Results are presented concerning the electrical characterization of p-Si/${\mathrm{Si}}_{1\mathrm{-}\mathit{x}}$${\mathrm{Ge}}_{\mathit{x}}$/Si quantum well (QW) structures by admittance spectroscopy, capacitance measurements, and deep-level transient spectroscopy (DLTS). The capture and emission processes of holes in QW structures are theoretically analyzed for equilibrium and nonequilibrium conditions taking into account external electric fields as well as local electric fields induced by confined charge carriers. The temperature dependence of potential barriers at the QW and of the Fermi level determines the activation energy ${\mathit{E}}_{\mathit{a}}$ of the conductance across the QW. Admittance spectroscopy data of QW’s with x=0.25 and thicknesses in the range from 1 to 5 nm are in fair agreement with the proposed theoretical model. Hole emission from the QW region was studied by DLTS investigations on ${\mathit{n}}^{+}$p mesa diodes for QW’s with x=0.17.

We present a theoretical study of the effect of longitudinal-optical phonons and surface-optical phonons on the electronic states of conduction electrons and donorlike excitons in a spherical semiconductor quantum dot. The effect of the quantum confinement is described by an infinitely deep potential well in the framework of the envelope-function approximation associated with two nondegenerate bands. The charge-carrier–phonon coupling is treated within the adiabatic approximation. A variational calculation of the ground-state energy of the donorlike exciton is performed using a one-parameter trial envelope function, which includes electron-hole correlation effects. The results show that in the case of a donorlike exciton located at the center of the microsphere, the effect of the lattice polarization gives rise to a lowering of the absolute value of the energy for all values of the microsphere radius R. The Huang-Rhys factor S, which is a measure of the charge-carrier–phonon interaction and the extent of the charge-carrier density, has been determined as a function of R. For the donorlike exciton, S reaches a minimum value, respectively, equal to ${\mathit{S}}_0$=0.2 for ${\mathit{R}}_0$=12.7 nm and ${\mathit{S}}_0$=3.1 for ${\mathit{R}}_0$=2.3 nm in the case of CdSe and CuCl.

Carrier equilibration is essential for semiconductor laser operation since carriers are injected into the active region at energies higher than the effective band edges. While the threshold current of the laser diode can be minimized by quantum confinement in extra dimensions, the quantum effects in carrier capture and thermalization become more pronounced. In this paper, a full treatment of the carrier thermalization in electronic systems of reduced dimensionality for injection conditions relevant to laser operation is given based on ensemble Monte Carlo simulations and the fundamental limits on modulation bandwidth are discussed. Results are presented for quantum wells, quantum wires, and quantum dots. The peculiarities of the relaxation process in each structure are elucidated. It is shown that the relaxation times increase from ≊1 ps in bulk, to ≊10 ps in quantum wells, ≊50 ps in quantum wires, and ≊200 ps in quantum dots. Since the intraband relaxation times determine the extent of gain nonlinearities in semiconductor lasers, the maximum modulation bandwidth imposed by the intrinsic process of carrier relaxation can be calculated via the dependence of the optical gain on the photon density in the laser structure. For a graded-index quantum-well laser structure, the calculated value of the nonlinear gain coefficient is 1.1×${10}^{\mathrm{-}17}$ ${\mathrm{cm}}^3$ with the maximum -3 dB modulation bandwidth of 78 GHz for a 100-μm cavity length. The nonlinear gain coefficient in quantum wires is enhanced in comparison with quantum wells, although the differential gain may be increased by as much as an order of magnitude with the exact value of the modulation bandwidth dependent on the details of the design of the quantum wire laser.

X-ray-absorption fine-structure spectroscopy and x-ray diffraction have been used to study the local atomic structure of four different short-period ZnTe/CdSe (001) superlattices grown by molecular-beam epitaxy. The data indicate Zn-Se and Cd-Te coordination numbers greater than those expected for sharp interfaces in the superlattice. Although strain would appear to suppress interdiffusion, it is shown that the results are consistent with an interchange of the Zn and Cd atomic planes across the Zn-Se interface and Se and Te atomic planes across the Cd-Te interface. Theoretically, this switch of atomic planes has been shown to be energetically favorable by comparing the changes in chemical bond-formation energies and strain energies between a perfect superlattice and one with the above-mentioned switching of atomic planes. Accommodation of this strain is shown through bond bending and bond stretching at the newly formed interfaces.

