AVS2001 Session SS3-FrM: Clean and Adsorbed Surfaces

Friday, November 2, 2001 8:20 AM in Room 122

Friday Morning

Start	Invited?	Item
8:20 AM		SS3-FrM-1 Structure and Electronic Properties of Planar and Faceted Ir(210) I. Ermanoski, M. Gladys, G.J. Jackson, T.E. Madey (Rutgers, The State University of New Jersey); J.E. Rowe (U.S. Army Research Office) The atomically rough Ir(210) surface is morphologically unstable: When Ir(210) is covered with more than 0.6 ML of oxygen and annealed, pyramidal facets develop on the initially planar surface. We have used a variety of methods to investigate the structure and electronic properties of planar and faceted Ir(210), including LEED, STM and high resolution soft X-ray photoelectron spectroscopy (SXPS) using synchrotron radiation. To prepare an oxygen-free faceted surface, we use catalytic CO oxidation at ~500 K to react the oxygen off the pre-prepared faceted surface. Cleanliness is verified both by AES and TPD. LEED and STM experiments show that the faceted surface is entirely covered with 3-sided pyramidal facets with dimensions of several nanometers. HRSXPS has been employed to investigate core-level features of all the surfaces mentioned. The Ir 4f7/2 core levels are fitted with Doniach-Sunjic lineshapes. Surface and bulk peak identifications are supported by measurements at different photon energies (different electron escape depths) and variable photoemission angles. All of the surface components (first, second and third layer peaks) are identified with core-level shifts positioned at higher binding energies with respect to the bulk. This result is in contrast to previous reports of binding energy inversion on Ir(100) and Ir(111) surfaces. The adsorption of oxygen onto the planar Ir(210) surface causes a suppression and shifting of the surface features. A comparison of planar and faceted surfaces reveals only minor differences in the SXPS core-level spectra. The effect of metallic overlayers on the morphological stability of Ir(210) will also be discussed. Supported by US DOE and ARO.
8:40 AM		SS3-FrM-2 Linking Stress to Surface Structure Using STM G.E. Thayer (UC Davis and Sandia National Labs); N.C. Bartelt, V. Ozolins (Sandia National Laboratories); A.K. Schmid (Lawrence Berkeley National Lab); S. Chiang (University of California, Davis); R.Q. Hwang (Sandia National Laboratories) Understanding the role of stress at solid surfaces is necessary to determine why surfaces have particular structures. While significant progress has been achieved in considering how local elastic interactions might contribute to the total surface energy a nd structure, it had not been possible to explicitly confirm this experimentally. The main reason for this is the difficulty in measuring stress fields on surfaces. Here, our approach has been to directly measure strain fields on a surface arising from l attice mismatch. Studying the phase diagram of CoAg/Ru(0001) single-monolayer films using STM, we found that annealed Co-rich films form an alloy with a structure that is not atomically mixed but instead consists of Ag droplets (15-30 atoms in si ze) with in a Co matrix. To quantitatively answer the question of how surface stress contributes to the formation of this structure, we have directly probed the stress fields on the surface. Analysis of atomically resolved images of the CoAg alloy has allowed us t o perform strain measurements over the surface. In our analysis we have compared strain measurements of about 800 Co atoms in a wide variety of Co-Ag neighborhoods with calculations of strain resulting from the Frenkel-Kontorova (FK) model and first-princ iples local spin density approximation (LSDA) calculations. The close agreement between the measurements and the calculations explicitly shows how stress due to lattice mismatch contributes to the formation of the droplet structure of the alloy. Within th e framework of the FK model we determine the relative forces acting on the surface by the measurement of strain fields and we are able to provide a direct link between surface stress and surface structure. We find the agreement even allows for the determination of details such as elastic spring constants from experimental measurements. Our success firmly demonstrates the possibility of using atomically resolved STM data to investigate surface stress.
