AVS1996 Session SS1-TuM: Structure of Surfaces with Adsorbates

Tuesday, October 15, 1996 8:20 AM in Room 204C

Tuesday Morning

Start	Invited?	Item
8:20 AM		SS1-TuM-1 Leed Intensity Analysis of ZnSe(100)-(2x2): Vacancy versus Rehybridized Structures C. Duke, A. Lazarides, A. Paton (Xerox Wilson Center for Research and Technology); A. Kahn (Princeton University) Low energy electron diffraction intensity analysis of the (100) surface of ZnSe has revealed the occurrence of slightly relaxed c(2x2) vacancy structures, either Zn or Se. Simplex searches of structures with relaxed top layer and second layer atoms yielded optimal structures. Contrary to expectations, these structures are not those predicted by various energy minimization calculations.\super 1\,\super 2\ In the case of the Zn vacancy structure we find a 10% contraction of the outer Zn layer relative to the Se layer below and a 5% contraction of the lateral Se-Se distance in the uppermost Se layer. For the Se vacancy structure, the outer layer contraction of Se relative to the next layer Zn is 1% and the lateral Zn-Zn distance is contracted by 2%. Local optimization of structures with nearly coplanar Zn and Se surface atoms similar to the c(2x2) structures predicted by energy minimization yielded structures with R-factors 40% larger than those of the globally optimized structure. The model description of the measured LEED intensities seem quite satisfactory as compared with other semiconductor surface structure analyses. Hence, the LEED results indicate the occurrence of vacancy structures which are quite different from the rehybridized versions predicted by energy minimization calculations, perhaps because of a high kinetic barrier to formation of the latter. 1. C.H. Park and D.J. Chadi, Phys. Rev. B 49(23), 16467 (1994) 2. A. Garcia and J.E. Northrup, J. Vac. Sci. Technol. B 12(4), 2678 (1994)
8:40 AM		SS1-TuM-2 Bulk and Surface Photoelectron Diffraction and Holography of W(110) J. Denlinger, B. Tonner (University of Wisconsin, Milwaukee); E. Rotenberg, S. Kevan (University of Oregon); P. Len, C. Fadley (University of California, Davis) We will discuss the current status and future prospects of photoelectron holography, a recently proposed method by which atomic images can be directly reconstructed from photoelectron diffraction data sets. Four current algorithms for obtaining atomic images will be compared and applied to both theoretically simulated single- and multiple-scattering data sets, as well as a large experimental data set for bulk and surface 4f core-level emission from W(110) recently obtained at the Advanced Light Source. The effects of data sampling density on these images will also be discussed.
9:00 AM		SS1-TuM-3 Photoelectron Holographic Study of Atomic Structure of Ge Adsorbed on the Si(100) Surface* J. Wu, M. Keeffe, C. Chen, G. Lapeyre (Montana State University) The experimentally determined image shows that Ge atoms replace Si atoms in the first surface layer rather than attach on the Si surface. The atomic image is obtained by the small-cone photoelectron holographic imaging (PHI) technique [1]. The sa mple is made by half monolayer deposition of Ge on a single-domain Si(100) 2x1 surface (4\super 0\ vicinal cut) held at 450\super 0\ C. The LEED pattern remains a single 2x1 domain with a slightly higher background. Angle-resolved, energy-dependent-photoe lectron spectra for the Ge 3d core-level emission were measured on an angular grid for one-fourth of the emission hemisphere, an irreducible symmetry element. The holographic inversion gives a real space image containing four strong spots which are in a p lane 1.2 \Ao\ below the emitter plane (coordinates within the plane are (1.8 \Ao\ in the perpendicular direction to the step edge and (0.5 \Ao\ in the parallel direction). Evaluation of the experimentally determined image leads to the model of Ge dimers in the Si surface. Because the two atoms in a dimer are in non-equivalent emitters, the diffraction from each inverts to two images which have the same origin. The reduction of the double image will be described in the presentation. Taking the second-laye r Si atom separation to be the bulk value, the observed dimer length is 2.8 \Ao\ and the Ge-Si bond length is 2.2 \Ao\. Because of phase shift considerations, the atom positions in the image could be off by several tenth \Ao\. These model independent val ue could be improved by trial and error model calculations. *Supported by NSF/DMR 1. H. Wu, G.J. Lapeyre, H. Huang and S.Y. Tong, Phys. Rev. Lett .71, 251 (1993); H. Wu and G. J. Lapeyre, Phys. Rev. B51, 14549 (1995).
