All energy-band calculations of Gd show seven very-narrow minority-spin 4f bands just above the Fermi energy. We show that this resonance is an artifact of the local-spin-density approximation and suggest that it is the reason that all calculations have yielded electronic densities of states at the Fermi energy larger than the experimental value.

A model for the description of the x-ray spectra of an infinitely heavy particle embedded in a sea of conduction electrons is extended to the case of many, randomly distributed local core electrons. The mean-field thermodynamics, i.e., the exact solution in d=∞, is presented and the Green’s function of the heavy electrons is thoroughly studied. The exact asymptotic limits of this function for short and long, imaginary and real times are obtained for all interaction strengths. Two important observations are made. First, the zero-temperature function exhibits for long imaginary and real times the edge singularity independently of the ground state of the conduction electrons. Second, the finite-time approximate schemes fail to reproduce the exact long-time asymptotics at any temperature and the Green’s function of the local electrons undergoes a nonanalytic crossover from short to long times. The exact long-time asymptotics are available only via the Wiener-Hopf or related nonperturbative techniques.

We report x-ray-absorption measurements near the Yb ${\mathit{L}}_{\mathrm{III}}$ absorption edge of ${\mathrm{YbAl}}_3$, ${\mathrm{YbCu}}_2$${\mathrm{Si}}_2$, ${\mathrm{YbAgCu}}_4$, and related compounds. Our results suggest that the bulk Yb valence changes with temperature in a manner consistent with predictions of the single impurity Anderson model. We discuss the relationship between these results and recent valence-band photoemission results in the same compounds where the existence of temperature dependence has been disputed.

We study the metal-insulator transition and magnetic ordering in the Hubbard model using the particle-hole mapping. The analysis simplifies near the ferromagnetic limit. We find that the two-dimensional Hubbard model has a charge excitation gap at half filling for any finite U in this region on both the bipartite square lattice and the nonbipartite triangular lattice. In some cases, the system goes through a first-order phase transition to become a paramagnetic metal as ${\mathit{S}}_{\mathit{z}}$ is lowered. We also discuss the extension to the doped case. We find that in the large U limit, a single doped hole has a bandwidth of order of J rather than t at ${\mathit{S}}_{\mathit{z}}$=0.

Based on the idea of charge-neutral atoimc spheres we have calculated the electronic structure of ordered and disordered Cu-Zn, Ni-Pt, and Al-Li alloys using the linear-muffin-tin-orbital (LMTO) method and the Korringa-Kohn-Rostoker coherent potential approximation (KKR CPA) method in the atomic-sphere approximation (ASA), respectively. The equilibrium lattice constants and the formation energies of ordered alloys obtained with the LMTO-ASA method show that the calculations done with charge-neutral atomic spheres are closer to the experimental results than the conventional equivolume atomic-sphere-type calculations. In the case of disordered alloys, we find that charge-neutral atomic spheres are essential for the stability of these alloys within the KKR-ASA CPA method where the Madelung-type contribution is neglected. Our results clearly indicate that for disordered alloys any future implementation of a full-potential method within the single-site CPA should be carried out with charge-neutral cells rather than the Wigner-Seitz cells.

A series of Au-Cu alloys of various stoichiometries and order have been studied using x-ray photoemission spectroscopy (XPS) and x-ray absorption spectroscopy (XAS). Significant electronic changes are associated with alloying and with changes in the local environment. The Au 5d electron charge depletion has been determined independently from XPS core-level and Mössbauer isomer shifts and from x-ray absorption near-edge structure (XANES) measurements. Strong correlations were found using both methods in the parameters investigated here, especially in the elucidation of the charge-transfer mechanism. It is found that at the Au site there is a loss of d charge, upon alloying, which increases as Au becomes more dilute in Cu. This along with XANES evidence of d charge gain at the Cu site, conduction (primarily 6s) charge gain at the Au site, and overall charge gain at the Au site verifies the charge compensation model in which Au loses d charge but is overcompensated by a gain of conduction charge. This observation is in line with electronegativity (Au is more electronegative than Cu) and electroneutrality (the overall gain of charge is small) arguments. It is also found that there is a strong linear correlation between the 5d hole count at the Au site and (1) the Mössbauer isomer shift, (2) alloy d-band width, (3) ‘‘Au’’ apparent spin-orbit splitting in the alloy d band. By comparing the ordered and disordered species (which have different local environments) it is found that the charge transfer depends primarily on the local environment (coordination number of like and unlike atoms) and not on atomic separation. The charge transfer (d-band depletion) was larger in the ordered species and can be explained by the number of Cu and Au nearest neighbors as compared to the disordered phases.

