Table of contents

The heats of formation of transition metal alloys are related to moments of the density of states within a tight-binding model for the d band which takes account of diagonal disorder. The contribution of charge transfer to alloy stability is argued to be small. The dominant effect comes from deformation of the bands as manifested by differences in the moments of order four and higher between the alloy density of states and a reference distribution obtained by superposing the pure metal constituents so that their Fermi levels coincide. The difference in fourth moment is trivial to calculate directly from the Hamiltonian and a simple fit to it reproduces to within 20% the results of some recent model calculations by Pettifor (see ibid., vol.7, p.613 (1977)).

Neutron scattering, transport and magnetic measurements are presented on two spin density wave hosts, Cr and Cr-Mo 2%, each containing 14% Fe. The results show that in both systems antiferromagnetism and spin glass ordering coexist at low temperatures.

The quasiparticle mass variation over the Fermi surface of Rb has been evaluated using approximations for the mass operator based on the density-functional formalism. A nonlocal approximation for the mass operator is found to yield a Fermi surface more nearly in accord with experiment than the usual local band theory approximation. However, both local and nonlocal mass operator approximations predict a much more anisotropic quasiparticle mass than has been inferred from the recent experiments of Paul and Springford (see ibid., vol.8, p.1713 (1978)).

The authors resolve the order of magnitude discrepancy between the measured low-temperature lattice thermal conductivity of K and that previously calculated using semiclassical transport theory. This is done by assuming that phonon-electron and phonon-dislocation scattering are dominant and that all other mechanisms for scattering phonons, in particular phonon-phonon scattering, can be ignored. This assumption also leads to excellent agreement between the measured and calculated phonon drag contribution to the high-field Nernst-Ettingshausen coefficient.

The authors present magnetisation measurements performed on a single crystal of TbCo₂ as well as spontaneous distortion studied by X-ray powder diffraction. At any temperature below T_c (240K) the easy magnetisation direction is (111) while the most difficult is (100). A huge rhombohedral distortion appears below T_c; at 4.2K the angular variation between the cube edges is epsilon =4.4*10^-3 rad. The anisotropy due to cobalt being negligible, the authors have analyzed the results in terms of the three contributions acting on the ⁷F₆ ground state multiplet of the Tb³⁺ ions: crystal field, exchange field and magnetoelastic energy. At 4.2K, as the anisotropy of Tb is very high, a 135 kOe field applied along the (100) or the (110) is not strong enough to align the magnetisation parallel to these axes. As in TbFe₂, the large spontaneous rhombohedral distortion results from a strong molecular field associated with strong crystal field effects. The molecular field leads to almost the fundamental level J_z=6 ( mu _Tb approximately=9 mu _B) that is a very anisotropic 4f distribution and the crystalline field essentially arises, as generally observed, from the d character of the conduction band.

The single-crystal elastic moduli C_ij and their pressure derivatives dC_ij/dP have been measured ultrasonically at 25 degrees C for six Ti-V-Cr, eight Nb-Mo and seven Ta-W solid solutions. The rigid-band model for the alloy electronic structure is known to work fairly well for these BCC metals in predicting the density of states at the Fermi energy and related properties. The results of this study show that changes in the composition dependence of the C_ij and dC_ij/dP appear to be related to specific topological changes in the Fermi surface which occur as the electron population varies in the rigid-band model. The relationship between the bandstructure and the C_ij is discussed and a qualitative explanation is given for the correlations between the moduli and the Fermi surface and for some of the differences between the three alloy series.

In the low-pressure limit, a helium atom embedded in the electron gas appropriate to metallic hydrogen is only weakly perturbed from its free space structure, and its interaction with the electron gas can be represented by an energy-dependent pseudopotential. A calculation is made for the Gibbs energy of mixing for H-He, using an Austin pseudopotential and a fluid state model based on hard-sphere structure factors. A large orthogonalisation hole is predicted, which may crudely represent the gradual transition from a helium atom to a screened alpha -particle as the electron density increases. The helium solubility is predominantly determined by the structure-independent interaction of the electron gas with the repulsive pseudopotential, and is probably least at the lowest pressures. In the hydrogen-helium planets (Jupiter, Saturn) the helium concentration may be at or close to saturation at the molecular-metallic hydrogen transition.

