Papers

AVS2001 Session SS2-MoM: Metal Clusters

Monday, October 29, 2001 9:40 AM in Room 121

Monday Morning

Start	Invited?	Item
9:40 AM		SS2-MoM-1 Mesoscopic Strain-Induced Magic Fe Nanostructures on Cu(100) N. Lin, A. Dmitriev (Max-Planck-Institut for Solid State Physics, Stuttgart, Germany); V.S. Stepanyuk, D.I. Bazhanov (Martin-Luther-University, Germany); J. Weckesser (Max-Planck-Institut for Solid State Physics, Stuttgart, Germany); J.V. Barth (EPFL, Switzerland); K. Kern (Max-Planck-Institut for Solid State Physics, Stuttgart, Germany and EPFL, Switzerland) We report a novel phenomenon in metal-on-metal heteroepitaxy, expressed in the arrangement of Fe islands grown at low temperature (<200K) on a Cu(100) surface. Scanning tunneling microscopy observations reveal that Fe atoms aggregate as perfectly oriented well-defined nanometer-scale arrays when Fe-Cu exchanging is inhibited. The basic unit of these nanoarrays is a tetramer of Fe atoms. More than 60 percent of the nanoarrays consist of four tetramers arranged in a square shape, aligning exactly to the high symmety directions of the Cu(100). The sharp distribution and distinct shapes of the nanoarrys can be understood in the framework of mesoscopic strain. STM and first principle calculations demonstrate that both the Fe isalnds and Cu substrate experience a large relaxation due to the mesoscopic strain: the Fe interatomic distance in the tetramers (2.0 Å) is 20% less than that in thin fcc Fe/Cu(100) films (2.5 Å), and the Cu substrate underneath the Fe nanoarrays is compressed 16.4% relative to the unrelaxed Cu surface (2.56 Å) . The giant compression and the high stress energy promote the formation of super-compressed Fe tetramers and their aggregation as magic nanoarrays with regular shape and orientation.
10:00 AM		SS2-MoM-2 Nanostructed Growth of Metal Clusters on Ultrathin Al₂O₃ Films on Ni₃Al(111) A. Rosenhahn (Lawrence Berkeley National Laboratory); A. Wiltner (Universitaet Bonn, Germany); K. von Bergmann (Universitaet Hamburg, Germany); J. Schneider (Lawrence Berkeley National Laboratory); C. Becker (Universitaet Bonn, Germany); B.S. Mun (Lawrence Berkeley National Laboratory); F.J.G. de Abajo (Centro Mixto CSIC-UPV/EHU); M.A. van Hove, C.S. Fadley (Lawrence Berkeley National Laboratory); K. Wandelt (Universitaet Bonn, Germany) Highly ordered, ultrathin alumina films can be grown on Ni₃Al(111) by simply exposing the alloy surface to oxygen at high temperatures.¹ Because of the conductibility of the thin films, the properties of metal aggregates on the surface of these oxide films can easily be studied without charging of the surface. Using STM, we have investigated the nucleation properties of different types of metals reaching from very noble ones like Ag, Au, and Cu to metals which are more reactive towards oxygen like V and Mn. All clusters show preferred nucleation on distinct sites of the oxide film.² As long as the oxide is not entirely covered by metal, two different mean cluster distances of 2.5 nm and 4.6 nm can be found. The distribution of the clusters on the surface reveals a hexagonal symmetry. The same periodicity can be found on the metal-free oxide film, depending on the applied bias voltage.³ As these superstructures reveal the same symmetry and periodicity as the cluster distribution, the nucleation is clearly determined by these sites. Because the superstructures are just at certain bias voltages visible, the prefered nucleation is not simply due to a modulation of the morphology. We will discuss, in which way the structure of the oxide film and the nucleation of the metal clusters are correlated. ¹ A. Rosenhahn, J. Schneider, C. Becker, and K. Wandelt, J. Vac. Sci. Technol. A 18(4), 1923 (2000). ² C. Becker, K. v. Bergmann, A. Rosenhahn, J. Schneider, K. Wandelt, Surf. Sci. Lett., in Press. ³ A. Rosenhahn, J. Schneider, J. Kandler, C. Becker, and K. Wandelt, Surf. Sci. 433-435, 705 (1999).
10:20 AM		SS2-MoM-3 Nanocatalysis:Tuning Efficiency and Selectivity Atom-by-Atom U. Heiz (University of Ulm, Germany) The preparation of a collection of supported metallic particles that are truly monodisperse, i.e. they all have exactly the same size, has long considered to be virtually impossible. As a consequence it has been difficult to recognize size effects for small supported clusters, as size distributions have been broad. For clusters in the gas phase, however, strong size-dependent chemical properties have been discovered during the last two decades and the obtained results lead to new concepts for understanding their chemical properties. Connections of these observations with real catalysis has often been stressed and suggestions for tuning efficiency and selectivity of a certain catalytic process by simply changing cluster size were made already in early days. Efficient and selective conversion is indeed important in catalysis, as most catalytic surfaces assist a variety of reactions. It is therefore of interest to study the factors affecting the size-dependent selective and efficient behavior of catalytic systems. We succeeded to prepare model catalysts consisting of a collection of metal particles of a single size. In these experiments metal atoms and small metal clusters are formed in the gas phase, size-selected and then deposited on thin MgO and TiO2 films grown on metal surfaces. Various chemical reactions on the obtained cluster-assembled materials are then investigated under UHV conditions by means of thermal desorption and infrared spectroscopies. Oxidation and polymerisation reactions are strongly dependent on cluster size and on the cluster-support interaction, and not only the number of product molecules per cluster is changed, but also the branching ratio of certain reactions. In many cases the measured reactivities are different from the ones obtained for the corresponding bulk systems. By combining the obtained experimental results with ab-initio calculations, a picture of the size-dependent behavior of cluster-assembled materials is now emerging.
