IntroductionGold has long been regarded as an "inert" surface and bulk gold surfaces do not chemisorb many molecules easily. However, in the last decade, largely through the efforts of Masatake Haruta, gold particles, particularly those below 5 nm in size, have begun to garner attention for unique catalytic properties (1-8). In recent years, supported gold particles have been shown to be effective as catalysts for low temperature CO oxidation (9), selective oxidation of propene to propene oxide (10), water gas shift (11), NO reduction (12), selective hydrogenation of acetylene (or butadiene) (13)

Using spin-density functional theory we investigated various possible structures of the hematite (0001) surface. Depending on the ambient oxygen partial pressure, two geometries are found to be particularly stable under thermal equilibrium: one being terminated by iron and the other by oxygen. Both exhibit huge surface relaxations (−57% for the Fe-and −79% for the O-termination) with important consequences for the surface electronic and magnetic properties. With scanning tunneling microscopy we observe two different surface terminations coexisting on single crystalline α-Fe2O3 (0001) films, which were prepared in high oxygen pressures.PACS numbers:68.35. Bs,61.16.Ch,68.35.Md,71.15.Ap Although metal-oxide surfaces play a crucial role for several profitable processes, good quality experimental and theoretical studies of their atomic structure and electronic properties are scarce. For example, α-Fe 2 O 3 appears to be the active catalytic material for producing styrene, 1 which was substantiated by recent reactivity studies performed over single crystalline hematite model catalyst films. 2 Other candidate applications are photoelectrodes 3 and non-linear optics materials 4 . Nevertheless, the surface properties of α-Fe 2 O 3 are basically unknown, and also for other metal oxides an understanding is developed only badly. The reason is the difficult preparation of clean surfaces with defined structures and stoichiometries, which, as in the case of hematite, can require high oxygen pressures not suitable in standard ultrahigh vacuum systems. Furthermore, electron spectroscopy techniques and scanning tunneling microscopy (STM) are hampered by the insulating nature of the material. We also note that surface-science techniques often do not probe a thermal equilibrium geometry but a frozen-in metastable state. Theoretical studies, on the other hand, have to deal with 3d electrons, oxygen with very localized wave functions, a rather open structure, unusual hybridization of wave functions, huge atomic relaxations, big super cells, and magnetism. This renders an ab initio study of α-Fe 2 O 3 surfaces a most challenging investigation. Some theoretical studies of the geometry of α-Fe 2 O 3 (0001) had been performed using empirical (classical) potentials 5,6 , and Armelao et al. 7 studied the electronic structure employing a cluster approach. In this paper we report spin-density functional theory (SDFT) calculations for a slab geometry (see Fig. 1). We use the generalized gradient approximation (GGA) 8 for the exchange-correlation functional and the full-potential linearized augmented plane wave (FP-LAPW) method 9,10 to solve the Kohn-Sham equations. The STM study was performed on a thin hematite film grown epitaxially onto a Pt (111) substrate.The identification of thermal equilibrium structures of surfaces is a prerequisite for an understanding of the endurance, electronic, magnetic, and chemical properties of the material. Upper half of the slab for the unrelaxed O3-terminated surface. The cross section of the upper half ...

Thin SiO 2 films were grown on a Ru(0001) single crystal and studied by photoelectron spectroscopy, infrared spectroscopy and scanning probe microscopy. The experimental results in combination with density functional theory calculations provide compelling evidence for the formation of crystalline, double-layer sheet silica weakly bound to a metal substrate. DOI: 10.1103/PhysRevLett.105.146104 PACS numbers: 68.35.Àp, 68.47.Gh, 68.55.Àa Silicon dioxide (SiO 2 ) plays a key role in many modern technologies and applications that range from insulating layers in integrated circuits to supports for metal and oxide clusters in catalysts. For better understanding of structureproperty relationships on silica-based materials, particularly of reduced dimensions, thin silica films grown on metal single crystal substrates are suggested as suitable model systems that allow the facile application of many ''surface science'' techniques. It has recently been shown that crystalline silica films and nanowires can be grown on Mo(112) [1][2][3][4][5]. The ultrathin film consists of a monolayer honeycomblike network of corner-sharing [SiO 4 ] tetrahedra, thus resulting in a SiO 2:5 stoichiometry of the film. The Si atoms in these films can be partly substituted by Al in the course of preparing metal supported aluminosilicate films [6], which is the first step towards experimental modeling of catalytic centers in zeolitelike materials. However, attempts to grow thicker silica films on the Mo substrates resulted in amorphous structures [7][8][9], most likely due to the formation of strong Si-O-Mo bonds at the interface that govern the growth mode [9]. Recently, the preparation of crystalline silica films on other supports such as Pd(100) [10] and Ni(111) [11] has been reported. However, the atomic structure of the films, film surface termination, and the nature of the silica-metal interface were not determined.In this Letter, we report on the preparation and the atomic structure of well-defined silica films on Ru(0001). The experimental results, obtained by photoelectron and vibrational spectroscopies and high-resolution scanning probe microscopy, are complemented by density functional theory calculations which together provide compelling evidence for the formation of a double-layer sheet silicate, with a SiO 2 stoichiometric composition, weakly bound to a metal support. The results open new perspectives for employing a ''surface science'' approach to understand the reactivity of silicate surfaces consisting of hydrophobic Si-O-Si bonds, such as those of microporous all-silica zeolites [12]. Also, these films can be used as model supports for catalytically active metal and oxide clusters [4,13].The experiments were performed in an ultrahigh vacuum chamber equipped with low energy electron diffraction (LEED) and Auger electron spectroscopy, x-ray photoelectron spectroscopy (XPS), infrared reflection absorption spectroscopy (IRAS), and scanning tunneling microscopy (STM). Atomically resolved atomic force microscopy (AFM) and STM image...

