We report a dielectric anomaly in several perovskites at intermediate frequencies 10≤f≤${10}^7$ Hz in the temperature range 400 °C ≤T≤800 °C. We have found this dielectric maximum in more than 100 samples of various shapes and textures. This anomaly is shown to arise from a Debye dielectric dispersion that slows down following an Arrhenius law. We have established a link between this dielectric relaxation and the conductivity. We propose that it can be achieved through a space-charge model that quantitatively agrees with the experimental results in the lower temperature range.

A gradientless correlation functional is derived within the Kohn-Sham density-functional theory (DFT) based on a spin-polarized pair correlation function of Colle-Salvetti type. The functional involves explicitly only the antiparallel spin correlation while parallel-spin correlation is taken into account indirectly in a manner providing self-interaction-free results. Combined with the local-spin-density (LSD) exchange, and with the nonlocal exchange functional of Becke, the resulting DFT schemes are tested on a set of atoms and small molecules. Compared to experiment, the calculated correlation energies of molecules are on average better than those obtained by other functionals employed in the deMon code [A. St.-Amant and D. R. Salahub, Chem. Phys. Lett. 169, 387 (1990)]: Becke-Perdew (88) and Perdew-Wang (91). For atoms the correlation functional gives slightly worse results than Perdew-Wang (91), but substantially better than the LSD approximation. The corresponding binding energies are on average slightly better than those obtained by the existing gradient-corrected schemes and in excellent agreement with experiment.

We present the microscopic derivation of a quasiparticle pseudo-Hamiltonian for an infinitesimally polarized electron liquid. The Hamiltonian is expressed in terms of suitably defined quasiparticle operators. Our approach is based on a canonical transformation which allows one to replace the bare Coulombic coupling between the interacting electrons with an effective interaction between quasiparticles in which collective charge and spin fluctuations are explicitly accounted for. The relevant matrix elements of the charge and spin-density operators enter our theory via linear-response functions: the charge response, the longitudinal and transverse spin responses, and the mixed charge-spin response. These susceptibilities are in turn expressed in terms of the appropriate many-body local fields. As a consequence our method can be seen as an attempt to satisfactorily include in a self-consistent manner the effects of the vertex corrections associated with charge and spin-fluctuations of the electron liquid. As a result useful expressions for the quasiparticle energy and the effective interaction between two quasiparticles are determined. These can, in turn, be employed in a microscopic determination of the parameters of the Landau theory of the Fermi liquid. The generalization of our results to a multicomponent system is also discussed.

We report a systematic study of the resistivity and magnetic susceptibility of pure ${\mathrm{V}}_2$${\mathrm{O}}_3$, the original Mott-Hubbard system at half filling, for pressures 0≤P≤25 kbar and temperatures 0.35≤T≤300 K. We also study (${\mathrm{V}}_{0.99}$${\mathrm{Ti}}_{0.01}$${)}_2$${\mathrm{O}}_3$ under pressure in order to elucidate the role of disorder on a metal-insulator transition in the highly correlated limit. Despite the low level of doping, we find that the two systems are very different. We observe a conventional collapsing of the Mott-Hubbard gap only for stoichiometric ${\mathrm{V}}_2$${\mathrm{O}}_3$; the Ti disorder stabilizes the long-range antiferromagnetic order and a magnetic Slater gap. Moreover, we discover different P-T phase diagrams for the two systems, with a decoupling of the charge and spin degrees of freedom at the approach to the T=0, pressure-driven metal-insulator transition in pure ${\mathrm{V}}_2$${\mathrm{O}}_3$.

The half-filled extended Hubbard model (containing nearest-neighbor interaction) is considered at strong coupling in fourth-order perturbation theory. The ground-state energies in the spin- and charge-density-wave phases and the resulting phase diagram are calculated in various dimensions: here d=1, d=∞, high dimensions (d≫1), and d=2,3 are discussed separately. In d=1 it is shown that, for U≳6, fourth-order perturbation theory leads to excellent agreement with existing Monte Carlo data. Second-order perturbation theory alone is valid only at unrealistically large values of U and V. In d=∞, one needs fourth-order perturbation theory to obtain any nontrivial contribution to the phase diagram at all. As a consequence the fourth-order corrections to the ground-state energy and to the phase diagram are large also in d=2,3. The formalism used to obtain the perturbative expansion is discussed in an appendix.

