Papers

AVS2000 Session SS3-WeA: Surface and Interface Structure I

Wednesday, October 4, 2000 2:00 PM in Room 210

Wednesday Afternoon

Start	Invited?	Item
2:00 PM		SS3-WeA-1 Adsorbate Structure Determination on Surfaces using Normal-Incidence X-ray Standing Waves D.P. Woodruff (University of Warwick, UK) X-ray standing waves (XSW) provide a particular simple way of obtaining quantitative information on adsorption geometries on well-characterised surfaces, and working a normal incidence (NIXSW) to the Bragg scatterer planes offers some specific advantages. In particular, under these conditions the experiment is extremely tolerant of crystal mosaicity and can be applied routinely to metal crystals. Normal incidence also implies lower photon energies, typically around 3 keV, making photoemission detection of the photoabsorption straight-forward. Photoemission detection offers two important advantages, namely easier access to low atomic-number species, and the possibility of obtaining chemical-state specific structural information through the use of 'chemical shift' in the photoelectron binding energies. On the other hand, photoemission detection requires the use of a modified interpretational scheme to take account of non-dipole effects in the angular dependence of the photoemission. These various aspects of the technique will be illustrated with recent results including: the direct measurement of non-dipole angular parameters; the use of chemical-shift NIXSW to determine the local geometries of coadsorbed molecular fragments (e.g. PF₃ species produced by X-ray fragmentation on Ni(111) and CH₃SH, CH₃S- and S on Cu(111) following surface reaction), and determinations of the temperature-dependent local structure of CO adsorbed on Ni(111).
2:40 PM		SS3-WeA-3 Oriented Quantum Dots By Buffer Layer Growth Process A.P. Baddorf, J.F. Wendelken (Oak Ridge National Laboratory); C.T. Venkataraman, T. Gog (Argonne National Laboratory) Quantum dots of Fe formed by a multi-step growth process are shown to be highly oriented on a Cu(100) substrate. Growth involved deposition of Fe on a buffer of 200 layers of condensed Xe. The Xe buffer layer, which promotes clustering, was subseque ntly removed by heating. In situ scanning tunneling microscopy results reveal formation of 3-D Fe clusters averaging 10 nm in diameter. An additional 20 nm thick cap of Cu was deposited over the Fe to allow ex situ x-ray studies at the Advanc ed Photon Source. Fe clusters were found to be highly oriented, with the close-packed [110] direction normal to the Cu(100) surface and with two in-plane orientations. The first in-plane structure has Fe \{-112\} planes aligned with Cu \{001\} and the se cond has Fe \{-110\} planes aligned with Cu \{001\}. These orientations are not observed in molecular beam epitaxy of Fe directly onto Cu(100) or in precipitate growth. Buffer layer growth bypasses the intermediate fcc Fe phase formed by direct deposition on fcc Cu and brings bcc Fe immediately into contact with fcc Cu. Sidestepping the fcc Fe phase may allow the newly observed orientations.¹ ¹ ¹ORNL is managed by UT-Battelle, LLC, under US DOE contract DE-AC05-00OR22725. The APS is supported by US DOE contract W-31-109-Eng-38.
3:00 PM		SS3-WeA-4 Direct Observations of Ordered Domain Structures and their Dynamics : Pb on Cu(111)¹ R. Plass, N.C. Bartelt, G.L. Kellogg (Sandia National Laboratories) There is considerable scientific interest in the spontaneous formation of two-dimensional periodic domain structures due to long-range interactions in two-phase systems. Theoretical investigations of the stability of periodic structures due to dipolar interactions predict a progression of droplet and striped phases as a function of area fraction² but experimental verification has been elusive. Using low energy electron microscopy (LEEM), we find that the growth of Pb on Cu(111) reproduces this domain evolution with surprising accuracy. Above 25C, Pb on Cu(111) follows the SK growth mode³ with a disordered surface alloy saturating at 0.4 ML Pb and an incommensurate (incom. hereafter) overlayer covering the surface at 1.0 ML Pb. As Pb deposition proceeds on the surface alloy held at 385C, the density of incom. structure droplets (about 90 nm in diameter) increases steadily. The droplets clearly repel each other, and pack together into a fairly well ordered lattice. They achieve maximum density near 0.3 area fraction (incom.) after which there is an abrupt transition to a striped phase characterized by lengthening bands of surface alloy and incom. regions. The stripes completely cover the surface at 0.5 area fraction, after which another abrupt transition occurs between the striped phase and a conjugate droplet phase where the surface alloy forms droplets in the incom. matrix. The conjugate droplet maximum density is near 0.7 area fraction. The domain structures' spatial dimensions, as well as the size of their thermal fluctuations are temperature sensitive. We use this sensitivity to probe the energetics responsible for the domain structures. ¹ ¹ Sandia is a multiprogram laboratory operated by Sandia Corporation, a Lockheed Martin company, for the U.S. Dept of Energy, contract #DE-AC04-94AL85000 ² K.-O. Ng and D. Vanderbilt, PRB 52 (95) 2177. ³ C. Nagl, et al., Surf. Sci. 321 (94) 237.
