Library

Notes: From an analysis of the low-energy electron diffraction (LEED) intensities we have determined the oscillatory distortion of the topmost interlayer spacings of the clean and the (2×1)H covered Pd(110) surface as well as the exact adsorption geometry in this latter structure, which is formed at T〈180 K at a coverage of θH≤1.0. An R-factor analysis was used for quantitative comparison with the experimental data. The oscillatory distortion of the clean surface— d12=−5.1±1.5%, d23=+2.9±1.5%, d34 at its bulk value of 1.37 A(ring) (with R factors RP=0.22 , RZJ=0.14 )— is found to be reduced by the H adlayer to d12=−2.2±1.5%, d23=+2.9±1.5%, and d34=bulk value. The H atoms are adsorbed on quasi-threefold sites with equal distances of 2.0±0.1 A(ring) to the two nearest Pd neighbors in the topmost and the closest Pd atom in the second layer, leading to an effective radius of the H atom of rH=0.6±0.1 A(ring). The long-bridge adsorption site, (octahedral) subsurface sites, or a hydrogen induced reconstruction via a lateral displacement of topmost Pd atoms by more than ±0.1 A(ring) can clearly be ruled out. These structural data, which are in good agreement with those of ordered Had structures on Ni(110) and Fe(110), characterize the (2×1)H structure on Pd(110) as being a typical adsorbate structure. There is no indication of either direct occupation of distinct subsurface or of near surface adsorption sites in this structure, nor does it open up channels for surface penetration, e.g., by a strong distortion of the topmost Pd substrate layers. In contrast to Ni(110) and Rh(110) the island growth of the (2×1)H structure on Pd(110) indicates predominantly attractive, indirect adatom–adatom interaction which, however, are of distinctly different nature than those causing island formation in the Pd–hydride phase.

Notes: From an analysis of the low-energy electron diffraction (LEED) intensities we have determined the adsorption geometry of the two ordered H adlayers formed at T〈270 K on Fe (110): a (2×1) and a (3×1) structure, with ideal coverages of θ= 1/2 and θ= (2)/(3) . Calculations were performed for different adsorption sites and structural models, taking the Fe–H bond length and the first Fe–Fe interlayer spacing as variable parameters. An R factor analysis was used for quantitative comparison with the experimental data. In both structures the H atoms are adsorbed on highly coordinated (i.e., quasithreefold) sites: The R factors of only the superlattice beams (RZanazzi–Jona=0.26, RPendry=0.55 in the (2×1) and RZJ=0.4, RP=0.58 in the (3×1) structure) are significantly lower than those from models with a long bridge adsorption site (RZJ=0.37, RP=0.66 and RZJ=0.6, RP=0.74). The on top site and the short bridge site can clearly be ruled out. For both structures the minima occur at the same Fe–H interlayer spacing of 0.9±0.1 A(ring), equivalent to an Fe–H distance of 1.75±0.05 A(ring) or rH=0.47±0.05 A(ring). From the R factor minimum of all beams(RZJ=0.23, RP=0.46) the first Fe–Fe interlayer spacing is found to be equal to its bulk value, like on the clean surface. In the (2×1) structure the only possible arrangement of the Had atoms consists of dense packed rows in [001] direction which are separated by a row of unoccupied sites, respectively, due to a delocalization of the H atoms over two neighboring threefold sites, short-range fluctuations can be envisaged. Their influence upon I/V curves and relative intensities of different superlattice beams was analyzed. As a result this effect could be excluded, large domains are required, in which only one type of threefold sites is occupied. For the (3×1) structure a model is favored in which the lateral distribution of the adatoms differs from a previous suggestion. It is shown that this model is more plausible in view of the H–H interactions. The higher density of threefold sites also has implications for the discussion of the 2D phase diagram of H/Fe (110), especially on the requirement of trio interactions.