The atomic geometry of the Sb/GaAs(110) interface has been determined by means of a grazing-incidence x-ray-diffraction study. The in-plane coordinates of the Sb atoms on GaAs(110) are found in perfect agreement with an epitaxial continued-layer structure and exclude an epitaxial on-top structural model. The chains of Sb atoms along the [11¯0] direction are characterized by a Sb-Sb bond length of 2.80 Å, close to the covalent Sb-Sb bulk value. The stable (1×1)-Sb/GaAs(110) interface presents a highly ordered structure, with a coherent domain size of 800 Å, larger than for the clean GaAs surface (∼600 Å), indicating an ordering role of antimony for the GaAs(110) surface.

The early stages of interface formation between ${\mathrm{CaF}}_2$ and Si(111) have been studied, in situ, by a combination of reflection high-energy electron diffraction, x-ray diffraction, and core-level photoemission. The results are combined with ex situ transmission-electron-microscopy measurements to show that the initial growth mode changes from Volmer-Weber to Stranski-Krastanow, depending on the substrate temperature. The crossover is correlated with a submonolayer transition from the Si(111)-(7×7) to a (3×1) reconstruction. This is accompanied by fluorine dissociation at the interface. Both initial growth modes can lead to a uniform ${\mathrm{CaF}}_2$ epilayer and subsequent growth on this surface is layer by layer. Using x-ray crystal truncation-rod analysis, we have examined the ${\mathrm{CaF}}_2$/Si(111) surface and interface structure. For films grown at temperatures above the (7×7)→(3×1) transition, the Ca atom in the CaF layer at the interface is located in a single ${\mathit{T}}_4$ bonding site. Finally, we have observed a structural transition at the interface from the as-grown structure to a (√3 × √3 )R30° reconstruction, which appears to be incommensurate. The dynamics of this transition and the possible mechanisms will be discussed.

The dispersion of several surface phonons on Si(111)1×1-As has been measured along the Γ¯ M¯ and Γ¯ K¯ direction using electron-energy-loss spectroscopy. In the acoustical surface-phonon range, good agreement with previous inelastic He-scattering data and calculations of the surface-phonon dispersion in the bond-charge model is found. Two new optical surface-phonon modes with frequencies of 355 and 520 ${\mathrm{cm}}^{\mathrm{-}1}$ at Γ¯ were followed throughout the surface Brillouin zone. The 355-${\mathrm{cm}}^{\mathrm{-}1}$ mode is found to be dipole active on Si(111)1×1-As, which is interpreted as being due to a decoration of this mode by the As atoms. The 520-${\mathrm{cm}}^{\mathrm{-}1}$ mode is identified with the Lucas mode.

We study how the influence of structural correlations in disordered systems manifests itself in experimentally measurable magnitudes, focusing on dc conductance of semiconductor superlattices with general potential profiles. We show that the existence of bands of extended states in these structures gives rise to very noticeable peaks in the finite-temperature dc conductance as the chemical potential is moved through the bands or as the temperature is increased from zero. On the basis of these results we discuss how dc conductance measurements can provide information on the location and width of the bands of extended states. Our predictions can be used to demonstrate experimentally that structural correlations inhibit the localization effects of disorder.

We report on the results of molecular-dynamics simulation studies of the classical two-dimensional electron crystal in the presence of disorder. Our study is motivated by recent experiments on this system in modulation-doped semiconductor systems in very strong magnetic fields, where the magnetic length is much smaller than the average interelectron spacing ${\mathit{a}}_0$, as well as by recent studies of electrons on the surface of helium. We investigate the low-temperature state of this system using a simulated-annealing method. We find that the low-temperature state of the system always has isolated dislocations, even at the weakest disorder levels investigated. We also find evidence for a transition from a hexatic glass to an isotropic glass as the disorder is increased. The former is characterized by quasi-long-range orientational order, and the absence of disclination defects in the low-temperature state, and the latter by short-range orientational order and the presence of these defects. The threshold electric field is also studied as a function of the disorder strength, and is shown to have a characteristic signature of the transition. Finally, the qualitative behavior of the electron flow in the depinned state is shown to change continuously from an elastic flow to a channel-like, plastic flow as the disorder strength is increased.