9:20 AM		SS1-TuM-4 Atomic Imaging with X-ray Holography P. Len (University of California, Davis); T. Gog (Oak Ridge National Laboratory and Brookhaven National Laboratory); C. Fadley (University of California, Davis); G. Materlik (HASYLAB and DESY, Germany) X-ray Fluorescence Holography (XFH)\super 1\ and Multi-Energy X-ray Holography (MEXH)\super 2\ have recently been demonstrated for the first time experimentally to be capable of producing three-dimensional atomic images. At a single photon energy, these methods are equivalent by virtue of the optical reciprocity theorem. However, in XFH scattering and diffraction of the outgoing fluorescent x-ray is used to produce the hologram, whereas in MEXH diffraction of the ingoing exciting x-ray generates the hologram, with the flourescent radiation being used simply to detect the excitng field strength at a given emitter. MEXH is thus more flexible in that holographic information can be recorded at many different energies, with the potential to suppress real-twin overlaps, aberrations, and artifacts in reconstructed atomic images. With variable polarization of the incident radiation, MEXH can also emphasize or de-emphasize different atomic planes in the structure. The results obtained to date and the future prospects for these two methods will be discussed, with particular emphasis on imaging simple Fe bcc and complex Fe\sub 2\O\sub 3\ bulk crystals, Si/Ge/Si(001) \delta\-layers, and more surface-specific structures.\super 1\M. Tegze and G. Faigel, Nature 380, 49 (1996), plus accompanying comments by C. S. Fadley and P. M. Len, Nature 380, 27 (1996). \super 2\T. Gog, P. M. Len, D. Bahr, G. Materlik, C. S. Fadley, and C. Sanchez-Hanke, Phys. Rev. Lett. 76, 3132 (1996).
9:40 AM	Invited	SS1-TuM-5 Structural Trends in Cu(110) Induced by Alkali Adsorption I. Robinson, R. Schuster (University of Illinois, Urbana) X-ray Diffraction is, by now, a well-established way to study the atomic structure of surfaces with a precision approaching 0.01\Ao\. In recent years the method has been extended to full three-dimensional structure determination, and so chemical effects, revealed in interatomic distances, can now be explored too. Alkali metals are strongly electropositive and therefore can induce significant changes in the structure of metal surfaces, both in the favored choice of atomic arrangement, and in the details of the optimum bonding distances. In this talk we will present structural data for Cs/Cu(110) that illustrates both kinds of induced change. The clean surface is unreconstructed, but attains a series of missing-row structures upon dosing with Cs. We identified three such structures and found them to have a greater number of missing Cu rows with increased Cs coverage. The nearest-neighbor distances around the top layer atoms were also increasingly compressed with coverage. These large changes in the first coordination shell cannot be explained by simple theories of the embedded-atom or effective-medium type, and require an explanation that includes the effect of the charge transfer from the alkali atom.
10:20 AM		SS1-TuM-7 Photoelectron Diffraction Study of Mn/Cu(001) C(2x2) J. Denlinger, E. Rotenberg, S. Kellar (Lawrence Berkeley National Laboratory); S. Kevan (University of Oregon); B. Tonner (University of Wisconsin, Milwaukee) The room temperature c(2x2) Mn/Cu(001) surface alloy is studied with X-ray Photoelectron Spectroscopy and Diffraction. Energy dependent photoelectron diffraction is used to confirm the large buckling of surface Mn atoms, originally proposed by an earlier LEED study [1]. Structural results from multiple-scattering simulations will be presented. In addition, weak forward scattering for high kinetic energy Mn 2p photoelectron diffraction indicates the presence of a small amount of subsurface Mn atomic sites for 0.5 ML deposition. Also, annealing of the c(2x2) surface alloy reveals the presence of at least two components in the lineshapes of Mn 2p, Mn 3p and Mn L-edge NEXAFS. These results will be compared to the Mn/Ni surface alloy [2]. [1] M. Wuttig, Y. Gauthier and S. Bl\um u\gel, Phys. Rev. Lett 70, 3619 (1993). [2] S. Banarjee, et. al, these proceedings.
10:40 AM		SS1-TuM-8 Structure Analysis of Ge Dimer on Si(001) by MEIS Blocking Profiles from Embedded Ge Layers K. Sumitomo (NTT Basic Research Labs, Japan); T. Nishioka (NTT Opt-electronics Labs, Japan); T. Ogino (NTT Basic Research Labs, Japan) We propose a novel technique to determine the surface structure quan titatively with higher accuracy by modification of the substrate, and apply it to the structure analysis of Ge dimer on Si(001) surface. Medium energy ion scattering (MEIS) is one of the most powerful methods to determine the a tomic displacement on sub-angstrom accuracy. However, in the case of recons tructed surface, complicated blocking profiles make the structure analysis d ifficult, because of existence of too many types of scattering-blocking pairs. In this paper, we describe the analysis of scattered ions from embed ded Ge layers. One or two atomic layers of Ge were embedded below 4 - 12 ML of Si overlayer by MBE, with atomic scale layer precision. The ions scatter ed from embedded Ge layers are blocked by the reconstructed surface atoms. Since the origin of the signals are restricted, the scattering-blocking pair s are uniquely assigned. In addition, the effect of thermal vibration becom es small because the distance between scattering and blocking atom is suitab le for analysis. Therefore, we observed sharp and unmixed blocking dips in the profile of embedded Ge signals, which directly reflected dimer structure s on the reconstructed surface. It is determined that the dimer is asymmet ric and Ge-Ge bond length is 2.4 angstrom. We demonstrate that this method can determine the reconstructed surf ace structure in a more accurate way.
