The self-consistent cluster-embedding method is discussed theoretically. A definition of the total energy for an embedded cluster has been introduced. The method has two advantages. (i) It can describe both localized and band properties, including their excitations. (ii) It can give a good description of the magnetic properties for both spin-ordered and spin-disordered states. The electronic structure of NiO is studied using a high-quality basis set to calculate the electronic structure of a small embedded cluster and an antiferromagnetic insulating ground state is obtained. The picture has both localized and band properties. A small energy gap separates the unoccupied and occupied nickel 3d orbitals which are well localized. Each 3d orbital is attached to a particular nickel ion. Below the 3d levels are two diffuse oxygen 2p bands, and above the 3d levels are oxygen 3s, nickel 4s, and oxygen 3p bands. Experimental data concerning photoemission and optical absorption can be interpreted naturally. The spin magnetic moment of the nickel ion is calculated correctly. The simulation of the spin-disordered state shows that NiO remains as an insulator in the paramagnetic state. The Néel temperature of NiO is calculated directly to give a reasonable result. The Hubbard U parameter for nickel 3d electrons is estimated. The calculation shows that the excited nickel 3d electrons are also well localized and the overlaps are less than 4.5%. We propose the following: The overlap of the excited 3d electrons is too small to form a metallic band, but the overlap is sufficient for the ‘‘hole’’ to migrate through the crystal. In this sense, NiO is a charge-transfer insulator with a gap of about 4 eV (mostly from oxygen to nickel). The calculated small energy gap (about 0.5 eV) provides the activation energy of NiO which is supported by the experimental results.

Recent progress in the processing of conducting polyaniline (PANI) protonated with functionalized sulfonic acids [such as camphor sulfonic acid (CSA)] has enabled the fabrication of high-quality, homogeneous films with excellent surface quality. We present measurements of the reflectivity of such PANI-CSA films over a wide spectral range (0.006–6 eV) at temperatures from 10 to 300 K. The reflectance spectra of PANI-CSA are characterized by metal-like signatures in the infrared (IR), including high reflectance in far IR and a plasma resonance around 1.2 eV. However, the optical conductivity σ(ω) and the real part of the dielectric function, ${\mathrm{\varepsilon }}_1$(ω), are not typical of a Drude metal; σ(ω) is suppressed below the Drude extrapolation as ω→0 with a peak around 0.2 eV, and ${\mathrm{\varepsilon }}_1$(ω)>0 below 0.2 eV. These features arise from disorder-induced localization (Anderson localization) in PANI-CSA. We present a quantitative analysis of σ(ω) and ${\mathrm{\varepsilon }}_1$(ω) in terms of the localization-modified Drude model. The analysis indicates that the mean free path is approximately 7 Å, comparable to the structural repeat unit along the PANI chain, implying that PANI-CSA is a disordered metal on the metal-insulator boundary.

We apply the model of a partially ionized gas to calculate the electronic magnetic susceptibility of Cs and Rb vapor along the coexistence curve. The deviations from the Curie law as obtained in the experiments of Freyland [Phys. Rev. B 20, 5104 (1979) and J. Phys. (Paris) Colloq. 41, C8-74 (1980)] are due to the formation of spin paired dimers. The enhancement of the electronic magnetic susceptibility at about 2–3 times the critical density in the domain of the expanded liquid can be explained by the formation of higher clusters such as ${\mathrm{Cs}}_2^{+}$ and ${\mathrm{Rb}}_2^{+}$. The implications on the mechanism of the metal-nonmetal transition in expanded liquid metals are discussed.