Notes: Combined UV and x-ray photoelectron spectroscopy, work function, and thermal desorption measurements are presented for the submonolayer to multilayer adsorption of xenon on a Ag(111) surface. The results compliment the detailed LEED studies of this system by Webb and co-workers. Successive layer-by-layer growth up to at least three Xe layers is seen from the variation of the XPS Xe (3d5/2) intensity as a function of exposure. The thermal desorption of xenon from the first and second layer is compatible with zeroth order desorption kinetics. This is suggestive of desorption from 2D solid Xe islands, in agreement with the first order phase transitions during population of the first and second layer observed by Webb et al. The adsorption energies of the first and second layer are estimated to be E1ad=5.2 and E2ad=4.3 kcal/mol, respectively. The dipole moment per xenon atom in the first layer has the coverage independent value of μ1(Xe)=0.2D. Particular emphasis is put on the discussion of the layer dependent photoemission behavior of xenon. The present work includes the first complete set of UPS and XPS electron binding energies for Xe on one and the same metal. Comparison of the layer values with those of other Xe/metal systems as well as their layer dependent shift provide twofold support for the notion that the electron binding energies (with respect to EF) of adsorbed xenon atoms "float'' with the local electrostatic surface potential of the respective adsorption site. This is further supported by some measurements for xenon adsorbed on a sputter-roughened Ag surface. Both the adsorption energy as well as the photoemission behavior of xenon are sensitive towards the surface topography, which in turn can be utilized for the characterization of surface defects induced by sputtering.

Notes: The interaction of oxygen with Al(111) was investigated by STM at temperatures between 350 and 530 K, by annealing an oxygen precovered surface and by adsorption of oxygen on the hot surface. For exposures up to 10 L and temperatures up to 470 K a considerable part of the oxygen exists still in the chemisorbed state, another part transforms into Al oxide. In contrast to 300 K chemisorbed Oad atoms are mobile at elevated temperatures, and compact, hexagonal (1×1)Oad islands develop by an ordinary nucleation and growth scheme. This evidences attractive interactions between the oxygen atoms on (1×1) sites. From the lateral distribution of Oad islands a diffusion barrier of 1.0–1.1 eV is derived. The imaging of the islands of the (1×1) phase by STM depends on their size, which is understood by a different imaging of the Oad/Al adsorbate complexes at the island borders. Defects in the islands and bright features at the edges are interpreted as nuclei of aluminum oxide. Additional features which appear as topographic holes may be attributed to nonconducting Al oxide grains. © 1998 American Institute of Physics.

Notes: The initial stages of oxide formation on atomically flat, monohydride terminated Si(111) surfaces by oxidation in 30% hydrogen peroxide solution (H2O2) were investigated by scanning tunneling microscopy and x-ray photoelectron spectroscopy. The reaction proceeds via homogeneous nucleation of small oxide clusters on the surface and subsequent lateral growth of these clusters within the surface bilayer. The oxidation of the topmost Si(111) bilayer in H2O2 solution is completed after 30 min, leading to a SiO1.2 average layer composition. For the next bilayer the oxidation rate decreases drastically—after 2 months only 60% of the second bilayer are oxidized. An inverse logarithmic rate of the second bilayer oxidation is consistent with a field assisted growth mechanism. The significant differences in the oxidation rates between the first and the second bilayer allow to produce well defined oxide layers of about 5 A(ring) thickness. © 1995 American Institute of Physics.

Notes: Roughness and general surface topography of polished Si wafers were systematically studied on an angstrom to nanometer scale by scanning tunneling microscopy (STM). Evaluation of a large number of STM images by using a simple classification scheme assures statistically relevant results without the disadvantages and loss of information connected with averaging the rms roughness or similar methods. For the flat parts of the surface, a rms roughness of 1.2–1.8 A(ring) is found, comparable to that derived from light scattering and diffraction measurements on similar surfaces. Significant amounts of the surfaces were found to exhibit more pronounced structures, contrasting results of other techniques. The reliability and applicability of STM measurements on technical surfaces, on an angstrom to nanometer scale, is discussed in the light of these results.

