Quantum states of hydrogen (H, D, T) atoms on Cu(1 0 0) and (1 1 0) surfaces

We investigate the quantum mechanical behavior of adsorbed hydrogen (H, D, T) on Cu(1 0 0) and (1 1 0) surfaces. We construct potential energy surfaces (PESs) for the motion of the hydrogen H atom on Cu(1 0 0) and (1 1 0) surfaces within the framework of density functional theory. The potential energy takes a minimum value on the hollow site of Cu(1 0 0) and on the short bridge site of Cu(1 1 0). Moreover, we calculate the quantum states of hydrogen atom motion on these calculated PESs. The ground state wave function of the hydrogen atom motion is strongly localized around the hollow site on the Cu(1 0 0) surface. On the other hand, the ground state wave function of the hydrogen atom motion on Cu(1 1 0) is distributed from the short bridge site to two neighboring pseudo-threefold sites. We finally show isotope effects on the quantum states of the motion of hydrogen on both surfaces.     

Diffusion of single adatom Cu on Cu (0 0 1) and (1 1 0) surfaces

With static relaxation, the surface diffusion activation energies of a single Cu adatom migrated by both atomic exchange and hopping mechanisms and the forces acted on the diffusing adatom from other atoms of Cu (0 0 1) or (1 1 0) surface are calculated by using the MAEAM. When adatom migrated on Cu (0 0 1) or (1 1 0) surface, the increment curves of the system energy by hopping mechanism are symmetrical and the saddle points are in the midpoints of the migration path, but the ones by the exchange mechanism are dissymmetrical and the saddle points are always close to the initial hole positions of the adatom and away from the initial equilibrium positions of the exchanged atom. From minimization of both the diffusion activation energy and the force acted on the diffusing adatom from other atoms, we found that, on Cu (0 0 1) surface the favorable diffusion mechanism is hopping mechanism, however, on Cu (1 1 0) surface, hopping via long bridge is easier than the exchange mechanism but the hopping via short bridge is more difficult than the exchange mechanism.     

Scattering of Cu(1 0 0) image state electrons from single Cu adatoms and vacancies: A comparative study

A theoretical study of the electron dynamics in image potential states on Cu(1 0 0) surfaces with different types of defects (Cu adatoms and Cu vacancies) is presented for low defect density at the surface. A wave packet propagation approach is employed for the electron scattering calculations, where the defect induced potentials are obtained from an ab initio density functional study. Scattering of the image state electron by a defect induces inter-band and intra-band transitions leading, respectively, to the population decay and to the dephasing of the image states. Comparison of the respective effects of adatoms and vacancies shows that Cu adatoms are much more efficient in inducing population decay and dephasing of the image potential states. Present results for the case of Cu adatoms are compared with available time-resolved two-photon photoemission data.     

Comparative investigation of the c(2 × 2)-Si/Cu(0 1 1) and (√3 × √3)R30°-Cu2Si/Cu(1 1 1) surface alloys using DFT

The electronic structure of the c(2 × 2)-Si/Cu(0 1 1) surface alloy has been investigated and compared to the structures seen in the three phases of the (√3 × √3)R30°Cu₂Si/Cu(1 1 1) system, using LCAO-DFT. The weighted surface energy increase between the alloyed Cu(0 1 1) and Cu(1 1 1) surfaces is 126.7 meV/Si atom. This increase in energy for the (0 1 1) system when compared to the (1 1 1) system is assigned to the transition from a hexagonal to a rectangular local bonding environment for the Si ion cores, with the hexagonal environment being energetically more favorable. The Si 3s state is shown to interact covalently with the Cu 4s and 4p states whereas the Si 3p state, and to a lesser extent the Si 3d state, forms a mixture of covalent and metallic bonds with the Cu states. The Cu 4s and 4p states are shown to be altered by approximately the same amount by both the removal of Cu ion cores and the inclusion of Si ion cores during the alloying of the Cu(0 1 1) surface. However, the Cu 3d states in the surface and second layers of the alloy are shown to be more significantly altered during the alloying process by the removal of Cu ion cores from the surface layer rather than by the addition of Si ion cores. This is compared to the behavior of the Cu 3d states in the surface and second layers of the each phase of the (√3 × √3)R30°-Cu₂Si/Cu(1 1 1) alloy and consequently the loss of Cu-Cu periodicity during alloying of the Cu(0 1 1) surface is conjectured as the driving force for changes to the Cu 3d states. The accompanying changes to the Cu 4s and 4p states in both the c(2 × 2)-Si/Cu(0 1 1) and (√3 × √3)R30°-Cu₂Si/Cu(1 1 1) alloys are quantified and compared. The study concludes with a brief quantitative study of changes in the bond order of the Cu-Cu bonds during alloying of both Cu(0 1 1) and Cu(1 1 1) surfaces.     