Selenium-doped silicon quenched after heat treatment at 800 °C exhibits deep luminescence similar to that previously reported in sulfur-doped silicon. The corresponding Se-related complex gives rise to two deep bound-exciton (BE) photoluminescence (PL) emissions, which at low temperature have no-phonon (NP) lines at 955.5 meV (${\mathrm{Se}}_{\mathit{A}}^0$) and 772.2 meV (${\mathrm{Se}}_{\mathit{B}}^0$), respectively. A series of equidistant phonon replicas is observed with the same phonon energy as in the sulfur case. By increasing the temperature in the PL experiments associated BE excited-state lines are detected higher in energy with a separation of 10.5 and 12.4 meV, respectively, from the lowest NP lines. The metastability of both the S- and Se-related defects is investigated and discussed as well as the influence of the excitation photon energy on the intensity of the PL emission. The metastability is also confirmed by optically detected magnetic-resonance results, which also show that the (${\mathrm{S}}_{\mathit{A}}^0$) and (${\mathrm{Se}}_{\mathit{A}}^0$) lines correspond to an excited spin-triplet state. Both defects are found to have a low symmetry. When they are in the A configuration the symmetry is monoclinic I with one axis tilted approximately 20° from the [11¯1] axis in the (110) plane.

Room-temperature photoluminescence (PL) spectroscopy of II-VI semiconductor alloys has been proposed as a useful tool for determination of spatial variations in alloy concentration and impurity concentration. However, the nature of the PL recombination can affect the emission peak energy, as well as line shape. These effects in the band-edge emission of the compound II-VI semiconductor CdTe are reported here. PL measurements at 300 K were recorded from (1) highly doped n-type CdTe films grown by molecular-beam epitaxy, (2) undoped high-resistivity (ρ∼${10}^8$–${10}^9$ Ω cm) bulk CdTe, and (3) as-grown nominally p-type (ρ∼${10}^4$–${10}^5$ Ω cm) bulk CdTe. The dependence of PL emission intensity and line shape over a range of excitation from 0.003–70 W/${\mathrm{cm}}^2$ was studied. As the excitation power density was increased, a redshift in PL peak position was observed from all samples. PL peak position, intensity, and line-shape analysis show the highly excitonic nature of the radiative recombination in these materials, even though the free-exciton binding energy in CdTe is about 1/3 of kT at 300 K. In addition, the PL peak position can be more than 7 meV higher from n-type CdTe as compared to undoped CdTe.

We present a systematic experimental investigation of the photoluminescence of heavily doped, compensated Si:P,B. The spectra broaden as the doping levels are increased, as expected. The spectral weights of the no-phonon peak and the phonon-assisted transitions vary with compensation and with excitation intensity. In contrast with the simple superposition of high-level and low-level spectra found for uncompensated material, increasing the input power for heavily doped, compensated Si produces a strong, continuous shift in energy of the entire spectrum in an essentially rigid way. A logarithmic power dependence found for the two most heavily doped samples is consistent with an exponential band tail. Deviations from this behavior are found for two samples with lower concentration and compensation. We suggest that these deviations may signal the onset of a minimum in the density of states that occurs when the acceptor (donor) band begins to separate from the host valence (conduction) band as the dopant concentration is reduced.

Cathodoluminescence has been investigated in diamond films, prepared by microwave plasma assisted chemical-vapor deposition. Results are presented for various diamond films differing in crystalline structure and doping. In the energy range 2.0–2.7 eV a series of 13 sharp luminescence lines has been observed at temperatures below 200 K. In a tentative interpretation they are attributed to donor-acceptor pair recombination. The Coulomb energy depends on the discrete values for the pair separation in the expected manner including a Bohr radius correction for the three closest pairs. The relative intensities of the lines depend on the growth conditions and on the film texture. Likely candidates for the pairs are the nitrogen donor and the boron acceptor. Both impurities are contained in the present samples. The sum of the binding energies is 3.57 eV, and inserting ${\mathit{E}}_{\mathit{a}}$=0.37 eV for boron yields ${\mathit{E}}_{\mathit{d}}$=3.2 eV for nitrogen.