Exciton binding energies and optical response in quantum wells and in multiple quantum wells of GaAs/${\mathrm{In}}_{\mathit{x}}$${\mathrm{Ga}}_{1\mathrm{-}\mathit{x}}$As/GaAs are computed by a variational envelope-function procedure using the four-band model and the simpler two-band model. The effect of hydrostatic and uniaxial strain are considered from a virtual-crystal stress Hamiltonian. The physical parameters used for the alloy (${\mathrm{In}}_{\mathit{x}}$${\mathrm{Ga}}_{1\mathrm{-}\mathit{x}}$As) are obtained by interpolating the parameter values of pure materials (GaAs, InAs). We verify that band-offset values in the range of 0.30–0.45 give exciton states and optical response in good agreement with experiments. The light-hole exciton energy is also well reproduced by theory and results are very close to the continuum states of the well, its binding energy being due to the attraction of the electron, localized inside the well.

We present a detailed investigation of the coherent dynamics of excitonic wave packets composed of heavy/light-hole, electron miniband, and Wannier-Stark states in GaAs/${\mathrm{Al}}_{\mathit{x}}$${\mathrm{Ga}}_{1\mathrm{-}\mathit{x}}$As superlattices. Using transient degenerate four-wave mixing, we study the dependence of Bloch oscillations and heavy/light-hole beats on the applied field, miniband width, lattice temperature, and excitation conditions. Bloch oscillations are observed in samples with minibandwidths varying from 13 to 46 meV and at lattice temperatures up to 200 K. Under certain excitation conditions, we observe higher harmonics of the Bloch oscillation frequency. Spectrally resolved transient four-wave mixing experiments show unambiguously that quantum-mechanical interference is the origin of the oscillations. The experimental four-wave mixing traces are compared with a theoretical model based on many-level third-order density-matrix theory.

We have used the empirical pseudopotential method to study the electronic and optical properties of [001] Si quantum wires with (110)–(1¯10) square cross sections ranging from 4×4 to 14×14 monolayers (7.7×7.7 to 26.9×26.9 Å, respectively). We present energy levels, band gaps, oscillator-strength, and charge-density distributions. To understand the electronic structure of these systems we calculate their properties in a stepwise process, considering (1) wires with a free surface but without hydrogen and (2) wires with hydrogen chemisorption on the surface. We find that (i) in both cases, the band gap between bulklike states increases as the wire size is reduced (due to quantum confinement). However, (ii) hydrogen chemisorption acts to reduce the gap. (iii) Whereas the low-energy states near the valence-band minimum are effective-mass-like, the near-band-gap states with or without H on the surface can be decisively non-effective-mass-like. The lowest conduction states are pseudodirect, not direct. (iv) The calculated energy dependence of the transition lifetimes is too strong to explain the observed low-energy ‘‘slow’’ emission band in porous Si purely in terms of transitions in an ideal wire. However, an alternative model, which introduces a mixture of wires and boxes, can account for the experimental slope.

The electronic energy bands of (ZnSe${)}_{\mathit{m}}$(CdSe${)}_1$ and (ZnSe${)}_{\mathit{m}\mathrm{-}1}$(${\mathrm{Cd}}_{0.5}$${\mathrm{Zn}}_{0.5}$Se${)}_2$ (m=2,3) superlattices are calculated with the empirical nonlocal pseudopotential method taking into account the strain effect. The energy gaps, effective masses along different directions, the splitting of heavy- and light-hole bands, and spin-orbit splitting, etc. are obtained. Assuming that the (ZnSe${)}_2$(CdSe${)}_1$ and (ZnSe${)}_3$(CdSe${)}_1$ superlattices in the corresponding [(ZnSe${)}_{\mathit{m}}$(CdSe${)}_{\mathit{n}}$${]}_{\mathit{N}}$-ZnSe (m=2 or 3, n=1) multiple quantum wells as a whole are one kind of potential-well materials, we calculated the quantum energy levels of the electron and the heavy hole, and the exciton binding energies as functions of period number N of the superlattice. The variations of energy gaps of the two superlattices with temperature are calculated. Taking into account the confined energy in the quantum well and the exciton binding energy, the energy positions of exciton peaks are compared with the luminescence experiments of the multiple quantum wells. The variations with temperature are in good agreement, and the discrepancies between the calculated results and the experimental data are suggested as being due to the interface alloy formation.

