The full-solid-angle (2π) intensity maps of the carbon-KVV Auger-electron emission at 260 eV from diamond (111), diamond (100), and highly oriented pyrolitic graphite (HOPG) (0001) samples were collected and compared with spherical-wave single-scattering cluster calculations. In contrast to Mg Kα excited C 1s patterns the C-KVV patterns show weak direct real-space information. We performed a detailed angular momentum analysis taking into account different l characters of the emitted Auger electrons and show that in diamond the C-KVV emission has mainly a d-like character while graphite emits a mixture of p- and d-like electrons with a minor contribution of s-like electrons.

The equilibrium crystal shape of lead develops an edge in the temperature range where surface melting is observed. The edge is usually explained as resulting from the coexistence of melted and unmelting orientations. This is shown to be incorrect. Localized step melting allows the average liquid thickness to change continuously with orientation and suppresses the formation of an observable edge. The melting of isolated steps is shown to result from the stress generated by a step. The edge is shown to result from a competition between a transition of the (111) facet between 275 °C and 300 °C, modeled as vacancy melting, and step melting. These two effects account for the unusual behavior of the lead equilibrium crystal shape. The thermal faceting dynamics of the surface melting free energy is investigated and predictions for the separated region size extracted.

A microscopic theory for the diffusive and vibrational motion of adsorbed atoms on a solid surface is presented. We evaluate and analyze the dynamical structure factor and the velocity autocorrelation function. The quasielastic scattering peak in the dynamical structure factor is found to consist of mixed contributions from diffusive and vibrational motion. The vibrational peak indicates a shift and broadening from anharmonic effects and frictional damping. As the temperature is lowered the diffusion constant crosses over from a continuous Brownian motion behavior to a thermally activated Arrhenius form. When the frictional coefficient is varied at low temperatures, the diffusion constant crosses over smoothly from the multisite jumping regime at low friction, via the transition-state regime, to the high-friction overdamped regime. Applied to Na on the Cu(001) surface, our theoretical predictions on the quasielastic-peak width and the temperature dependence of the shift of the vibrational mode agree quantitatively with recent experimental measurements. For adatoms on anisotropic bcc (110) surfaces, we also study the diffusion anisotropy as well as the coupling of the two frustrated translational modes.

The asymptotic elastic-displacement field of a surface step is shown to result from a planar force distribution which is composed of two point elastic dipoles, one oriented along the surface of arbitrary magnitude and one normal to the surface with dipole moment ag where a is the step height and g is the surface stress. The analytic form of the dipole displacement field for an isotropic material is presented. Experimental and simulated TEM images of the asymptotic displacement field are presented which demonstrate the dipolar nature of the force distribution. The normal dipole moment of a monatomic step on Si(111)(7×7) is computed to be 0.58±0.04 eV/Å. The tangential dipole moment is measured to be 1.46±0.3 eV/Å. The stress and strain tensors for an isolated step and the displacement field for a stepped surface are presented.

A one-dimensional, step-induced image state is found at vicinal Cu(100) surfaces with inverse photoemission and two-photon photoemission. Its energy nearly coincides with the bottom of the two-dimensional image-state band of the flat surface. The one-dimensional nature of the step state manifests itself in a lack of dispersion perpendicular to the steps. Parallel to the steps it exhibits free-electron-like dispersion, making this type of state an ideal model for one-dimensional electron physics.

Ultrathin metallic films deposited on amorphous substrates show island growth until electrical percolation occurs. Scaling theory makes rather precise predictions about the deposition-time dependence of the island-size distributions and related properties. It is therefore important to be able to analyze experimental data in a quantitative way in order to be able to make a detailed comparison with theory. This paper addresses two aspects of the data analysis of island-size distributions. In the first, a robust method of presenting measurements in the form of continuous distributions (as opposed to histograms) is examined. In the second, we obtain a prescription for transforming chord distributions into diameter distributions. This transformation will be useful in cases where the islands are (or approximate to) regular droplets and when it is easier to measure chords than diameters. Relations between the moments of the chord and diameter distributions are obtained. To demonstrate the first method, it is used to extract the island-size distribution within a Pt/a-C film. A remarkably good match to a log-normal distribution is found without making any of the usual priori assumptions. Comments on implications for scaling theory are made.

