We investigate stable structures and electronic properties for isolated ${\mathrm{C}}_{70}$ and nonisolated ${\mathrm{C}}_{80}$, by performing an ab initio molecular-orbital calculation based on a nonlocal density-functional formalism and a Harris-functional approximation. We first show the stable structure for ${\mathrm{C}}_{70}$ and give the explanation for the experimental inconsistency in structure by revealing the existence of a metastable structure. Concerning the seven isomers of ${\mathrm{C}}_{80}$ satisfying an isolated pentagon rule, it is found that the isomer with nearly ${\mathrm{D}}_2$ symmetry (${\mathrm{D}}_{2\mathrm{-}}$${\mathrm{C}}_{80}$) has a large bond energy and highest occupied molecular-orbital–lowest unoccupied molecular-orbital (HOMO-LUMO) energy gap which are comparable to those for isolated higher fullerenes. Therefore we propose that ${\mathrm{D}}_{2\mathrm{-}}$${\mathrm{C}}_{80}$ should be detectable in an experiment. Since ${\mathrm{C}}_{80}$ does not belong to the sequence of fullerenes with magic numbers, we discuss the problem of whether it is really unstable or not from the viewpoint of the electronic properties. Further, we consider the structures of ${\mathrm{C}}_{80}$ ions and show the great stability of the ${\mathrm{C}}_{80}^{6\mathrm{-}}$ with nearly ${\mathit{I}}_{\mathit{h}}$ symmetry which has a larger HOMO-LUMO gap than both ${\mathrm{C}}_{60}$ and ${\mathrm{C}}_{70}$.

The morphology of multiple-shell fullerenes is investigated by ab initio calculations using Yang’s O(N) method. It is found that for large single-shell fullerenes with ${\mathit{I}}_{\mathit{h}}$ symmetry, the spherical morphology has lower energy than that of polyhedrons. The formation energy per atom follows a simple scaling law. Including an estimate of intershell van de Waals interactions leads to the conclusion that spherical multiple-shell fullerenes are likely the most stable structure of large carbon clusters. These results are in good agreement with recent experiments.

The van der Waals interaction between two ${\mathrm{C}}_{60}$ fullerene molecules is computed using a discrete-dipole formalism in which ${\mathrm{C}}_{60}$ is viewed as a rigid cluster of 60 polarizable, interacting carbon atoms. The dispersion energy is obtained from the shift in zero-point energy of the ground-state dipolar fluctuations. The intermolecular attraction potential deduced from this approach is compared with the Girifalco potential [J. Chem. Phys. 96, 858 (1992)] which includes only averaged, pairwise additive interactions. The results are also compared with the potential of two continuum dielectric shells of finite thickness [Ph. Lambin et al., Phys. Rev. B 46, 1794 (1992)]. The dispersion energy between ${\mathrm{C}}_{60}$ and graphite is also computed with the same discrete-dipole molecular model and by treating the substrate as a continuous dielectric medium. The adsorption energy is compared with the result of summing discrete 1/${\mathit{r}}^6$ C-C interactions. Similar calculations are presented for a ${\mathrm{C}}_{60}$ molecule on semiconductors (Ge, Si, and GaAs) and insulators (LiF and MgO). Lastly, the substrate-mediated dispersion energy between two adsorbed molecules is discussed. The calculations include all orders of dipole-dipole interactions between the molecules and with the substrate. The comparison with pairwise additive dispersion energies brings out the non-negligible role of many-body contributions.

Carbon thin films exhibiting a different structure have been obtained using low-energy neutral-cluster beam deposition. The first-order Raman scattering of such materials exhibits some characteristic features related to the phonon density of states of diamond. A careful analysis shows that the electronic structure is not explained by a classical random network model such as Polk’s model and an alternative film structure is proposed. In addition, the correlation between the free-cluster properties and the film properties seems to confirm the existence of the fullerenes in the 20–32-atom size range.

A scheme of real-space calculation from first principles is developed for electronic structures in large disordered systems by combining the tight-binding linear-muffin-tin-orbital method with the forced-oscillator method. The scheme has the following advantages: (1) the electronic structure can be calculated self-consistently in large disordered systems of several thousands of atoms in a reasonable amount of computer time, which is linearly proportional to the system size; (2) the calculation can be performed exclusively within a particular energy range; (3) the energy resolution of the resultant electronic structure can be controlled explicitly. We show examples of the method applied to bcc and amorphous pure Fe systems.

