Table of contents

The authors show that longitudinal magnetic excitations are responsible for an additional T^3/2 temperature dependent term in the decrease of the magnetisation of amorphous ferromagnets at low temperatures. This provides a possible explanation for the discrepancy on the values for the spin wave stiffness as measured by magnetisation and neutron scattering experiments. The discrepancy is compared to the anomalous temperature dependence of the resistivity, another effect of these longitudinal magnetic excitations.

The Mean Spherical Approximation is solved exactly for a model liquid metal potential using the Baxter factorisation of the Ornstein-Zernike equation. The solution reduces to solving four simultaneous nonlinear algebraic equations and an outline of a method for numerically solving this system is given. Explicit expressions for the energy and virial pressures are given, as are the methods for calculating structure factors and correlation functions.

The formation energies of vacancies in Li, Na and K have been calculated as functions of temperature using pair potentials determined from first principles. Included in the calculation is a term describing the change in entropy upon formation of the defect. This term is of critical importance in obtaining the right temperature dependence. When combined with earlier calculations of migration energies, the results agree very precisely with the measured self-diffusion energies, giving strong support to the single-defect mechanism for diffusion. On the other hand the calculated formation energies do not agree particularly well with the published measurements, leading to the suggestion that a re-interpretation of these experimental results would be useful.

The authors have measured the electrical resistivity rho (T) of potassium to a relative accuracy of 0.2% in the temperature range 200K<T<330K. Comparison with the theoretical results of Shukla and Taylor (1976) (corrected for the effect of the Debye-Waller factor, multiphonon processes and non-local screening in the form factor) reveals an exponential increase in the measured rho (T) as T approaches the melting temperature. Assuming that this phenomenon is due to vacancy formation the authors are able to deduce a value of the formation enthalpy H_v which is in good accord with the theoretical prediction of Jacucci and Taylor (1979). However a careful error analysis shows that the combined effects of theoretical and experimental uncertainties are sufficiently large to place H_v at any point in the range from 0.2 eV to 0.6 eV. They therefore conclude that measurement of the high-temperature electrical resistivity of a metal is not a useful tool for investigating vacancy formation.

Mossbauer effect measurements are reported with sources of ¹³³Xe implanted in Mo, W and Ta by means of the Leuven Isotope Separator. Information about the direct environment of an implanted radioactive probe atom is obtained through the hyperfine interaction of the daughter nucleus (¹³³Cs) with its surroundings. Each of the sources is subjected to an isochronal annealing sequence in order to study the recovery of each metal. For Mo and W four visible components are detected in the Mossbauer spectra. Identification of the components on the basis of their Mossbauer parameters shows that the components represent substitutional xenon atoms and xenon atoms associated with one, two and three vacancies. The observed population of the visible sites leads to the necessity of the existence of at least one type of 'invisible' site with a recoilless fraction very close to zero. Vacancy trapping at xenon impurities occurs in Mo around 425K and in W around 550K. The authors show that the associated vacancy migration is largely a free migration, indicating that recovery stage III in Mo and W is due to vacancy migration. The free-vacancy migration in Mo and W is shown to be due to single vacancies. For Ta the results are less clear due to considerable overlap of lines in the Mossbauer spectra. However, the spectra suggest vacancy mobility around or below room temperature.

A generalisation of the Lang-Ehrenreich-Heine model for the transition metals is applied to give specific formulae for their electronic Gruneisen parameters within normal band theory. Numerical values are calculated both with and without consideration of electron-phonon interaction with the aid of data from band calculations. Agreement with the experimental values available is variable but a simple framework for discussing the factors affecting the parameter is developed from the results.

The linear response theory of electromigration as proposed by Kumar and Sorbello (1975) and developed by Schaich (1976) is applied to a number of binary alloys. The authors have compared detailed model-potential calculations with experiment for the systems Pb-Bi, Sn-Pb, Na-K and Na-Rb. For the polyvalent systems the comparison is reasonably favourable but for the others the formulation in terms of model potentials appears to fail because the d wave scattering from K and Rb ions is not described correctly. In contrast, phase shift calculations appear to yield a correct semiquantitative description. The success of Ashcroft and Langreth in calculating the resistivity of Na-K using simple model potentials thus appears, at least in part, to be a consequence of the number of freely variable parameters in their theory, rather than of a correct description of the basic physics.