Self-trapped triplet excitons (STE’s) on the ${\mathrm{Cl}}^{\mathrm{-}}$ or ${\mathrm{F}}^{\mathrm{-}}$ sublattices of SrFCl have been studied using quantum clusters embedded within point ions or self-consistently within shell-model ions. The unrestricted Hartree-Fock method of calculation is used to treat the clusters. Three possible STE’s are studied. When the in-plane [100] ${\mathrm{Cl}}_2^{\mathrm{-}}$ STE is displaced from an on-center position the total energy decreases and it becomes basically a one-center species with a hole on one ${\mathrm{Cl}}^{\mathrm{-}}$ ion and the excited electron on the incipient ${\mathrm{Cl}}^{\mathrm{-}}$ vacancy. The Cl-Cl bond length becomes 3.46 Å, reflecting the one-center character of this species. The ${\mathrm{Cl}}_2^{\mathrm{-}}$ out-of-plane STE behaves similarly with displacement in the [111] direction. The ${\mathrm{F}}_2^{\mathrm{-}}$ STE remains as a molecular species with displacement in the [110] direction. The hole becomes localized on the ${\mathrm{F}}^{\mathrm{-}}$ ion nearest the incipient ${\mathrm{F}}^{\mathrm{-}}$ vacancy and on which the excited electron becomes localized. There is a large barrier for electron-hole separation for all of these STE’s. The presence of an adjacent halide vacancy allows the electron and hole to separate into a ${\mathit{V}}_{\mathit{K}}$ center and an F center without going through an intermediate H center, as occurs in alkali halides.

A comparison of experimental results and predictions of an effective-medium-theory model by Mantese, Curtin, and Webb has suggested that while the electrical conductance, in the regime where a continuous metallic network exists, is dominated by conduction within this network, the resistance noise is dominated by intergrain tunneling. A critical examination of the assumptions made in their model indicates that this suggestion is correct if the high resistance nature of the overlap regions between the individual grains is not considered. It is shown that by the consideration of these regions (which is done here within the framework of effective-medium theories) one can account for the experimental data by a viable alternative which suggests that both the resistivity and the resistance noise are dominated by the conduction throughout the continuous metallic network.

We present first-principles calculations of the electronic structure of crystalline fullerite (${\mathrm{C}}_{60}$) and of several metal-doped ${\mathit{A}}_{\mathit{x}}$${\mathrm{C}}_{60}$ fullerides (A=K,Ca,Sr,Ba and x=3, and ${\mathrm{Ca}}_5$${\mathrm{C}}_{60}$). Whereas the alkali-metal-doped ${\mathit{A}}_3$${\mathrm{C}}_{60}$ phases are metallaic for A=K,Rb,Cs, the alkaline-earth-doped ${\mathit{A}}_3$${\mathrm{C}}_{60}$ fullerides with A=Ca,Sr,Ba are narrow-gap semiconductors. This corresponds to complete electron transfers to the ${\mathrm{C}}_{60}$ molecules and the formation of ${\mathrm{C}}_{60}^{6\mathrm{-}}$ hexa-anions. The ${\mathrm{Ca}}_5$${\mathrm{C}}_{60}$ phase is metallic and shows a strong hybridization of the Ca ions over the tetrahedral and octahedral interstices of the face-centered cubic lattice of the ${\mathrm{C}}_{60}$ molecules. Detailed calculations of the photoemission and inverse photoemission spectra are presented.

Using a plane-wave-expansion method, we obtained expressions for the effective dielectric constants of composite periodic dielectric materials as the long-wavelength limit of the photonic-band-structure problem. The effective dielectric constants of several classes of photonic-band-gap structures are computed and we found that in the long-wavelength limit, they can be isotropic, uniaxial, or biaxial, depending on the symmetry of the structure under consideration. We also found that the scalar wave approximation gives poor results for effective dielectric constants while the Maxwell-Garnett theory offers very good approximations in the low-filling-ratio regimes.

Gradient corrections have been shown to improve the accuracy of density-functional theory (DFT) when applied to homonuclear dimers, small molecules, and bulk properties of transition metals (such as Fe). A more thorough evaluation is needed before adopting these corrections for large-scale computations in solids. We investigate a broad range of different systems by using the most recently proposed gradient correction to DFT by Perdew and Wang, which is expected to give a marked improvement over earlier attempts. We find that when this correction is used to calculate the energy of atoms it gives better agreement with experiment than the local-density approximation (LDA). When applied to bulk properties, this correction gives results which are not consistently better and tend to overcorrect LDA results.