9:00 AM		SS3-FrM-3 In-Situ STM Study of the Au(111) Herringbone Reconstruction Under Applied Stress O. Schaff (Fritz-Haber-Institut der Max-Planck-Gesellschaft, Germany); A.K. Schmid (Lawrence Berkeley National Laboratory); N.C. Bartelt, J. de la Figuera, R.Q. Hwang (Sandia National Laboratories) Surface stress is generally invoked as the driving force for the remarkably well-ordered and stable dislocation networks or reconstructions found in the surface layers of many epitaxial thin films, as well as in many metal single-crystal surfaces. To test this conjecture, the effect of uniaxial applied stress on dislocation networks present in the atomic surface layer of Au(111) was studied. The measurements were made using a novel instrument combining ultrahigh vacuum scanned-probe microscopy with an in-situ stress-strain testing machine. The technique provides microscopic information, up to atomic resolution, about the large-scale plasticity of surface layers under applied loads. The herringbone reconstruction of the Au(111) surface is a classic example of a strain stabilized dislocation network. We find that under 0.5% uniaxially applied compressive strain a dramatic restructuring of the network takes place. The three-fold orientational degeneracy of the system is removed and threading edge dislocations are annihilated. By considering the energetics of the herringbone reconstruction in the context of the Frenkel-Kontorova model, we are able to explain why these changes take place at this value of the applied strain.
9:20 AM		SS3-FrM-4 Core Level Shifts and Stress at the Ni/W(110) Interface^* R.T. Franckowiak (Utah State University); N.D. Shinn (Sandia National Laboratories); B. Kim, K.J. Kim, T.-H. Kang (Pohang Light Source, South Korea); D.M. Riffe (Utah State University) Stress can play an important role in determining the structure and stability of heterogeneous interfaces. A recent STM and strain study¹ of Ni overlayers on W(110) identified a pseudomorphic (1x1), and incommensurate (8x1) and (7x1) phases that exhibit compressive, tensile, and compressive stress, respectively. Interfacial electronic structure was suggested as the origin of the unexpected compressive stress in the (1x1) phase, rather than simple lattice mismatch in the adsorbed Ni film. This hypothesis was tested by measuring the W(4f) core-level binding energy shifts of interfacial W atoms during Ni overlayer growth. Photoemission spectra were obtained at the National Synchrotron Light Source, Brookhaven National Laboratory, using photon energies between 60 at 110 eV at a resolution of 150 meV. For the compressive pseudomorphic phase, interface W atoms exhibit a shift (compared to the bulk W) of -210 meV, whereas the 8x1 and 7x1 phases induce much smaller shifts of -90 and -120 meV, respectively. These shifts, which cannot be interpreted in terms of simple W-Ni coordination arguments, suggest that the interface tungsten electronic structure is very different in the pseudomorphic phase compared to the denser phases, and thus may be the source of the stress observed in the pseudomorphic phase. The similar W(4f) shifts in the 8x1 and 7x1 phases, which exhibit tensile and compressive stress, indicate that these stresses result from the Ni adlayer, as is expected from simple strain arguments based on the bulk Ni lattice constant. ¹D. Sander, C. Schmidthals, A. Enders, and J. Kirschner, Phys. Rev. B 57, 1406 (1998). ^*Supported by the DOE-BES Division of Materials Sciences. Sandia is a multiprogram laboratory operated by Sandia Corporation, a Lockheed Martin Company, for the United States Department of Energy under contract DE-AC04-94AL85000.
9:40 AM		SS3-FrM-5 Metal Nucleation and Adhesion on Ionic-Oxide Terraces* D.R. Jennison (Sandia National Laboratories); A. Bogicevic (Ford Research Laboratory); A.E. Mattsson (Sandia National Laboratories) We present two topics: First, it has recently been proposed¹ that oxide surfaces with a substantial number of oxygen vacancies may show greatly increased adhesion. It has also been observed that metal islands nucleate not only at line defects, but also on oxide terraces. Using first-principles DFT, we previously² studied the Pt dimer on MgO(100), and found that neither the O nor the Mg vacancy increased its stability, and in fact the dimers were only marginally stable at room temperature. We have now extended this study to include the dimers of eight different metals, both on the clean surface and at the O vacancy. We find that not all metals behave similarly! In fact, the dimers of Ag and Au are much more tightly bound on the clean surface than are those of Pd or Pt, and actually increase their stability at the defect. We understand these results based on dimer electronic structure. The second topic concerns the accuracy of DFT for studies of metal/oxide interfaces where the metal binds mostly by polarization, i.e. where there are no significant chemical bonds. This bonding is much stronger than van der Waals and it is measurable by analysis of the Wulff shapes of metal nanoparticles.³ In the cases of Pd and Cu on alumina films, DFT in the LDA predicts a work of adhesion close to experiment,³ while the GGA fails badly (and differs by about 50%). We have now understood this unusual failure of GGA in terms of the surface self-exchange error in DFT, as evidenced recently for another metal.⁴ In fact, the experimental result, vs. the computational numbers, supports our previous assertion⁵ that the interfacial binding of most metals on ionic oxides is mainly due to polarization. ¹Y. F. Zhukovskii, et al., PRL 84 (2000) 1256. ²A. Bogicevic and D. R. Jennison, Surf. Sci. 437 (1999) L741. ³K. H. Hansen, et al., PRL 83 (1999) 4120; T. Worren et al., Surf. Sci. 477 (2001) 8. ⁴K. Carling, et al., PRL 85 (2000) 3862. ⁵C. Verdozzi, et al., PRL 82 (1999) 4050; A. Bogicevic and D. R. Jennison, PRL 82 (1999) 799. *Sandia is a multiprogram laboratory operated by Sandia Corporation, a Lockheed Martin Company, for the United States Department of Energy under Contract DE-AC04-94AL85000; www.sandia.gov/surface_science/drj/.Sandia is a multiprogram laboratory operated by Sandia Corporation, a Lockheed Martin Company, for the United States Department of Energy under Contract DE-AC04-94AL85000; www.sandia.gov/surface_science/drj/.