When two objects of subwavelength size interact in the presence of a light beam, a spatially confined electromagnetic field arises in a small spatial region located at the immediate proximity of the particles. In scanning probe microscopy, such induced short-range interactions change the magnitude of the forces interacting between the probe tip and the substrate. Depending on the frequency of light excitation with respect to those of the gap modes associated with the tip-sample junction, these inductive forces act to pull the probe toward the surface. Such an effect can be used to record optical adsorption of various samples with an atomic-force microscope. In this paper we show that the accurate description of the physical processes responsible for these forces can be analyzed within the framework of the localized field-susceptibility method. Practical solutions for the light-inductive force were found by discretization of the probe apex in real space. All multiple interactions including reflections with a substrate of arbitrary profile were accounted for by self-consistent procedures. We can therefore present simulations performed on systems of experimental interest.

We report the transport properties of thin Mo-C films which show the weak-localization behavior as the thickness is varied from 8.6 to 20.6 Å. The inelastic and spin-orbit scattering times were determined from magnetoresistance measurements; a crossover is observed at T∼30 K at which we argue results from ‘‘antilocalization.’’

We present a restricted solid-on-solid Hamiltonian for fcc(110) surfaces. It is the simplest generalization of the exactly solvable body-centered solid-on-solid model which is able to describe a (2×1) missing-row reconstructed surface. We study this model by mapping it onto a quantum spin-1/2 chain of the Heisenberg type, with second and third neighbor ${\mathit{S}}_{\mathit{i}}^{\mathit{z}}$${\mathit{S}}_{\mathit{j}}^{\mathit{z}}$ couplings. The ground-state phase diagram of the spin-chain model is studied by exact diagonalization of finite chains up to N=28 sites, as well as through analytical techniques. We find four phases in the phase diagram: two ordered phases in which the spins have a Néel type of long-range order (an unreconstructed and a missing-row reconstructed phase, in the surface language), a spin-liquid phase (representing a rough surface), and an intermediate dimer phase which breaks translational invariance and has a doubly degenerate ground state, corresponding to a disordered flat surface. The transition from the (2×1) reconstructed phase to the disordered flat phase belongs to the two-dimensional Ising universality class. A critical (preroughening) line with varying exponents separates the unreconstructed phase from the disordered flat phase. The possible experimental signatures of the disordered flat phase are discussed.

Scanning-tunneling-microscopy (STM) data for the Cu(111) surface have been calculated theoretically, based on Bardeen’s approximation and a layer-Korringa-Kohn-Rostoker-like approach to describe the sample. The onset of the surface state gives rise to a steplike increase in the conductivity in very good agreement with recent experimental data, which demonstrates the ability of Bardeen’s approximation to describe surface states in spite of its failure regarding surface states in one dimension. The STM conductivity can be approximated in terms of the local density of states (LDOS) of the sample at the voltage V, plus a background term which involves states at all energies, and which is frequently neglected in the interpretation of STM images. In the presence of a step edge, the LDOS of the Cu(111) surface state displays standing wave oscillations. In this case, the background term shows oscillations of the Friedel type which drop as ${\mathit{J}}_1$(2kx)/kx for a two-dimensional electron gas. This has important implications for determining the surface state dispersion relation and for the interpretation of standing wave patterns in terms of the LDOS.