Clear as glass: The atomic structure of a metal-supported vitreous thin silica film was resolved using low-temperature scanning tunneling microscopy (STM). Based on the STM image, a model was constructed and the atomic arrangement of the thin silica glass determined (see picture). The total pair correlation function of the structural model shows good agreement with diffraction experiments performed on vitreous silica.

The effect of the suppport on oxidative dehydrogenation activity for vanadia/ceria systems is examined for the oxidation of methanol to formaldehyde by use of well-defined VO(x)/CeO(2)(111) model catalysts. Temperature-programmed desorption at low vanadia loadings revealed reactivity at much lower temperature (370 K) as compared to pure ceria and vanadia on inert supports such as silica. Density functional theory is applied and the energies of hydrogenation and oxygen vacancy formation also predict an enhanced reactivity of the vanadia/ceria system. At the origin of this support effect is the ability of ceria to stabilize reduced states by accommodating electrons in localized f-states.

Monolayer gold islands with a thickness of one or a few monolayers on an FeO(111) substrate are found to exhibit identical CO adsorption behavior as large gold particles (see picture). This result suggests that particle thickness does not play a significant role in CO adsorption, and therefore size effects seen for the low‐temperature CO oxidation reaction are not related to quantum size effects.

Metallic nanoparticles finely dispersed over oxide supports have found use as heterogeneous catalysts in many industries including chemical manufacturing, energy-related applications and environmental remediation. The compositional and structural complexity of such nanosized systems offers many degrees of freedom for tuning their catalytic properties. However, fully rational design of heterogeneous catalysts based on an atomic-level understanding of surface processes remains an unattained goal in catalysis research. Researchers have used surface science methods and metal single crystals to explore elementary processes in heterogeneous catalysis. In this Account, we use more realistic materials that capture part of the complexity inherent to industrial catalysts. We assess the impacts on the overall catalytic performance of characteristics such as finite particle size, particle structure, particle chemical composition, flexibility of atoms in clusters, and metal-support interactions. To prepare these materials, we grew thin oxide films on metal single crystals under ultrahigh vacuum conditions and used these films as supports for metallic nanoparticles. We present four case studies on specifically designed materials with properties that expand our atomic-level understanding of surface chemistry. Specifically, we address (1) the effect of dopants in the oxide support on the growth of metal nanoclusters; (2) the effects of size and structural flexibility of metal clusters on the binding energy of gas-phase adsorbates and their catalytic activity; (3) the role of surface modifiers, such as carbon, on catalytic activity and selectivity; and (4) the structural and compositional changes of the active surface as a result of strong metal-support interaction. Using these examples, we demonstrate how studies of complex nanostructured materials can help revealing atomic processes at the solid-gas interface of heterogeneous catalysts. Among our findings is that doping of oxide materials opens promising routes to alter the morphology and electronic properties of supported metal particles and to induce the direct dissociation and reaction of molecules bound to the oxide surface. Also, the small size and atomic flexibility of metal clusters can have an important influence on gas adsorption and catalytic performance.

Hydrogenation of unsaturated hydrocarbons occurs efficiently on noble-metal catalysts, such as platinum, rhodium, and palladium.[1] The reaction mechanism first proposed by Horiuti and Polanyi [2] in 1934 proceeds by a) hydrogen dissociation on the metal surface, b) alkene adsorption, c) subsequent hydrogen addition to alkene and, finally, d) desorption of the product (alkane). Real hydrogenation catalysts represent very complex systems for studying reaction mechanisms at the molecular level. Therefore, model systems with a reduced complexity have been invoked ranging from single crystals to metal particles deposited on oxide films. [3][4][5][6][7][8] The conclusions regarding reaction mechanism and structural sensitivity are often based upon experiments on single metal crystals.[3] In particular, hydrogenation of alkenes has been shown to be structure insensitive.Herein, we report results showing that alkene hydrogenation reaction under low-pressure conditions, which does not occur on Pd(111) single crystal, proceeds efficiently on palladium nanoparticles. We show that the formation of weakly bound "subsurface" hydrogen is a key factor for hydrogenation to occur efficiently. The subsurface hydrogen exists in both Pd systems. However, the nanoparticle dimensions are such that this hydrogen is accessible to the adsorbed alkene, and hydrogenation occurs. In contrast, for crystals, the hydrogen atoms diffuse so deep into the bulk that they are not accessible to an adsorbed alkene, and therefore hydrogenation does not occur.We have studied the surface chemistry of ethene and different pentene isomers on both Pd(111) single crystal and Pd particles deposited on a thin alumina film (Figure 1). The particles studied are approximately 5 nm in diameter and consist primarily (% 90 %) of (111) facets [8] (% 10 % are (100) facets). The experiments were performed in ultrahigh vacuum on clean and well-defined systems. Using the temperatureprogrammed desorption (TPD) technique, we have observed that a number of hydrocarbon transformations, such as dehydrogenation and H-D exchange, occur on both palla-