The surface structures formed by segregated silicon on a bcc ${\mathrm{Fe}}_{96.5}$${\mathrm{Si}}_{3.5}$(110) surface have been studied by scanning tunneling microscopy (STM). Additionally Auger-electron spectroscopy and low-energy ion-scattering spectroscopy have been used to obtain the chemical and structural information necessary to decide between possible configurations suggested by the STM images. Near the expected saturation coverage a substitutional c(1×3)Si superstructure weakly disordered by 〈110〉 aligned domain boundaries was observed. Their arrangement is equivalent to the incommensurate domain-wall structures of chemisorption systems like H/Fe(110), which allow a continuous increase of the surface coverage by successive insertion of equidistantly arranged heavy domain walls. Our observations suggest a possible increase of the silicon coverage above the 33.3% of the perfect c(1×3) structure at least up to 40% corresponding to a c(1×5)Si commensurate domain-wall structure. Correlation analysis of the disordered phase at 15% coverage, obtained by short annealing, allows a qualitative determination of the interaction forces among individual silicon atoms on the surface, yielding attraction of third-nearest neighbors and repulsion of nearest neighbors.

High resolution measurements are reported of the surface core-level shift of the 3d level for the Rh(111), Rh(110), Pd(111), Pd(110), and Ag(111) single-crystal surfaces. These measurements and earlier ones for the Mo(110), Rh(100), and Pd(100) surfaces are analyzed by ab initio calculations of the surface core-level shift. The calculations are found to reproduce well the trends of the experimental shifts with the 4d metal and with the crystal plane. The comparison between these experimental and theoretical results demonstrates the importance of proper inclusion of final-state effects for accurate calculations of surface core-level shifts. A core hole in a surface atom is found to be better screened than one in a bulk atom for the 4d metals to the left of Pd in the Periodic Table. The use of the Z+1 approximation to describe the core hole is investigated both by explicit use of this approximation and by performing calculations for 1s and 3d core holes, respectively. The Z+1 approximation is found to be well obeyed in the case of Ag whereas for the rest of the 4d transition metals it is less precise, introducing errors of typically 0.1 eV.

The sensitivity and utility of circular dichroism in the angular distribution of photoelectrons (CDAD) as a probe of molecular orientation is demonstrated for adsorbed CO and NO molecules. A comparison between measured CDAD spectra and calculated values for spatially oriented CO and NiCO clearly confirms the well-known perpendicular adsorption for CO on Ni(100), whereas for CO adsorbed on Fe(100) a tilted adsorption geometry was found. For NO/Ni(100) and for NO on the oxygen-preadsorbed Ni(100) surface, an average tilt angle of α=40±10° was observed. In the case of the oxygen-preadsorbed Ni(100) surface, a higher fraction of NO molecules was found to be in a tilted orientation than on the clean surface.

An adsorption system of metal-on-metal growth is presented. The adsorption of alkali-metal atoms on monolayers of dissimilar alkali metals preadsorbed on the Cu(001) surface has been studied by Auger-electron spectroscopy, work-function change, and low-energy electron-diffraction intensity analysis. We have examined the following three adsorption systems on Cu(001) at 180 K; Na on a full monolayer of Li, Li on a full monolayer of Na, and Na on a submonolayer of Li, namely, the c(2×2) structure of Li adatoms. It is found that a simple overlayer formation of Na takes place on the full Li monolayer, while Li atoms substitute Na adatoms for Li adsorption on the full Na monolayer. For adsorption of Na on the Li-c(2×2) structure whose coverage is 5/8 of the full Li monolayer, it is found that Na atoms compress the Li adlayer to become denser monolayers. The Na atoms do not intermix with Li adatoms, but they form islands of the c(2×2) structure on Cu(001).

The influence of surface excitations on electrons elastically backscattered from solid surfaces is investigated. Elastic-scattering differential cross sections are calculated using the partial-wave expansion method and the finite difference techique for solid atoms with the Hartree-Fock-Wigner-Seitz potential. An extended Drude dielectric function which allows the characteristic oscillator strength, damping constant, and critical-point energy for each subband of valence electrons is employed to estimate electron inelastic mean free paths for volume excitations. The same dielectric function is applied to evaluate the probability of surface excitations for incident and escape electrons by including the recoil effect without the small-angle approximation. Results of Monte Carlo simulations on the elastic reflection coefficient and the angular distribution of electrons elastically backscattered from Cu and Ag surfaces are presented. It is revealed that surface excitations significantly reduce the elastic reflection coefficient for low-energy electrons, but less significantly influence the angular distribution for large escape angles. Our results agree very well with experimental data.

