Table of contents

On the basis of a detailed calculation for a substitutional impurity in BCC iron, it is pointed out that the presence of an impurity resonant state at the dip position in a host BCC transition metal is fairly general, almost regardless of the parameters associated with an impurity atom. It is proposed that the presence of this resonant state is the key to understanding the striking difference in the solubility of hydrogen between V (or Nb) and Cr (or Mo).

The well known first-order variational expression of the phonon-limited electrical resistivity is examined for the alkali metals by taking into account the correct energy dependence of the electron distribution function. It appears that, at low temperatures, the energy dependence of the electron distribution function is enhanced by the umklapp process and thus leads to a large reduction in the electrical resistivity.

The relationship between theories of dislocation electrical resistivity due to Basinski et al. (1963) and Brown (1977) is explored and the predictions of the two theories are compared.

A comparison has been made of the electrical resistivity measurements made on single crystals of dilute Cd-Mg alloys and those made on similar crystals of pure Cd in the temperature range 2-300K. It is concluded that an extra term arises in the resistivity of the alloys about 200K having all the salient features expected for electron scattering from a phonon local mode. Although such an effect has been predicted and previously searched for in polycrystalline specimens, it has not been recognised in earlier data.

Binding energies for the core levels of scandium metal and its oxide and hydride are recorded. Plasmon energy-loss peaks are observed for the metal and show a large shift on hydriding, approximately 3.7 eV, compared with a shift of approximately 0.2 eV for the 2p core levels. The large plasmon shift is consistent with the view that scandium forms a metallic hydride in which the H 1s electrons contribute to the Sc valence band.

The X-ray photoemission spectra of the valence bands of Be, Na and Mg have been measured. The analysis shows that after taking into account the partial density of states the core hole-conduction electron coupling and the energy-loss structure, there is good agreement between the experimental spectra and the calculated density of states curves.

It is shown that enormous numerical difficulties occur for a simple metal like lithium when realistic wavefunctions are used in Pick's formulation (1970) of lattice dynamics based on the Fourier transform of the ionic potential. A new formalism is suggested in which the ionic potential is separated in a term which is essentially nonzero near the nuclei and a weakly varying part. In contrast to Pick, only the latter is described by a Fourier expansion. The contributions of the higher bands can be summed exactly yielding a term which contains only the density of the conduction electrons. As a consequence, it is sufficient to know the inverse microscopic dielectric function epsilon ^-1(q+K, q+K') for about a hundred different reciprocal lattice vectors K and K' respectively. As a by-product it is found that the local field corrections of the microscopic dielectric function epsilon (q+K, q+K') are of decisive importance for the electron-phonon interaction in lithium. The numerical results for lithium are in satisfactory agreement with experiment.

A formalism, based on the use of pseudopotential theory and the Gibbs-Bogoliubov inequality, is used to investigate solid-liquid coexistence in the Cd-Mg and Na-K systems. Qualitative agreement with experiment is obtained for the phase separation curves.

The isotope effect for the diffusion of hydrogen, deuterium and tritium in BCC niobium has been measured using the Gorsky effect. In the temperature range 190-310K the data can be represented by Arrhenius relations. The ratio D_H/D_D is much larger than the classical value of square root 2 and shows a strong temperature dependence. The ratio D_D/D_t is slightly larger than square root 3/2 and is relatively temperature-independent. These results are not explicable on the basis of current theory.

A model of the actual potential seen by an electron in an ion is used as a basis for calculations of Heine-Abarenkov type weak model potentials. The energy dependence of the model potential, adjustment of the energy scale for the solid, and inclusion of other contributions to the potential in the solid are easily handled by this method. The parameters of a modified Heine-Abarenkov potential are derived for fifteen elements, designed to minimise nonlocality and to give almost k-independent effective masses. The adequacy of first-order perturbation theory in the calculation of depletion holes is checked by comparison with exact values obtained from numerical integration.