Notes: We have used variable temperature, ultrahigh vacuum scanning tunneling microscopy (STM), in both static and time-dependent experiments, to study the chemistry of the ethylene/Pt(111) system. Images of ethylene which exhibit long-range order have been obtained at a sample temperature of 160 K. The conversion of ethylene to ethylidyne has been observed directly in STM images. This conversion reaction is observed to occur in a "patchy'' manner across the surface at saturation coverage. As the reaction proceeds, well-defined islands of unreacted ethylene continue to be clearly observed. Further dehydrogenation of the ethylidyne formed from ethylene leads to carbon containing particles dispersed randomly across the sample. After annealing the ethylidyne covered sample to 500 K, the surface is uniformly covered with carbon containing particles which exhibit a bimodal distribution of heights (one and two atomic layers) consisting of an average of ten and twenty carbon atoms, respectively. Further annealing to 700 K results in the formation of larger particles which appear to be a single atomic layer high and 10–15 A(ring) in diameter in the STM images. These particles contain an average of thirty-four carbon atoms. Pt steps do not show any preferential reactivity for these reactions. Annealing the carbon particle covered surface to higher temperatures results in the formation of monolayer thick graphite islands, which eventually accumulate at the Pt steps.

Notes: The interaction of oxygen with Al(111) was studied by scanning tunneling microscopy (STM). Chemisorbed oxygen and surface oxides can be distinguished in STM images, where for moderate tunnel currents and independent of the bias voltage the former are imaged as depressions, while the latter appear as protrusions. An absolute coverage scale was established by counting O adatoms. The initial sticking coefficient is determined to so=0.005. Upon chemisorption at 300 K the O adlayer is characterized by randomly distributed, immobile, individual O adatoms and, for higher coverages, by small (1×1) O islands which consist of few adatoms only. From the random distribution of the thermalized O adatoms at low coverages a mobile atomic precursor species is concluded to exist, which results from an internal energy transfer during dissociative adsorption. These "hot adatoms'' "fly apart'' by at least 80 A(ring), before their excess energy is dissipated. A model is derived which explains the unusual island nucleation scheme by trapping of the hot adatoms at already thermalized oxygen atoms. Oxidation starts long before saturation of the (1×1) O adlayer, at coverages around aitch-thetaO(approximately-equal-to)0.2. For a wide coverage range bare and Oad covered surfaces coexist with the surface oxide phase. Upon further oxygen uptake both chemisorbed and oxide phase grow in coverage. Oxide nucleation takes place at the interface of Oad islands and bare surface, with a slight preference for nucleation at upper terrace step edges.Further oxide formation progresses by nucleation of additional oxide grains rather than by growth of existing ones, until the surface is filled up with a layer of small oxide particles of about 20 A(ring) in diameter. At very large exposures up to 5×105 L they cover the entire surface as a relatively smooth, amorphous layer of aluminum oxide. The difference in Al atom density between Al metal and surface oxide is accommodated by short range processes, with no indication for any long range Al mass transport. Based on our data we discuss a simpler two step model for the interaction of oxygen with Al(111), without making use of an additional subsurface oxygen species. The complex spectroscopic data for the O/Al(111) system are rationalized by the wide coexistence range of bare and Oad covered surface with surface oxide and by differences in the electronic and vibronic properties of the surface atoms depending on the number of neighboring O adatoms in the small Oad islands.

Notes: We report on the fabrication of nanometer-sized structures by scanning tunneling microscope induced decomposition of Mo(CO)6 precursor molecules on Si(111)-(7×7). Based on the analysis of characteristic properties of these structures in the region of the deposition threshold the mechanism leading to Mo(CO)6 fragmentation under these conditions and subsequent formation of nanostructures can be identified. This mechanism consists of an electron induced excitation of the molecule to an antibonding state with subsequent spontaneous separation of a CO ligand. © 2002 American Institute of Physics.

Notes: The evolution of the tantalum/diamond interface upon room-temperature Ta deposition on the (100) surface of a boron doped, synthetically grown diamond single crystal was monitored by Auger electron spectroscopy (AES), ionization loss spectroscopy (ILS), and electron energy loss spectroscopy (ELS). Characteristic loss peaks indicate carbide formation at the interface from very low coverages on, reflecting the strong interaction between tantalum and carbon. Thicker layers of TaC are formed during subsequent thermal annealing by diffusion of carbon into the tantalum film, at the same time the topmost diamond region is transformed into poorly ordered graphitic carbon. © 1996 American Institute of Physics.