The effects of 1-MeV ${\mathrm{Si}}^{+}$ implantation on the optical band gap ${\mathit{E}}_{\mathit{g}}$ of pure (a-Si) and hydrogenated amorphous silicon (a-Si:H) have been determined. An identical decrease in ${\mathit{E}}_{\mathit{g}}$ of 0.46±0.02 eV upon ion implantation is observed for a-Si and a-Si:H. ${\mathit{E}}_{\mathit{g}}$ decreases because the valence and conduction bands broaden upon ion irradiation. Raman scattering shows that the broadening of the energy bands is accompanied by an increase in the average bond-angle variation. Hydrogenation results in a separate increase of ${\mathit{E}}_{\mathit{g}}$ by 0.22±0.02 eV for 11 at. % hydrogen.

A Monte Carlo path-integral simulation of crystalline silicon has been performed with the empirical interatomic potential developed by Stillinger and Weber. Several finite-temperature properties (potential energy, radial distribution function, and quantum delocalization) have been calculated and compared with experimental data and with classical simulation results. The employed quantum method leads to an adequate description of quantum effects such as zero-point vibrations, and reproduces the crossover to the classical limit at high temperatures. Deviations of the simulated vibrational energies from those derived from experiment are due to the limitations of the potential model, which overestimates the vibrational frequencies of the solid.

The structure of the low-coverage Al phases on the Si(100)2×1 surface was determined by scanning tunneling microscopy (STM) with varying bias voltage and bias polarity. Surface structures of 2×2, 2×3, and 2×5 phases formed at below 350 °C consist of Al-dimer lines perpendicular to the underlying Si-dimer rows. The STM images of the Al-dimer lines taken at positive and negative bias between 1 and 3 V agree with those of the theoretical simulation by assuming the parallel Al-dimer structure. Moreover, we found that filled states of Al-Si backbonds and empty states of Al-Al dimer bonds of the parallel Al-dimer lines are observed prominently at -3 and +1 eV at positions on the underlying Si-dimer rows and between Si-dimer rows, respectively. Atomic configurations at the ends of the Al-dimer lines combined with the underlying Si missing dimer defects are discussed on the basis of the observed STM images.

Phonons in parabolic-quantum-well structures fabricated with semiconductor alloys are studied. We calculate frequencies, displacement patterns, and densities of states for lattice vibrations in a (GaAs${)}_{\mathit{N}1}$(${\mathrm{Ga}}_{1\mathrm{-}\mathit{x}}$${\mathrm{Al}}_{\mathit{x}}$As${)}_{\mathit{N}2}$(001) superlattice where the GaAs layers are replaced by a mixed crystal ${\mathrm{Ga}}_{1\mathrm{-}\mathit{x}\mathit{s}}$${\mathrm{Al}}_{\mathit{x}\mathit{s}}$As with Al fraction ${\mathit{x}}_{\mathit{s}}$ varying quadratically with atomic layer s. The influence of disorder, composition, and the two-mode behavior of optical phonons are investigated within a random-element-isodisplacement model, generalized to layered structures with arbitrary composition profiles.

A simplified microscopic model of optical phonons in dimensionally confined structures is formulated and applied to calculate electron–optical-phonon scattering rates in GaAs/AlAs quantum wells. For this simplified model which circumvents performing a complicated ab initio calculation of the force constants at the interface, it is demonstrated that the resulting dispersion relation and scattering rates for electron–optical-phonon interactions agree very well with those obtained from detailed ab initio studies. It is also shown that for GaAs/AlAs structures, the macroscopic dielectric continuum model provides a good approximation to the scattering rate predicted by the microscopic models.

Hot-electron cooling dynamics in photoexcited bulk and quantum-well GaAs structures were determined using time-correlated single-photon counting of photoluminescence (PL) decay. Hot-electron cooling curves were generated from analyses of the time-resolved PL spectra. The time constant characterizing the hot-electron energy-loss rate, ${\mathrm{\tau }}_{\mathrm{avg}}$, was then determined, taking into account electron degeneracy and the time dependence of the quasi-Fermi-level. This analysis was also applied to earlier data obtained by Pelouch et al. with the same samples, but based on PL up-conversion experiments with <80 fs temporal resolution. Both sets of experiments and analyses show that the hot-electron cooling rate can be much slower in GaAs quantum wells compared (at the same photogenerated carrier density) to bulk GaAs when this density is above a critical value. This critical density was found to range from high ${10}^{17}$ to low ${10}^{18}$ ${\mathrm{cm}}^{\mathrm{-}3}$, depending upon the experimental technique; at the highest carrier densities, values of ${\mathrm{\tau }}_{\mathrm{avg}}$ for quantum wells were found to be many hundreds of ps.

Transverse electron focusing is reported for hot electrons. Differential transresistance measurements show anomalous behavior at high electron energies. Corresponding dc electron focusing data reveal that the differential characteristics arise from the electron energy dependences of the mean free path and specularity coefficient. The variation of mean free path with energy is derived from the experimental data.

The interaction between the step array and the terrace reconstruction of vicinal surfaces is evaluated by measuring with reflection high-energy electron diffraction the variation with the misorientation angle of the regularity of the steps and of the stability of the reconstruction. We demonstrate on GaAs≊(001) surfaces that the configurations for which the mean terrace width is a multiple of the reconstruction unit mesh, predicted by theoretical calculations to be energy minima, are a sharp optimum for both the step array and the reconstruction. A few minutes of arc away from this optimum, the steps modify the domain of existence and the coherence length of the reconstruction and the reconstruction modifies the step arrangement. This is clear experimental evidence that such configurations also minimize the mutual frustration of the two-ordered structures, step array, and terrace reconstruction.