We consider the energy needed to form a spherical hole or void in a simple metal, modeled as ordinary jellium or stabilized jellium. (Only the latter model correctly predicts positive formation energies for voids in high-density metals.) First we present two Hellmann-Feynman theorems for the void-formation energy 4π${\mathit{R}}^2$${\mathrm{\sigma }}_{\mathit{R}}^{\mathit{v}}$(n¯) as a function of the void radius R and the positive-background density n¯, which may be used to check the self-consistency of numerical calculations. They are special cases of more-general relationships for partially emptied or partially stabilized voids. The difference between these two theorems has an analog for spherical clusters. Next we link the small-R expansion of the void surface energy (from perturbation theory) with the large-R expansion (from the liquid drop model) by means of a Padé approximant without adjustable parameters. For a range of sizes (including the monovacancy and its ‘‘antiparticle,’’ the atom), we compare void formation energies and cohesive energies calculated by the liquid drop expansion (sum of volume, surface, and curvature energy terms), by the Padé form, and by self-consistent Kohn-Sham calculations within the local-density approximation, against experimental values. Thus we confirm that the domain of validity of the liquid drop model extends down almost to the atomic scale of sizes. From the Padé formula, we estimate the next term of the liquid drop expansion beyond the curvature energy term. The Padé form suggests a ‘‘generalized liquid drop model,’’ which we use to estimate the edge and step-formation energies on an Al (111) surface.

We present an asymptotically exact solution of the infinite-dimensional Hubbard model at a special interaction strength ${\mathit{U}}_{\mathit{T}}$ corresponding to the strong-coupling Fermi-liquid fixed point. This solution is intimately related to the Toulouse limit of the single-impurity Kondo model and the symmetric Anderson model in its strong-coupling limit.

Using a semiempirical molecular-orbital technique, incorporating Car-Parrinello molecular dynamics, the structure and properties of the common aggregates of nitrogen atoms in diamond have been investigated. Nitrogen-vacancy complexes have been examined in the same way, showing that, as is observed, it is energetically favorable for the nitrogen to form aggregates and for the vacancies to be trapped at the nitrogen, whatever its state of aggregation. The activation energy for migration of the vacancy was found to be 2.4 eV, in close agreement with experiment, and the processes by which the nitrogen complexes trap vacancies were modeled. The mechanisms involved in the migration of the single nitrogen atom and its aggregation to form pairs are explored. It is found that a vacancy-assisted mechanism has an activation energy of 4.5 eV, and a direct atomic exchange of the nitrogen with its carbon neighbor has an energy of 6.3 eV. These values straddle the experimental estimate of 5 eV and would indicate why radiation damage enhances the nitrogen aggregation. The mechanisms involved in the further aggregation of nitrogen pairs to form well-defined clusters of four nitrogen atoms were also modeled, and an activation energy of 7–8 eV found.

Faraday and Voigt magneto-optic effects in undoped ${\mathrm{Cd}}_{1\mathrm{-}\mathit{x}}$${\mathrm{Mn}}_{\mathit{x}}$Te, a diluted magnetic semiconductor, modify nonlinear beam coupling during photorefractive two-wave mixing. Three magnetophotorefractive geometries using transmission holographic gratings are identified: (1) the longitudinal Faraday geometry; (2) the transverse Faraday geometry; and (3) the Voigt geometry. In the magnetophotorefractive phenomena, magnetic-field-induced circular and linear birefringence alters the polarization and phase of coupled modes propagating in the presence of electro-optic gratings generated by photoinduced space charge. Experimental and theoretical behavior for each of the geometries is presented as functions of incident beam polarization and magnetic-field strength. The magnetic field controls the magnitude and the direction of nonreciprocal energy transfer between the two beams.

The antisite-acceptor recombination process in as-grown Zn-doped InP has been studied by optically detected magnetic resonance (ODMR) at 16.6 GHz. A model of dynamic polarization of P nuclei at the excited antisites ${\mathrm{P}}_{\mathrm{In}}^{+}$ by partially unthermalized antisite electron spins is proposed to explain the difference in the intensity of the two hyperfine lines of the antisite resonance. The trends of this difference with microwave and laser power are also studied. We suggest that nuclear spin polarization may be responsible for similar phenomena observed in a variety of ODMR spectra of antisite resonances.