The specific heat of Pd_0.775Cu_0.06Si_0.165 alloys has been measured between 30 mK and 1 or 2K successively in glassy and polycrystalline phases obtained by annealing at 400 and 550 degrees C. Well defined T terms appear between 0.15 and 1K in each case, not very dependent upon the structure, but systematically higher in the glassy state. Below 80 mK, the specific heat is dominated by a hyperfine T^-2 term, almost independent of the structure. This second point leads to the conclusion that the short-range order remains almost unchanged from the crystalline to the glassy states.

The linear augmented plane wave (LAPW) method of Takeda and Kubler (1979) and the augmented spherical wave (ASW) method of Williams and others (1979) are extended to the relativistic case. The relativistic LAPW and ASW methods make use of two eigenfunctions of Dirac's equation with a spherically symmetric effective one-particle potential. In the latter method, furthermore, the total angular momentum representation for the spin-orbit coupling is used in the expansion of wavefunctions describing electronic states. In both methods, the augmentation is applied to both large and small components of wavefunctions in bispinor form. By this method of the augmentation, radial derivatives of radial wavefunctions are made continuous at the muffin-tin sphere surface. Formulae are given for self-consistent relativistic LAPW and ASW methods which can be numerically treated just like the nonrelativistic counterparts.

Results are calculated for the effects of both uniaxial stress and pressure on the principal de Haas-van Alphen orbits in magnesium. Agreement with experimental work is satisfactory. The results are expressed in relatively simple physical terms as the sum of a kinetic energy contribution and a potential energy contribution. Changes in their relative magnitudes are shown to account for the wide range of experimental values. Both the kinetic and the potential contributions are the product of an intrinsic factor and an extrinsic factor. The intrinsic factor depends only on the bandstructure, and not the stress, whereas the extrinsic factor depends primarily on the stress. All the results may be understood in terms of these factors, which are easy to understand in themselves. The intrinsic factors contain all of the information about the changes in shape of the orbits. These are illustrated.

For pt.I see ibid., vol.3, p.1941 (1973). Earlier work by the same authors is completed to give formulae from which may be calculated the dependence on any symmetry conserving stress of any principal extreme cross sectional area in beryllium, magnesium, zinc and cadmium. Comparison with de Haas-van Alphen experiments will be made elsewhere.

A theory of electrical conductivity for disordered systems (doped semiconductors, liquid metals and binary alloys) is developed using the Kubo formalism within a one-orbital tight-binding approximation. The density of states is calculated within the self-consistent single-site approach to structural and energetic disorder. The theories of Matsubara, Toyozawa (1961), Ishida and Yonezawa (1973) are critically reviewed. Particular attention is paid to the problem of decoupling the averaged product of the Green functions appearing in the conductivity formula. The influence of the Coulomb interaction is considered and calculated within the Hubbard alloy analogy. The density dependence of the conductivity is used as a criterion for the theory in systems where it is likely to be applicable (doped semiconductors, liquid metals and binary alloys). The extension of the theory to systems like caesium vapour is discussed briefly.

An expression is derived for the dependence of the electrical resistivity on both long- and short-range order and composition. The effects of static atomic and thermally induced displacements are included and found to affect only the short-range order term. The effects of inhomogeneous order are also discussed and the behaviour found to depend upon the dimensions of the ordered regions compared to the electron mean free path.