10:00 AM		SS3-FrM-6 Structure Analysis of Oxygen-adsorbed and Annealed W(001) Surface at Liquid Nitrogen Temperature S. Kamimizu, K. Hara, T. Kokubun, T. Haga, K. Sakamoto, H. Yamazaki (The University of Electro-Communications, Japan) The structure of oxygen-adsorbed and annealed W (001) surface was investigated at the liquid nitrogen temperature. We have already reported the structure of this system at the room temperature (H. Yamazaki, et al., Surf. Sci. 447 (2001) 174). The result obtained previously is that the 2x1 LEED pattern was observed, and the most probable calculated structure consists of missing rows of W and double rows of oxygen atoms adsorbed on one of two three-hold hollow sites of W (011) facet which was appeared by losing the W-atoms. The Pendry R-factor for this structure is 0.25. As the energy difference between those two three-hold hollow sites is so small as 12 meV by a rough calculation, the adsorbed oxygen atoms should occupy those two sites by a certain ratio. It is considered that the oxygen atoms fall into the more stabilized adsorption site when the sample is cooled down to the liquid nitrogen temperature. Then the experiment was performed on the same sample at the liquid nitrogen temperature, and the most probable structure was determined with the Pendry R-factor of 0.20. The structure agrees well with that at the room temperature within an error. We will present the experimental and calculating procedure, and the most probable structure compared with that at the room temperature.
10:20 AM		SS3-FrM-7 The Local Adsorption Structure of CH₃- on Cu(111) M. Pascal, C.L.A. Lamont (University of Huddersfield, UK); J. Robinson, R.L. Toomes, J.-H. Kang, D.P. Woodruff (University of Warwick, UK); L. Constant, S. Bao, M. Kittel, J.-T. Hoeft, M. Polcik (Fritz-Haber-Institut der Max-Planck-Gesellschaft, Germany) Methyl is an important intermediate in many surface reactions, but there is a dearth of information on its local adsorption geometry. We present the results of both experimental and theoretical studies of the structure of CH₃- on Cu(111). The experiments were based on C 1s scanned-energy mode photoelectron diffraction (PhD) from a surface formed either by exposure to CH₃- ions from a Bent-type azomethane (CH₃-N-N-CH₃) cracking source, or from surface dissociation of methyl iodide (CH₃I) on the surface. The coadsorbed I , which occupies the fcc hollow sites (directly above a third layer Cu atom) appears to have little effect on the geometry of the methyl which also occupies an fcc hollow site but shows a marginally (0.04±0.03 Å) smaller layer spacing relative to the substrate in the presence of the coadsorbate. In the pure methyl layer the C-Cu outermost layer spacing is 1.66±0.02Å. Density-functional theory results (using the CASTEP code in GGA with ultrasoft pseudopotentials) reproduce the preference for hollow site adsorption (albeit with a very small preference for the hcp hollow) but also reveal a strong preference for an azimuthal orientation in which the C-H bonds are oriented along the Cu-Cu close packed directions, placing the H atoms closer to near-neighbour Cu atoms.