Calculated pre-exponential factors and energetics for adatom hopping on terraces and steps of Cu(1 0 0) and Cu(1 1 0)

We have calculated the vibrational dynamics and thermodynamics for Cu adatom hopping on terraces and near step edges on Cu(1 0 0) and Cu(1 1 0), using the embedded atom method for the interatomic potential. The local vibrational densities of states were calculated using real space Green’s function formalism and the thermodynamical functions were evaluated in the harmonic approximation. The calculated diffusion energy barriers for six specific local environments on Cu(1 0 0) agree well with experimental and previous theoretical results. Contribution of vibrational entropy to the change in the free energy of the system as the adatom moves from the equilibrium configuration (hollow site) to the saddle point, is found to be as much as 55 meV (144 meV) at 300 K (600 K). The prefactors for all 13 cases are found to be of the order of 10⁻³ cm²/s, almost independent of temperature, and the respective activation energy barriers.     

Neutralization of Li ions scattered by the Cu (1 0 0) and (1 1 1) surfaces: A localized picture of the atom-surface interaction

Large and face dependent neutral fractions have been found recently in the scattering of Li⁺ by Cu(1 0 0) and Cu(1 1 1) surfaces. These results for high work function surfaces are unexpected within the ‘traditional’ picture of a Li⁺ ion departing from a jellium surface model. In the present work the Li⁺/Cu(1 0 0) and Li⁺/Cu(1 1 1) interacting systems are described by a previously developed bond-pair model based on the localized interactions between the projectile ion and the atoms of the surface, and on the extended features of the electronic band structure through the surface local density of states. By only including the resonant neutralization to the Li atom ground state we explained the face and energy dependences of the measured neutral fractions for large outgoing energy values. We found that the downward shift of the Li ionization level below the Fermi level caused by the short range chemical interactions, is the main responsible of a high neutralization by the resonant mechanism. The remaining differences between theory and experiment values can be explained in terms of the energy gaps and image potential states appearing in these surfaces. The calculated distance behaviours of the energy levels corresponding to the first excited (Li-1s²2p) and the negative (Li-1s²2s²) atomic configurations indicate that they can also participate in the ion-surface charge exchange process.     

para-Sexiphenyl thin film growth on Cu(1 1 0) and Cu(1 1 0)-(2 × 1)O surfaces

We present a reflectance difference spectroscopy (RDS) study of para-sexiphenyl (p-6P) thin film growth on Cu(1 1 0) and Cu(1 1 0)-(2 × 1)O substrates. The RDS spectra show pronounced anisotropies for p-6P films formed on both substrates at room temperature, demonstrating that the molecules are uniaxially aligned within the films. Based on the RD spectra and the evolution of the optical transitions with p-6P coverage the growth mode on both substrates could be identified. From the dominating RDS feature, assigned to the lowest energy HOMO-LUMO transition, the orientation of the molecular chain can be determined. On Cu(1 1 0), the p-6P molecular chains align in the direction, i.e., along the Cu atomic rows, whereas on the Cu(1 1 0)-(2 × 1)O surface, the molecules are oriented in the orthogonal [0 0 1] direction, i.e., along the “added” Cu-O rows of the Cu(1 1 0)-(2 × 1)O surface. The energetic position and line shape of the main RDS feature differs for the two substrates and varies with p-6P coverage. This fine structure is discussed in terms of different molecular conformations, adlayer structure and vibronic replicas.     

