A. Wander

An automated tensor LEED analysis of the Ni{111}-c(4×2)-2CO structure

Abstract The Ni{111}-c(4×2)-2CO system has been studied using fully dynamic low-energy electron diffraction. CO is found to occupy both hcp and fcc threefold hollow sites. This is in contrast to earlier conclusions from vibrational spectroscopy where assignment was made to bridge sites. This study also represents an improvement over earlier SEXAFS and photoelectron diffraction studies in that the full surface geometry is obtained revealing significant bucklings of the nickel first and second surface layers and points to possible bends and tilts of the CO molecules.

Breakdown of adsorbate site assignment from vibrational frequencies. NO on Ni(111) re-visited by tensor LEED

Abstract The c(4×2)NO-Ni(111) system has been studied using fully dynamic low-energy electron diffraction. NO is found to occupy both hcp and fcc threefold hollow sites on the surface as proposed in an earlier SEXAFS study. This is in contrast to earlier conclusions from vibrational spectroscopy where assignment was made to the bridge sites, and to a recent X-ray photoelectron diffraction study which led to an assignment to all fcc like sites. Both of these structures can be ruled out on symmetry grounds alone. The current assignment underlines a dramatic failure of site assignment by vibrational spectroscopy, indicating that adsorbate site assignment for NO cannot be simply based on group frequencies obtained from organometallic compounds.

Ambiguities in adsorbate site assignment from vibrational frequencies. A TLEED structural study of (2 X 1)COPd(110)

Abstract The structure of the (2 × 1)COPd(110) surface phase has been determined by LEED intensity analysis. The CO molecule is found to be adsorbed in an atop site, tilted by 11° ± 4° with respect to the surface normal, with a CO bond length of 1.16 ± 0.04 A. Interestingly, the CO vibrational frequency for this system (2003 cm −1 ) is virtually identical to the frequency observed for the (2 × 1)CONi(110) surface phase (1998 cm −1 ) which a previous LEED study has shown involves bridge bound CO molecules. The result indicates that care must be taken in assigning site symmetries on the basis of CO stretching frequencies alone.

Surface crystallography of three catalytically important structures in the Au{111}-Pd system

Automated Tensor LEED has been used to determine the detailed crystallography of structures formed during Pd deposition on Au{111}. At 200 K, deposition of 0.2 monolayer Pd results in a complete lifting of the clean surface Au{111} reconstruction, with excess Au atoms being squeezed into a disordered adlayer with the added Pd adatoms. At monolayer coverage at this temperature a disordered structure is formed, which orders on heating to 300 K to a (1 × 1) structure with a predominantly Pd top layer. Further deposition of Pd produces a disordered overlayer. On annealing to ⩾ 500 K, this converts to an ordered alloy structure with (3×3)R30° periodicity. This structure consists of at least 2 layers of Pd2Au alloy. It is unique in showing a constant splitting of all Bragg beams, attributed to facetting into domains oriented only 0.75° from the planar Au{111} surface. This rumpled structure possesses domains which are larger than 100 A in width. The results are discussed in relation to the known catalytic activity for the various structures in the cyclisation of ethyne to benzene.

A tensor LEED structural study of the coverage-dependent bonding of iodine adsorbed on Rh{111}

Using automated tensor LEED (ATLEED), the structure of I on Rh{111} has been studied in the coverage range from 0.12 to 0.33 monolayers, including the ordered (3 × 3)R30° phase. Jodine adatoms are bonded in fcc three-fold hollow “lattice” sites with an iodine-rhodium interplanar separation of 2.22 ± 0.06 A and a first to second rhodium interlayer contraction of 3.7%. An iodine-to-rhodium nearest neighbour bond length of 2.71 A indicates an iodine “effective radius” of 1.37 A, very close to the covalent radius of 1.35 A assuming a rhodium metallic radius of 1.34 A. Adsorption in hcp hollow sites yielded a minimum Pendry R-factor of 0.40. The possibility of site mixing was examined and an R-factor minimum of 0.27 was obtained for 85 ± 15% fcc site occupation. The structure at lower coverages (0.25 and 0.12 ML), where only a diffuse p(1 × 1) LEED pattern is observed was examined by diffuse I–V analysis. Diffuse I–V spectra were found to be practically indistinguishable from spectra from the (3 × 3)R30° structure for identical parallel momentum transfer. An R-factor analysis indicates no coverage-dependent site switching and a coverage independent RhI bond length in the coverage range studied (0.12 ≤ θ ≤ 0.33 ML).

