Table of contents

In connection with the recent criticism by Hafner (see Non-Cryst. Solids, vol.61-62, p.175, 1984) of the Henderson-Ashcroft mean density approximation (MDA), the Helmholtz free energy is calculated through a different thermodynamic relation. The resultant expression for the free energy is clearly beyond that of the random phase approximation (RPA). By comparing both theories on the same footing, it is concluded that the MDA still exceeds the RPA.

A high-resolution neutron-diffraction investigation of the planar helical-paramagnetic phase transition in Dy observed the helical order parameter to be continuous both in the vicinity of T₂ approximately=171K where Amitin et al. (1984) observed a c-axis strain discontinuity and at the Neel temperature where a recent renormalisation group calculation predicts a first-order transition.

The hyperfine parameters of ¹⁵⁹Tb in TbFe₂ and TbFe₃ have been measured by nuclear magnetic resonance at liquid helium temperatures. The parameters presented remove previous uncertainties about the values in TbFe₂. The new observations on TbFe₃ show resolved spectra for the two crystallographically inequivalent Tb sites.

The dynamic scattering functions S(Q,E) experimentally determined for liquid lead at 623 and 1173K and for liquid bismuth at 578K have been used to derive the corresponding intermediate scattering functions F(Q,t). The procedure involved a fitting of a sum of gaussians to the measured S(Q,E). Theoretical checkpoints are that the zeroth and the second energy moments are satisfactorily fulfilled. It is concluded that the F(Q,t) obtained are of sufficiently good quality to serve as a basis for deriving various memory functions for comparison with modern kinetic theories.

The effect of hydrogen absorption on the atomic structures of some CuTi glasses has been investigated by X-ray diffraction. The dependence on composition has been studied using three alloy compositions, Cu_0.50Ti_0.50, Cu_0.58Ti_0.42 and Cu_0.66Ti_0.34, and the influence of loading increasing amounts of hydrogen investigated for the latter two compositions. It is confirmed that these glasses can accept significant quantities of hydrogen to form amorphous hydrides and that the process is limited by a dissociation reaction involving the formation of titanium hydride TiH₂ and free copper. Quite characteristic changes in the structure factor are produced by hydrogen loading and these are reflected in definite changes in the radial distribution function, similar to those observed elsewhere for other metallic glass-hydrogen systems. The results obtained in this study are in broad agreement with the conclusions of independent bulk measurements on CuTi-H glasses.

A theory of the impurity-induced interatomic forces between atoms of a metal host is formulated in terms of the linear response function of the inhomogeneous electron gas containing a dissolved atom, hydrogen in the present case. Asymptotic properties of the response function, reflecting the presence of a both bound and scattered electronic states, are analysed. Numerically results are shown for a simplified model of the hydrogen impurity in a simple metal, for which the solvent-impurity-solvent triplet ion forces could be determined explicitly. The impurity-induced forces between host atoms near the solute hydrogen are found to be as strong as either the direct hydrogen-host-atom interaction or the host-host interaction in the pure metal; moreover, they have comparable large central and non-central (tangential) components. Angular analysis shows a strong isotropic component, as in a 'spherical solid' model, but the angular dependence is also important for host atoms with a small core radius. The relevance to lattice deformation and local force-constant changes at the defect is discussed.

Crystallisation of (Mo_0.6Ru_0.4)_100-xB_x glasses takes place in three steps: the first step corresponds to the precipitation of the sigma phase Mo₅Ru₃ which decomposes at higher temperatures, the second step is associated with the formation of an HCP solid solution of Mo in Ru, and in the last step the remaining amorphous matrix crystallises in an FCC boride. Detailed electrical resistivity measurements taken below the crystallisation temperature reveal an excess resistivity above the usual linear temperature dependence predicted by the Ziman theory. The increase in resistivity is associated with the onset of long-range compositional inhomogeneity (spinodal decomposition) and the decrease with the onset of crystallisation. The electrical behaviour for samples pre-annealed for 12 hours at various temperatures suggests that boron migration is partly responsible for this excess resistivity. The changes in the radial distribution function of (Mo_0.6Ru_0.4)₇₈B₂₂ during annealing at 525 degrees C are also reported. The Mott s-d scattering model does not seem to be applicable to these amorphous alloys. The data suggest instead that an important contribution to the conduction comes from the d electrons.

Simple approximations are derived from Roth's effective medium approximation (EMA) written in a locator-propagator representation. The EMA integral equation for the propagator is solved in an approximate way in order to make the calculation much less time-consuming. Two different approximations are proposed (AI and AII). The very simple AI is no more difficult than the non-self-consistent and quasi-crystalline approximation and much more accurate in the strong coupling case (transition metal). The AI satisfies exactly the non-overlap condition and provides the exact density-of-state second energy moment. It is superior in both respects to the self-consistent Gyorffy-Koringa-Mills approximation. The AII is no more difficult than the Ishida-Yonezawa theory but compares favourably with the EMA. It provides the first three energy moments and satisfies approximately the non-overlap condition. The reliability of the new approximations is studied numerically for a tight-binding and a muffin-tin model for liquid nickel. In both cases the AII density of states was found to compare well with the EMA one.

