements and calculations for threewell-defined adsorbate and coadsorbate layers containing O and H atoms:(1 × 1)-H,(2 × 2)-(O+H)and (2 × 2)-(O+3H) on Ru(0001). As H is stabilized in fcc sites in the prior two structures and in hcp sites inthe latter, this enables us to not only study coverage and coadsorption effects on theadsorbate-induced SCLSs, but also the sensitivity to similar adsorption sites. Remarkablygood agreement is obtained between experiment and calculations for the energies andgeometries of the layers, as well as for all aspects of the SCLS values. The additivity of thenext-neighbor adsorbate-induced SCLSs is found to prevail even for the coadsorbatestructures. While this confirms the suggested use of SCLSs as fingerprints of the adsorbateconfiguration, their sensitivity is further demonstrated by the slightly differentshifts unambiguously determined for H adsorption in either fcc or hcp hollowsites.

The geometries of coadsorbate layers of O and H on Ru(0 0 1): How well can quantitative LEED see hydrogen atoms?

Using a CCD LEED system for the collection of IV data with low beam damage, and full dynamical as well as tensor LEED calculations, we have determined the geometries of the (2 × 2)-(O + 3H) and the (2 × 2)-(O + H) coadsorbate structures on Ru(0 0 1). We show that here quantitative LEED can locate the H atoms very well. Not only their sites (hcp in the first, fcc in the second case), but also the Ru–H spacings and changes in the first two substrate layers are clearly determined. We argue that this success is due to the relatively large data range and to the smaller H mobility compared to pure H layers caused by their repulsive lateral interactions with the oxygen atoms.

Structure and lateral interactions in binary and ternary coadsorbate layers of O, H and CO on Ru(001)

The mutual interactions and adsorption properties of the ternary Ru(001)-(2×2)-(O+H+CO) coadsorbate layer and some binary combinations thereof have been investigated using infrared absorption spectroscopy and thermal desorption spectroscopy. Common to all layers investigated is a (2×2)-O overlayer into which the other components have been incorporated afterwards. Hydrogen adsorption as well as H 2 thermal desorption are found to be strongly influenced by a site-blocking effect by coadsorbed CO which prevents any hydrogen uptake (when CO has been preadsorbed), or increases the H 2 desorption temperature by about 75 K (when CO has been post-adsorbed). This is why the preparation of the ternary system requires a definite adsorption sequence. For the binary Ru(001)-(2×2)-(O+ nH) system well-ordered layers are found for n=1 and n=3. A revised structural model is proposed to explain the apparent zeroth to first order desorption kinetics of the α 1 H 2 desorption peak ( Θ H>0.25 ML) which contradicts earlier subsurface hydrogen models; instead, the formation of localized high H-density areas accompanied by a site change (fcc→hcp) is suggested.

Vibrational characterization of a high-densityRu(001)−(2×2)−(NO+3O)phase

High-density $(\mathrm{NO}+\mathrm{O})$ coadsorbate layers on Ru(001) have been studied for oxygen precoverages, ${\ensuremath{\Theta}}_{\mathrm{O}},$ between 0.5 and 1 ML by means of high-resolution electron energy loss spectroscopy and temperature programmed desorption. In this oxygen coverage range NO adsorption is possible on any remaining hcp threefold-coordinated site, to a saturation coverage of ${\ensuremath{\Theta}}_{\mathrm{O}}+{\ensuremath{\Theta}}_{\mathrm{NO}}=1.$ On the well-ordered $(2\ifmmode\times\else\texttimes\fi{}2)\ensuremath{-}3\mathrm{O}$ oxygen layer $({\ensuremath{\Theta}}_{\mathrm{O}}=0.75\mathrm{ML})$ NO molecules adsorb at 90 K with a high sticking coefficient close to unity up to the saturation coverage of 0.25 ML. The NO sublayer is ordered as the NO's occupy the threefold sites of the $(2\ifmmode\times\else\texttimes\fi{}2)$ hole structure within the $(2\ifmmode\times\else\texttimes\fi{}2)\ensuremath{-}3\mathrm{O}$ mesh conserving the same symmetry. For this well-ordered $\mathrm{Ru}(001)\ensuremath{-}(2\ifmmode\times\else\texttimes\fi{}2)\ensuremath{-}(\mathrm{NO}+3\mathrm{O})$ layer the external and internal NO stretching modes show downward dispersions of 16 and $23{\mathrm{cm}}^{\ensuremath{-}1},$ respectively, from the $\overline{\ensuremath{\Gamma}}$ point to the ${K}^{\ensuremath{'}}$ point at the boundary of the surface Brillouin zone. The dispersion of the internal mode can be completely described by dynamical dipole-dipole coupling. This coupling is also dominant for the external mode dispersion for which additional substrate-mediated contributions exist. Based on this understanding of the dynamical coupling the chemical shift of the NO internal and external stretch is determined for various $(\mathrm{NO}+\mathrm{O})$ structures. It can be related to the occupation of nearest- and next-nearest-neighbor sites. The internal mode shows chemical shifts between 12 and $30{\mathrm{cm}}^{\ensuremath{-}1}$ per neighboring NO or O but is insensitive to the structure beyond the nearest neighbors. For the external mode significant chemical shifts due to the occupation of the next-nearest-neighbor sites have been determined.