In this paper we present a general recursive scheme which provides the basis for an efficient calculation of the local Green’s function in the resolvent-matrix approach. To test the validity and the efficiency of the scheme, we applied it to calculate the bulk density of states of a crystalline GaP. Since the eigenstates in a crystalline solid are all extended, this case study should provide a stringent test for the efficiency of the convergence procedure. The result of the case study indicates that the general recursive relation eliminates the redundancy in the calculation of the local Green’s function, thus improving the efficiency of the convergence procedure. Furthermore, the case study also demonstrates that the convergence procedure can yield results as accurately as one desires even under the most unfavorable conditions.

Silicon codoped with aluminum and iron is investigated by electron paramagnetic resonance. Three spectra of trigonal, two of orthorhombic and two of monoclinic symmetry are detected and interpreted as being associated with iron- and aluminum-related defects. Most of the spectra are known but the analysis presented here supplemented by the observation of two additional spectra of trigonal and monoclinic symmetry, respectively, suggests alternative interpretations. We give evidence that the three trigonal spectra arise from three charge states of one and the same trigonal Fe-Al pair. Particularly, the trigonal spectrum, that can be described by a paramagnetic center with a spin S=1, is associated with the corresponding acceptor state. A calculation of effective g’ values shows that the ground states of the neutral orthorhombic, the neutral trigonal, and the positively charged trigonal Fe-Al pairs are orbital doublets ${(}^4$E${,}^5$E). Both monoclinic spectra belong to the same charge state of a complex consisting of two iron atoms and one aluminum atom. The resonances are consistently explained as transitions within the two energetically lowest Kramers doublets of a spin 5/2 system with a large zero-field splitting.

The paper is devoted to the statistics of semiclassical charge transport in semiconductors subject to high electric fields. Two main approaches are distinguished: one is mostly analytical and makes use of the notion of a mean free path, and it led to the so-called lucky-drift model; the other one is based upon Monte Carlo computer simulations and has the ability to include a realistic band structure. The present paper aims to show that in spite of the differences in wording and viewpoint, both approaches are fundamentally equivalent. We first review the lucky-drift model in its simplest form, and then present some basic features of the statistical approach used in numerical simulations of transport. On the one hand, statistical concepts are introduced into the lucky-drift model, and the notion of lucky-drift trajectories is criticized and replaced by that of energy autocorrelation. On the other hand, the results obtained in numerical simulations are reproduced in the framework of a generalized lucky-drift model tailored to allow for strong nonparabolicities. The statistical description of transport is applied to the determination of the rate of inelastic collisions undergone by a hot electron in the presence of an impact-excitable impurity. It is argued that such a rate is a better indicator of the fraction of carriers attaining a certain energy than the band-to-band impact-ionization rate. The previous expression of the impact-excitation rate, derived on the basis of the picture of lucky-drift trajectories, is revised, and a comparison with experiment is attempted.

It is demonstrated how the heavy-hole dispersion of bulk GaAs can be obtained along specific high-symmetry directions from hot-electron to neutral acceptor photoluminescence spectroscopy. This is done by analyzing the degree of linear polarization in terms of a detailed line-shape calculation to identify that part of the emission which is due to carriers with a particular wave-vector direction. A systematic plot of ${\mathit{E}}_{\mathit{h}}$(k) with sub-meV accuracy is thus made possible over a range of wave vectors in the central Γ valleys of direct-gap semiconductors. This method exploits the hot-electron momentum orientation by linearly polarized light, which is systematically investigated, together with the extent to which such momentum orientation persists in a hot-electron population either after one-phonon emission or when it is subjected to significant carrier-carrier scattering.

Intrinsic deep-defect-related recombination process has been investigated in a series of undoped hydrogenated amorphous silicon (a-Si:H) films grown under different deposition conditions. Steady-state photoconductivity (${\mathrm{\sigma }}_{\mathrm{ph}}$) was measured as a function of deep-defect density ${\mathit{N}}_{\mathit{d}}$, Urbach energy ${\mathit{E}}_{\mathit{u}}$, and dark Fermi energy ${\mathit{E}}_{\mathit{F}}$. It was found that ${\mathrm{\sigma }}_{\mathrm{ph}}$ strongly depends on these parameters while ${\mathit{E}}_{\mathit{F}}$ stays at the energy levels lower than 0.82 eV below ${\mathit{E}}_{\mathit{c}}$, but it is nearly independent of those if ${\mathit{E}}_{\mathit{F}}$ stays above 0.82 eV. A possible distribution of deep-defect-related recombination centers and recombination process has been suggested.