The formation energies and electronic structures of three kinds of Si impurities in diamond are studied using the modified neglect of differential overlap method and the discrete-variational local-density-functional method with cluster models. The results show that it is difficult for all of the silicon impurities (interstitial, substituting, and pair-interstitial) to be formed in diamond, and that there is local density of states within the band gap for the interstitial silicon and the pair interstitial, but not for the silicon substitution impurity in diamond. It is suggested that the silicon impurities in diamond destroy locally the ${\mathit{sp}}^3$ hybrid of diamond, which results in changes in some intrinsic properties of diamond.

The tight-binding method is used to calculate the electronic densities of states of various models of a-C. Both models generated using Tersoff’s potential and quantum mechanically derived forces show π and ${\mathrm{\pi }}^{\mathrm{*}}$ states in the σ gap in agreement with experimental data. The former models, however, have an additional large peak at the Fermi level, which is shown to derive mainly from misoriented ${\mathit{p}}_{\mathrm{\pi }}$ orbitals and dangling bonds. The models do not support the presence of large aromatic ring structures. We find that embedding the local π electron systems of small clusters in the rigid ${\mathit{sp}}^3$ network leads to reduced level splittings and is thus consistent with the observed small optical gaps ≤2.5 eV.

We report on the optical band gap of the filled tetrahedral semiconductor LiZnAs [viewed as a hypothetical zinc-blende (ZnAs${)}^{\mathrm{-}}$ lattice partially filled with He-like ${\mathrm{Li}}^{+}$ interstitials]. The band-structure nature of LiZnAs is found to be direct with a forbidden gap of 1.51 eV at 300 K. The fundamental absorption edge is shifted to the higher energy by 100 meV at 77 K. In the photoluminescence study at 12 K an emission associated with a donor-acceptor pair is observed at around 814 nm (1.52 eV), indicating that the estimated band gap is a reasonable value, but the band-edge emission is not found. We suggested that the larger band gap of LiZnAs, as compared to that of GaAs is attributed to the higher ionicity.

Overlapping bulk and two-dimensional excitons have been observed in modulation-doped heterostructures. Their differing behavior in a magnetic field allows them to be studied separately. Both free and bound excitons, associated with the two-dimensional electron gas, have been observed in the photoluminescence spectra of these structures. These excitons show a reduced binding energy due to screening and are found in a region of the structure where band-gap renormalization occurs due to many-body interactions. The two-dimensional excitons are identified from measurements conducted in applied magnetic fields.

We report the observation of above-band-gap radiative recombination from highly doped n-type CdTe:I epilayers grown by molecular-beam epitaxy. Radiative emission involving phonon absorption processes was observed at sample temperatures up to 30 K with excitation power densities as low as 300 mW/${\mathrm{cm}}^2$ using conventional phase-sensitive signal detection. The momentum-conserving phonon absorption photoluminescence (PL) process in the epilayers involves excitons with an average kinetic energy of 5 meV. Even though phonon emission bands of donor-related recombination have not been observed in CdTe, we see a clear indication of phonon absorption PL due to coupling with donor states. The dependence of PL intensity on incident excitation power density for near-band-edge and above-band-gap emissions reveals that phonon absorption processes increase at a slightly faster rate, indicative of the strong electron-phonon coupling in CdTe.

We present a self-consistent field theory for the infrared absorption coefficient of an array of narrow antiwires in a perpendicular external magnetic field. A detailed study is made of the way in which the collective mode changes from a cyclotron mode when the confining potential is weak to tunneling coupled modes for intermediate antiwire potential strength and then to edge and one-dimensional lattice magnetoplasmon modes for strong potentials. Our numerical calculations show that at low magnetic fields, there is appreciable electron tunneling between quantum wires. However, as the magnetic field is increased, the electron tunneling is suppressed. The suppression of electron tunneling between wires is observed when the tunneling coupled modes emerge into cyclotron modes in the strong magnetic field regime. The edge mode excitation energy oscillates as a function of the electron density. These oscillations corespond to soft or hard potential walls for which the electron states are extended and localized, respectively.

The quantum transport of low-dimensional systems of electrons is studied with generalized-Langevin-equation methods. The non-Markovian behavior caused by geometrical confinement and by dissipative mechanisms is analyzed. The ballistic conductance quantization of waveguides is shown to be caused by the long-lived correlations existing in the device. Diffusive boundary scattering is shown to be a general consequence of lateral quantization, and to cause phase-breaking effects when it interacts with elastic scattering centers.

We study theoretically the electronic structure of periodically Si-δ-doped GaAs subject to a homogeneous electric field applied along the growth direction. The space-charge potential due to δ doping is obtained by means of the Thomas-Fermi approach. Analyzing the change in the density of states in the superlattice introduced in the electric field, we observe a set of equally spaced sharp peaks corresponding to Stark-ladder resonances. Intrinsic broadening of resonances turns out to be smaller than the level spacing in the whole range of the electric field we consider. We use the inverse participation ratio to evaluate the spatial extent of electron wave functions, and we find that the Stark-ladder spectrum is related to a strong-localization regime at high field.

