Systematic calculations on planar clusters and monolayers of lithium are performed to study geometries and stabilities of the clusters as well as their convergence behavior with increasing cluster size. The calculations are based on ab initio methods using density-functional theory within the local-spin-density approximation for exchange and correlation. The optimized nearest-neighbor distances ${\mathit{d}}_{\mathrm{NN}}$ of the ${\mathrm{Li}}_{\mathit{n}}$ clusters, n=1,...,25, of both hexagonal and square geometry increase with cluster size, converging quite rapidly towards the monolayer results. Further, the cluster cohesive energies ${\mathit{E}}_{\mathit{c}}$ increase with cluster size and converge towards the respective monolayer values that form upper bounds. Clusters of hexagonal geometry are found to be more stable than square clusters of comparable size, consistent with the monolayer results. The size dependence of the cluster cohesive energies can be described approximately by a coordination model based on the concept of pairwise additive nearest-neighbor binding. This indicates that the average binding in the ${\mathrm{Li}}_{\mathit{n}}$ clusters and their relative stabilities can be explained by simple geometric effects which derive from the nearest-neighbor coordination.

Second-order susceptibilities of small metallic particles are calculated by the quantum size effect, in which the wave functions of conduction electrons are constrained in a small sphere. The electric-quadrupole term, which plays an essential role in the second-harmonic generation in metals, varies inversely with different powers of the radius. The magnitude of the quadrupole susceptibility is enhanced by the quantum size effect as the particle size diminishes below 10 nm, and approaches the results evaluated from the free-electron-gas model as the particle radius goes to infinity. The enhancement of the second-harmonic reflectivity generated from small metallic particles, as calculated in this work, is in general agreement with experimental results.

Surface precipitates, resulting from the compositional deviation in the growth of compound films, can be annihilated by impingement of elements that are necessary for the precipitates to assimilate into the underlying stoichiometric compound, and such phenomena are found to occur at a substrate temperature above the melting point of the surface precipitates. We report on the annihilation of Pd-rich Te precipitates on PdTe(001) films through the impingement of a ${\mathrm{Te}}_2$ molecular beam. The Pd-rich Te precipitates, which are liquid above 620 K, are formed by the impingement of excess Pd atoms on the Pd-terminated PdTe(001)-(√3 × √3 )R30° surface. When the ${\mathrm{Te}}_2$ molecular beam impinges onto the surface at above 620 K, ${\mathrm{Te}}_2$ molecules are captured on the Pd-terminated (√3 × √3 )R30° surface and the liquid precipitates with the sticking probability of unity. The excess Pd liberated from the liquid precipitates moves across the surface to form a planar Pd-terminated PdTe(001) surface until the precipitate assimilates into the underlying stoichiometric compound. After the assimilation, the Te-terminated PdTe(001)-(1×1) surface appears. Thus the alternating impingement of Pd and ${\mathrm{Te}}_2$ beams onto the PdTe substrate repeats the process of generation and annihilation of the liquid precipitates; consequently precipitate-free epitaxial growth occurs. A possible growth mechanism is proposed.

The height-height correlation function H(r) from a scale-invariant surface is compared with the corresponding power-spectrum function W(q) using a variety of mathematical scaling functions. We show that for a non-self-affine surface with the roughness exponent α≥1, one of the asymptotic scaling relations, either H(r)∼${\mathit{r}}^{2\mathrm{\alpha }}$ or W(q)∼${\mathit{q}}^{\mathrm{-}2\mathrm{\alpha }\mathrm{-}\mathit{d}}$, can be violated. An inconsistency in the values of α also exists between H(r) and W(q) when α>0.9. The impact on the data analysis using different experimental methods is discussed.

The influence of anions on the underpotential deposition of Cu on Au(100) electrode surfaces has been investigated by in situ scanning tunneling microscopy. In pure ${\mathrm{Cu}}^{2+}$-containing ${\mathrm{H}}_2$${\mathrm{SO}}_4$ solution a (1×1) lattice is observed in the entire potential range and interpreted in terms of an increasing population of the energetically most favorable lattice sites by mobile Cu adatoms with weakly repulsive interactions. In the presence of ${\mathrm{Cl}}^{\mathrm{-}}$ anions a one-dimensionally incommensurate (n×2) structure is formed, coincident with a sharp peak in the current-potential curve. In this structure the adatoms are arranged in a quasihexagonal pattern with a spacing of 3.6 to 4.4 Å, which is close to the diameter of the ${\mathrm{Cl}}^{\mathrm{-}}$ ion. The difference in the observed adsorption behavior corresponds well to the different shape of the corresponding electrochemical current-potential curves.

The simultaneous emission of two electrons from W(100) after excitation by a normally incident primary electron is studied as a function of the primary electron energy (10–22 eV relative to the Fermi level) and emission angle. We found that the coincidence rate as a function of the primary electron energy had a low-intensity tail in the region below 18 eV with an energy threshold at 11.5–12 eV. The most probable angle between two scattered electrons was found to be about 80° at 20.6-eV primary electron energy. The maximum sum energy of correlated electrons is found to agree with the energy conservation law limit. Experimental results are in fair agreement with calculations based on a two-parabolic-band approximation.