10:40 AM		SS3-FrM-8 Cu(100)c(2x2)-N: a New Type of Adsorbate-Induced Surface Reconstruction S.M. Driver (University of Warwick, UK); J.-T. Hoeft, M. Polcik, M. Kittel, R. Terborg (Fritz-Haber-Institut der Max-Planck-Gesellschaft, Germany); R.L. Toomes, J.-H. Kang, D.P. Woodruff (University of Warwick, UK) Using a combination of N 1s scanned-energy mode photoelectron diffraction and scanning tunnelling microscopy the Cu(100)c(2x2)-N surface is shown to undergo a symmetry-lowering reconstruction with a large amplitude (0.34 Å) periodic distortion of the outermost Cu layer perpendicular to the surface. This contrasts with the more usual surface layer density changes or parallel distortions which are the primary characteristic of other metal surface reconstructions. Atomic resolution imaging under varying tip conditions shows that in the c(2x2) phase STM appears to always image as asperities the Cu atoms, and not the N atoms, and that the N induces the unusual rumpling of the outermost Cu layer. This structural modification is quantified by the PhD data. Our STM images also provide a clear demonstration of the dangers of an over-simplistic interpretation of such data in terms of adsorbate atomic coordinates. The presence of the rumpling reconstruction, which we attribute to N-induced compressive surface stress, allows one to understand many detailed aspects of the mesoscopic c(2x2) island structures observed in this system (and reported in earlier STM investigations). Not only can the general self-organisation be attributed to the minimisation of the long-range elastic strain field energy, but similar elastic strain arguments arising from local rumpling can account for the N-N attraction needed for the island formation. In addition, the symmetry-lowering nature of the reconstruction provides a simple explanation for the systematic width variations of the inter-island boundaries.
11:00 AM		SS3-FrM-9 Phases of Oxygen on Cu(100) Imaged by Low Energy Electron Microscopy C.L.H. Devlin (Air Force Research Laboratory); Y. Sato, S. Chiang (University of California, Davis) Low energy electron microscopy (LEEM) was used to perform a detailed study of three phases of oxygen on the Cu(100) surface, including two new phases. These phases occurred when the sample was heated above 600°C. Dark field imaging was used to identify regions corresponding to particular low energy electron diffraction (LEED) patterns. The structure of the steps was also examined in the LEEM images of the different phases. At low coverage, the familiar (√2x2√2)R45° structure was observed. When annealed to 600°C, the steps changed from being gently curving to being highly bunched and completely straight with 90° kinks, resulting from a disorder-order transition. At higher oxygen coverage, <1.8ML, a new, complicated LEED pattern emerged, consisting of the (√2x2√2)R45° pattern plus a centered rectangular structure. Images of this surface included decoration of step edges and bright, sometimes hatched, areas on the terraces. Distinct bunching of steps, which curved gently over many microns, occurred. At higher coverage, <3.2ML, another new LEED pattern occurred. The "12-spot" hexagonal pattern corresponds to a hexagonal phase with two domains. The lattice constant of this structure agrees well with that for Cu₂O, suggesting that this phase corresponds to multilayers of cuprous oxide. LEEM imaging was used to follow structural transitions among the three phases.
11:20 AM		SS3-FrM-10 Temperature Dependence of Oxides on Titanium Surface in UHV Y. Mizuno (Chiba Institute of Technology, Japan); A. Tanaka (ULVAC-PHI, Inc.); Y. Takakuwa (Tohoku University, Japan); T. Momose (Miyagi National College of Technology, Japan); Y. Yamauchi (Chiba Institute of Technology, Japan); T. Homma (Chiba Institute of Technology) It is well known that titanium has a good ability to passivate a surface, and as a result to exhibit a high degree of immunity against attack by acids and chlorides, although the titanium surface is very active such as a gettering property. Titanium has stable oxides which are Ti₂O, TiO, Ti₂O₃, Ti₂O₅, Ti_nO_2n-1(42. The temperature dependence of titanium oxides was investigated in-situ with using Auger electron spectroscopy (AES) with quadrupole mass spectrometer (QMS) and X-ray photoelectron spectroscopy (XPS) at a temperature range RT-500 °C. The surface of specimen was prepared by electro-polishing, and on the surface thin titanium oxide layer was formed, although the oxide was mainly TiO₂. As the temperature of titanium surface increases, over 300 °C, O_KLL Auger intensity decreases in a drastic way, and in contrast Ti_LMM Auger intensity increases. Over 450 °C, no oxygen was detected on titanium surface in UHV. However, outgassing rates of O₂, CO, and CO₂ from the titanium surface did not show large changes in QMS measurements below 450 °C. This behavior is explained such as a gettering property, which is an enhancement diffusion of oxygen from a surface to bulk. In order to investigate temperature dependence on the chemical states of titanium, XPS analysis was performed in-situ at a temperature range between RT - 500 °C. The results showed drastic changes of titanium oxides which the suboxides began to decrease at 150 °C, even the amount of TiO₂ decreasing over 200 °C. Titanium has a very sensitive oxide surface depending on temperature in UHV.