Tunnelling in symmetric superconducting tunnel junctions made of two superconductors with the same energy gaps ( Delta ₁= Delta ₂), as well as in asymmetric junctions with different energy gaps ( Delta ₁ not= Delta ₂), having PbIn alloy as the base electrode in both cases show sharp current steps, the so-called subharmonic gap structures (SGS), in their I-V characteristics. The voltage locations, the changes due to applied microwaves, the RF power dependence, the temperature and magnetic field dependence, the lineshapes, the matrix element dependence, and the ratios of the relative amplitudes of the SGS have been studied to distinguish between two given explanations, namely the self-coupling and the multiparticle tunnelling mechanisms. The author's results on shorted junctions, like many other previous works, support the self-coupling mechanism as a possible explanation for the SGS. On the other hand the results on PbIn alloy base junctions free of any metallic shorts show that, unlike as often stated, the multiparticle tunnelling is a plausible explanation for the SGS in such junctions.

AuFe alloys containing 2 to 12 at.% Fe are produced by quench condensation from the vapour phase onto a cold substrate. The paramagnetic susceptibility chi is measured in situ in an AC magnetic field of 22 Oe and 1100 Hz at temperatures between 2.5K and 150K. The paramagnetic Curie temperature, the effective magnetic moment and the temperature T_f of the maximum of chi depend on the Fe concentration and on annealing, T_f increases with annealing. Films annealed at 1100K show nearly the same T_f values as bulk AuFe. Most of the increase of T_f with annealing can be attributed to the rising mean free path of the conduction electrons. A rounded peak of chi is observed for films annealed at 80K; a sharper peak after annealing at 295K or 1100K. Small DC fields reduce only the susceptibility peak of films annealed at 295K or at higher temperature. A sharp peak of chi and its reduction in a DC field is apparently due to Fe clusters produced during annealing.

Studies of the low field (H_rms<or approximately=14 A m^-1) initial AC susceptibility chi _i of unannealed and annealed gadolinium around the Curie temperature (T_c approximately 290K) are reported. The frequency dependence of chi _i is found to be characteristic of an impurity diffusion magnetic after effect of mean relaxation time tau approximately 0.3 s. A comparison of thermal modulation experiments for annealed and unannealed samples shows that annealing greatly reduces both the observed temperature hysteresis of chi _i and the thermal relaxation of domain properties in the temperature region of domain nucleation.

Results concerning a study of the influence of a dislocation density of the value of 10¹²-10¹⁶ m^-2 on the Mossbauer spectrum, particularly on a linewidth and on an area under the line are presented for Fe-Cr alloys containing 1.20, 3.63 and 7.98 wt.% of chromium. No evidence of a change of these two parameters has been registered. In addition, such hyperfine interaction parameters as the magnetic field and the isomer shift have been revealed to be independent of the dislocation densities from the range given above. The experimental results, which were obtained in this study are not in disagreement with a theoretical formula given by Kvashnina and Krivoglaz (1967). Its experimental verification, however, was not possible because the error in determination of the linewidth was of the same order of magnitude as the line broadening expected from that formula. The authors' experimental results do not agree with those reported by Elsnar and Reid (1974).

The crystal structure and magnetic properties of the pseudobinary system Gd(Co_xIr_1-x)₂ are reported. The compounds crystallise in the C15 Laves phase with a sudden lattice contraction at x=0.4. Magnetic measurements indicate that Co develops a magnetic moment for x>0.4. The onset of this moment is coincident with a sudden contraction of the lattice. Electron spin resonance measurements of g shift and linewidth also show sharp increases at x=0.4. The correlated changes in the various properties are interpreted as due to a movement of the Fermi level as Co is substituted for Ir.

The dependence of the dynamic and static structure factor and the temperature-dependent part of the electrical resistivity on the spatial correlations of the lattice vibrations is investigated. For this purpose an amplitude correlation function for the lattice vibrations is introduced and expressed in terms of a correlation length. The authors' numerical model calculations show that a negative temperature coefficient of the electrical resistivity is favoured by a short correlation length.