When analyzing data from scanning tunneling spectroscopy (STS) it is generally assumed that the density of states and local density of states (LDOS) are simply related with the current-voltage (I-V) curves. This assumption has been derived phenomenologically within the framework of theories developed for a variety of situations. We continue here with the development of a theory of STS with concrete applications for semiconducting samples, going beyond the traditional Bardeen approximation. As in previous works, our present model provides a current that is a sum over tip states of semiconductor local density of states modulated by tip-dependent coefficients. For a tungsten tip and silicon sample, we find that these coefficients are nonzero only for energies close to the Fermi energy. This makes the I-V curve closely follow the LDOS. On the other hand, we have found that the conductance-voltage curves present features derived from the Van Hove singularities of the semiconductor.

A monolayer of ${}_{}{}^{57}{}_{}{}^{}\mathrm{Fe}$ is used as a probe layer in molecular-beam-epitaxy-grown Ag/Fe(001)/Ag(001) structures. Different Fe sites can be distinguished in the iron multilayer based on the isomer-shift (i.s.) and quadrupole-splitting (QS) values of different components in the Mössbauer spectrum. This makes it possible to determine the concentration of ${}_{}{}^{57}{}_{}{}^{}\mathrm{Fe}$ atoms in the different atomic Fe layers and obtain information about the growth processes for different regions of the Fe multilayer. It was found that after deposition of about 5-Fe ML the growth changes from a Fe multilayer formation mode to a quasi-layer-by-layer mode. The values of the i.s. and QS are the same for Fe atoms located at the top and bottom Ag/Fe interface. ${\mathit{H}}_{\mathrm{hf}}$ values, on the other hand, are different by 9 kOe at the top and bottom for a sample in which the upper Ag(001) layer is covered by another Fe(001) layer. This result may be due to a contribution to the top ${\mathit{H}}_{\mathrm{hf}}$ value from Ruderman-Kittel-Kasuya-Yosida-like spin-density oscillations that couple the ferromagnetic Fe multilayers across the nonmagnetic Ag spacer.

The dynamical excitations in a commensurate monolayer solid of ${\mathrm{N}}_2$ molecules adsorbed on graphite have been studied using molecular-dynamics simulations. Velocity and rotational correlation functions as well as coherent intermediate scattering functions and dynamical structure factors have been calculated at several temperatures and wave vectors. Effects of out-of-plane motions and the formation of pinwheel-like structural elements associated with the orientational-disorder transition are observed both in the equilibrium order parameters and in the time-correlation functions. The calculated temperature dependence of the Brillouin-zone-center frequency gap in the acoustic phonon branches up to the melting point agrees well with the experimental observations and gives strong support to a model of the substrate corrugation energy. Diffusive motion in the fluid just above the melting point has been demonstrated from the time dependence of the mean-square displacement of the molecules and the two-dimensional diffusion coefficient is estimated to be 2–3×${10}^{\mathrm{-}5}$ ${\mathrm{cm}}^2$/s at 75–80 K.

In order to calculate the surface response of a solid, it is modeled by a periodic modulated ionic density perpendicular to the surface and a translationally invariant density parallel to the surface. This model is designed for the purpose of physical systems like the layered crystals of the transition metal dichalcogenides. A Green’s function procedure in position space is presented which allows us to determine all the response quantities used in the standard jellium calculations. The algorithm is applied to a system with parameters adapted to ${\mathrm{TiS}}_2$. In particular, the normalized response field and density near the surface are calculated as functions of the photon energy for the range below the plasma frequency. The results are compared with the jellium case.

Ferromagnetic/ferromagnetic Ni/Co superlattices have been grown by molecular-beam epitaxy on ${\mathrm{Al}}_2$${\mathrm{O}}_3$(112¯0) substrates. X-ray diffraction and surface analysis techniques (low-energy electron diffraction and Auger-electron spectroscopy) show that the superlattices grow epitaxially, both Ni and Co in the fcc structure, and with the (111) direction normal to the surface. The superlattices are single crystalline in the growth direction, have an approximately square chemical modulation wavelength, and show four different types of in-plane domains. The lattice parameters measured in the (111) direction are consistent with Co and Ni maintaining their bulk lattice parameters. In the plane of the superlattice, the two materials are lattice matched.

We develop a hydrodynamic model for the calculation of second-harmonic generation (SHG) at the surface of conductors with arbitrary equilibrium electronic density profiles ${\mathit{n}}_0$(${\mathit{r}}_{\mathrm{\perp }}$). We apply our model to simple profiles and calculate the linear surface conductivity s(ω) and the nonlinear surface susceptibility tensor ${\mathrm{\chi }}_2^{\mathit{s}}$(2ω=ω+ω) for all ω, and we obtained the way they scale with the relevant bulk and surface parameters. The conductivity s(ω) displays a peak that corresponds to the dipolar surface plasmon at a frequency ${\mathrm{\omega }}_{\mathit{d}}$, which depends on the profile shape and width. The susceptibility (${\mathrm{\chi }}_2^{\mathit{s}}$${)}_{\mathrm{\perp }\mathrm{\perp }\mathrm{\perp }}$ has very large resonances at ${\mathrm{\omega }}_{\mathit{d}}$ and ${\mathrm{\omega }}_{\mathit{d}}$/2. The SHG efficiency is enhanced by several orders of magnitude at these resonances, suggesting that SHG spectroscopy might be a useful probe of surface collective modes.