The Hall coefficient and the dc resistivity were measured in the basal plane of stage-2 ${\mathrm{SbCl}}_5$ intercalated graphite between 4.2 and 300 K. The Hall coefficient ${\mathit{R}}_{\mathit{H}}$ during the first cooling of a sample is 0.57×${10}^{\mathrm{-}8}$ ${\mathrm{m}}^3$ ${\mathrm{C}}^{\mathrm{-}1}$ and is from the carriers in the two basic graphitic energy bands. ${\mathit{R}}_{\mathit{H}}$ is larger and has a maximum between 100 and 150 K in subsequent coolings. The change is attributed to the truncation of the basic Fermi surface by the periodicities of the in-plane intercalant structures. The resistivity at room temperature is (8±2)×${10}^{\mathrm{-}5}$ Ω cm. It decreases as the temperature is lowered from the change in electron-phonon scattering. There is also a large decrease of resisitivity in a small temperature range during the first cooling of some samples. This is attributed to an initial change from disorder to order of the in-plane structure. The resistivity anisotropy is 1.6×${10}^4$.

The argument derived in S. Ghosh and V. Kumar’s paper [Phys. Rev. B 46, 7533 (1992)] for direct evidence for the negative-U nature of the DX center in ${\mathrm{Al}}_{\mathit{x}}$${\mathrm{Ga}}_{1\mathrm{-}\mathit{x}}$As is based on a wrong interpretation of their experimental result since they have not taken into account the very small capture rate of the DX center at 50 K. Their conclusion is therefore not convincing.

The hydrogen-desorption process from Si(100) surfaces quenched from gas-source molecular-beam epitaxy using silane has been investigated by the temperature-programmed-desorption method for the hydrogen coverage 0.06<Θ<0.97. An order-plot analysis indicated a reaction order of 1.59, which is in contrast to the first-order desorption kinetics reported for the monohydride phase of Si(100) surfaces formed by atomic hydrogen adsorption. A model calculation involving both first- and second-order desorption kinetics was performed, which successfully reproduced the intermediate reaction order. The presence of atomic roughness on the grown surface is suggested to inhibit formation of paired monohydride dimers, the precursor state for the desorption, by either limiting the migration of surface hydrogens or reducing the number of surface dimers.

We study the Coulomb drag between two spatially separated electron systems in the quantum Hall regime. At a fixed temperature, the drag is peaked near the transitions between plateaus. The temperature dependence of the drag is a probe of the dynamics of the Hall system in a frequency and wave-vector regime inaccessible to transport measurements in a single layer. Most strikingly, in the transition regions it directly measures an exponent $\eta$, discussed by Chalker, that is characteristic of the fractal structure of the critical eigenstates.

We examine the low-temperature magnetotransport properties of a lateral semiconductor quantum dot defined electrostatically on a GaAs-${\mathrm{Al}}_{\mathit{x}}$${\mathrm{Ga}}_{1\mathrm{-}\mathit{x}}$As two-dimensional electron-gas (2DEG) heterostructure. We depopulate the point contacts to form tunnel barriers, at a series of fixed gate biases, by varying the magnetic field. With both dot and ungated 2DEG in the ν=2 filling-factor regime we observe periodic oscillations in the magnetoresistance resulting from single-electron magnetodepopulation of the dot. We substantiate the Coulombic nature of this characteristic, which is peculiar to high magnetic fields and relatively large (∼2 μm diameter) dots, by developing a theory for Coulomb-blockade-regulated magnetoconductance oscillations. Based on our theory, and on self-consistent calculations of our device classical capacitances and potential profile, we predict an oscillation period which is in reasonable agreement with the observations.

Effects of in-plane magnetic fields on the channel resistance of double quantum wells are studied for the case where impurities are introduced only in one of the wells. The resonant increase of the channel resistance caused by the gate-induced resonant coupling of two levels is found to be quenched by an in-plane magnetic field and it is successfully ascribed to the field-induced mixing of the resonantly coupled states.

Using energies obtained from empirical-potential calculations, we study the relative stability between abrupt, ordered, and randomized interfaces in thin-layer superlattices. For most of the substrate parameter values, both ordered and randomized layers are more stable than the abrupt interface. The ordered geometries have even lower formation enthalpies than random structures. We examine the implications of these results in view of experimental studies of annealing and interdiffusion.