Spin-resolved and high-energy-resolution x-ray photoelectron spectroscopy (XPS) measurements of the 3s and 2s levels of metallic cobalt are reported. The high-resolution 3s spectrum shows a multiplet splitting Δ${\mathit{E}}_{3\mathit{s}}$ of 3.3±0.3 eV. The metallic value and literature Δ${\mathit{E}}_{3\mathit{s}}$ values for ${\mathrm{CoF}}_2$ and ${\mathrm{CoF}}_3$ display a linear variation with (2S+1) in accord with the Van Vleck model. This finding, combined with the absence of core-level XPS satellites in Co metal attributable to localized final-state 3d charge fluctuations, implies that intra-atomic exchange dominates the Co 3s spectrum, and indicates that the creation of a core hole does not alter the 3d electron population in metallic Co, ${\mathrm{CoF}}_2$, or ${\mathrm{CoF}}_3$. The large role played by intra-atomic exchange is explicitly revealed by spin-resolved XPS (SRXPS) investigation of the Co 3s level. The SRXPS data show a predominantly minority-spin character for the main (low-binding-energy) 3s component, and a strong majority-spin character for the smaller 3s component at higher binding energy. Although high-resolution Co 2s XPS measurements give no indication of multiplet structure, SRXPS Co 2s measurements reveal a 0.65±0.3 eV exchange splitting between majority-spin and minority-spin components, as well as a significant spin dependence to the ${\mathit{L}}_1$${\mathit{L}}_{2,3}$${\mathit{M}}_{4,5}$ Coster-Kronig transition probability.

The role of electron spin in photoelectric emission from the 2s, 2${\mathit{p}}_{1/2}$, 2${\mathit{p}}_{3/2}$, and 3p core levels of cobalt metal is investigated through a combination of spin-resolved x-ray-photoelectron spectroscopy (SRXPS) and high-energy-resolution XPS (HRXPS) studies. The sensitivity of the Co core-level spectra to a variation in the valence spin polarization is revealed by comparing the Co-metal data with SRXPS and HRXPS spectra from the amorphous cobalt-based ferromagnetic glass ${\mathrm{Co}}_{66}$${\mathrm{Fe}}_4$${\mathrm{Ni}}_1$${\mathrm{B}}_{14}$${\mathrm{Si}}_{15}$. The results demonstrate explicitly the role of intra-atomic exchange in deep-core-level photoemission from Co, and reveal core-hole spin dependencies to core-hole-induced electron-hole-pair excitation, to the ${\mathit{L}}_1$${\mathit{L}}_{2,3}$V Coster-Kronig transition rate, and to the rates of Auger transitions involving the valence band. The data also suggest that the reduction of the cobalt magnetic moment in the ${\mathrm{Co}}_{66}$${\mathrm{Fe}}_4$${\mathrm{Ni}}_1$${\mathrm{B}}_{14}$${\mathrm{Si}}_{15}$ glass results from the transfer of majority-spin 3d electron density into the minority-spin 3d band caused by a hybridization-induced weakening of the 3d electron-electron interaction.

We have previously used molecular dynamics to simulate the force oscillations experienced by a model atomic-force-microscope tip brought near a surface under a Lennard-Jones liquid. Here we perform these simulations for additional tip radii. We also apply an Ornstein-Zernicke-type integral equation theory to this system, and obtain force-distance curves for several different state points, tip radii, and surface-liquid potentials. We find this theory to be in good agreement with simulation results for tip-wall separations greater than one molecular diameter. We conclude that the magnitude of the force oscillations experienced by an atomic-force-microscope tip is a linear function of the effective tip radius (at constant temperature) and that measurement of these force curves with a standard solvent could provide a method of estimating the relative radii of different tips.

The transfer of an Al atom between tip and sample in the scanning tunneling microscope is analyzed theoretically. The bias required to reduce the activation barrier for transfer to zero is a strong function of tip-sample separation, changing from ∼5 to ∼0.5 V over a separation range of 2 Å. The degree of bias-induced ionization of the atom is found to be small for the range of separations studied.