Lateral interactions in coadsorbate layers: Vibrational frequency shifts

The effect of coadsorbed argon, hydrogen, and oxygen on the internal vibration of CO on Ru(001) has been studied by infrared absorption spectroscopy in order to disentangle electrostatic and chemical frequency shifts. Ar is expected to lead only to the former, H only to the latter, and O to a combination. In all cases, intermolecular interactions among CO molecules are avoided by working at very low CO coverages (0.01–0.03 ML). Interestingly, the observed frequency shifts are discrete rather than continuous which is attributed to a local interaction. Density functional calculations for suitable clusters have been used to model the frequency shifts, arriving at good agreement with experiment. Analysis of these theoretical results is then used to quantify the contributions of electrostatic fields and of chemical effects on these shifts. It is shown that, despite very different signatures of the various coadsorbate species, the observed C–O frequency shifts are largely of electrostatic origin, provided one uses the electrostatic field generated by the coadsorbate and not an effective constant field.

Lateral interactions in (NO+O) coadsorbate layers on Ru(001): fundamental and overtone modes

A systematic investigation of various (NO+O) coadsorbate layers on Ru(001) is presented. By using infrared absorption spectroscopy, fundamental and overtone bands were studied in a wide temperature range to extract dynamical information on lifetime broadening (low-temperature limit) and dephasing behavior (thermal broadening and line shifts). The layers studied were (1) 3NO-(2×2)/Ru(001), (2) (2NO+O)(2×2)/Ru(001) and (3) (NO+2O)(2×2)/Ru(001). It is found that replacing hollow-site NO by oxygen atoms causes characteristic frequency shifts of ν Ru–NO and ν N–O of linearly bond on-top NO which is present for all three layers. In agreement with earlier results on CO/Ru(001) and with theoretical prediction, the ν N–O″ overtone band displays rapid broadening with increasing temperature (about four times faster than the respective localized ν N–O fundamental band) while the temperature-dependent line shift and the 0 K linewidth is twice as high as observed for ν N–O. An unusually large bandwidth close to 100 cm −1 is found for the ν N–O mode of the row-like (NO+O)(2×1)/Ru(001) overlayer.

A systematic investigation of the geometrical structures of four oxygen/nitric oxide coadsorbate layers on Ru(001)

LEED IV analysis has been used to determine the detailed geometries of four well-defined ordered coadsorbate structures which can be formed by the interaction of NO with (2×1)-O and (2×2)-O layers on Ru(001), and which have been characterized previously by various surface spectroscopies in this laboratory. New high-resolution XPS measurements are also reported which provide exact information on the coverages and chemical types of the species concerned, information which is helpful for the selection of model structures. Post-adsorbing NO below 150 K onto the well-developed (2×1)-O row structure leads to a layer with equal amounts of O and NO consisting of alternating rows of oxygen atoms and NO molecules, i.e. a (2×1)-(O+NO) structure. All adsorbates sit in hcp sites. In terms of geometry as well as electronic and bonding properties, the NO (upright orientation, NO bond length r e=1.20 Å, Ru–N vertical layer distance z e=1.32 Å) is essentially identical to the electronegative v 1 NO species sitting in hcp and fcc sites in the pure NO layer reported previously [M. Stichler, D. Menzel, Surf. Sci. 391 (1997) 47]. The first-to-second Ru layer distance is considerably expanded ( d 12=2.22 Å), while that from the second to the third layer contracted ( d 23=2.08 Å). The oxygen parameters are virtually unchanged from those of the (2×1)-O layer. Heating this layer to 300–450 K leads to the desorption of half the NO and restructuring of the residual coverage. The resulting well-ordered (2×2)-(2O+NO) layer contains a honeycomb layer of O atoms, half of them having switched to fcc sites. The remaining NO molecules sit upright on the top sites surrounded by the O honeycombs, and have properties ( r e=1.12 Å, z e=1.76 Å) which are very similar to the electropositive v 2 species of the pure NO layer. There is a clear difference between the hcp and fcc Os ( z e=1.20 Å and 1.39 Å, respectively); they can also be distinguished in XPS. The average distances d 12 and d 23 are less changed than in the (2×1)-(O+NO) layer, but there is considerable buckling. Starting from a (2×2)-O layer with O on hcp sites, one or two NO molecules can be incorporated per unit mesh. In the (2×2)-(O+NO) layer, the NO sits upright on the top site, with essentially v 2 parameters. In the (2×2)-(O+2NO) layer, one NO is roughly identical, and the second sits on the fcc site with slightly changed parameters compared to v 1 ( r e=1.22 Å, z e=1.39 Å). We discuss conclusions from this systematic series of structures for the reliability of geometry determinations by quantitative LEED, and for the chemistry of these layers. Using Badger's rule we can estimate the internal bond orders of the NO molecules, 2.0–2.2 for the v 1(O)-NO and 2.5–2.7 for the v 2(O)-NO species, which are in line with expectations from a simple frontier orbital argument.

