We have studied the dynamic susceptibility in a quantum spin glass which is described by an Ising model in a transverse field, the latter introducing quantum tunneling in an otherwise classical problem. An effective single-spin Hamiltonian is obtained on the basis of an extant thermofield dynamic approach. The coupling to a dissipative heat bath is designed to affect thermal as well as quantal fluctuations and to yield Glauber kinetics in the absence of the transverse field. A resolvent expansion of the underlying time-development operator is then set up and the dynamic susceptibility calculated to leading order in perturbation theory. It is shown that the frequency of the peak of the susceptibility shifts towards higher values of ω, its amplitude is reduced and peaks are broadened as quantum effects strengthen. Our results are in qualitative agreement with the experimentally studied crystal-field-split system of ${\mathrm{LiHo}}_{\mathit{x}}$${\mathrm{Y}}_{1\mathrm{-}\mathit{x}}$${\mathrm{F}}_4$ by Wu et al.

We have succeeded in doping hydrogen into the ‘‘semiheavy’’ ${\mathrm{UPd}}_2$${\mathrm{Al}}_3$ up to a stable concentration of ${\mathrm{UPd}}_2$${\mathrm{Al}}_3$${\mathrm{H}}_{1.3}$. Magnetism is depressed in ${\mathrm{UPd}}_2$${\mathrm{Al}}_3$${\mathrm{H}}_{\mathit{x}}$: ${\mathit{T}}_{\mathit{N}}$ falls from 14.5 K with an initial slope of about 55 mK/%H, χ(1.8 K) initially remains relatively constant but rises substantially for x>0.34 to 21.4 memu/mole at x=1.30, while the effective moment inferred from the Curie-Weiss behavior in the susceptibility above 100 K falls monotonically with hydrogen doping from 3.39${\mathrm{\mu }}_{\mathit{B}}$ for x=0 to 2.82${\mathrm{\mu }}_{\mathit{B}}$ for x=1.30. Superconductivity is totally suppressed by approximately x=0.5. At the same time as the antiferromagnetic peak in the specific heat is being depressed to lower temperatures with increasing x, the specific heat divided by temperature data show more and more of an upturn below 10 K, reminiscent of the behavior seen in heavy fermion systems like ${\mathrm{CeCu}}_2$${\mathrm{Si}}_2$. Whether or not this upturn in C/T is magnetic in character was checked via measurements of the magnetization at 1.8 K, which showed no signs of saturation up to 5.5 T, and of the field dependence of the low-temperature (0.3 K≤T<1.5 K) specific heat in 14 T, which showed approximately a 25% decrease. Such a decrease in the specific heat with field is significantly smaller than that seen in, e.g., ${\mathrm{CeCu}}_6$ (65% suppression in 14.5 T) where magnetic correlations are thought to be strong, and is similar to the size effect seen in nonsuperconducting samples of ${\mathrm{CeCu}}_2$${\mathrm{Si}}_2$ (35% suppression in 10 T). However, whether this upturn in C/T in ${\mathrm{UPd}}_2$${\mathrm{Al}}_3$${\mathrm{H}}_{1.3}$ is indicative of an electron effective mass increased by a factor of 4 vs pure ${\mathrm{UPd}}_2$${\mathrm{Al}}_3$, or due to magnetic correlations/short-range magnetic order remains an open question.

It is pointed out that Kondo-effect rare-earth compounds with insulating ground states must, except for a small probability, be mixed valent. The lack of low-energy magnetic correlations and the optical properties are discussed in a mixed-valent model. The curious properties of FeSi are also discussed.

We have studied the influence of the interface roughness on the giant magnetoresistance (GMR) of polycrystalline Fe/Cr superlattices, grown by molecular beam epitaxy on polycrystalline substrates. The interface quality was changed systematically by varying the growth temperature between 0 and 400 °C and by the use of a Cr seed layer. X-ray-diffraction spectra, combined with resistivity data, show that a moderate step density at the interfaces can enhance the GMR, whereas interdiffusion and important interface roughness strongly suppress the GMR.

