Search

A brief survey is given on the current state-of-the-art of this surface structural technique based on photoelectron spectroscopy, with particular emphasis on the progress that has been made recently by routinely measuring full-hemispherical intensity distributions. We limit the discussion to the photoelectron forward focusing regime, which is attained at electron kinetic energies of a few hundred eV. Surface bond directions are directly revealed as pronounced maxima in the angular distributions from subsurface atoms, while characteristic interference features are observed for surface species. For both cases the dependence on the atomic type is weak enough so that these features provide a fingerprint of the local bonding geometry. For surface and near-surface species, this may then serve as a starting point for a structure refinement using single-scattering cluster calculations. Selected examples are given for illustrating these procedures.

The spatial distribution of quasi-elastically back-scattered electrons from single-crystalline YBa 2 Cu 3 O x (001), a high- T c superconductor ( T c = 91 K), has been measured at 2 keV energy with high angular resolution (1°). The data are presented as a full-hemispherical intensity map. It is characterized by a fourfold symmetry and resembles the Kikuchi pattern of a bcc crystal. These data are compared with single-scattering cluster simulations performed with structural models for the near-surface layers identical to those in the bulk and for different types of the crystal termination. Rather good agreement is found when the cluster is sufficiently large (more then 700 atoms) indicating that at room temperature the investigated near-surface layers have a local structure similar to the bulk one.

Natural diamond (001) and (111) surfaces as well as highly oriented pyrolithic graphite (HOPG) (0001) have been analyzed by X-ray induced photoelectron diffraction (XPD). The measured 2π patterns of C Is emission (964 and 1450 eV) and of KVV Auger emission (260 eV) are compared to single scattering cluster (SCC calculations, and excellent agreement is found. The comparisons show that photoelectron forward focusing is much less prominent in carbon solids than in all previously studied, heavier elements, and a direct interpretation of the data is therefore more difficult. The highly textured nature of HOPG presents itself as a circular pattern with concentric rings centered at the surface normal ( c -axis). The aim of this work is to show the advantage of the XPD technique as a diagnostic tool for the investigation of epitaxial diamond growth with monolayer sensitivity.

Electrons with kinetic energies between 600 eV and 1.5 keV have been diffracted at the surface of a Cu(111) crystal. The intensity of the electron wavefield within the crystal has been measured using low-energy Cu M 2,3 VV Auger electrons integrated over a large solid angle. The results are well reproduced by single-scattering calculations. We show that this technique is chemically selective as well as sensitive to the crystal structure and that it is essentially the time reversal of X-ray photoelectron diffraction but in general more surface-sensitive than the latter. It is therefore well suited to study epitaxial growth and interdiffusion.

High resolution angle resolved photoemission spectra from potassium adsorbed on a single-domain Si(100)2 × 1 surface show K-derived states near the Fermi level. The metallization occurs when half a monolayer of K is deposited and a maximum decrease in the work function is observed. At slightly lower K coverage, the surface is still semiconducting. The metallic K-derived state presents non-parabolic-like dispersions in the directions along and perpendicular to the dimer rows, showing a behavior rather different to the one observed for alkali adsorption on metals.

While photoelectron spectroscopy has been demonstrated to be useful for the determination of Fermi surfaces of two-dimensional systems we discuss the possibility to characterize a fully three-dimensional metal in this manner. Angular distributions of valence photoelectrons at UV energies integrated over a small energy window at the Fermi energy are measured over a large part of the 2π hemisphere for the case of Cu(001), Cu(111) and Cu(110). They are compared to the model of direct transitions from bulk bands crossing the Fermi energy to empty bands. The observed patterns show the proper translational symmetry with respect to the Brillouin zone (BZ). We conclude that the measurements represent a well defined cut through the Fermi surface.