The effect of correlation has been included in the bandstructure calculation for thorium and osmium using the RAPW method. The complete bandstructure and Fermi surface cross sections are obtained for both the metals using the two potentials, the one (V_s) which includes only the Slater exchange and the other (V_H) which includes both the exchange and correlation through Overhauser's formula (1971). A comparison of the authors' results with those of the previous calculations using the traditional approximations with regard to the exchange leads to the conclusion that V_H yields results in better agreement with the experimental data than those obtained with V_s although the latter may still be considered fairly reliable without any violent disagreement.

For previous work see ibid., vol.7, p.1467-75 (1977). The atomic structure factor for metallic aluminium is calculated from bandstructure eigenfunctions obtained using an improved Heine-Abarenkov type model potential. The effect of the nonlocal and energy-dependent nature of the model potential is investigated and found to be important. Corrections due to the depletion hole have been evaluated and are found to be comparable to the original model valence scattering factors. The electron density within the solid is calculated and compared with first-order pseudoatom predictions. The 'bandstructure' density is more free-electron-like and implies weaker screening of the bare-ion potential than that predicted by the pseudoatom method.

The temperature dependence of the electrical resistivity of amorphous metals is calculated using two complementary models of the lattice vibrations. The model which describes uncorrelated Einstein oscillators leads to a very satisfactory fit of the experimental data for CuSn alloys at T>or approximately=10K. The negative temperature coefficients are explained by the temperature dependence of the elastic scattering. The results are interpreted as an indication of strong vibrational disorder. For very low temperatures (T<or approximately=2K), a positive temperature dependence of the resistivity proportional to T² is predicted as a result of coherent long-wavelength phonons.

Using semiclassical transport theory, exact expressions are obtained for the low-temperature phonon-drag contribution to the thermoelectric tensor of a metal in a high transverse magnetic field, H. It is shown that the off-diagonal elements of the thermoelectric tensor to leading order to as 1/H and, under certain symmetry conditions, are independent of the impurity scattering. It is also shown that under the same conditions the diagonal elements to leading order go a 1/H².

A mathematical formula for the quadrupole part of the nuclear spin relaxation rate in liquid metals is given. Preliminary calculations indicate that the formula may be capable of explaining the difference in variation with temperature of the rate in liquid metals and alloys.

The Hall coefficients of liquid bismuth alloys (Bi-Cd, Bi-In and Bi-Sn) have been measured from their melting temperatures to about 360 degrees C. Their values are negative but deviate greatly from free-electron ones. These deviations become larger in the order Sn, In and Cd.

For previous work see ibid., vol.7, p.981 (1977). Some new results on the resistivity behaviour of ytterbium in its semiconducting phase are reported. In the high temperature region (T>150 degrees C) the temperature coefficient of resistance is strongly positive like that in a degenerate semiconductor at high pressures and high temperatures. Measurements of the resistivity against pressure at room temperature and above show a distinct change in the slope accompanying the metal to semiconductor transitions previously reported to be unobservable by other workers.

Electric potential measurements have been made with nine probes along the length of a PbIn cylinder which carries a transport current in the presence of an axial field. The resistance is identified as being due to flux flow and the flow pattern is not uniform throughout the cylinder. Localised resistive oscillations are correlated with magnetisation oscillation at frequencies up to 5 kHz. The measurements help to specify the nature of the yet unexplained longitudinal flux flow resistance process.

The five independent single-crystal elastic constants of hexagonal Gd-40%Y have been measured in the temperature range 4.2-300K and in magnetic fields of up to 7 T. This temperature and magnetic field range covers the different magnetic states of the material and has allowed the magnetic phase diagram to be constructed from the anomalies present in the elastic constants and associated ultrasonic attenuation. At low temperatures and low fields the material does not follow Dy and Tb-50%Ho in transforming from an antiferromagnetic to a ferromagnetic phase but exhibits a variety of more complicated magnetic structures.