The possibility of more than one magnetic state in cluster systems has been explored using the local spin density linear combination of Gaussian orbitals method. Fixed moment calculations were done and their converged states were used to determine input potentials to find as many stable states as possible for the V and Cr clusters having body-centered-cubic (bcc) symmetry. Several calculations were made at different atomic spacings for ${\mathrm{V}}_9$ and ${\mathrm{Cr}}_9$ clusters. It was found that for some atomic spacings, as many as four or five magnetic states exist for a cluster. In addition to the states reported previously with small average magnetic moments of 0.33${\mathrm{\mu }}_{\mathit{B}}$ and 0.67${\mathrm{\mu }}_{\mathit{B}}$ per atom, states with moments as large as 2.78${\mathrm{\mu }}_{\mathit{B}}$ and 3.78${\mathrm{\mu }}_{\mathit{B}}$ per atom are also found to be possible for ${\mathrm{V}}_9$ and ${\mathrm{Cr}}_9$ clusters, respectively, as well as intermediate-moment states in some cases. The ${\mathrm{Cr}}_{15}$ cluster is found to have a state with 0.80${\mathrm{\mu }}_{\mathit{B}}$ per atom in addition to the 0.40${\mathrm{\mu }}_{\mathit{B}}$ state reported previously. The local moments for the 0.80${\mathrm{\mu }}_{\mathit{B}}$ state in ${\mathrm{Cr}}_{15}$ clusters are found to be very close to that which has been reported by other groups.

Adatom diffusion on stepped square lattices has been studied by means of Monte Carlo modeling. Steps consisting of two adjacent rows running in the x direction have been separated by two terrace rows and it has been assumed that one of the step rows is located on the up side and the other on the down side of a lattice step. In order to account for differences between up and down sides of steps, an extra adsorption energy ${\mathit{cphi}}_{\mathrm{st}}$ has been added or subtracted for adatoms on down-side or up-side step rows. Pairwise interaction energies between nearest neighbors ${\mathit{cphi}}_{\mathrm{NN}}$ have also been taken into account. It has been assumed that these couplings are not effective between adatoms on up-side and down-side step sites. At low surface coverages θ and temperatures T, most of the adatoms are trapped by down-side step sites. Both parallel and perpendicular diffusion require detachment from such step sites, and therefore diffusion is almost isotropic. As coverage increases the down-side step sites become highly populated and then act as barriers for perpendicular diffusion. The up-side step sites constitute channels for fast parallel diffusion so that anisotropy increases. In the case of nearest-neighbor repulsion ${\mathit{cphi}}_{\mathrm{NN}}$<0 the activation energies for surface diffusion are generally smaller. At low temperatures T≤${\mathit{T}}_{\mathit{c}}$c(2×2) ordered lattice gas domains are observed for θ≊0.5 with disordered regions nucleating at step rows in between. The effect of ordering on surface diffusion is described.

We apply an efficient algorithm for density-functional calculations based on conjugate-gradient minimization to the problem of interacting adatoms on a surface. Using a jellium model for sodium, we calculate the interaction energy of two rows of adatoms as a function of separation. We find a long-ranged repulsion which is electrostatic in origin. This repulsive interaction is well modeled by a dipole-dipole repulsion consistent with the large dipole moment of the adatoms.

A theory for the effect of a strong intra-atomic Coulomb repulsion U on the nonadiabatic transfer of charge between a metallic surface and a moving atomic species is presented. Using slave bosons and a nonequilibrium Green’s-function technique, we solve the equations appropriate for the U=∞ problem in the case when either the atom-surface hopping matrix element is small, or the number of degenerate atomic states is large. We generalize the earlier treatment of Langreth and Nordlander (LN) to include off-diagonal self-energies and present a general numerical scheme for the exact solution of the Dyson equations. We verify that our scheme gives the correct answer in several limiting cases where an exact solution is known, and give quantitative predictions of when deviations from these limits become important. These limits include (1) the simple master equation limit for low velocities and weak coupling, (2) the generalized master equation of LN for larger velocities and atom-surface coupling, (3) the approach to thermal equilibrium when the time dependence is removed, and (4) the maintenance of local thermal equilibrium when the energy parameters vary sufficiently slowly. From a calculation of the instantaneous (nonequilibrium) spectral function of the level on the scattering atom, we are able to study the rate of formation of the Kondo and mixed valent resonances near the Fermi level. We find a slow formation rate for such resonances relative to that of the broader parts of the spectral density centered near the bare atom level positions.