The dc electrical conductivity (σ) of activated carbon fibers (ACF’s) heat treated at temperatures (${\mathit{T}}_{\mathrm{HT}}$) ranging from 300 to 2500 °C has been measured from 4.2 K to room temperature, and the magnetoresistance of these same fibers has been measured for magnetic-field strengths up to 0.8 T at 4.2 K. With heat treatment below ∼1200 °C, the heat-treated ACF’s exhibit similar transport properties to those of the as-prepared ACF’s with no heat treatment, in that the electrical conductivities for these ACF’s all show the exp[-(${\mathit{T}}_0$/T${)}^{1/2}$] temperature dependence, characteristic of granular metallic systems. A modified Coulomb-gap variable-range-hopping conduction model is proposed to explain this temperature dependence in ACF’s. As ${\mathit{T}}_{\mathrm{HT}}$ reaches ∼1200 °C, an electronic transition takes place, as evidenced by an increase in the absolute value of σ at 300 K by almost two orders of magnitude, a drastic change in the temperature dependence of σ, a reversal of the dependence of σ on disorder, a sign change from positive to negative in the magnetoresistance, and the emergence of anisotropy in the magnetoresistance for the heat-treated ACF’s. Most of these changes in the transport properties are consistent with the metal-insulator transition observed in similarly disordered systems, signifying that nanoporous materials, with their porosity controllable by heat treatment, can be used to study electronic behavior in the strong localization regime as well as near the metal-insulator transition.

For the spherically averaged Hamiltonian approximation an accurate nonvariational calculation is reported of the energies of the one-dimensional bands split off from the edge of the valence band by the shear strain fields of screw dislocations in various cubic semiconductors. The calculation gives significantly greater binding energies for holes than previous variational calculations and also a different order of the first two levels when the ratio of light- and heavy-hole masses is less than 0.19. The calculation was performed on the basis of the effective-mass and deformation-potential approximations. This approach is justified because the resulting binding energies are much smaller than the energy gaps. The results may be used for the interpretation of various dislocation-related phenomena such as luminescence, absorption, microwave conductivity, and combined resonance (a kind of electric-dipole spin resonance).

The decay kinetics of isoelectronic bound excitons at hole-attractive defects in silicon is studied by means of time-resolved photoluminescence. From these measurements it is found that their radiative decay generally follows a multiexponential relationship due to fully unthermalized excited singlet (S=0) and triplet (S=1) states. In the case of the C(0.79 eV) line the intrinsic transient is described by two exponentials whereas in cases of the P(0.767 eV) and H(0.926 eV) lines a triexponential decay is found. From the temperature dependence of the time constants and amplitudes the energetic splittings of the singlet and triplet levels, the relevant singlet-to-triplet transfer times, as well as the radiative decay times of the singlet levels are determined. In all cases consistent results are found only if nearly equal initial excitation densities for the singlet state and for each of the three triplet substates are assumed. In the case of the P line thermalization of the triplet substates around 14 K is observed. In the case of the C line the intrinsic transition is superimposed by nonradiative, sample-dependent, excitation losses which are explained in terms of excited electron tunneling to surrounding defects which also obscures the third exponential.

Defects in diamond films, produced by the filament-assisted chemical-vapor deposition of methane and hydrogen as a function of total gas pressure and substrate temperature, were investigated by electron-paramagnetic-resonance measurements. We found an isotropic g value (2.0028±0.0002) independent of growth conditions. The peak-to-peak linewidth increases with pressure from 3.5 to 5 G and the spin density decreases with increasing pressure and temperature and varies between ${10}^{19}$ and ${10}^{17}$ spins/${\mathrm{cm}}^3$. The line shape was found to be the superposition of two components, a narrower Lorentzian and a broader Gaussian, suggesting exchange narrowing and a nonuniform distribution of paramagnetic defects. The line shape was analyzed using Van Vleck’s theory of moments and a model regarding the distribution of the dominant paramagnetic center in diamond films was proposed and compared with structural studies on the same films. The temperature dependence of the spin-lattice relaxation rate, evaluated by the saturation method, was also investigated.