We have performed positron-lifetime and Doppler-broadening experiments before and after illumination of undoped semi-insulating GaAs with 0.7–1.5-eV photons. When the EL2 defect is transformed to the metastable state by 1.1–1.3-eV photons, we observe a metastable vacancy defect. Its concentration is proportional to the EL2 concentration. The metastable vacancy has smaller open volume than isolated Ga and As vacancies and it is connected to a negative charge. The vacancy is generated with an optical cross section of 2×${10}^{\mathrm{-}18}$ ${\mathrm{cm}}^2$ at the photon energy of 1.15 eV and it recovers at 120 K. As these properties are identical to those of the metastable state of EL2, we conclude that the observed vacancy is involved in its atomic configuration. The results are in good agreement with the isolated As antisite model for the structure of the EL2 defect.

We have measured the dependence on pressure of the low-temperature (10 K) direct exciton optical absorption of Ge up to 12.3 GPa. The sharp exciton peak at the direct energy gap (${\mathit{E}}_0$) of Ge is found to broaden significantly with increasing pressure. This effect, which is attributed to intervalley scattering via electron-phonon interaction, is most pronounced for pressures above ∼0.6 GPa, where the X valley becomes the lowest conduction-band minimum. From the pressure-induced exciton broadening we determine the Γ to X point intervalley deformation-potential constant ${\mathit{D}}_{\mathrm{\Gamma }\mathit{X}}$=2.2(3) eV/Å and an upper bound of ${\mathit{D}}_{\mathrm{\Gamma }\mathit{L}}$=4.5 eV/Å for scattering from Γ to the L valleys. The deformation potential ${\mathit{D}}_{\mathrm{\Gamma }\mathit{X}}$ of Ge is about 50% smaller compared to isoelectronic GaAs. This difference is attributed to the fact that interatomic matrix elements between s and d states of the Γ and X conduction-band minima as well as the d character of the X minimum are larger in GaAs.

The widely used GW approximation for the self-energy operator of a system of interacting electrons may, in principle, be improved using an approximate vertex correction Γ. We estimate Γ using the local-density approximation. We report the results of a comparable series of GW calculations for the band structure of silicon, in which such a vertex correction is (i) excluded entirely, (ii) included only in the screened Coulomb interaction W, and (iii) included in both W and the expression for the self-energy. We also discuss the symmetry properties of the exact vertex correction and how they may be retained in further improvements.

A generalized mathematical description of thermally stimulated luminescence (TL) and thermally stimulated conductivity (TSC) is presented in terms of a formulation that replaces the quasiequilibrium (QE) and kinetic-order approximations by two new functions Q(T) and P(T), respectively. These functions are related and can be described in terms of the physically meaningful processes of charge trapping, release, and recombination. With the Q(T) and P(T) functions and their relations to the reaction rates, we show that slow-retrapping (first-order) processes will be dominant in systems for which QE is satisfied, and that QE and non-first-order kinetics are incompatible. We develop general expressions for TL and TSC in terms of the Q(T) and P(T) functions and arrive at analytical solutions to the rate equations for the first-order case without the QE approximation and without the presence of Q(T) in the final equations. These general first-order equations are then parametrized and compared to curve shapes generated both from the Randall-Wilkins expressions and from numerical solutions to the rate equations. We show that the general first-order equations reproduce the numerical solutions over a wider range of parameter cases than do the Randall-Wilkins equations. In addition, we suggest a simple method for the extraction of activation energies from experimental TL and TSC peaks, independent of the QE approximation.

The energetics and band structures of the elemental semiconductor Si in two fourfold-coordinated clathrate structures have been studied. For the Si(34) and Si(46) structures we predict indirect band gaps near 1.9 eV, or 0.7 eV wider than that of Si in the well-known ${\mathit{sp}}^3$ diamond structure. We find that the energy differences between the clathrate structures and the diamond structure are about 0.07 eV, which is nearly three times smaller than those for more compact high-pressure structures of Si. The calculations are done using an ab initio tight-binding-like molecular-dynamics method, and further checked by the fully self-consistent plane-wave method. The possible origin of the small energy differences between Si in the clathrate and diamond structures is discussed in terms of strain energy. We compare the results for Si with those for similar structures involving carbon.