Many recent ion-surface scattering experiments, e.g., ${\mathrm{He}}^{+}$→Pb, show dramatic oscillations of the neutralization probability as a function of the inverse ion velocity. These oscillations are qualitatively similar to the Stueckelberg oscillations that occur during quasiresonant charge exchange in atom-atom collisions. We present here a general many-body approach for the description of such charge-exchange processes near surfaces. We show that for the proper understanding of the oscillatory behavior it is crucial to include intra-atomic correlation effects. The interference between different tunneling channels influences both the amplitude and the frequency of these oscillations.

Recent Raman-scattering investigations of visible- or ultraviolet-irradiated solid ${\mathrm{C}}_{60}$ film have revealed that the film consists of photopolymerized fullerene molecules with covalent intermolecular bonding in contrast to the weak van der Waals type bonding found in the crystal. In this report various ${\mathrm{C}}_{60}$ dimer geometries are studied to determine the most stable dimer configuration, using a recently developed generalized tight-binding molecular-dynamics technique. Comparison of calculated properties with experimental results strongly suggests that the linkage between ${\mathrm{C}}_{60}$ monomer units is across parallel double bonds reforming into a four-sided ring, giving a ‘‘dumbbell’’ structure.

Eight isomers of the toroidal cage form of graphitic carbon ${\mathrm{C}}_{240}$ and ${\mathrm{C}}_{250}$ that contain pentagonal chains composed of hexagons of carbon atoms are proposed. The separation between the parallel fringes in these structures is close to that observed experimentally for nanometer-scaled carbon tubes (0.34 nm). The isomers consist of pairs of pentagons and heptagons with different orientation and position along with hexagons of carbon atoms. The cohesive energies and thermal stabilities of the isomers are studied by molecular dynamics. The proposed structures are found to be thermodynamically stable, although their cohesive energies are slightly higher than the spherical fullerene ${\mathrm{C}}_{240}$. Furthermore, the electronic states for the isomers are calculated by molecular-orbital calculations. Highest-occupied-molecular-orbitals–lowest-unoccupied-molecular-orbitals gaps for the tori obtained by simple Hückel-type calculations indicate that toroidal structures tend to be chemically stable when the pentagons and heptagons are radially aligned, i.e., the inner and outer hexagon tubes have the same orientation.

It is demonstrated that the Gaussian-broadening scheme for the explicit inclusion of fractional occupation numbers in density-functional calculations of crystals, as originally given by Fu and Ho [Phys. Rev. B 28, 5480 (1983)], contains an additional term to the total energy which makes the energy functional variational and exactly cancels a correction to the density-functional forces originating from changes in the occupation numbers.

We report the valence-band and Co 2p photoemission spectra of ${\mathrm{LaCoO}}_3$ obtained at different temperatures (100, 300, and 573 K). Analysis of the core-level spectrum in terms of a configuration interaction model suggests that both low- and high-spin states coexist at the low temperature (100 K). It also indicates that there is a decrease in the low-spin contribution at 573 K related to local structural changes. Photoemission spectra of the valence-band region further support this interpretation.

Electrical conductivity, magnetic susceptibility, and electron-spin-resonance studies are reported in polyaniline and its derivatives poly(o-toluidine), poly(o-ethylaniline), and poly(o-propylaniline). It is shown that substituents on the ${\mathrm{C}}_6$ rings of the polymer backbone lead to decreased conductivity, decreased Pauli susceptibility, and an increased number of Curie spins. Also reflecting these features are electron-spin-resonance line shapes, which transform from Lorentzian for polyaniline toward Gaussian for poly(o-propylaniline). These results are interpreted in terms of increased electron localization caused by increased interchain disorder. Furthermore, evidence of electron-electron correlations is presented for poly(o-propylaniline).