The symmetry of the EL2 center in n-type liquid-encapsulated Czochralski GaAs is investigated through numerical fitting of capacitance transients measured under uniaxial stress. From experimental data which superficially appears compatible with ${\mathit{T}}_{\mathit{d}}$ symmetry, we extract reproducible defect energy-level splitting under uniaxial stress and conclude that EL2 has ${\mathit{C}}_{3\mathit{v}}$ symmetry, supporting the ${\mathrm{As}}_{\mathrm{Ga}\mathrm{-}}$${\mathrm{As}}_{\mathit{i}}$ pair model with ${\mathrm{As}}_{\mathit{i}}$ weakly bound to ${\mathrm{As}}_{\mathrm{Ga}}$.

Thermally stimulated polarization (TSP) and depolarization current measurements in as-cut as well as in oxidized magnesium-doped aluminum oxide show a huge peak at around 265 K, another one at 245 K, and a broadband at higher temperatures. They are related to the presence of ${\mathrm{Mg}}_{\mathrm{Al}}^{\mathit{x}}$ centers (substitutional magnesium ions with a hole trapped at a neighboring oxygen ion) and originate from the thermally activated ionization of these centers, which release holes. The current-voltage curves and the TSP results related to the main polarization process show that this is due to the formation of blocking contacts at the sample-electrode interfaces.

Donor states in GaAs and InP at megagauss magnetic fields are studied experimentally and theoretically. The data allow us to investigate the magnetodonors for values of γ=ħ${\mathrm{\omega }}_{\mathit{c}}$/2 ${\mathrm{Ry}}^{\mathrm{*}}$≊10, unattainable until now in these materials. The observations are described by a theoretical model, which uses variational atomic-magnetic wave functions and accounts for the band nonparabolicity within an effective two-band k⋅p formalism. The theory describes well the absolute values and field dependences of energy shifts between the cyclotron resonance (CR) transitions and the corresponding magnetodonor impurity cyclotron resonance (ICR) transitions in both materials. Spin doublets of CR and ICR transitions are considered; they are shown to be related to a variation of the spin g value with the energy. It is argued that the spin-doublet splittings at high fields should be considerably larger in GaAs than in InP. This conclusion agrees with the observations.

The paramagnetic resonance absorption spectra of ${\mathrm{Bi}}_{2\left(1\mathrm{-}\mathit{x}\right)}$${\mathrm{Gd}}_{2\mathit{x}}$${\mathrm{Te}}_3$ grown by the Bridgman technique have been measured at room temperature. The magnetization measurements have been performed in magnetic-field intensities up to 65 kOe at temperatures of 1.8, 4.2, 16, and 25 K. The EPR data show that the ${\mathrm{Gd}}^{3+}$ site has a ${\mathit{C}}_{3\mathit{v}}$ symmetry, indicating a true substitution of ${\mathrm{Bi}}^{3+}$ atoms by ${\mathrm{Gd}}^{3+}$. The curve fitting of the magnetic-field dependence of the magnetization has been performed using a Brillouin function representing isolated magnetic ions. An extra term was introduced to represent the contribution of the pairs. The results indicate a weak antiferromagnetic coupling between the ${\mathrm{Gd}}^{3+}$ ions. This is similar to the one found in the case of ${\mathrm{Pb}}_{1\mathrm{-}\mathit{x}}$${\mathrm{Gd}}_{\mathit{x}}$Te. The bond lengths as well as the bond angles in the case of ${\mathrm{Bi}}_{2\left(1\mathrm{-}\mathit{x}\right)}$${\mathrm{Gd}}_{2\mathit{x}}$${\mathrm{Te}}_3$ and ${\mathrm{Pb}}_{1\mathrm{-}\mathit{x}}$${\mathrm{Gd}}_{\mathit{x}}$Te are compared. The similarity of the ${\mathit{J}}_{\mathit{P}}$/${\mathit{k}}_{\mathit{B}}$ values (-0.5 K) in the two systems is explained in terms of superexchange interaction via anions between ${\mathrm{Gd}}^{3+}$.

Experimental evidence for light-induced annealing of defects in a-Si:H has recently been reported. However the experimental data about the removal of the defects under illumination are so limited that considerable uncertainty remains about the form of the light-induced term in the defect rate equation and its thermal activation energy. We here analyze the kinetics of light-induced creation and annealing of metastable defects in terms of a dispersive model. Our analysis suggests that the light-induced annealing term is proportional to the product of defect density and the carrier density. The activation energy of the light-induced annealing prefactor is about 0.45 eV.