The local density of states of a concentrated alloy can be calculated by an extension of the matching Green function method. This involves the inversion of a matrix with the form of tight binding between bonds. The method is applied to calculating the bandstructure and local density states of a square-well model of ordered LiMg. At low energies near the bottom of the band the local density of states is fairly uniform throughout the crystal, but at higher energies it approaches the bulk value for each atom. The relationship between this approach, which stresses the local environment, and nearly-free-electron perturbation theory is explained. Finally the implications of the work for soft X ray results and charge transfer are considered.

The effects of uniaxial stress ( sigma ₃₃) and pressure on the principal sections of the Fermi surface of zinc are calculated. Agreement with experimental work is good. The results are expressed as the sum of a kinetic energy contribution and a potential energy contribution. It is shown that the kinetic contribution is generally rather similar to the free-electron prediction, and that where results differ from free-electron theory this is due mainly to the influence of the potential contribution. Both the kinetic and potential contributions are expressed as the product of an intrinsic factor, which depends on the bandstructure, and an extrinsic factor which depends primarily on the stress. This separation makes it possible to see how different stresses differ in their effects on a given section, and how sections differ between themselves for a given stress. The changes of shape produced by stress are illustrated.

The electronic structure of Rh, Pd, Ir and Pt has been investigated by double beam photoemission. Data are shown to be interpreted as being due to direct optical (laser) transitions, associated with indirect transitions.

Using the multiscattering approach the electron Green's functions of clusters of atoms approximating extended solids are calculated in r space. Exploiting the point symmetries of the system and separating the Green's function into structure- and energy-dependent parts allows calculations for clusters consisting of up to 200 atoms. In these self-consistent calculations the influence of the cluster size on the local densities of states is discussed and the results are compared to those of bandstructure work. As examples the authors discuss the systems BCC Nb the rocksalt structure compound YS and, as a representative for a complex structure, the compound Th₄H₁₅. The results are applied to the calculation of superconducting transition temperatures.

The model of two unhybridised bands is used for the 50-50% substitutional FeTi alloy. The dependence of the electrical conductivity for the 4s and 3d bands and the magnetic moment per atom on the long-range order parameter is obtained by means of the Green functions method within the coherent potential approximation (CPA).

The charge transfer, electrical resistivity, thermoelectric power and Hall mobility of binary alloys are calculated as functions of alloy composition in the coherent potential approximation. The inter-relation among the composition dependences of these physical quantities is discussed on the basis of the numerical results and shown to be consistent within a model.

Electrical resistivity rho _s of metal containing ordered (spin glass) magnetic impurities is calculated. Two limiting cases are considered: (i) tau _e/ tau _i<<1 ( tau _e and tau _i are the times of collisions of electrons with each other and with nonmagnetic impurities); (ii) square root ( tau _i/ tau _ph)<<1 ( tau _ph is the electron-phonon collision time). In both cases at solution of the kinetic equation it was assumed that at the first stage magnetic impurities do not take part in the conduction electron relaxation. It is shown that dependence of rho _s on electric field (or on current density) is strongly nonlinear irrespective of the relaxation regime, and derivatives of rho _s with respect to field or current contain a characteristic cusp which can be easily recorded experimentally. The calculated effect can be observed at relatively small current densities j approximately 10⁵-10⁶ A cm^-2.

The hydrodynamic model incorporating spatial dispersion is used to investigate the nonretarded plasmon dispersion for a bimetallic interface and for two arbitrary metal half-spaces separated by a vacuum. For the bimetallic interface, the full curve for the well known interface plasmon, starting at omega ²=¹/₂( omega _p1²+ omega _p2²), is given. It is shown that a similar plasmon mode exists when the metals are cleaved and separated, but which corresponds to a different form of density fluctuation and has a dependence on the separation distance. In addition, a new 'image' plasmon omega _im is predicted. This arises from the image of the denser metal 2 ( omega _p2> omega _p1) forming as a result of the density response of the bulk electrons in metal 1, and is independent of the parameters of the less dense metal. The generalisations of the well known symmetric ( omega ₊) and antisymmetric ( omega _-) surface plasmons, seen for two identical metals, are also derived. The generalised upper branch omega ₊ only appears if omega _p1> omega _p2/ square root 2 and forms the continuation of omega _im below the bulk continuum region of the less dense metal.

The generalised orbital magnetic susceptibility chi (q) given by Hebborn and March (1976) is numerically calculated at small wave vectors q in order to estimate the uniform orbital magnetic susceptibility chi _orb given by Hebborn and Sondheimer (1960) from the extrapolation of chi (q) to q=0. For BCC 3d series transition metals the dependence on the electron per atom ratio of chi _orb is obtained in the rigid band assumption. The energy band structure and the wavefunctions required in the calculation are obtained in the APW method from the crystal potential of chromium. It is found that chi _orb is positive in the middle of the series, shows a large peak at vanadium and becomes negative only near the ends of the series. The calculated values of chi _orb for vanadium and chromium are 195 and 106*10^-6 EMU/mol, respectively, which are in good agreement with the recent experimental results. Numerical results for the behaviour of chi (q) at small q are also shown.