It is shown that the critical concentration for the onset of ferromagnetism in the CuNi system is 47.6 at% Ni and that in general the onset of ferromagnetism in all 'giant moment' alloys is necessarily predicted by a mictomagnetic regions.

A special case of a two-band model is proposed to describe d electrons in transition metals. It is shown that this new model is a generalisation of the modified Zener model. The spin wave spectrum for small wavevectors in the ferromagnetic case is calculated within the random phase approximation from the poles of the intraband part of the transverse susceptibility.

The magnetic hyperfine field, B_hf, has been determined for cadmium as an impurity element in Heusler alloys of the types Pd₂MnY (Y=In,Sn,Sb) and PtMnY (Y=Sn,Sb) at a temperature of 77K. The measured fields were -19.4(1) T and -22.5(2) T for Cd in Pd₂MnSn and Pd₂MnSb, respectively, whereas the average field for Cd in Pd₂MnIn (which is antiferromagnetic) was zero with a distribution of Delta B_hf=1.2(1) T. For Cd in PtMnSn and PtMnSb, positive fields were observed, +9.5(5) T and +16.1(1) T, respectively. The positive field can be explained if it is assumed that Cd atoms in these alloys are not going into Y positions as usual, but rather into interstitial or 'empty' positions in the lattice, giving a higher screening charge.

The spin-wave dispersion relation of ferromagnetic Pd₃Fe has been measured by neutron inelastic scattering methods at 4.2, 156, 231 and 296K. At small magnon wavevectors the dispersion relation is isotropic with an exchange stiffness of 66+or-4 THz AA² at 4.2K. At large wavevectors the dispersion relation is very anisotropic and there is strong evidence in the (00 zeta ) and ( zeta zeta 0) directions of intrinsic spin-wave lifetime effects. At higher temperatures the spin-wave modes are reduced in frequency and display temperature-induced broadening effects.

Spin echo measurements have been used to establish the way in which the magnetic moments in GdAl₂ adjust themselves to the presence of dysprosium ions introduced substitutionally into the lattice. The dysprosium moments favour a (100) easy magnetisation direction in contrast to the (100) of the GdAl₂ lattice. The results suggest that the gadolinium moment directions change gradually from the (111) direction and a model is examined for spherical volumes of gadolinium ions about each dysprosium.

Optical constants n( omega ) and k( omega ) were calculated for gadolinium metal in the spectral range 1.5-5 Ev from a polarimetric method. For various temperatures corresponding to the ferromagnetic and paramagnetic states, reflectance measurements were performed, at a pressure near 10^-9 Torr, on very thick opaque films. Two essential structures were found in the optical conductivity near 1.5 eV and 2.5 eV. The influence of magnetic ordering was discussed in the energy range investigated. The abscissa of the peak located near 1.5 eV in the ferromagnetic state is displaced towards 1.85 eV in the paramagnetic state. In the range which was investigated, no additional feature occurs between these two states as it is the case for lowest energies. The absorption peaks near 1.5 eV and 2.5 eV were interpreted in terms of bandstructure calculations and attributed to maxima in the joint density of states.

It is demonstrated that an accurate approximation of the screening function is crucial for a quantitative evaluation of the XPS asymmetry index alpha . It is also shown that the dynamical aspects of the valence electron response are important in order to get an accurate description of the XPS spectrum at finite distances from the threshold.

The wavelengths of the strongest M X-ray lines of plutonium have been measured and compared with available experimental values and with those extrapolated from neighbouring elements. New energy determinations have been given for six atomic levels.

Measurements of the photoemitted electron distribution curves from polycrystalline metallic tin samples are reported for photon energies in the range 7.7-10.8 eV. The bandstructure as obtained by Ament and de Vroomen (1974) is applied to a calculation of the photoemission spectra. The direct and nondirect models for optical excitation between bulk electronic states are considered. Certain features in the experimental spectra can only be interpreted in terms of (k conserving) transitions.