The first self-consistent spin-polarised band-structure calculation of the ferromagnetic compound MnBi in its low-temperature phase has been performed. The spin-orbit interaction was treated as a perturbation using scalar-relativistic wavefunctions. On the Mn atoms an effective 3d^5.5 configuration is found, with a magnetic moment of 3.6 mu _B. A part of the d electrons is rather delocalised and strongly hybridised with Bi 6p states. This leads to a low density of states at the Fermi level in agreement with experiment. The calculated band structure is very similar to that of MnSb. The physical properties of both compounds are compared, and the dependence of the band structure on the direction of the spin magnetic moment is discussed in relation to the magnetic anisotropy and the magneto-optical polar Kerr effect.

A rigid shift of a single muffin-tin potential does not generally satisfy the Friedel sum rule. An additional scattering potential of a simple form, which basically comes from the screening effect is taken into account so that the whole scattering effect satisfies the Friedel sum rule. Some problems of the rigid-muffin-tin approximation and the modifications arising from the present method are discussed with some example calculations for Al, Cu, Nb and Mo.

The authors have developed a self-consistent transport equation for electrons in disordered solids using the method of projection operators and an effective-medium approximation for the Wigner distribution function. The form of the equation is identical with the Boltzmann equation but contains a generalised scattering kernel T(k,k') and a generalised 'driving' or source term, both of which have to be calculated self-consistently. The phenomenon of '2k_F' scattering is found in the weak scattering limit but the theory is applicable to more general situations. The magnitude of the '2k_F' scattering is found to be in disagreement with results obtained by other authors.

Using the theory of '2k_F' scattering and localisation of electrons in disordered systems the authors have carried out calculations of the resistivity of liquid potassium containing halide ions across the entire concentration range from pure potassium to the molten salt and have obtained remarkable agreement with the experimental results.

The authors present, within the framework of the linearised Boltzmann equation and for a metal with a spherical Fermi surface, a theoretical analysis of the electrical resistivity of wires in the presence of normal electron-electron scattering, Umklapp electron-electron scattering and electron-impurity scattering when surface effects are not negligible. They give both approximate analytic and exact numerical results which, besides showing the general behaviour of the electrical resistivity with varying temperature, are able to explain from first principles of the Gurzhi effect and the Knudsen effect in metals. A comparison of the theoretical results with recent experimental data on the electrical resistivity of thin wires of potassium shows very satisfying agreement from both a qualitative and a quantitative point of view.

Enhancements of the low-temperature thermopower of several La-based and Ca-based glassy alloys are compared with the predictions for the electron-phonon enhancement. The agreement is reasonably good for the La-based alloys, but unexpected structure is observed for the Ca-based alloys. The measurements suggest that the low-frequency alpha ²( omega )F( omega ) is proportional to omega ^beta with beta >1 for these alloys.

In itinerant ferromagnets at high temperatures the finite electronic lifetime can significantly broaden the Stoner continuum. The magnons can then decay into short-lived electron-hole pairs. This broadens the magnetic response of the system in neutron scattering experiments. A simple model calculation illustrates this. The data of Lynn (1975, 1984) for iron at high temperature can be explained in this way, and other applications are suggested.

It is shown that published magnetic susceptibility and magnetisation data for PdNi alloys are consistent with a homogeneous model in which the Ni atoms have a spin fluctuation temperature T_s approximately 55K and are coupled by ferromagnetic Ni-Ni interactions. Using a molecular-field model the equivalent classical RMS moment per Ni atom square root ( mu ²) is shown to be concentration independent for concentrations between 1.3 and 16.7 at.% Ni and for temperatures between the ordering temperature T_c and 700K. square root ( mu ²) exhibits a temperature-dependent exchange enhancement at high temperatures, and at low temperatures falls to zero due to intrinsic spin fluctuations. The exchange enhancement and the ferromagnetic coupling between Ni moments are shown to be related in accordance with the indirect-coupling model of Takahashi and Shimizu (1965). The critical concentration for the onset of ferromagnetism and the concentration dependence of the ordering temperature T_c are determined from the curve for square root ( mu ²) and are shown to be in excellent agreement with experiment. A simple model is presented which qualitatively reproduces the magnetisation curves for paramagnetic and ferromagnetic alloys.

Magnetic and Mossbauer measurements have been used to study the superparamagnetic behaviour of a magnetic fluid containing iron particles. Electron micrographs show the particles to form a Gaussian distribution with median diameter D_p=65 AA, sigma =12 AA whilst magnetic analysis gives the median diameter D_vm=60+or-5 AA at 80K and D_vm=55+or-5 AA at 300K. Superparamagnetic behaviour was evident for the magnetic measurements at both 80 and 300K. The Mossbauer spectrum at 86K contained two lines which were associated with the superparamagnetic iron core and the FeO coating. The FeO line showed no evidence of magnetic ordering when the coating was thin ( approximately=10 AA) but this was apparent when the coating thickness was increased by prolonged oxidation.