Electronic structure and surface reconstruction of adsorbed oxygen on copper(0 0 1)

The electronic structure of the c(2 × 2) and the missing row phases of chemisorbed O on Cu(0 0 1) at a coverage of 0.5 monolayers has been calculated using a full-potential semi-infinite embedding technique. Calculations are made over a range of Cu-O layer spacings, and from the change in the work function, the effective charge on O is obtained. The effective charge is the same for both phases of the O/Cu(0 0 1) surface with a value of −0.3|e| on O, and consequently too small to drive any surface instability. The angular momentum-resolved density of states and energy-resolved charge densities are used to describe the binding and the spatial electronic overlap at the surfaces. In the reconstructed phase the O and the surface Cu atoms undergo displacements, which optimises the bonding.     

Surface resonance transition of roughened Cu(1 1 0)

The temperature dependence of the reflection anisotropy spectroscopy (RAS) of a Cu(1 1 0) surface has been studied over the temperature range 700-1000 K. Because of the roughening transition at 900 K, the bimodal feature at 4.2 eV for a clean surface shifted to 4.3 eV on annealing. A significant decrease in intensity of the same energy level was also observed with increasing annealing temperature. In the annealing temperature range 700-1000 K, anharmonic behavior is expected to be the predominant process of atomic disordering at the surface. Changes in the RAS of Cu(1 1 0) as a result of thermal processing can be understood in terms of the associated changes in surface states. The RAS signal for a surface resonance transition at 4.2 eV is associated with monoatomic [0 0 1] steps.     

Broadening mechanisms for the adsorbate-induced resonance in the Na/Cu(1 1 1) and Cs/Cu(1 1 1) systems

Adsorption of alkali atoms on the (1 1 1) and (1 0 0) noble metal surfaces has been shown recently to induce long-lived resonances located inside the surface projected band gap. However, the width of these resonances, as it appears in two-photon photo-emission experiments, is much larger than the inverse of their lifetime. We report on a theoretical study of some broadening mechanisms of these resonance lines in the Na/Cu(1 1 1) and Cs/Cu(1 1 1) systems at low coverage, including the homogeneous natural line broadening and the inhomogeneous statistical broadenings due to the distribution of adsorption heights associated to the quantal vibration of the alkali adsorbate and to the lateral disorder of alkali adsorption on the surface. The inhomogeneous mechanisms are shown to induce a very large broadening of the resonance line, in quantitative agreement with experimental results. The most important broadening effect appears to be the effect of the distribution of alkali adsorption heights.     

LCAO study of the Cu(1 1 0)-p(2 × 1)O surface

We have performed semi-empirical LCAO calculations of the electronic structure of the Cu(1 1 0)-p(2 × 1)O surface. This has been done accounting for the Cu-Cu interactions by means of a recently proposed set of parameters, which give very good results for the bulk as well as for the surfaces of lowest Miller indices. Furthermore, the O-O interactions, which have been neglected in the preceding similar studies, have been taken into account. The resulting surface bands are in very good agreement with the overall set of the available experimental data. Several issues concerning the physical properties of this surface are addressed in the present paper: the changes induced on the clean surface bands by the adsorption and the reconstruction; the arrangement of the Cu and O atoms in the added rows; the position of the p_y antibonding band of the oxygen. In particular, we have found that the latter has an energy of −0.2 eV at the point. This result confirms an experimental indication in the same direction previously reported by Courths et al. [R. Courths, S. Hüfner, P. Kemkes, G. Wiesen, Surf. Sci. 376 (1997) 43].     

Quantum-well states and spin polarization in thin Ni films on Cu(0 0 1)

The unoccupied quantum-well states in thin Ni films on Cu(0 0 1) have been studied by spin- and angle-resolved inverse photoemission. Three quantum-well features are clearly resolved with exchange splittings of up to 70 meV. As a function of the wave vector parallel to the surface, the quantum-well states follow the corresponding sp band dispersion and evolve into surface resonances upon approaching the band-gap boundary.     