The influence of electronegative coadsorbates on the geometry of benzene on Ru(001)

The geometrical structures of the Ru(001)-p(3 × 3)-C 6D 6 + 2O and Ru(001)-p(3 × 3)-C 6D 6 + 2NO coadsorbate layers have been determined by a detailed LEED IV analysis. The benzene molecule as well as the coadsorbates are found to be bound on hcp sites. The most conspicuous result is that the molecular symmetry is different for the two coadsorption layers. In the O coadsorbate layer it is essentially C 3v( σ v), as for the pure benzene layer, while NO coadsorption induces a rotation by ∼20° towards C 3v( σ d). In addition, the benzene molecule in the NO mixed layer has a distinct triangular distortion towards Kekulé benzene. These changes are explained by spatial restrictions induced by the repulsive interaction between the benzene and the NO molecules, and by the electronic interactions between the benzene and the electronegative coadsorbate. Both coadsorbates occupy two inequivalent hcp sites, which are discriminated by different local environments with respect to the benzene molecule. We suggest that this leads to anisotropic indirect (through-substrate) interactions between the benzene molecule and the respective coadsorbate. In addition, there may also be direct attractive interactions between the coadsorbates.

Quantum-resolved angular distributions of neutral products in electron-stimulated processes: NO desorption from and NO 2 dissociation on Pt(111)

We present the first quantum-resolved angular distributions of ground-state neutral molecules which are products of electron stimulated desorption (ESD) and electron stimulated dissociation. Laser resonance-enhanced multiphoton ionization (REMPI) and two-dimensional imaging have been used to obtain angular distributions of NO desorbed by 350 eV electrons from O-precovered Pt(111). In a similar fashion, we have measured angular distributions for the NO product of NO 2 dissociation on clean and O-precovered Pt(111). In all cases, we observe narrow widths which are roughly the same as ion distributions determined by ESDIAD (ESD ion angular distributions). The angular distribution for NO ESD is sharply peaked (7° half-width at half maximum) along the surface normal for an O coverage (θ o) of 0.25 monolayer (ML). The angular distribution of the NO product from dissociation of side-bonded NO 2 on clean Pt(111) is unexpectedly peaked about the surface normal, and thus does not reflect dissociative forces parallel to the surface or the ∼ 25° off-normal ground-state bond direction. On O-precovered Pt(111), where NO 2 is N-bonded, ∼ 6° off-normal beams are observed. When the substrate is precovered with θ o > 0.5 ML, local disorder creates asymmetric site geometries which result in multiple peaked angular distributions with both normal and off-normal (∼9–10°) components; similar effects for NO ESD are observed. In all these studies, the NO angular distributions are invariant to rotational or vibrational state. This implies that the lateral translational degrees of freedom are essentially de-coupled from the internal modes of the molecule. The results are discussed in terms of desorption mechanisms, dissociative forces, site geometries, and disordered coadsorbate layers.

Coadsorption of CO and H on Ni(110): evidence for a strong local COH interaction

We have investigated differently prepared coadsorbate layers of H and CO on Ni(110) by thermal desorption, low-energy electron diffraction, ultraviolet photoemission and inverse photoemission. In addition we have carried out cluster calculations in order to model the electronic structure of the mixed COH layer on this surface and to investigate effects of H coadsorption on bond lengths and the corresponding stretching vibrations. We find evidence for a strong local COH interaction. No global H-induced effects, such as d-band filling, need to be invoked for the interpretation of the photoemission spectra. Moreover, neither changes in the long-range order of the coadsorbate layer nor a surface reconstruction seem to profoundly affect the character of the COH interaction. This supports the assumption of a local interaction mechanism.

X-ray photoelectron diffraction: a sensitive probe of adsorbate and surface structure

Recent experiments are discussed in which forward-scattering intensity enhancements observed in X-ray photoelectron diffraction (or in Auger electron diffraction) yield structural information about surfaces. Two types of surface structural problems are highlighted which are particularly suitable for studies using this technique: (1) adsorbate and coadsorbate layers on surfaces, and (2) thin epitaxial metal films in the early stages of epitaxy. In the former one can determine molecular orientations to accuracies approaching ± 1° and also gain information on the molecular vibrations (‘wagging’ modes arising from frustrated molecular motion) in the adsorbate layer. The transition observed on Ni(110) from perpendicular CO is tilted CO with increasing coverage, induced by adsorbate ‘crowding’, provides a good example of the technique's capabilities in this regard. In epitaxial films the growth mode in the early stages (up to several layers) can be determined, e.g. layer-by-layer growth can be distinguished from clustering. The crystal structure and degree of elastic strain in such films can also be studied. Examples of such studies include Co and Cu films on Ni(001).

Electronic structure and interactions in well-defined coadsorbate layers

A synopsis of results and models for weakly and strongly interacting, well-defined coadsorbate systems is presented. Some experimental examples are given for various adsorbates on Ni(111) representing weak (different N2 states, Ar/Xe+N2) as well as strong (N2+K) interaction. Models for coadsorbate interaction, in particular between molecules (CO, N2) and alkali atoms, are discussed.