Both the L- and K-edge x-ray-absorption near-edge spectra for the five compounds Si(${\mathrm{OCH}}_3$${)}_{\mathit{x}}$(${\mathrm{CH}}_3$${)}_{4\mathrm{-}\mathit{x}}$ (x=0 to 4) are reported. The spectra of the two end members of the series, Si(${\mathrm{OCH}}_3$${)}_4$ and Si(${\mathrm{CH}}_3$${)}_4$, are compared to the spectra of their respective solid-state analogs, ${\mathrm{SiO}}_2$ and SiC. The L-edge spectra of gaseous Si(${\mathrm{OCH}}_3$${)}_4$ and solid ${\mathrm{SiO}}_2$ are qualitatively identical; while the L-edge spectra of gaseous Si(${\mathrm{CH}}_3$${)}_4$ and its solid-state counterpart, SiC, show strong similarities. MS-Xα calculations were performed for the two species Si(${\mathrm{OCH}}_3$${)}_4$ and Si(${\mathrm{CH}}_3$${)}_4$ for the Si 2p, 2s, and 1s regions and used to assign the spectra for both compounds. By comparison of molecular-orbital diagrams of Si(${\mathrm{OCH}}_3$${)}_4$ and ${\mathrm{SiO}}_2$, the first two transitions in the L-edge spectra of ${\mathrm{SiO}}_2$ have been assigned to transitions from the Si 2p to orbitals of ${\mathit{a}}_1^{\mathrm{*}}$ and ${\mathit{t}}_2^{\mathrm{*}}$ symmetry. In addition to ${\mathrm{SiO}}_2$, the L-edge spectra of the isoelectronic species ${\mathrm{PO}}_4^{3\mathrm{-}}$, ${\mathrm{SO}}_4^{2\mathrm{-}}$, and ${\mathrm{ClO}}_4^{\mathrm{-}}$ are also assigned in a similar manner. Compared to the L edge, the K-edge region of the gas-phase compounds is found to have less in common with the spectra of the solid-state analogs. In the K-edge region, peaks from the extended band structure and/or multiple scattering are evident in the solid-state compounds. These peaks are not present, as expected, in the gas-phase models or in the theoretical calculations on the gas-phase compounds. These multiple-scattering peaks are not seen in the L-edge solid-state spectra because of the strong shape resonances peaks in this region.

We present calculated average multiplet energies for spin-dependent strong-field ${\mathit{e}}^{\mathit{m}}$${\mathit{t}}_2^{\mathit{n}}$ configurations of ${\mathit{d}}^{\mathit{N}}$ ions in ${\mathit{T}}_{\mathit{d}}$ or ${\mathit{O}}_{\mathit{h}}$ point symmetry in terms of the Griffith parameters a,b,. . .,j. The calculations have analogies with the effective-crystal-field approach of Fazzio, Caldas, and Zunger for the spin-independent case [Phys. Rev. B 30, 3430 (1984)]. By using empirically determined parameters from the observed multiplet splittings the total-energy separations of the spin-dependent configurations can be derived. The expressions for the average energies are compared with those from a mean-field spin-polarized model, similar to the way in which Slater compared spin-dependent average multiplet energies of the free ion with spin-polarized hyper-Hartree-Fock values [Phys. Rev. 165, 655 (1968); 165, 658 (1968)]. Hence the relation of the single-electron and total energies derived in the mean-field model to the strong-field multiplet average energies is found. The results are also compared with other models found in the literature. As an example we discuss ZnS:Co and the chemical trends in the exchange splittings of the 3d orbitals in ZnS doped with transition metals. We also show how a consistent spin-dependent modification of the crystal-field multiplet theory can be introduced. However, the large number of disposable parameters to be determined in a semiempirical approach is a limitation of the procedure.

First-order Raman scattering of polycrystalline ${\mathrm{Ge}}_{1\mathrm{-}\mathit{x}}$${\mathrm{Si}}_{\mathit{x}}$ alloys is studied from T=300–900 K. The Raman shifts of the three optical phonons are modeled by considering thermal expansion and coupling to two phonons. The shift of the Si-Si mode near 500 ${\mathrm{cm}}^{\mathrm{-}1}$ has a temperature dependence similar to that of c-Si for x>0.7, but has a slower temperature variation for more Ge-rich alloys. The Ge-Ge mode near 300 ${\mathrm{cm}}^{\mathrm{-}1}$ has a temperature dependence similar to that of the optical-phonon mode in c-Ge for all compositions studied. The Raman shift of the Ge-Si mode has a dependence in between that of the Ge-Ge and Si-Si modes. Also, the Raman linewidths of the Si-Si and Ge-Ge modes have a temperature dependence similar to that of c-Si and c-Ge, respectively.