Employing the stabilized-jellium model [J. Perdew et al., Phys. Rev. B 42, 11 627 (1990)] and using the bonding-valence and interstitial-density parameters [J. H. Rose and H. B. Shore, Phys. Rev. B 48, 18 254 (1993)], the binding energy and bulk modulus of simple metals are calculated. Using the same approach the bulk moduli of d metals are also calculated and they are in quite good agreement with experiment and with more sophisticated calculations.

After first arguing that linear-response theory is an appealing method for performing many molecular-dynamics calculations for nearly-free-electron metals, we make a careful derivation of the result written down by Cohen in 1962. In so doing we find a subtle mistake in Cohen’s result and extend the theory to include all many-body effects insofar as they can be included in the static dielectric function.

We calculate the maximum acceptor concentrations (acceptor solubilities) for BeSe, MgSe, CaSe, and BeTe doped with Li and Na acceptors. All of these compounds but BeTe have lower acceptor solubilities than ZnSe, which can explain the low-acceptor concentrations achieved to date in ZnMgSSe cladding layers used in ZnSe-based lasers.

In this paper two GaAs samples were investigated; one was a very pure sample grown by chemical-vapor deposition, the other was grown by molecular-beam epitaxy. The dominant optical transition in the high-purity sample was the donor-bound-exciton transition. Phonon sidebands associated with both the free exciton and the donor-bound exciton were observed. The active phonons were the longitudinal-optical (LO) and the transverse-optical (TO) modes associated with both the free exciton and the donor-bound exciton at the Γ point in k space; the TO mode from the donor-bound exciton at the X point, the LO from the free exciton at the L point and/or X point, and the longitudinal-acoustical mode from the free exciton at the X point and the L point. These phonon-coupled exciton transitions were not observed in the sample grown by molecular-beam epitaxy; however, phonon sidebands associated with free-to-bound and bound-to-bound transitions were observed. The transverse-acoustical mode at the X point was observed along with lower-energy modes that were attributed to impurity and defect structures in the material.

We calculate the low-temperature dependence of the electrical conductivity of a Kondo insulator (described within the mean-field large-N theory) in which a small number of rare-earth vacancies are present. Treating the vacancies within the coherent-potential approximation (which becomes exact at small concentrations), we obtain the resistivity as a sum of an activated part and an impurity-band part. For small enough concentrations the resistivity shows a peak as a function of inverse temperature with a maximum value that can be two orders of magnitude larger than the unitarity limit.

The conductivity between 0.05 and 3.0 K of five samples of insulating CdSe:In containing dif- ferent dopant concentrations 0.75${\mathit{n}}_{\mathit{c}}$<n<0.95${\mathit{n}}_{\mathit{c}}$ is consistent with two-parameter scaling, σ=${\mathrm{\sigma }}^{\mathrm{*}}$(n)f[T/${\mathit{T}}^{\mathrm{*}}$(n)], where the parameter ${\mathit{T}}^{\mathrm{*}}$ goes to zero and ${\mathrm{\sigma }}^{\mathrm{*}}$ tends toward a finite value ${\mathrm{\sigma }}_{\mathit{c}}^{\mathrm{*}}$ at the metal-insulator transition. This form implies a minimum metallic conductivity, and must break down as n→${\mathit{n}}_{\mathit{c}}$ if the transition is continuous.

Two-terminal resistance (R) measurements are used to investigate the low-temperature (T<1.4 K) Hall effect in GaAs/${\mathrm{Al}}_{\mathit{x}}$${\mathrm{Ga}}_{1\mathrm{-}\mathit{x}}$As quantum wires with a geometrical width smaller than 400 nm, in order to eliminate the possible invasive influence of (Hall) voltage leads in standard four-terminal (Hall-bar-type) configurations. Experiment shows that although R assumes a temperature- and length-independent value near integral filling factors, i.e., ${\mathit{R}}_{\mathit{x}\mathit{x}}$→0, the remaining Hall resistance R=${\mathit{R}}_{\mathit{x}\mathit{y}}$ exhibits a finite slope vs B unlike two-dimensional systems. The results suggest that long disordered electronic wires can exhibit an ordinary Hall effect in the vicinity of integral filling factors.

Through a quantum-mechanical analogue of the Ramakrishnan and Yussouff theory of freezing for classical systems, Wigner crystallization is studied here for a two-dimensional electron gas in a uniform neutralizing background. The prediction of the phase transition of the electron fluid to a Wigner solid corresponding to a triangular lattice at low electron density is shown to compare well with the corresponding computer-simulation results of Tanatar and Ceperley.