Islanding or random growth? The low coverage growth modes and structure of NO on Ni{111} studied by diffuse ATLEED

Abstract A novel technique, utilising DLEED intensity measurements, has been used to determine the growth modes of NO on Ni{111}. Random adsorption was found at low coverages, switching to clustering above 0.25 ML until the half monolayer c(4×2) phase was reached. AnI–V analysis of the 0.25 ML NO on Ni{111} phase has also been performed, using diffuse ATLEED. The optimum structure consists of 70% fcc and 30% hcp three-fold hollow site adsorption, with a large buckling found in the first nickel layer. From these results a new explanation of previous RAIRS spectra is given.

Structural transitions in ultra-thin nickel films on Rh{111}

Abstract For the first three monolayers, ultra-thin nickel films on Rh{111} grow in a close to layer-by-layer fashion. A nickel monolayer film is “pseudomorphic” with the underlying substrate. An automated tensor LEED I(V) analysis indicates the clean Rh{111} surface to be laterally unreconstructed with a small amplitude damped oscillatory relaxation of the outermost two atomic layer spacings of Δdz12 = −2.7 ± 1.4% and Δdz23 = + 0.4 ± 1.4%, resulting in a Pendry R-factor of 0.16. For the nickel adsorption site is found to be the fcc threefold hollow, with a nickel to rhodium interlayer spacing of 2.06 A, slightly shorter than the sum of metallic hard sphere radii, with the outermost rhodium interlayer spacing relaxed to Δdz12 = −1.4 ± 3.2%. The minimum Pendry R-factor of 0.28 was obtained. Spot profile analysis indicated significant beam broadening upon initiation of second layer growth with asymmetry at high parallel momentum transfer consistent with formation of small domains of incommensurate nickel with a nearest-neighbour separation intermediate between rhodium and nickel, finally yielding a complex pattern due to multiple scattering between the pseudomorphic p(1 × 1) slab and the incommensurate film for thicknesses beyond 1 ML. Model pair potential calculations indicate a layer dependent gradual relaxation of the nickelickel interlayer spacing towards that of bulk nickel in the outermost three atomic layers. Alloying occurs on an experimental time scale at temperatures above 600 K leading to formation of a substitutionally disordered RhNi alloy of variable composition.

Fitting dozens of coordinates by LEED: automated determination of complex surface structures

Abstract We present the first comprehensive discussion of automatic optimization procedures for surface-structure determination by LEED. These procedures combine numerical search algorithms with efficient methods of determining the diffracted intensities for varying structures. Such approaches can reduce the computer time required for an entire structure determination by many orders of magnitude, while fitting many times more unknown structural parameters. Thereby, relatively complex structures, with typically 10 adjustable atoms (or 30 adjustable coordinates), can be readily determined on todays workstations. We list over two dozen structures so determined, many as yet unpublished.

Buckled layer structure for atomic adsorption on W{100}: the (√2 × √2)R45° nitrogen structure from ATLEED

Adsorption of 0.5 monolayer of N adatoms on W{100} results in a sharp (√2 × √2)R45° LEED pattern. The only previous quantitative LEED study of this system gave a simple overlayer model with a Pendry R-factor of 0.55. An exhaustive search has been made of possible structures, including a novel vacancy reconstruction, displacive reconstructions and underlayer adsorption. From this work a new overlayer structure is derived with an Rp value of 0.22, displaying a considerable buckling of 0.27 ± 0.05 A within the second W layer and consequently involving large changes in the interlayer spacings of the surface. The N adatom is pseudo-five-fold coordinated to the W surface, bonding to a second-layer W atom with a nearest-neighbour bond length of 2.13 A and with the four next-nearest-neighbour W atoms in the surface plane at 2.27 A. The structure does not resolve the work function anomaly observed on this surface.

Surface science lettersAn adsorbate-stabilised vacancy structure for Cu on W{100}: a surface alloy

Deposition of 0.5 ML of Cu on W(100) leads to the formation of a sharp c(2×2) structure when the surface is annealed at 800 K. A LEED intensity analysis reveals that the Cu atoms are adsorbed displacively into W sites, forming an ordered 2D surface alloy. Due to the lattice mismatch between copper and tungsten, a substantial buckling of the first layer of the alloy is also observed. The clean, bulk terminated W(100) surface is only just stable relative to the c(2×2) vacancy covered W(100) surface. This relative stability of the vacancy structure explains the driving force behind the formation of this alloy.