An 8 monolayer Ag film has been deposited in situ on Cu(001) at room temperature. Its structure has been investigated by X-ray photoelectron diffraction, presented as full solid-angle intensity maps, and by low-energy electron diffraction. The formation of a compact Ag(111)-like overlayer with one preferential domain has been observed. Annealing the Ag film at about 300 o C creates Ag islands on the Cu(001) substrate. The system is put into the context of eight other noble-metal (Ag, Au and Pt) films deposited on different substrates (Cu, Pt, Cu 3 Au, Bi 2 Sr 2 CaCu 2 O 8+x ). The crystalline orientation of the overlayers with respect to the substrate and the formation of a surface alloy are discussed

We have applied XPD (X-ray photoelectron diffraction) to the surfaces of W(001)c(2 × 2)-Au and -Ag. Au4f and Ag3d XPD patterns have been measured for the two c(2 × 2) surfaces at electron take-off angles of 4°–22°. It is found that both Au 4f and Ag 3d XPD patterns are very similar to each other showing that the two surfaces have essentially the same structure. Consideration of the general properties of XPD, coverage and bond-lengths of the constituent atoms led to two plausible models (surface-alloy and overlayer models) for the surfaces. We have tested the two models by kinematical analyses of the XPD patterns for the W(001)c(2 × 2)-Au surface using either an s-symmetry and spherical-wave formula or a formula with spherical and dipole-selected final states. It has turned out that although the surface-alloy model is more plausible, the kinematical calculations for the two models do not reproduce the details of the experiment indicating the importance of multiple scattering of electrons for these surfaces consisting of high-Z atoms.

The dissolution of Pt/Cu(111) and the kinetic blocking of the dissolution process which results in the formation of a surface alloy have been studied. We use AES, PES, XPD, and LEED to characterise the non-classical dissolution behaviour. Chemical shifts on the Pt 4f72 and the Cu 3p core levels are measured during dissolution. The top layer of the surface alloy is Cu3Pt for an annealing temperature of 315°C. For the higher annealing temperature of 350°C the Pt concentration is smaller. The alloy composition is therefore temperature dependent and the formation of a relatively stable alloy in local equilibrium over the surface layers is determined by the segregation and ordering behaviour. The chemical ordering of the surface alloy and its electronic structure remains to be verified.

We have carried out a combined X-ray photoelectron diffraction (XPD) and low-energy electron diffraction (LEED) study of the interaction of oxygen with Ni(001) at ambient temperature from the c(2 × 2) structure up to the saturated oxide. Several structural conclusions are possible based on an R-factor comparison of an extensive series of azimuthal- and polar-O 1s XPD data to theoretical simulations based on a single-scattering cluster (SSC) model with spherical wave scattering. A new method for normalizing experimental and theoretical intensities for XPD R-factor analyses is also used. For the c(2 × 2) structure, we find that the oxygen sits in the fourfold hollows site, at a vertical distance (z) of approximately 0.75 A above the first Ni plane, in excellent agreement with the most recent LEED determination. There is also strong evidence from the XPD results that oxide nucleation occurs very early in the chemisorption region (much earlier than observable by LEED) and this effect could explain the proposals for in-plane bonding or pseudobridge bonding previously reported in the literature. The saturated oxide that forms at ambient temperatures is found from LEED and XPD to form a strained superlattice that is expanded by 1 6 with respect to the underlying Ni(001). We also find that the saturation oxide coverage is much larger than that previously reported in the literature of 2–3 monolayers (ML), and that it is in fact 4–5 ML. Other conclusions concerning the nature of the strain in the oxide are also possible, including a vertical expansion in lattice constant from the first to second plane of 0.1–0.2 A.