The infinite-range M-component models of quantum rotors and anharmonic oscillators coupled by random exchange interactions with Gaussian distributions are considered. Both models exhibit a quantum phase transition separating the spin glass and disordered phases where the quantum effects are regulated by a parameter Δ (inverse momentum of inertia for rotor model and inverse mass for oscillator model) parametrizing the kinetic energy. The models allow, in the limit M→∞, for an exact solution in both the spin glass and disordered phases. Employing the quantum finite-temperature real-time description, we examine the static and dynamic properties of the systems without making use of the n-replica method. We find that the dynamical solution is marginally stable in the spin-glass phase with a constant static susceptibility below the spin-glass freezing temperature. A closed-form solution for the averaged free energy has been obtained and various thermodynamic functions are derived. In particular, the entropy is always positive in the spin-glass plane and goes to zero with a ${\mathit{T}}^3$ dependency in the low-temperature region. This is in contrast to the classical M-component Heisenberg model in the M→∞ limit.

A transfer-matrix-scaling technique is developed for randomly diluted systems, and applied to the site-diluted Ising model on a square lattice in two dimensions. For each allowed disorder configuration between two adjacent columns, the contribution of the respective transfer matrix to the decay of correlations is considered only as far as the ratio of its two largest eigenvalues, allowing an economical calculation of a configuration-averaged correlation length. Standard phenomenological-renormalization procedures are then used to analyze aspects of the phase boundary which are difficult to assess accurately by alternative methods. For magnetic site concentration p close to ${\mathit{p}}_{\mathit{c}}$, the extent of exponential behavior of the ${\mathit{T}}_{\mathit{c}}$×p curve is clearly seen for over two decades of variation of p-${\mathit{p}}_{\mathit{c}}$. Close to the pure-system limit, the exactly known reduced slope is reproduced to a very good approximation, though with nonmonotonic convergence. The averaged correlation lengths are inserted into the exponent-amplitude relationship predicted by conformal invariance to hold at criticality. The resulting exponent η remains near the pure value (1/4) for all intermediate concentrations until it crosses over to the percolation value at the threshold.

We have measured the magnetoresistance of high-vacuum-sputtered Co/Cu(100) multilayers grown on Cu buffer layers. The magnetoresistance in the first antiferromagnetic- (AF-) coupling peak is very sensitive to the buffer layer thickness. We find a linear dependence between the actual measured magnetoresistance and the fraction of AF coupling, as determined by magnetization measurements. We compare our Co/Cu(100) magnetoresistance data at 4 K of completely antiparallel-aligned multilayers with the quantum model of giant magnetoresistance of Levy, Zhang, and Fert. This reveals evidence for strong spin-dependent interface scattering, whereas the spin dependence of the bulk scattering in Co is small.

The magnetocrystalline anisotropy energy of a free-standing [111] Co monolayer has been calculated from first principles by means of the linear muffin-tin orbital method in the atomic-spheres approximation. The results are used to analyze those obtained previously for [111] Co/${\mathrm{Pd}}_2$ and Co/${\mathrm{Ag}}_2$ multilayers. Although the magnetization of both multilayers is predicted to be oriented perpendicular to the plane of the multilayers, the easy axis of the Co monolayer is predicted to lie in plane. These results can be understood in terms of the energy bands at the high-symmetry points of the Brillouin zone only. Co/${\mathrm{Pd}}_2$ is found to have a perpendicular anisotropy because the Fermi energy is located much closer to states which have mainly Co ${\mathit{d}}_{\mathit{x}}^2$-${\mathit{y}}^2$ and ${\mathit{d}}_{\mathit{x}\mathit{y}}$ character than it does in the free-standing monolayer. These states are degenerate in the absence of spin-orbit coupling as a consequence of the threefold symmetry. The anisotropy energy is enhanced by the hybridizaton of the Pd d and Co d states and the large value of the Pd d spin-orbit coupling parameter. The smaller perpendicular magnetic anisotropy of a Co/${\mathrm{Ag}}_2$ multilayer can be attributed to the increased band filling of the Co d band structure, as well as to a smaller hybridization between Ag and Co d states.

Using the augmented-plane-wave method, we performed total-energy calculations for all the metals crystallizing in the hcp structure and for all the 5d metals in the hcp structure. Also, using the predicted equilibrium lattice constants from our present and previous total-energy calculations, we carried out band-structure calculations that explore possibilities for superconductivity and magnetism in all the transition metals in the fcc, bcc, and hcp structures.