We report a theoretical study of the noise properties of an ultrasmall resonant-tunneling structure. An analytical expression is derived for the noise power spectrum with the equation-of-motion method. We study the noise behavior as a function of both the bias voltage and the barrier characteristics, and find that the correlation between electrons induced by the Coulomb repulsion can suppress the noise significantly, and should be observable in a noise measurement.

We consider the ground and excited states of ${\mathit{D}}^{\mathrm{-}}$ centers, i.e., neutral donors ${\mathit{D}}^0$ trapping an extra electron, in the spin-singlet and spin-triplet states in GaAs/${\mathrm{Ga}}_{1\mathrm{-}\mathit{x}}$${\mathrm{Al}}_{\mathit{x}}$As quantum wells (QW’s) in quantizing magnetic fields B>6 T. Taking into account the effects of band nonparabolicity and magnetopolaron corrections, we derive the energies of the dipole-allowed transitions and their line strengths. Very good agreement (e.g., within 1–3 ${\mathrm{cm}}^{\mathrm{-}1}$ below the region of the resonant magnetopolaron splittings) with the experimental data of Huant et al. for the singlet transitions is found.

A model of a classical electron liquid without external disorder is applied to two-dimensional electrons in a strong magnetic field. Computer modeling gives a quantitative explanation for the recently observed gap in the tunneling current of a double-quantum-well structure. We find that both the Coulomb gap in the single-well density of states and the correlation of electron motion in the two wells are responsible for the tunneling gap. We show that the classical liquid model provides an accurate description of the low-temperature compressibility obtained from a magnetocapacitance experiment.

A self-consistent pseudopotential technique is used to calculate the electron potential for adatom, stepped, and atomically smooth Al (111) surfaces with a perpendicular applied electric field of 3 V A${\mathrm{̊}}^{\mathrm{-}1}$. We calculate rate constants for the field ionization of inert gas atoms as a function of their position above the surface, finding significant enhancement above the protruding surface atoms. We discuss the implications of our results for understanding the mechanism of image formation in the field-ion microscope.

Binding energies of metal clusters confined to two dimensions along different crystallographic directions have been calculated using self-consistent ab initio molecular-orbital theory. The results reveal a nonmonotonic dependence of the energies on the cluster size, with conspicuous maxima corresponding to close atomic packing. This behavior persists even when the interaction with a substrate is included.

In a recent paper [N. M. Rosengaard and H. L. Skriver, Phys. Rev. B 47, 12 865 (1993)], it has been demonstrated by means of the linear-muffin-tin-orbital theory in the atomic-sphere approximation that the energies for the fcc stacking faults may be reasonably estimated in a nearest-neighbor-interaction model from the energy differences between the fcc and hcp phases. It is now shown by ab initio pseudopotential calculations that the model itself remains essentially valid when abandoning the atomic-sphere approximation, although the errors of the model are intensified and the quantitative results may be considerably influenced by this abandonment.

Ionic pseudo-Hamiltonians, which replace core electrons in atomic calculations, are used to build a jelliumlike model, which describes electronic ground state and excitation properties of atomic clusters. As an application, we successfully describe the plasmon resonance in lithium clusters for which recent rexperimental data have shown the failure of the conventional jellium model.

Recent literature data for the thermal conductivity κ of single-crystal ${\mathrm{C}}_{60}$ are analyzed using a standard model. It is shown that in the intermediate range 100–250 K, below the rotational transition, κ can be described extremely well by a model taking into account phonon and point defect scattering only. The magnitude derived for the phonon-phonon scattering term agrees very well with an estimate from a simple model.

Ferrón obtained a reentrant layer-by-layer growth in his recent Monte Carlo study. His suggestion of a relation to the experimentally observed reentrant layer-by-layer growth for the system Pt/Pt(111) is wrong due to neglect of thermal adatom mobility and of diffuse He scattering.

In this work we analyze the two main points addressed in the preceding Comment. We show that the thermally induced Cu mobility is not negligible at the reentrant temperature. The importance of the large diffuse scattering cross section for He in connection with both our results and the similarities between reflection high-energy electron diffraction and thermal energy He atom scattering is also addressed.

We have investigated the decay of the charge-density-wave (CDW) polarization induced by the application of electric-field pulses of various lengths. Our results show the same features as the energy relaxation in low-temperature specific-heat experiments. We argue that both phenomena can be well described in the framework of linear-response theory. Though there are similarities to the aging effects in spin glasses, in the case of CDW’s these simply arise from the nonexponential nature of the relaxation. We show that in the case of charge-density waves there is no change in the barrier distribution when the system is allowed to age.

A mechanically controllable break junction is used to study the transition between vacuum tunneling and contact for three different metals. In the tunnel regime a faster-than-exponential behavior is seen at close distances, followed by a jump to contact, both interpreted as being due to metal bonding forces. We show that stable contacts of a single atom can be formed. For Cu, the conductance value for a one-atom contact is very close to 2${\mathit{e}}^2$/h. For Al this value is less well defined but of similar magnitude, while for Pt it is noticeably higher, implying that the electronic structure of the atoms is relevant to the one-atom conduction process.