A theoretical study of the local field inside a metallic quantum well placed on top of a dielectric substrate is presented in the case where the optical diamagnetic response is the dominating one. Taking as a starting point a fundamental integral equation for the local field, the prevailing sand p-p p-polarized case it is demonstrated that the so-called slave approximation gives a result in complete agreement with the exact numerical calculation. In the slave approximation it is assumed that (i) the field-induced motion of the conduction electrons parallel to the plane of the film is independent of the local field across the quantum well and (ii) the motion of the particles across the well is driven by the background field plus the local field caused by the motion of the carriers along the quantum well. On the basis of the homogeneous part of the fundamental integral equation the self-sustaining oscillations in the local field, i.e., the local-field eigenmodes, are investigated. The basic theory for the local field is used to calculate the s- and p-polarized amplitude-reflection coefficients of the quantum well/substrate system, and it is shown that for thin quantum wells a macroscopic two-layer model carrying surface currents as well as surface charges at the vacuum/substrate interface can account for the optical-reflection properties once the surface currents and charges have been determined from microscopic considerations. Numerical calculations of the local field inside the quantum well, the local electric displacement field, the s-polarized energy-reflection coefficient, and the surface-wave dispersion relation are presented for superthin niobium films deposited on crystalline quartz. It is demonstrated that our theory is in excellent agreement with experimental results for the s-polarized reflectivity of the Nb/quartz system recently obtained by Alieva et al. [Phys. Lett. A 152, 89 (1991)].

A simple variational method has recently been used to obtain the bulk-plasmon dispersion relation in a metal. In the present work we investigate the eigenmodes of the valence electrons in a metal cluster considering a semiclassical version of the method presented by Andō and Nishizaki. As a variational function we consider the Slater determinant ‖φ〉 which is related to the Slater determinant ‖${\mathrm{\phi }}_0$〉, describing the ground state, by means of the unitary transformation ‖φ〉=${\mathit{e}}^{(\mathit{i}/\mathrm{ħ})\mathit{S}}$‖${\mathrm{\phi }}_0$〉, where S(x,p,t)=χ(x,t)+1/2[p⋅s(x,t)+s(x,t)⋅p]. We use a polynomial approximation to determine the dynamical fields χ(x,t) and s(x,t). It is shown that the eigensolutions satisfy the energy weighted sum rule and the cubic energy weighted sum rule. The spectrum of excited energies, as well as transition densities and currents are obtained for the sodium, aluminum, and silver.

The ab initio full-potential linearized-augmented-plane-wave method for a free-slab geometry was used to calculate the electronic structure and geometry of a clean ${\mathrm{TiO}}_2$ (110) rutile surface. Surface induced states were found in the density of states, such as an s-like surface state at -15 eV. Band bending states of width 0.5 eV appear just below the Fermi energy, in agreement with photoemission experiments. The positions of the atoms in the surface and subsurface layers and the corresponding change of Ti-O bond lengths were derived by total-energy minimization. In general, downward relaxations were obtained for which the fivefold-coordinated Ti experienced the largest relaxation of -0.180 Å, whereas the second most important relaxation effect, -0.156 Å, occurred for the surface O. The calculated Ti-O bond lengths are in very good agreement with experimental data for the ${\mathrm{TiO}}_2$ (100) surface. The calculated work function 6.79 eV compares favorably with the experimental result of 6.83 eV. Based on an extension of density-functional theory to excited states the valence- and conduction-band gap was calculated to be 1.99 eV, which is in reasonable agreement with the experimental gap of 2.6 eV when compared to the one-particle band gap of 0.65 eV.

The azimuthal-angle dependence and the frequency dependence of second-harmonic generation (SHG) measured at the interband transition region of a Ag(111) surface have been examined using a theoretical model developed by Sipe, Moss, and van Driel [Phys. Rev. B 35, 1129 (1987)]. The analyses of the experimental results provide strong evidence that the dramatic variation of the second-harmonic signal with frequency at 2ħω=3.87 eV is primarily caused by the sharp change of the dielectric function. The dispersion of the dielectric function affects both the azimuthal angle and the frequency dependence through different radiation efficiencies of the isotropic and anisotropic components of the generated second-harmonic field. The observed SHG features at 3.9 eV can be adequately accounted for by the presence of the interband transitions, and it is not necessary to invoke other surface electronic transitions for their assignments.