Heat-pulse and time-resolved-luminescence experiments are performed on bound excitons in dilute GaP:N. A prominent broad luminescent band called the V band, is shown to arise from both zero-phonon and phonon-assisted recombination of excitons from the lower bound exciton state. A quantitative explanation of the shape of the V band is given in terms of resonantly enhanced second-order phonon-assisted recombination involving the upper bound exciton state, taking due account of the averaging of the phonon-induced strain over the exciton wave function. From the time-resolved luminescence experiments we derive a time of 300 ns for exciton energy transfer to disturbed sites.

Low-temperature polarized photoluminescence and polarized photoluminescence-excitation spectroscopies are used to probe the electronic structure of ${\mathrm{GaInP}}_2$ epilayers grown by organometallic vapor-phase epitaxy. The epilayers were grown at various sample temperatures and exhibit different degrees of ordering. A systematic study of the absorption edge and luminescence line shapes provides evidence for a distribution of order parameters, η, within the more ordered samples of this study.

Single-crystal films of (001)${\mathrm{Zn}}_{1\mathrm{-}\mathit{x}}$${\mathit{A}}_{\mathit{x}}$Se (A=Mn, Fe, Co) (0≤x<0.14) grown by molecular-beam epitaxy on (001)GaAs have been studied by spectroscopic ellipsometry in the 3.5–5.5 eV photon-energy range. Using fits of the ${\mathit{E}}_1$ and ${\mathit{E}}_1$+${\mathrm{\Delta }}_1$ peaks with a standard analytic expression, we find that the linewidths increase with x for all samples, the energies increase with x for ${\mathrm{Zn}}_{1\mathrm{-}\mathit{x}}$${\mathrm{Fe}}_{\mathit{x}}$Se and ${\mathrm{Zn}}_{1\mathrm{-}\mathit{x}}$${\mathrm{Co}}_{\mathit{x}}$Se, and the energies decrease with x for ${\mathrm{Zn}}_{1\mathrm{-}\mathit{x}}$${\mathrm{Mn}}_{\mathit{x}}$Se. A model describing the effects of the sp-d exchange interaction on the L point band-gap energy is developed and applied. We find that the strength of the energy correction due to this interaction, which is proportional to the product of the square of the exchange integrals and the magnetic susceptibility of the material, is largest in Mn-doped and smallest in Co-doped ZnSe. While the sp-d exchange interaction model is consistent with the composition dependence of the ${\mathit{E}}_1$ and ${\mathit{E}}_1$+${\mathrm{\Delta }}_1$ band-gap energies in ${\mathrm{Zn}}_{1\mathrm{-}\mathit{x}}$${\mathrm{Mn}}_{\mathit{x}}$Se, it does not describe the behavior observed in ${\mathrm{Zn}}_{1\mathrm{-}\mathit{x}}$${\mathrm{Fe}}_{\mathit{x}}$Se and ${\mathrm{Zn}}_{1\mathrm{-}\mathit{x}}$${\mathrm{Co}}_{\mathit{x}}$Se. We show that an sp-d hybridization model, which includes the location of the energy levels of the magnetic impurity d levels, can account for the composition dependence of ${\mathit{E}}_1$ and ${\mathit{E}}_1$+${\mathrm{\Delta }}_1$ band-gap energies of all three materials.

Liquid and amorphous silicon networks are generated by a tight-binding molecular-dynamics method. The structural, electronic, and vibrational properties are calculated. The generated amorphous silicon networks are mostly fourfold coordinated with dangling bonds (3.2%) and floating bonds (12.5%). The calculated results are in excellent agreements with previous ab initio results. The results are further compared to the experimental results and those of other classical models.

Thin films of ${\mathrm{SrF}}_2$ and ${\mathrm{BaF}}_2$ grown on clean GaAs(100) surfaces have been studied by high-resolution photoemisison spectroscopy. Similar to the previously studied ${\mathrm{CaF}}_2$/GaAs(100) interface, there is evidence of cation-substrate bonding after annealing at temperatures near 600 °C. For ${\mathrm{BaF}}_2$, Ba was found to react with As at the interface with an associated loss of Ga and F. For ${\mathrm{SrF}}_2$, the Sr 3d core-level shift also suggests cation-substrate bonding. The valence-band maxima for GaAs and the respective alkaline-earth fluorides (${\mathrm{CaF}}_2$, ${\mathrm{SrF}}_2$, ${\mathrm{BaF}}_2$) were measured directly by linear extrapolation of the high-kinetic-energy edge of the respective valence bands. The resulting valence-band offsets were found to follow the same trend for coverages up to 2–4 monolayers.