11:00 AM		SS2-MoM-5 Ion Beam Deposition of Size-selected Supported Metal Clusters M. Aizawa, S. Lee, S.L. Anderson (University of Utah) We describe an experimental setup that allows us to investigate physical and chemical properties of supported metal clusters on various substrates at low impact energies(<10eV/atom). The metal cluster ions are generated by a laser ablation source, mass selected, and deposited on substrates. The performance of this instrument is illustrated with supported metal clusters (V, Ni, and Ir) on titania and graphite. Electronic structure of the supported metal clusters and chemical reactions over them will be studied by XPS and TPD, respectively, in terms of cluster size, impact energy and substrate defects.
11:20 AM		SS2-MoM-6 Decay Characteristics of Surface Nanostructures: (100) vs (111) Surfaces¹ J.F. Wendelken (Oak Ridge National Laboratory); M. Li, B.-G. Liu, E.G. Wang (Chinese Academy of Sciences); Z. Zhang (Oak Ridge National Laboratory) The stability of nanostructures after their creation is a critical issue for nanotechnology. Here we study the fundamental mechanisms of atomic scale mass transport on surfaces with regard to the instability of surface nanostructures. Pyramidal mounds were created on Cu(100) and (111) surfaces at 297 K by molecular beam deposition of Cu atoms. The stability of the nanoscale mounds was then observed with a scanning tunneling microscope for up to 24 hours. Movies produced from sequential scans show that the mounds are unstable and the decay process is profoundly different for the (100) and (111) surfaces. Decay of the (100) mounds proceeds with removal of atoms from the base of the mounds and subsequent transport to the bottom of pyramidal holes with the result that the mound walls become steeper with time. In contrast, the (111) mound decay is characterized by loss of atoms on all terrace levels producing a constant average slope. The mechanism for the decay on both surfaces at 297 K involves the diffusion of islands or terraces by periphery diffusion² to an edge where a rapid decay³ or avalanche process may take place. Direct observation shows that this avalanche process is site selective on the (100) surface, but is not selective on the (111) surface. Kinetic Monte Carlo simulations⁴ at 400 K show that the observed behavior is a consequence of selective vs. non-selective edge diffusion barriers and does not depend on the mechanism by which atoms are delivered to the edge. ¹ Supported in part by the U.S. DOE through contract DE-AC05-00OR22725 with UT-Battelle, LLC, and in part by the NSF of China. ² W. W. Pai, A. K. Swan, Z. Zhang, and J. F Wendelken, Phys. Rev. Lett. 79, 3210 (1997). ³ M. Giesen, G. Schulze Icking-Konert, and H. Ibach, Phys. Rev. Lett. 80, 552 (1998). ⁴ Maozhi Li, J. F. Wendelken, Bang-Gui Liu, E. G. Wang, and Zhenyu Zhang, Phys. Rev. Lett. 86, 2345 (2001).
11:40 AM		SS2-MoM-7 Reactions on Free Platinum Clusters: Adsorption of Oxygen and Hydrogen and Formation of Water M. Andersson, A. Rosén (Chalmers University of Technology and Göteborg University, Sweden) We use a cluster beam experiment to investigate chemical reactions on the surface of small unsupported metal clusters. A pulsed beam of metal clusters is generated with a laser vaporization source, in which metal atoms are vaporized into a flow of helium gas and condense in small clusters. After expansion into vacuum, the cluster beam passes through two reaction cells. The cell pressure is varied over a range where the clusters make from less than one up to a few collisions with the reactive molecules. The clusters are detected with laser ionization and time-of-flight mass spectrometry. By measuring the abundance of pure clusters and reaction products as a function of reaction cell pressure, the reaction probability in a cluster-molecule collision can be determined. For platinum clusters with more than 6 atoms we measure stable reaction products with both oxygen and hydrogen. The reaction probability with oxygen is for most sizes between 0.2 and 0.3, and appears lower with hydrogen though difficult to quantify since the Pt isotope distribution limits the mass resolution. If we let the clusters collide with both hydrogen and oxygen molecules, the resulting mass spectrum deviates significantly from a co-adsorption spectrum where the respective contributions are added. Instead, if, for example, the clusters first react with oxygen and then hydrogen we observe fewer oxide products and more pure clusters compared without hydrogen collisions. The only reasonable explanation for this is that water molecules are formed on the clusters and desorb. The efficiency of the reaction is high on all cluster sizes measured (8-30 atoms), with only a weak dependence on cluster size.