The electronic structures of small sulfur clusters, ${\mathit{S}}_2$–${\mathit{S}}_{12}$, are calculated using the ab initio self-consistent discrete variational local-density-functional method, and the calculated ionization potential curve is in good agreement with that from experiment in the range from ${\mathit{S}}_2$ to ${\mathit{S}}_8$, and the calculated ionization potentials of clusters from ${\mathit{S}}_9$ to ${\mathit{S}}_{12}$ are reliably predicted where measurements are lacking. We suggest that the different mixtures of atomic orbitals 3s and 3p on the highest occupied molecular orbitals for different clusters are related to the change of ionization potential with the cluster size. In addition, we point out that the two sulfur dimers in ${\mathit{S}}_4$ are bound by the van der Waals force.

Layer-resolved quantum-well states (QWS’s) corresponding to islands of different thicknesses for ultrathin Ag films on graphite have been observed by angle-resolved ultraviolet photoemission. The spectral linewidth dependence on island thickness and on binding energy is well described within a phase-accumulation model, taking into account the specific quantum-well barriers involved. Thus high-resolution photoemission of QWS’s provides an interesting tool to investigate the morphology and island-height distribution of ultrathin metal films.

Lattice gas models have been used to study the growth of icosahedral and fcc clusters. The simulations include adsorption, desorption, and diffusion of atoms to the nearest and next nearest lattice sites. A general survey of cluster growth is given from a diffusion dominated low pressure case to the adsorption dominated high pressure case. In general, clusters seem to grow monotonically without spending longer times at any cluster size. Compact clusters are observed only close to the limits of low and high pressures, and the growth is controlled either by diffusion or adsorption, accordingly. Both of these mechanisms can produce layer by layer type growth. When the layer by layer growth takes place only along (111) surfaces in fcc clusters, a reglar pattern is obtained in the abundance spectrum. This has the same frequency as the octahedral growth, even though the individual clusters are not perfect octahedra.

Total-energy calculations for 13 different fullerene clusters, including 8 toroidal fullerenes, are performed using a parametrized tight-binding functional, and in select cases, an all-electron first-principles local-density functional. The tori considered here are composed of graphitelike six-membered rings, with five-membered rings on the outer surface of the torus providing positive Gaussian curvature, and seven-membered rings on the inner toroidal surface providing negative Gaussian curvature. Comparisons are made with energies calculated from a Stillinger-Weber–type potential for graphite, which is found to be inadequate for addressing relative energetic stability among the tori. Two empirical strain relations are tested against our tight-binding energy calculations. A simple elastic theory is found to give reasonably good energy orderings. Charge-density calculations on the tori reveal that electron density is enhanced for atoms in five-membered rings, and depleted for atoms in seven-membered rings. First-principles total-energy calculations indicate that a single potassium atom is bound to the center of a ${\mathrm{C}}_{144}$ torus by 0.5 eV. It is proposed that atoms with higher orbital angular momentum (e.g., Zr, etc.) may bond more tightly to the torus, and thus facilitate formation of ${\mathrm{C}}_{120}$ or similar tori.

We present results of simulations of deposition growth at finite temperatures using a ballistic deposition rule with continuous hopping of surface particles. We examine the pattern of growth as a function of the physical parameters of hopping rate and substrate temperature, and identify the different regions of layer-by-layer, fractal-cluster, multilayer, and scaling growth as functions of the growth parameters.

Tractable schemes for calculating the d parameters that characterize surface optical response are derived and evaluated. Simple models of metals at the extremes of tight binding and nearly free electrons are considered, and in both limits the effect of surface states is explicitly illustrated. We find that the screening process for a field normal to the surface is very efficient for excitation energies less than the bulk-plasmon energy, suppressing single-particle spectral structure.

We show that the nonlinear optical response reflects sensitively the electronic structure of transition-metal surfaces and interfaces. d and s electrons may contribute rather differently to the second-harmonic-generation (SHG) signal. This results from the different sensitivity of d and s electrons to surface and symmetry changes. Consequently, SHG for noble metals shows a by far larger dependence on the polarization of the incoming light than for transition metals like Fe, Co, Ni, in particular at lower frequencies. The theroetical results are compared with recent measurements. We conclude that the SHG yield is, in addition to the nonlinear magneto-optical Kerr effect, a sensitive fingerprint of the electronic structure at surfaces and interfaces.