3:20 PM		SS3-WeA-5 Normal Incidence X-ray Standing Wave and Medium-energy Ion Scattering Investigation of the Structure of Ultra-thin Films on Cu(111) M.D. Crapper, M.T. Butterfield (Loughborough University, U.K.) Many of the techniques available for surface structural determination fall into two broad categories. The first of these is the family of diffraction techniques, where long-range order is required to allow investigation of a structure. In samples that are not completely ordered, a diffraction based technique will often preferentially reveal information about the regions of the sample that are ordered. The second category is the so-called local probes, which sample the environment around a particular species but do not reveal information about the long-range order. Two techniques that do not clearly fall into either category, but have aspects of each are medium energy ion scattering (MEIS) and the normal incidence X-ray standing wave method (NIXSW). Both of these methods can yield valuable information where imperfect order is present, but on a scale larger than merely local. We report the novel application of a combination of these methods to investigate the structure of ultra-thin overlayers of Co, Fe and Mn on Cu(111). In the case of Co on Cu(111), pure fcc growth does not continue after the first two layers, but there is no one single growth mode. Instead there is a gradual inclusion of stacking faults with evidence of both hcp and fcc twinning. The initial growth of Fe on Cu(111) is also fcc but with a transition of the entire film to bcc at around 5-8 ML rather than a gradual transition. Overlayers of Mn show no site coherence, even below one monolayer. Upon annealing, however, an alloy is formed to a depth of around five layers that has an effective fcc structure but with increased d-spacing over that found in Cu(111) and with Mn substituting some Cu sites.
3:40 PM		SS3-WeA-6 Low Energy Dynamics for S/Cu C.J. Hirschmugl (University of Wisconsin, Milwaukee); M.V. Pykhtin, S.P. Lewis (University of Georgia) Dissociated H₂S on Cu(100) at 300K has been studied by a combination of AES, LEED, and Far-IRAS in the 200-2200 cm^-1 frequency range using synchrotron radiation. Density functional theory (DFT) calculations for a p(2x2) ordered overlayer of S on Cu have also been completed. In the Far-IRAS experiments, diffusive scattering of substrate electrons from the adsorbates gives rise to a broadband infrared absorption. This behavior was induced by sulfur coverages up to and including the p(2x2) overlayer (0.25 monolayers (ML)). DR/R at high frequency (2200 cm^-1) changes monotonically with increasing coverage up to 0.125 of a monolayer, and then remains constant. For the low-coverage linear regime, we calculate a scattering cross section σ of sulfur for the substrate free electrons and the e-hole pair damping rate η for the S hindered translation. We apply expressions arising from a "friction" model proposed by Persson to determine these parameters from the broadband infrared absorption data. In addition, we determine both σ and η for the ordered overlayer from previously published inverse photoemission results and DFT calculations, using a complementary theory proposed by Persson. The scattering cross section for the disordered overlayer (below 0.125 ML coverage of S) is determined to be approximately 20 Å ² from the infrared results. The scattering cross section for the ordered overlayer is dramatically smaller, approximately 3 Å ² as determined from both the photoemission results and DFT calculations. The non-linear behavior in the background change for S adsorption is attributed to a changing cross section, or a disorder to order effect.
4:00 PM		SS3-WeA-7 Does Adsorbed Oxygen Change the Electron Density in Cu? E.F. McCullen, C.-L. Hsu, R.G. Tobin (Tufts University) We present new evidence that thin Cu films can have significantly lower conduction electron density n than pure bulk Cu, and that the resistivity increase caused by oxygen adsorption is due in part to a further reduction in n. This finding contradicts the prevailing model of adsorbate-induced resistance, which attributes the resistance increase to electron scattering. In a series of experiments we measured simultaneously the infrared reflectance and dc resistance changes of epitaxial Cu(100) films during oxygen dosing. Earlier experiments challenged the scattering model by showing that the reflectance-resistance change ratio is adsorbate-dependent.¹ Adsorbate-induced changes in n could account for the results, but only if n in the films was significantly below the bulk value. The present work shows that the films indeed have reduced electron density. We find that the reflectance-resistance change ratio for adsorbed oxygen depends upon the clean-film conductivity, which varies from sample to sample. This dependence can be understood if the variations in clean-film conductivity are due in part to variations in n. We use these data to set limits on the films' electron density. The results are consistent with a model in which the resistance change produced by oxygen is caused by equal contributions from scattering and reductions in n, and in which each oxygen atom localizes about two conduction electrons. ¹ ¹ E.T. Krastev, D.E. Kuhl and R.G. Tobin, Surf. Sci. Lett. 387, L1051 (1997); C.-L. Hsu, E.F. McCullen and R.G. Tobin, Chem. Phys. Lett. 316, 336 (2000).