We derive the nonlinear-optical coefficients of insulators using a fully quantum-mechanical theory of the electron-photon interaction. Using the minimal-couping interaction, we find an alternative interpretation for the absence of the vector-potential-squared (${\mathit{A}}^2$) term in the second-order response. Specializing to the case of up-conversion, we use the resulting expression, together with empirical tight-binding band-structure calculations, to compute the second-order susceptibility ${\mathrm{\chi }}^{\mathrm{}\left(2\right)\mathrm{}}$ of bulk semiconductors in the static limit and of semiconductor heterostructures at resonance. These nonlinear-optics calculations are based on the empirical tight-binding model without additional parameter fit beyond the band-structure model.

The calculated bulk values of ${\mathrm{\chi }}^{\mathrm{}\left(2\right)\mathrm{}}$(0) are much smaller than available experimental values. This provides an independent and sensitive test on the accuracy of the higher tight-binding conduction states and also reveals the need for more conduction states. Calculations for the heterostructures were done for GaAs/${\mathrm{Al}}_{\mathit{x}}$${\mathrm{Ga}}_{1\mathrm{-}\mathit{x}}$As and HgTe/${\mathrm{Hg}}_{1\mathrm{-}\mathit{x}}$${\mathrm{Cd}}_{\mathit{x}}$Te quantum wells. The ‘‘macroscopic’’ asymmetry of the quantum structures is reflected in the optical properties. Our calculations of intersubband ${\mathrm{\chi }}^{\mathrm{}\left(2\right)\mathrm{}}$ for the ${\mathrm{Al}}_{\mathit{x}}$${\mathrm{Ga}}_{1\mathrm{-}\mathit{x}}$As system confirm similar results obtained using one-band calculations and agree with recent experimental results. The results predict interesting differences with respect to the ${\mathrm{Al}}_{\mathit{x}}$${\mathrm{Ga}}_{1\mathrm{-}\mathit{x}}$As system due to the inverted nature of the HgTe band structure. We obtain a nonzero interconduction subband c1-c3 optical coupling in a symmetric HgTe quantum well, even at the zone center of the Brillouin zone. Our analysis of ‘‘symmetric’’ quantum wells predicts the possibility of second-order nonlinearity due to the tetragonal ${\mathit{D}}_{2\mathit{d}}$ symmetry. This possibility had been overlooked in previous discussions of nonlinearity based on the quantum-well picture. The contribution due to interconduction subband transitions in the symmetric wells is negligible. On the other hand, for the p-doped wells of both material systems, we predict a susceptibility that is two orders of magnitude larger than typical bulk values and, hence, should be amenable to experimental verification.

Interfacial roughness in ${\mathrm{Si}}_{1\mathrm{-}\mathit{x}}$${\mathrm{Ge}}_{\mathit{x}}$/Si superlattices grown by molecular-beam epitaxy on vicinal Si(001) surfaces has been investigated using low-angle x-ray-diffraction diffuse-intensity measurements and atomic-force microscopy. The vertically correlated-interfacial roughness is, in specific situations, highly anisotropic and oriented with respect to the substrate miscut. The lateral length scale of the roughness is many times greater than the average separation of the substrate steps. The presence of the anisotropy depends on Ge concentration. A thermodynamic model for the interface morphology is presented.