The spectral momentum density of the valence band of arc evaporated amorphous carbon has been measured by (e,2e) spectroscopy with significantly improved energy resolution relative to earlier studies. The valence band has been studied over a range of momenta from 0 to 1.6 a.u. with a resolution of 0.15 a.u. and over a range of binding energies from 9 eV above to 35 eV below the Fermi energy with a resolution of 1.5 eV. As seen in earlier studies, two major peaks are observed in the spectral momentum density which previously have been associated with σ and π bands in graphite. A third feature in the spectra, a weak shoulder approximately 4 eV below the Fermi energy, is observed. A heuristic model is introduced based on the assumption that the spectral momentum density of evaporated amorphous carbon is an angular average of the spectral momentum density of graphite. The behavior of the strongest feature in the experimental spectra is described well by this model, but the other two features, which are in the energy range of the graphitic π, ${\mathrm{\sigma }}_2$, and ${\mathrm{\sigma }}_3$ bands, are poorly represented by the model. It is suggested that the poor agreement is due to rehybridization of these graphitic bands.

Low-dimensional systems have attracted a great deal of attention during the past decade. The lowered symmetry and coordination number give rise to new and interesting electronic and magnetic phenomena. Several ab initio numerical methods have been developed to calculate the electronic and magnetic structure of materials, but these calculations require a great deal of computational effort. In order to study complex systems it is of interest to be able to make use of simpler approximative methods. For transition metals such an alternative approach is provided by the tight-binding approximation. Within the conventional parametrized tight-binding approach the lowered dimensionality of surfaces gives rise to an unphysical filling of the surface d orbitals (sp-d charge transfer), which in turn gives rise to lowered surface magnetizations. Results obtained by applying the linear muffin-tin orbital (LMTO) method to a repeated sequence of slabs and empty spheres show the existence of a spillover coming from the s and essentially p surface orbitals, with the d-band occupation remaining nearly the same as in the bulk materials. We suggest in this work a simple way of parametrizing the tight-binding Hamiltonian in such a way that the characteristics observed in LMTO calculations are preserved and the simplicity of a tight-binding approach remains valid. This is obtained by only adding a new layer of orbitals on the surface in order to simulate the spillover. We compare in this contribution results for Rh, Fe, and Cu (001) monolayers and five-layer slabs obtained using LMTO and an unrestricted Hubbard Hartree-Fock Hamiltonian with the surface parametrization.

Monolayers (ML’s) of Ni and Co grow at equilibrium on {110} W and Mo substrates, with which they bond strongly and have large negative misfits, pseudomorphically until completion of the pseudomorphic monolayer. When excess atoms are deposited atop the pseudomorphic monolayer, it transforms to a misfitting, more or less close-packed (cp) monolayer with misfit dislocations at the monolayer-substrate interface. In nonequilibrium growth, the pseudomorphic-to-close-packed transition occurs in monolayer islands before completion of the pseudomorphic ML. The main objectives of the paper are (i) to show that the pseudomorphic ML is more stable than the close-packed monolayer and (ii) to develop an understanding of the pseudomorphic-to-close-packed transition. In paper I we introduce force models and derive expressions for (a) equilibrium zero-order energies of complete and incomplete misfit dislocations having displacement vectors equal to a lattice and half a lattice vector, respectively, and (b) equilibrium first-order energy for incomplete misfit dislocations (assumed to yield the lowest energy configuration), these two being the candidates for minimum energy of the close-packed monolayer. The numerical computation of parameters (using embedded-atom methods) and derived energies (including anharmonic pseudomorphic strain energy), needed to answer the relevant questions, are left for paper II.

The aim of this investigation is (a) to confirm by numerical quantification of analytical expressions in paper I, involving embedded-atom methods (EAM’s), that a pseudomorphic, rather than a close-packed, Ni monolayer (ML) on Mo{110}, is stable, and (b) to establish the physics of the pseudomorphic-to-close-packed transition. Of the two close-packed configurations identified by a rigid-model approach, the one with incomplete misfit dislocations is shown, using transformed-stiffness constants (intralayer interaction) and EAM-calculated Fourier coefficients (interlayer interaction), to be the stable one with an average energy per Ni atom of 0.297 eV after final relaxation. An EAM calculation, presumably accounting for anharmonicity at 26% of pseudomorphic strain and for substrate proximity, yields the lower value 0.258 eV for the average energy per Ni atom in the pseudomorphic monolayer, and thus confirms the relative stability of the pseudomorphic monolayer. The pseudomorphic-to-close-packed transition–apparently from a lower energy pseudomorphic to a higher energy close-packed configuration–has been explained on kinetic grounds: the rapid formation of the close-packed configuration by penetration of excess atoms under nonequilibrium conditions of growth. Under equilibrium conditions, individual excess atoms reach the ML periphery by surface migration to effect continued pseudomorphic growth.