We have obtained azimuthal X-ray photoelectron diffraction (XPD) data with a high angular resolution of ± 1.5° for S2p emission from the well-defined surface structure of c(2 × 2)S on Ni(001). The relatively high position of the adsorbate with respect to the substrate makes this a stringent test case of the structural sensitivity of forward-scattering-dominated XPD. With this higher resolution, the data are nonetheless found to be sensitive to atomic structure, including in particular both the vertical height of S above Ni ( z ) and the first-to-second layer Ni interplanar spacing ( d 12 ). A single scattering cluster (SSC) theoretical analysis using R -factors to judge goodness of fit yields z = 1.39 ± 0.05 A and d 12 = 1.86 ± 0.05 A , in excellent agreement with other recent experimental and theoretical studies. This analysis also indicates that clusters of up to at least 25 A in radius (200–250 atoms) are needed to accurately describe all of the diffraction fine structure observed; thus, although XPD is primarily a short-range order probe, high-resolution data provides sensitivity to order that may go out as far as 10–15 neighbor shells. For takeoff angles with respect to the surface of less than about 10°, multiple scattering effects appear to become more important, as verified by fully converged multiple scattering cluster (MSC) calculations; however, for takeoff angles larger than 10°, these effects fall away rapidly, making a single-scattering analysis of such data still a useful approach. Finally, we have analyzed our experimental data and SSC simulations of it using recently suggested Fourier-transform holographic inversion methods. Although our data are too limited to permit fully accurate holographic imaging, features associated with the nearest neighbor S atoms in the adsorbate overlayer are seen in both experimental and theoretical images. In addition, the theoretical calculations indicate that the atomic images can be improved if: the solid angle of the hologram is limited so as to exclude the strong forward scattering features at low takeoff angles; effects due to non-constant scattering factor amplitudes and phases are corrected out using the scattered-wave-included Fourier-transform method of Saldin et al., and/or the hologram range is further limited so as to avoid the overlap of twin and real images. Several interesting directions for further study with such high-resolution data, SSC R -factor analyses, and holographic imaging, are thus suggested.

Real-space images along the nearest-neighbour directions are calculated from measured photoelectron diffraction maps of Cu(001) using Huygen's theorem. The images are corrected for the phase shifts and anisotropies of the electron scattering using the method of the scattered-wave-included Fourier transform (SWIFT). Measurements taken at 7 different kinetic energies (768–1740 eV) allow us to distinguish symmetry and interference effects. It is shown that for photoemission from substrates even the corrected images are dominated by symmetry and little indication of atomic positions is found.

Ultrathin films of one monolayer (ML) or less of Au on Cu(001) have been prepared by in situ deposition at 300 K in order to investigate the formation of a surface alloy. The local order within these films was determined by X-ray photoelectron diffraction, measuring full solid angle Au4d 5 2 intensity distributions. For deposition of between 0.2 and 0.6 Ml we confirm that a surface alloy is formed, with a mixed c(2 × 2) Au/Cu layer on top. In contrast to earlier low-energy electron diffraction measurements we find a measurable amount of substitutional Au in the 2nd and 3rd layer. At a coverage of 1.2 ML all azimuthal structure in the angular distribution is lost, consistent with the growth of an incommensurate pseudo-hexagonal Au overlayer with no short-range orientational order. Still the presence of narrow rings of higher intensity reveals the presence of subsurface Au.

The electron wave field near the photoemitter is calculated from measured photoelectron angular distribution maps of Cu(001) using Huygen's theorem. A novel formalism for these calculations is introduced. The resulting real-space images exhibit considerable structures along the nearest and next-nearest neighbour directions. By comparing images taken'at 7 different kinetic energies (768–1740 eV) it is shown that these structures cannot be readily assigned to atomic positions but are more probably related to the crystal symmetry.

Multiple-scattering (MS) calculations are presented for polar intensity scans of Mg Kα excited A12s emission from linear atomic Al chains, and compared to corresponding scans of Cu LVV Auger emission from linear Cu chains. Such model calculations permit an assessment of the importance of multiple-scattering effects in substrate emission along low-index directions. Intensity reductions in forward scattering due to MS defocusing effects are seen for both Al and Cu. However, for Al they are found to be significantly slower in turning on with distance along the chain. A systematic investigation revealed that this difference between Al and Cu is caused by their different scattering potentials. This material dependence should be important in analyzing such diffraction data in a quantitative way.