Occupied d-like surface states were observed for the close-packed surfaces of Tm(0001) and Yb(111) metal films grown epitaxially on W(110) and Mo(110) substrates, respectively. In the case of Tm metal, the surface state exhibits almost no dispersion and is extended over the whole surface Brillouin zone, while in the case of Yb metal, it is restricted to a narrow region around the Γ¯ point. The surface shifts of the 4f binding energies on these close-packed surfaces were determined to be 0.47±0.05 eV for Tm(0001) and 0.45±0.03 eV for Yb(111). Both values are substantially smaller than reported previously for polycrystalline surfaces of the two metals. State-of-the-art band-structure calculations, based on density-functional theory, predict a flat surface state slightly above the Fermi energy. There is no obvious explanation for this inconsistency at the present time.

Spectra of scattered ions and ion-induced electron emissions are used to investigate the electronic processes that occur during 0.4–5.0-keV ${\mathrm{Li}}^{+}$ bombardment of clean and alkali-atom-covered Al(100). The results show that Li 1s electrons are promoted during hard Li-Al collisions, but not during Li-K or Li-Na collisions. Consequences of the inner-shell electron promotions are evident in the inelastic loss, neutralization behavior, and electron emissions of the scattered Li particles. Spectra of scattered ${\mathrm{Li}}^{+}$ ions exhibit discrete-loss features, which are resistant to the usual increase in resonant neutralization that accompanies the deposition of alkali atoms on the surface. The loss features are due to Li 1s excitation via electron promotion, while the production of ions away from the surface via autoionization is responsible for their lack of response to alkali-atom adsorption. Spectra of ion-induced electron emissions confirm that ${\mathrm{Li}}^{\mathrm{*}}$(1s2${\mathit{s}}^2$) is produced and that it undergoes autoionization.

The detailed imaging process of subwavelength objects deposited on a planar surface is studied within the framework of a three-dimensional model of scanning near-field optical microscope. The model consists of a truncated pointed fiber approaching a planar surface on which a three-dimensional protrusion is deposited. For this geometry, Maxwell’s equations are solved exactly by applying the field-susceptibility method in the direct space. The technique provides precise evaluations of the physically relevant near and far fields. In order to refine the understanding of the imaging process of subwavelength objects, we present simulated images of low-symmetry protrusions for two different modes of polarization and as a function of the approach distance. These simulations show clearly that subwavelength surface defects induce confined optical near-field distributions that are directly related to the shapes of the objects. We conclude that the central problem of near-field optical microscopy is the optimal detection of the confined fields that are set up by the objects themselves.

We report on soft-x-ray photoelectron spectroscopy (SXPS) of a ${\mathrm{TiO}}_2$(110) surface during deposition of Fe in the monolayer regime. At low fractional monolayer coverages, the adsorbed Fe atoms are oxidized and Ti cations at the interface become reduced due to Fe adsorption. SXPS from shallow core levels and valence bands show that Fe starts to exhibit metallic character at a coverage of approximately 0.7 equivalent monolayers. Two well separated defect states appear in the band gap of ${\mathrm{TiO}}_2$ at iron coverages well below one monolayer. We use resonant photoemission to obtain information on the partial density of states, and we assign these defect states as being Fe-derived and Ti-derived states, located at the Fe and Ti sites, respectively. We suggest that a position change of oxygen is involved in the bonding of Fe on the ${\mathrm{TiO}}_2$(110) surface.

Core-level 4${\mathit{f}}_{7/2}$ photoemission spectra have been measured from a single, bifacial W crystal, which has both a flat W(110) and a vicinal, stepped W(110) [W(320)] surface. This procedure reduces uncertainties in the quantitative description of peaks in the spectra from W(320). Various analyses, including nonlinear least-squares curve fitting, show that the average surface core-level shift (SCS) for W(320) is only ∼-140 meV, compared to -310 meV for W(110) and that, at a maximum, only two of five terrace rows are isoelectronic to W(110) surface atoms. The absence of a large SCS for the step-edge atoms contradicts earlier interpretations of W(320) core-level spectra and departs significantly from expectations based on atomic-coordination models or tight-binding calculations of a bulk truncated surface. We suggest that systematic errors are responsible for the differences in reported core-level shifts for W(320). Implications of possible step-edge-driven atomic rearrangements are discussed.