The computational advantage and accuracy of the Harris method is linked to the simplicity and adequacy of the reference-density model. In an earlier paper, we investigated one way the Harris functional could be extended to systems outside the limits of weakly interacting atoms by making the charge density of the interacting atoms self-consistent within the constraints of overlapping spherical atomic densities. In the present study, a method is presented for augmenting the interacting atom charge densities with symmetrized partial-wave expansions on each atomic site. The added variational freedom of the partial waves leads to a scheme capable of giving exact results within a given exchange-correlation approximation while maintaining many of the desirable convergence and stability properties of the original Harris method. Incorporation of the symmetry of the cluster in the partial-wave construction further reduces the level of computational effort. This partial-wave cluster method is illustrated by its application to the dimer ${\mathrm{C}}_2$, the hypothetical atomic cluster ${\mathrm{Fe}}_6$${\mathrm{Al}}_8$, and the benzene molecule.

The Kubo-Greenwood formula is used to numerically evaluate the conductance for the Anderson model of three-dimensional noninteracting electrons in a uniformly distributed random potential. We construct the trajectory function β(g)=d lng/d lnL, verify that it depends on the single parameter g, and use it to calculate the critical exponent as ν=1.5±0.2. A new trajectory function is defined and evaluated that describes the scaling behavior of the renormalized ac conductivity. We numerically verify the ${\mathrm{\omega }}^{1/3}$ frequency dependence of the ac conductivity.

We investigate the validity of the Fourier transformations that have long been used to transform a given Hamiltonian written in configuration space into momentum space. We show that although these transformations are valid for ideal (perfect) lattices, their use for lattices with vacancies poses mathematical problems and renders inaccuracies in the form of the Hamiltonian by not conserving the number of degrees of freedom in the Hilbert space. Consequently, transformations that avoid this problem are found. We also investigate another solution to this problem that was proposed in the literature. We study the difference between these three methods by investigating the ground-state energy of an electron system in the context of the s-band tight-binding model due to removing (a) one site and (b) two sites. The three methods give similar expressions for the ground-state energy for both of these cases, but they predict different binding energies of electrons at the vacancies.

Recently, Engel and Vosko have proposed an alternative form of a generalized gradient approximation (GGA) [Phys. Rev. B 47, 13 164 (1993)] within density-functional theory. This version of the GGA (EV93) was obtained by optimizing the exchange-correlation potential ${\mathit{V}}_{\mathrm{xc}}$ instead of the corresponding energy ${\mathit{E}}_{\mathrm{xc}}$. We have applied this functional to various solids and found that equilibrium lattice constants generally are far too large in comparison to experiment, so that the total-energy aspect is unsatisfactory. However, EV93 describes FeO and CoO as antiferromagnetic insulators, while the local-spin-density approximation or the Perdew-Wang version of the GGA yields a metallic ground state. EV93 favors magnetism, e.g., the spin susceptibility of V is increased, but Pd remains nonmagnetic in agreement with experiment.

The Hartree-Fock exchange energy of the electron gas is calculated at nonzero temperatures. The calculation is done self-consistently in that the Hartree-Fock self-energy is included self-consistently in the Fermi-Dirac occupation numbers. We also calculate the first and second derivatives of the Hartree-Fock exchange energy with respect to density, which are useful for density-functional calculations at nonzero temperatures.

We present a comparison between calculations based upon the local-density approximation (LDA) and the generalized- gradient approximation (GGA) of the ground-state properties of the f-electron metals Ce, Th, Pa, U, Np, and Pu in their α phase. For α-Ce, Th, Pa, U, and Np the calculations have been performed for the observed crystal structures. Studies for the hypothetical fcc crystal structure have also been done for Pa, U, Np, and Pu. Equilibrium volumes and bulk moduli for these f-electron metals have been calculated for both the LDA and GGA formulation of the exchange-correlation energy functional in the full-potential implementation of the linear-muffin-tin-orbital method. The equilibrium volumes, which generally are too low in LDA calculations, are improved considerably by the gradient correction for these elements. Similarly the bulk moduli are decreased substantially when LDA is replaced by GGA and thereby they come closer to the experimentally observed values.