4:20 PM		SS3-WeA-8 Soft X-ray Photoelectron Spectroscopy Studies of Faceting and Alloying for Ultra Thin Films of Ruthenium on W(111) and W(211) G.J. Jackson (Rutgers University); J.E. Rowe (North Carolina State University); T.E. Madey (Rutgers University) High resolution soft X-ray photoelectron spectroscopy (using synchrotron radiation) and low energy electron diffraction (LEED), have been used to study alloying and faceting of Ru dosed onto W single crystal surfaces. W 4f core-level photoemission spectra and valence band spectra have been measured at various photon energies for W(111) and W(211), this photoemission being collected along the surface normal and at shallow grazing angles. The ultrathin film growth and evolution before and after annealing, on both W surfaces, has been investigated for coverages ranging from 0 to greater than 3 physical monolayers. After annealing multilayers of Ru on W(211) at low temperature (600 K), core-level shifts in the W 4f photoemission indicate formation of a Ru/W alloy that may be subsurface. Upon annealing to higher temperatures (600-1600 K), the core-level shifts reveal a concentrated Ru/W alloy. For the W(111) surface, previous studies from several other 4d and 5d dosed metals have shown that the ultrathin film covered surface forms nanoscale pyramidal facets with (211) faces. However, our LEED observations provide no evidence that Ru induces faceting of W(111). W 4f core-level shifts and their relative photoelectron intensities have indicated that clustering may occur on the W(111) surface after annealing. The results from studies of the Ru investigations are closely compared with previous studies from Pt,Pd,Ir,Rh dosed onto W(111) and W(211). In these cases it has been shown that fractional monolayer coverages do not form alloys and that for coverages exceeding 1 ML, the W atoms can dissolve into the adsorbed layer. Our results indicate some differences in the thermal stability of Ru/W, compared with Pt,Pd,Ir, Rh on W.
4:40 PM		SS3-WeA-9 Oxygen Induced Faceting of Ir(210) I. Ermanoski, K. Pelhos, T.E. Madey (Rutgers, The State University of New Jersey) As a part of a larger program to study the morphological stability of adsorbate covered metallic surfaces, we have investigated the adsorption of oxygen on fcc Ir(210) and the oxygen induced faceting of Ir(210). The techniques we used include low energy electron diffraction (LEED), temperature programmed desorption (TPD), Auger electron spectroscopy (AES) and scanning tunneling microscopy (STM). The atomically rough Ir(210) surface, when exposed to more than ~0.9L of oxygen and annealed to temperatures higher than 600K, experiences significant morphological restructuring: pyramid-like structures (facets) are formed on the initially planar surface. Our high temperature LEED measurements show that these pyramidal facets exhibit a quasi-reversible behavior upon annealing to higher temperatures. The surface reverts to its planar state at temperatures above 850K but, provided the maximum annealing temperature is below the desorption temperature of oxygen, facets reappear upon cooling to temperatures below 800K. LEED measurements show that these facets have a different structure than the original ones. Furthermore, we are able to remove the oxygen from the surface via catalytic oxidation of CO at 480K, while preserving the faceted structure. TPD and AES have shown that residual adsorbed oxygen and CO are negligible after this procedure. The faceted clean surface is stable up to 600K, but irreversibly reverts to the planar state when annealed above 600K. These experiments indicate that the clean, faceted, metastable Ir(210) surface provides an ideal substrate to study thermal relaxation of nanometer-scale surface features.
5:00 PM		SS3-WeA-10 Evolution of Ni(110) Surface with Low Energy Ion Sputtering S.-J. Kahng, B.-Y. Choi, J.-Y. Park, Y. Kuk (Seoul National University, Korea) We have studied the spatio-temporal evolution of the Ni(110) surface structure induced by the low energy ion sputtering with scanning tunneling microscopy. In order to have better understanding of the sputtering kinetics, we performed sputtering and homoepitaxial growth experiments at 295 K. In the early stage of sputtering, erosion pits and adatom islands, elongated along the [110], are present with asymmetric adatom kinetics. They become seeds for the ripple structure under the strong diffusion bias along the [001] direction. The surface roughness increases linearly with sputtering time. In the late stage of sputtering, however, symmetric mounds replace the ripple structure and the surface roughness saturates to certain point, showing good agreement with the recent numerical simulation on the basis of nonlinear Kuramoto-Sivashinsky equation.