We present experimental results that demonstrate the nonergodic nature of charge transport in the insulating regime of indium oxide samples. These results include an anomalous field effect described in detail by Ben-Chorin et al. and persistent photoconductivity created by exposure to light. The similarity of the temporal dependence of the conductance after excitation due to a burst of light and that due to charging the sample by a nearby gate suggests that in both processes the electronic system is excited and the time it takes the system to reach thermal equilibrium is much longer than the Maxwell time. We offer a simple theoretical model that ascribes all of these effects to nonequilibrium transport phenomena peculiar to the hopping regime. It is argued that exciting a hopping system out of thermal equilibrium leads to a conductivity that is higher than in equilibrium. The excited state is long lived and similar in nature to that observed in the phenomenon of persistent photoconductivity of various semiconductors measured at low temperatures. The sluggish equilibration process of the electronic system is ascribed to the inhomogeneous nature of charge transport and to the slow energy relaxation which are inherent features of disordered solids.

In this paper, we use the coherent potential approximation to study the role quantum hopping plays on the Coulomb gap of insulating strongly correlated d=2 and d=3 dimensional systems. We find that substantial increase in the density of states at the Coulomb gap occurs only when the ratio between the hopping integral t and the gap width B exceeds a critical value. We estimate that the hopping integral corresponding to the experimental condition n<${\mathit{n}}_{\mathit{c}}$/3 satisfies t/B<0.05. For such values of t/B, quantum hopping brings about little change in the single particle density of states. The classical Coulomb gap therefore remains intact in both three and two dimensional systems, in contrast to a previous claim that the gap disappears for d=2 systems. The implication of these results on experiments on doped semiconductors is discussed.

We have investigated the photoluminescence, Raman, and infrared spectra of ${\mathrm{Cd}}_{1\mathrm{-}\mathit{x}}$${\mathrm{Mg}}_{\mathit{x}}$Te and ${\mathrm{Cd}}_{1\mathrm{-}\mathit{x}\mathrm{-}\mathit{y}}$${\mathrm{Mg}}_{\mathit{x}}$${\mathrm{Mn}}_{\mathit{y}}$Te bulk crystals with the zinc-blende structure. From the band-gap (${\mathit{E}}_{\mathit{g}}$) photoluminescence peak we deduce ${\mathit{E}}_{\mathit{g}}$=1.595+1.607x+1.592y eV at 10 K. The Raman spectra of the quaternary alloys exhibit a classic three-mode behavior with MgTe-like, MnTe-like, and CdTe-like LO-TO pairs of zone-center optical phonons, in decreasing order of frequency. Their composition dependence follows the modified random-element isodisplacement model. For very small x and/or y, the infrared absorption spectra recorded with a Fourier-transform spectrometer clearly reveal the local modes of ${\mathrm{Mn}}^{2+}$ and ${\mathrm{Mg}}^{2+}$. ${\mathrm{Mn}}^{2+}$ exhibits one and ${\mathrm{Mg}}^{2+}$ three local modes corresponding to their isotopic abundances; the latter have frequencies proportional to 1/ √${\mathit{M}}_{\mathrm{Mg}}^{2+}$ .

Molecular-dynamics simulation has been employed to investigate a previously proposed statistical-mechanical model [F. H. Stillinger, Phys. Rev. B 46, 9590 (1992)] for dimer buckling on the reconstructed Si(001) surface. Using model parameters suggested by low-temperature scanning-tunneling-microscopy observations, the calculations reveal a weakly first-order phase transition from a low-temperature c(4×2) pattern of buckled dimers, to a disordered high-temperature phase in which dimers have the untipped state as their most probable configuration. Alterations of input parameters establish that the thermal behavior of the generic model is sensitive to their specific values. Cases of strong first-order and of higher-order (Ising-like) phase transitions have also been simulated, as well as a case devoid of transitions.

We present a comprehensive experimental study of the tunneling and transport characteristics of split-gate ‘‘leaky’’ one-dimensional (1D) electron waveguides implemented in ${\mathrm{Al}}_{\mathit{x}}$${\mathrm{Ga}}_{1\mathrm{-}\mathit{x}}$As/GaAs heterostructures. In a leaky electron waveguide, electrons can tunnel out of the 1D channel through a thin side wall barrier into an adjacent 2D electron bath. A sharp peak and valley structure is observed in the 1D-to-2D tunneling current as the carrier concentration is modulated in the 1D waveguide through the field-effect action of the split gates. A semiclassical model confirms that the tunneling features originate from the 1D subbands in the channel.