The thermal expansion and the forced volume magnetostriction have been measured for the Invar type Fe-Pt alloys. Derived values of the spontaneous volume magnetostriction are large and the same order of magnitude as those for the Fe-Ni Invar alloys. Their dependence on concentration and chemical ordering can be explained by the change of Fe-Fe atom pairs and their temperature dependence satisfies the empirical relation, omega _s(T)=CM²(T). The forced volume magnetostriction observed at 4.2K is about one order smaller than that for Fe-Ni Invar alloys. This suggests the presence of strong ferromagnetism in Fe-Pt Invar alloys, although this contradicts the observation of large values for the high-field susceptibility in these alloys.

Electron spin resonance (ESR) measurements have been performed on Gd_0.5Dy_yY_0.5Ir₂ with y<or=0.007 in order to investigate the effect of dilute concentrations of Dy on the Gd resonance. The ESR bottleneck is opened due to a large relaxation rate of conduction electron spins to Dy. This reveals the significance of anisotropic interactions associated with Dy. Addition of Dy also increases the unbottlenecked limits of the g shift and the temperature gradient of the linewidth.

Magnetic measurements and Mossbauer spectrometry under magnetic field have been performed on amorphous Fe_xGe_1-x alloys of various compositions. For x=0.35, no ferromagnetic ordering is observed down to 1.6K. For x=0.45 and x=0.55, magnetic order exists at 4.2K. From the hyperfine field distribution measured by Mossbauer spectrometry, the distributions of magnetic moments born by the Fe atoms has been calculated, as well as the value of saturation magnetisation expected for colinear spin structure. Fair agreement is obtained between these calculated values and the experimental ones measured directly by an extraction method. The very large widths of the distributions are attributed to hybridisation effects between the d orbitals of iron and the sp orbitals of germanium, whose effect on the magnetic moment of a given iron atom depends on its variety of near neighbour environments. A large number of near neighbour Ge atoms will widen the d levels of the iron atom and make it either non-magnetic or weakly magnetic; on decreasing the number of Ge neighbours the d levels get narrower and higher values of localised moments are obtained.

The quadrupole frequency e²qQ/h for ¹⁷⁵Hf in a Lu single crystal has been determined by nuclear orientation at temperatures down to 4 mK as +307_-90⁺⁵⁰ MHz. With the known groundstate quadrupole moment of Q=+2.7(4) barn for ¹⁷⁵Hf the electric field gradient of HfLu is deduced to be eq=+(4.7_-1.5^+1.0)*10¹⁷ V cm^-2. The magnitude and the sign fit well into the frame of a recently proposed correlation of electric field gradients in hexagonal transition metals.

A transmission spectrometer utilising quartz fibre light pipes and operating at photon energies between 0.64 and 5.4 eV is described. The sample chamber of the spectrometer consists of an ultra-high-vacuum cryostat allowing us to produce metastable alloy films. Measurement of the films at low temperature and annealing at temperatures below 600K are feasible in situ. As a result the authors present the optical properties of metastable, polycrystalline CuFe films in the concentration range 0-100 at.% Fe. The films are evaporated at 77K and subsequently annealed at 350K. Transmission spectra are taken between 1.2 and 5.4 eV. For Fe concentrations between 0-20 at.% the square of the plasma frequency depends linearly on concentration as predicted by the theory of virtual bound states. The interband transitions at 2.2 eV and 5 eV do not shift in this concentration range, whereas the parallel band absorption at 4.8 eV disappears.

Calculations of the L alpha soft X-ray emission spectra of Cu-Ni alloys using the muffin-tin form of the coherent potential approximation (KKR-CPA) are presented. The transition matrix elements are determined exactly within the KKR-CPA, and lifetime broadening is included. The results of new experiments on the L alpha spectra are also reported and combined with X-ray photoemission measurements of the core state binding energies, so that the relative positions of the Cu and Ni emission bands may be determined. Good agreement between theory and experiment is found, confirming the KKR-CPA description of the local densities of states in Cu-Ni alloys. Striking non rigid-band behaviour is observed. It is suggested that soft X-ray studies of the type the authors describe should be a useful probe of alloy potential functions.