Photoemission study of glycine adsorption on Cu/Au(1 1 1) interfaces

The adsorption of glycine on Au(1 1 1) pre-deposited with different amounts of Cu was investigated with both conventional X-ray photoelectron spectroscopy (XPS) and synchrotron-based photoemission. In the Cu submonolayer range, glycine physically adsorbs on the Cu/Au(1 1 1) surfaces in its zwitterionic form and completely desorbs at 350 K. The C 1s, O 1s and N 1s core level binding energies monotonically increase with Cu coverage. This indicates that, in the Cu submonolayer range, the admetal is alloyed with Au rather than forming overlayers on the Au(1 1 1) substrate, consistent with our recent experimental and theoretical results [X. Zhao, P. Liu, J. Hrbek, J.A. Rodriguez, M. Pérez, Surf. Sci. 592 (2005) 25]. Upon increasing the amount of deposited Cu over 1 ML, part of the glycine overlayer transforms from the zwitterionic form to the anionic form (NH₂CH₂COO⁻) and adsorbs chemically on the Cu/Au(1 1 1) surface with the N 1s binding energy shifted by −2.3 eV. When the amount of deposited Cu is at 3.0 or 6.0 ML, the intensity of the N 1s chemisorption peak increases with aging time at 300 K. It indicates that glycine adsorption induces Cu segregation from the subsurface region onto the top layer of the substrate. Judging from the initial N 1s peak intensities, it is concluded that 64% and 36% of the top layer are still occupied by Au atoms before glycine adsorption even when the amounts of deposited Cu are 3.0 and 6.0 ML, respectively. On the Au(1 1 1) surface pre-dosed with 6.0 ML of Cu, part of the chemisorbed glycine will desorb and part will decompose upon heating to 450-500 K. In addition, about 20% of the glycine exists in the neutral form when the glycine overlayer was dosed on Cu/Au(1 1 1) held at 100 K.     

Density functional theory investigation of CN on Cu(1 1 1), Ni(1 1 1) and Ni(1 0 0)

The adsorption of CN on Cu(1 1 1), Ni(1 1 1) and Ni(1 0 0) has been investigated using density functional theory (DFT). While experimental studies of CN on Cu(1 1 1) show the molecular axis to be essentially parallel to the surface, the normally-preferred DFT approach using the generalised gradient approximation (GGA) yields a lowest energy configuration with the C-N axis perpendicular to the surface, although calculations using the local density approximation (LDA) do indicate that the experimental geometry is energetically favoured. The same conclusions are found for CN on Ni(1 1 1); on both surfaces bonding through the N atom is always unfavourable, in contrast to some earlier published results of ab initio calculations for Ni(1 1 1)/CN and Ni(1 0 0)/CN. The different predictions of the GGA and LDA approaches may lie in subtly different relative energies of the CN 5σ and 1π orbitals, a situation somewhat similar to that for CO adsorbed on Pt(1 1 1) which has proved challenging for DFT calculations. On Ni(1 0 0) GGA calculations favour a lying-down species in a hollow site in a geometry rather similar to that found experimentally and in GGA calculations for CN on Ni(1 1 0).     

Growth and crystalline structure of Co layers on Cu(0 0 1)

The crystalline structure of Co layers deposited on the Cu(0 0 1) surface was investigated with the use of the directional elastic peak electron spectroscopy (DEPES). For clean Cu(0 0 1) the experimental DEPES profiles obtained for different energies of the primary electron beam exhibit intensity maxima corresponding to the close packed rows of atoms. The Auger peak kinetics recorded during continuous Co deposition suggest the layer-by-layer growth mode. The DEPES profiles recorded for 10 monolayers (ML) of Co on Cu(0 0 1) reflect a short-range order in the adsorbate. Intensity maxima observed in the DEPES profiles for Co along [1 0 0], [0 1 0], and [1 1 0] azimuths of Cu(0 0 1) are characteristic of the face centered cubic (fcc) Co(0 0 1) layers. Low-intensity reflections and considerable background intensities were found in the low energy electron diffraction (LEED) patterns recorded from 10 ML of Co, which indicates a weak long-range order in the adsorbate. The adsorption of about 20 ML of Co results in considerable background contribution to DEPES. No reflections but a large background were observed with the use of LEED for this layer. The heating of the Co/Cu(0 0 1) system at T = 770 K leads to an increase of the short- and long-range order in the overlayer, observed in the DEPES profiles and LEED patterns, respectively. The theoretical DEPES profiles were obtained with the use of a multiple scattering approximation. A very good agreement between experimental and theoretical scans was found for the clean and covered copper substrate. The latter proves the epitaxial growth of Co layers on Cu(0 0 1).     