We theoretically investigate optical-absorption spectra of ${\mathrm{C}}_{60}^{6\mathrm{-}}$ and ${\mathrm{C}}_{70}^{6\mathrm{-}}$, and discuss relations with the optical properties of alkali-metal-doped fullerides ${\mathit{A}}_6$${\mathrm{C}}_{60}$ and ${\mathit{A}}_6$${\mathrm{C}}_{70}$. This is a valid approach for systems where molecular-exciton (Frenkel-exciton) effects are dominant. We use a tight-binding model with long-ranged Coulomb interactions and bond disorder. Optical spectra are obtained by the Hartree-Fock approximation and the configuration interaction method. We find that the Coulomb interaction parameters, which are relevant to the optical spectra of ${\mathit{A}}_6$${\mathrm{C}}_{60}$ (${\mathit{A}}_6$${\mathrm{C}}_{70}$) in order to explain the excitation energies and relative oscillator strengths of absorption peaks, are almost half of those of the neutral ${\mathrm{C}}_{60}$ (${\mathrm{C}}_{70}$). The reduction of the effective Coulomb interactions is concluded for the heavily doped case of ${\mathrm{C}}_{60}$ and ${\mathrm{C}}_{70}$. This finding is closely related to the experimental fact that dielectric constants of fullerides, which are maximumly doped with alkali metals, become about twice as large as those of the neutral systems.

We present a generalization of the Li, Nunes, and Vanderbilt density-matrix method to the case of a nonorthogonal set of basis functions. A representation of the real-space density matrix is chosen in such a way that only the overlap matrix, and not its inverse, appears in the energy functional. The generalized energy functional is shown to be variational with respect to the elements of the density matrix, which typically remains well localized.

The successive coupled-cluster approximation up to the third level is applied to study the ground state of the linear E-e Jahn-Teller effect. Over the whole range of the coupling parameter, our results for the ground-state energy are not only in good agreement with the exact numerical diagonalization results, but they are also better than those from earlier analytical studies. The mathematical treatment in this work is quite simple and could be easily extended to the studies of the more complicated Jahn-Teller systems.

Substitutional ${\mathrm{Ni}}^{\mathrm{-}}$ in synthetic diamonds has been investigated by electron-paramagnetic resonance (EPR) and photo-EPR. The photo-EPR investigation shows that the ${\mathrm{Ni}}^{\mathrm{-}}$ EPR signal can be diminished by optical illumination, and the threshold energy of 2.47±0.02 eV suggests that the acceptor level is located at 3.03 eV above the valence band. The energy dependence of the optical cross section for the electron ionization ${\mathrm{\sigma }}_{\mathit{n}}$ was determined from an analysis of the EPR transients.

A hybrid experimental system in which the inhomogeneous magnetic-field distribution near a superconducting gate layer is projected down onto the two-dimensional electron gas beneath it has been the subject of a large number of recent investigations. Studies have focused on both the transport of carriers in such a ‘‘magnetic’’ mesoscopic system as well as the dynamics of vortices in the superconductor, itself. In particular, the weak-localization correction to the classical conductivity in an array of flux vortices has been shown to be qualitatively different from the homogeneous case. We present here the results of a complete numerical calculation of the conductivity for different vortex models where the separation between the electronic and superconducting systems has been explicitly accounted for and we are able to obtain good agreement with recent experiments. Remaining discrepancies at very low magnetic fields arise because of the assumption that the vortices are distributed homogeneously on an ordered lattice, which is certainly not the case in the experiment, and we show how relaxing this approximation can yield much better agreement.

We study the falloff of the wave function of an electron in a strong magnetic field and a one-dimensional periodic potential with period larger than the magnetic length. If the energy of an electron state is far away from the center of the Landau level, the falloff is described by a simple exponent, i.e., the logarithm of the wave function depends linearly on the tunneling distance. The characteristic length is proportional to the inverse period of potential and is linear in the reciprocal magnetic field.

We have developed a method for the local measurement of surface stress using transmission electron microscopy. Applying this technique to the clean Si(111) surface during 7×7-1×1 phase coexistence, we determine the stress difference between the phases to be 30±5 meV/A${\mathrm{̊}}^2$. This relatively small difference is consistent with the presence of adatoms in the 1×1 surface. The method can be easily extended to other surfaces and interfaces that have well-defined domain boundaries.

We show that measurement of the absorption of uniform electric fields at frequencies of order 500 kHz can provide information about the transition between the Wigner crystal and fractionally quantized Hall states of a two-dimensional electron gas in a high magnetic field. In addition, absorption at somewhat higher frequencies (10 MHz–1 GHz) can provide information about the Wigner-glass state observed at higher magnetic fields.