11:00 AM		SS1-TuM-9 High Resolution XPS Studies of Ni(111)-2-butyne C. Hirschmugl, K. Schindler (Fritz-Haber-Institut der Max-Planck-Gesellschaft, Germany); G. Paolucci, S. Lizzit, F. Esch (Sincrotrone Trieste SCpA, Italy) The adsorption of 2-butyne on the Ni(111) surface has been investigated with high-resolution fast XPS using soft x-rays from the SuperESCA undulator beamline at Elettra. Simultaneous growth of two distinct C1s emission lines (FWHM approximately 0.3 and 0.5 eV) with a separation of 0.5 eV was observed during uptake measurements at 150 K. In addition, preliminary energy scan photoelectron diffraction measurements have been performed for the monolayer coverage. The two C1s emission lines exhibit dissimilar intensity modulations indicating different local geometries for the respective carbon atoms. Assignment of these emission lines to carbon atoms in the molecularly adsorbed 2-butyne will be discussed. Furthermore, the thermal decomposition of the monolayer has been monitored up to 800 K. The initial reaction, which proceeds at about 350 K with complex kinetics, leads to a broad unresolvable C1s emission feature (FWHM >1.5 eV). Subsequent decomposition above 500 K converts the broad feature into two well resolved lines (FWHM .7-.8 eV) with a 1.2 eV separation. High resolution fast XPS proved to be a powerful technique to reveal the evolution of species during the adsorption and thermal decomposition of a complex molecule at a metal surface.
11:20 AM		SS1-TuM-10 The Bonding Configuration and Phase Diagram of the H/Be(0001) System K. Pohl, P. Hofmann (University of Tennessee & Oak Ridge National Laboratory); S. Christensen (Aarhus University, Denmark); E. Plummer (University of Tennessee & Oak Ridge National Laboratory) The interaction of H with beryllium is one of the simplest chemisorption systems and so an ideal test case for first principle calculation; it may also prove to be technologically important for Be lined inner walls of fusion reactors. We have used LEED to observe the long range order, HREELS and high resolution core level spectroscopy to determine the local bonding sites for the H/Be(0001) system as a function of coverage and temperature. The absolute coverage was determined using Nuclear Reaction Analysis. In contrast to earlier studies, the adsorption of atomic H produces at least two superstructures: A (\sr\3x\sr\3)R30 phase associated with saturated adsorption of 1 monolayer (ML) at low temperature and a (1x3) structure with the annealed surface at 2/3 ML hydrogen coverage. Our quantitative structural determination by LEED I-V of the (\sr\3x\sr\3)R30 phase reveals a H induced reconstruction of Be surface in which 1/3 of the Be toplayer atoms are removed to form a honeycomb structure of vacancies each decorated with three H adatoms, bonded in tilted bridge sites, in agreement with first principles calculations by Stumpf and Feibelman [1]. Our LEED-IV analysis was highly sensitive to the lateral H adsorption position. The effects of H on the large surface core level shifts observed for clean Be(0001) [2] were investigated and correlated with the local bonding configuration observed with ELS. We will present a complete phase diagram for this system. [1] R. Stumpf and P. J. Feibelman, Phys. Rev. B 51, 13748 (1995) [2] L.I. Johannson, H.I.P. Johansson, J.N. Andersen, E. Lundgren, and R. Nyholm, Phys. Rev. Lett. 71, 2453 (1993)
11:40 AM		SS1-TuM-11 Bilayer Growth of In on Si(001) M. Evans, J. Glueckstein, J. Nogami (University of Wisconsin, Milwaukee) In grows as metal adatom dimers on Si(001) at room temperature. The densest possible packing of these dimers results in a 2 x 2 phase, which covers the entire surface at 0.5 ML. Other studies have shown that coverages above 3 ML lead to the formation of 3D islands. These studies have assumed that after the formation of a complete 2 x 2 layer at 0.5 ML, additional In would fill in the first layer creating a 1 ML density 2 x 1 surface structure similar to clean Si(001). We have found instead that the In above 0.5 ML forms a second layer on top of the 2 x 2 displaying a combination of "2 x" and "3 x" structure. We will discuss possible models for the second layer, and compare this behavior with other metals that form a 2 x 2 surface at 0.5 ML.