Electron dynamics in H/Na/Cu(1 1 1) collisions

Surface adsorbates induce strong local perturbations in the electronic structure and potentials in their surroundings. Consequently, charge transfer processes between projectiles and adsorbate-covered surfaces are strongly affected. The theoretical calculations and experiment measurements reported herein are focused on the H⁻/Na/Cu(1 1 1) system. The electron dynamics at the Na/Cu(1 1 1) surface and the influence of Na adsorbates on the H⁻-Cu(1 1 1) charge transfer are treated and discussed in detail. The ion fractions are mainly influenced by the ion exit trajectories. At low Na coverage, they exhibit a maximum near the 60° exit angle from surface. The calculations and experimental data are in good agreement.     

Photoemission study of the Li/Ge(1 1 1)-3 × 1 reconstruction

In this article we report our findings on the electronic structure of the Li induced Ge(1 1 1)-3 × 1 reconstruction as determined by angle-resolved ultraviolet photoelectron spectroscopy (ARUPS) and core-level spectroscopy using synchrotron radiation. The results are compared to the theoretical honeycomb-chain-channel (HCC) model for the 3 × 1 reconstruction as calculated using density functional theory (DFT). ARUPS measurements were performed in both the and directions of the 1 × 1 surface Brillouin zone at photon energies of 17 and 21.2 eV. Three surface related states were observed in the direction. In the direction, at least two surface states were seen. The calculated band structure using the single-domain HCC model for Li/Ge(1 1 1)-3 × 1 was in good agreement with experiment, allowing for the determination of the origin of the experimentally observed surface states. In the Ge 3d core-level spectra, two surface related components were identified, both at lower binding energy with respect to the Ge 3d bulk peak. Our DFT calculations of the surface core-level shifts were found to be in fair agreement with the experimental results. Finally, in contrast to the Li/Si(1 1 1)-3 × 1 case, no double bond between Ge atoms in the top layer was found.     

Quasiclassical studies of Eley-Rideal and hot-atom reactions on a surface at 94 K: H(D) → D(H) + Cu(1 1 1)

Reactions and reaction dynamics of gas-phase H(or D) atoms with D(or H) atoms adsorbed onto a Cu(1 1 1) surface have been investigated by the quasi-classical molecular dynamics method. To simulate the H(D) → D(H) + Cu(1 1 1) system at a 94 K surface temperature, D(or H) adsorbates were disseminated arbitrarily on the surface of Cu(1 1 1) to form 0.50, 0.28 and 0.18 ML of coverages. The interaction of hydrogen atoms and the surface system is worked out by an LEPS function. LEPS parameters have been determined by using the total energy values which were calculated by a density functional theory (DFT) method and the generalized gradient approximation (GGA) for the exchange-correlation energy for various configurations of one and two hydrogen atoms on the Cu(1 1 1) surface. The Cu(1 1 1) surface, imitated by an embedded-atom method which is a many-body potential parameterized by Voter-Chen, is formed as a multilayer slab. The slab atoms are permitted to move. Various processes, trapping onto the surface, inelastic reflection of the incident projectile and penetration of the adsorbate or projectile atom into the slab, are examined. The dependence of these mechanisms on isotopic replacement has also been analyzed. Considerable contributions of the hot-atom pathways for the product formations are consequently observed. The rate of subsurface penetrations is obtained to be larger than the sticking rate onto the surface.     

Structure and dynamics of the missing-row reconstruction on O/Cu(0 0 1) and O/Ag(0 0 1)

The geometry and the vibrational properties of missing row reconstructed O/Cu(0 0 1) and O/Ag(0 0 1) surfaces are investigated by means of density functional theory and density functional perturbation theory, using the local density and the generalized-gradient approximations. Our results predict very similar structural and vibrational properties for the two reconstructed surfaces. In the case of copper our calculations reproduce quite accurately the experimental results, while for the missing row reconstructed O/Ag(0 0 1) surface the agreement between theory and experiment is less satisfactory.