We consider the Hubbard model on the infinite-dimensional Bethe lattice and construct a systematic series of self-consistent approximations to the one-particle Green’s function, ${\mathit{G}}^{\left(\mathit{n}\right)}$(ω), n=2,3,.... The first n-1 equations of motion are exactly fullfilled by ${\mathit{G}}^{\left(\mathit{n}\right)}$(ω) and the nth equation of motion is decoupled following a simple set of decoupling rules. ${\mathit{G}}^{\mathrm{}\left(2\right)\mathrm{}}$(ω) corresponds to the Hubbard-III approximation. We present analytic and numerical results for the Mott-Hubbard transition at half filling for n=2,3,4.

We present self-consistent calculations for the self-energy and magnetic susceptibility of the two-dimensional (2D) and 3D symmetric Anderson lattice Hamiltonian in the fluctuation exchange approximation. At high temperatures, strong f-electron scattering leads to broad quasiparticle spectral functions, a reduced quasiparticle band gap, and a metallic density of states. As the temperature is lowered, the spectral functions narrow and a pseudogap forms at the characteristic temperature ${\mathit{T}}_{\mathit{x}}$ at which the width of the quasiparticle spectral function at the gap edge is comparable to the renormalized activation energy. For T≪${\mathit{T}}_{\mathit{x}}$, the pseudogap is approximately equal to the hybridization gap in the bare band structure. The opening of the pseudogap is clearly apparent in both the spin susceptibility and the compressibility.

We have calculated the electron energy-loss function for NiO and Ni using the random-phase approximation. Experimental results are presented for NiO, and it is shown that theory is in rather good agreement with experiment. The local-field effects are shown to be small for the loss function, although the system is very inhomogeneous, while they are large for the static dielectric function. We demonstrate that the size of the local-field effects are governed by a certain Coulomb integral w which is determined by the important transitions. These are 3d→4f transitions for the loss function, where ${\mathit{w}}_{\mathit{d}\mathit{f}}$ is small, and 3d→3d and 4s→4s for the static dielectric function, where ${\mathit{w}}_{\mathit{d}\mathit{d}}$ is large. We discuss how the local-field effects reduce the static screening. We show that this reduction is so efficient for the 3d→3d contribution that the 4s→4s contribution plays a substantial role for the static dielectric function, although the 4s density of states is much smaller than the 3d density of states.

Inelastic scattering of electrons in a crystalline environment may be represented by a complex non-Hermitian potential. Complete generalized expressions for this inelastic-electron-scattering-potential matrix, including virtual-inelastic scattering, are derived for outer-shell electron and plasmon excitations. The relationship between these expressions and the general anisotropic dielectric-response matrix of the solid is discussed. These generalized expressions necessarily include the off-diagonal terms representing effects due to departure from translational invariance in the interaction. Results are presented for the diagonal band-structure-dependent inelastic and virtual-inelastic scattering potentials for Si, from a calculation of the inverse dielectric matrix in the random-phase approximation. Good agreement is found with experiment as a function of incident energies from 10 eV to 100 keV. Anisotropy effects and, hence, the interaction delocalization represented by the off-diagonal scattering-potential terms are found to be significant below 1 keV.

Semiempirical tight-binding electronic-energy-band structures of the following ${\mathrm{II}}_3$${\mathrm{V}}_2$ semiconductors are presented: ${\mathrm{Cd}}_3$${\mathrm{P}}_2$, ${\mathrm{Zn}}_3$${\mathrm{P}}_2$, ${\mathrm{Cd}}_3$${\mathrm{As}}_2$, and ${\mathrm{Zn}}_3$${\mathrm{As}}_2$. The zinc-blende approximation of the real crystal structure was used. The calculations were carried out within the nearest- and second-nearest-neighbor approximation and the single s and p atomic-orbital basis set. The tight-binding parameters were obtained empirically by fitting calculated density of valence states and reflectivity spectra to the empirical spectrum. The presented parametrization of the band structures provides an accurate representation of the valence bands and an adequate description of the lowest and most important conduction bands of ${\mathrm{II}}_3$${\mathrm{V}}_2$ semiconductors.