A thermodynamic model of the rotational phase transition in solid ${\mathrm{C}}_{60}$ is presented. The phase transition which appears at 257 K is a first-order phase transition in which a competition occurs between entropy gain by rotation and energy gain by intermolecular attraction. The anomaly in the specific heat, and the entropy and enthalpy changes observed at the critical temperature are well reproduced in the present numerical results.

We study the response of the interacting electron gas in one-dimension (1D) to a local external potential. In the low-frequency limit the power-law singularities are essentially modified by backscattering effects which, in the case of zero forward scattering, result in the universal critical exponent depending only on the Luttinger-liquid interactions. The results obtained may be used to describe singular responses of the 1D boson chains.

We calculate the tunneling rate of a heavy particle in a metal, which can hop between two sites at ±a with a frequency rate Δ/(2ħ) and can interact with the conduction electrons with the screened potential G. The tunneling rate presents an asymmetric peak, centered at the reduced energy ${\mathrm{\Delta }}^{\mathrm{*}}$, that decreases with increasing G. The integrated tunneling rate has a finite value for the phase shift at the Fermi level ${\mathrm{\delta }}_0$=π/2. Our results point to no localization of the heavy particle in this model.

In recent years the self-interaction-corrected (SIC) density-functional formalism has been applied to many physical systems. We discuss the Kohn-Sham-like SIC equations within the framework of the density-functional theory. We prove that in the SIC method the kinetic energy of the noninteracting electron gas ${\mathit{T}}_0$[${\mathit{n}}_{\mathrm{\uparrow }}$,${\mathit{n}}_{\mathrm{\downarrow }}$] is evaluated erroneously. Since the kinetic energy is, in general, a large part of the total energy, this fact represents a crude approximation in the method and may lead to appreciable errors in the final results.

The theory of magneto-optic Kerr effects is formulated on the basis of classical interaction of light and matter in a coordinate system attached to the light wave. Induced polarization of the magnetic medium along the propagation axis of the electromagnetic wave is described in terms of an equivalent electric field and the effects are incorporated in the boundary conditions and Fresnel equations for the reflectivity tensor are modified accordingly. The results are given up to second order in the magnetic field.

The structural relaxation process occurring in ion-implanted a-Si is investigated through optical-absorption measurements. The strain reduction associated with the relaxation process leads to an increase in the band gap of the material and to sharpening of the band edge. Changes occurring in the sub-gap region of the absorption spectra point in the previously suggested direction of point-defect annihilation being associated with the relaxation phenomenon.

A measure of the Fe 3d partial density of states (DOS) in the valence bands of ${\mathrm{Cd}}_{0.89}$${\mathrm{Fe}}_{0.11}$Se single-phase single crystals has been obtained using synchrotron-radiation photoemission. Based on an analysis of the Fe 3d DOS spectrum in terms of a configuration-interaction calculation on a ${\mathrm{Fe}}^{2+}$(${\mathrm{Se}}^{2\mathrm{-}}$${)}_4$ model cluster, we find that the spectral intensities in the top 6 eV are predominantly due to transitions to ${\mathit{d}}^6$L final-state configuration, where L represents a ligand hole. On the other hand, those at 7–15 eV originate mainly from transitions to the ${\mathit{d}}^5$ configuration. The model calculation provides the Fe 3d Coulomb correlation energy of 4.5 eV.

Specific-heat measurements at low temperatures (0.07≤T≤1.6 K) are reported for Ge:Sb with donor concentrations between 0.6 and 4.0×${10}^{17}$ ${\mathrm{cm}}^{\mathrm{-}3}$ spanning across the metal-insulator transition at ${\mathit{N}}_{\mathit{c}}$=1.44×${10}^{17}$ ${\mathrm{cm}}^{\mathrm{-}3}$. The electronic coefficient γ of the specific heat C is roughly twice as large as expected for donor electrons with the Ge conduction-band effective mass. Below about 0.3 K an additional contribution to C is found which is attributed at least partly to exchange-coupled clusters of local moments. This is supported by measurements of C in magnetic fields where this contribution turns into a Schottky-like anomaly. Local moments persist also in the metallic region of Ge:Sb as previously inferred from magnetic susceptibility measurements. The dependence of the density of local moments on donor concentration is similar to that found previously in Si:P.