Table of contents

The electronic transport coefficients of the liquid alkaline earth metals are computed within the framework of the Ziman theory by using a model potential previously used in calculations of the resistivity and thermopower of simple liquid metals and alloys. The results obtained are in fair agreement with experimental data. Attention has been paid to the role of the energy dependence of the form factors and to the ability of the hard-sphere structure factors to give correct answers for rho and Q.

Resistivities of Cu-Ir, Au-Ta and Au-Hf dilute alloys were measured between 1.5K and 4.2K. It was found that the resistivities had no dependence on temperature and that the values obtained for residual resistivities per atomic percentage concentration agreed with those of 4d dilute alloys.

The author discusses the volume dependence of the interionic potential in sodium, potassium and aluminium as described by X alpha pseudopotential theory. The volume effect is shown to be most important in the region of the first neighbour shell, but less important for larger distances. The theoretical potentials are compared with pair potentials derived from the observed structural data of the liquid metals. Electronic exchange and correlation effects (both valence-core and valence-valence interactions) are investigated. The pair potentials are applied to calculate the elastic shear constants and their hydrostatic strain derivatives.

A previously proposed six parameter phenomenological screened shell model is applied to the calculation of the phonon dispersion curves of copper. Three of the six parameters are determined from the elastic constants and the remaining three from a fit to the experimental phonon dispersion curves. One set of calculations was performed using the Thomas-Fermi screening parameter and the other set used the Bohm-Pines value. A significantly better fit was obtained using the Bohm-Pines value.

Phonon dispersion relations have been measured in the Pd_0.96Fe_0.04 and Pd_0.90Fe_0.10 single crystals along directions (100), (110) and (111). Comparison of the phonon frequency shifts for Pd_0.90Fe_0.10 with theoretical predictions based on the mass and simple force defect theory has shown that the theory is invalid in this case. An analysis of obtained data in terms of the Born-von Karman force constants is also presented. Anomalies observed in the slope of the (110) T₁ phonon branches are discussed.

The effective Debye temperature for high temperatures is obtained, which is valid with no limitation on the magnitude of the impurity-host force constant. The result is approximately consistent with that of Maradudin and Flinn (1962) for small changes in the force constants, but for large changes in the force constants the difference is large. The changes in the force constants between impurity and host atoms in the case of a ⁵⁷Fe nucleus as a dilute impurity in various host lattices are estimated using the effective Debye temperature. The values thus obtained are in agreement with the values reported by others.

The nonretarded Van der Waals energy, E_VdW, between two metal slabs is calculated using the random phase approximation and an infinite potential barrier at each surface. It is shown that E_VdW comes from the change in surface plasmon zero point energy, and making use of this, asymptotic results are obtained which show when the Lifschitz results (1956) become valid. With an electron density corresponding to Al, E_VdW at the smallest separation compatible with the infinite barrier model is -260 erg cm^-2, much less than the Lifschitz value and ¹/₁₀ of the surface exchange-correlation energy. However, by a separation of 10 au, E_VdW is within 10% of the Lifschitz result.

The Korringa-Kohn-Rostoker (KKR) method of calculating energy bands is applied to determine the electron density rho (r, E) below energy E in the conduction band of body centred cubic lithium metal. This is done not only at the Fermi level E_F but for a number of energies throughout the occupied part of the conduction band. The local density of states sigma (r, E) is also plotted for a number of energies. The off diagonal generalization of sigma can also be calculated from the tabulation required to determine sigma (r, E). A detailed comparison is made between these KKR results for the crystal and a model of superimposed spherical pseudo-atoms. Some brief comments are finally made on the relation to experiment; X-ray intensities at the Bragg reflections, Fermi surface and point defect problems are referred to.

The selfconsistent charge density is calculated by means of the density functional formalism using a local exchange-correlation potential for a simple jellium model of a divalent impurity and a vacancy in a monovalent host. The resulting charge density around the impurity does not lend itself to an interpretation in terms of a surface dipole layer, though the dependence of the total charge transfer on the density is as predicted by the dipole model. The charge density inside a vacancy is found to be higher than that obtained by folding over a free surface; the calculated formation energy is an order of magnitude lower than typical experimental values. The validity of the Thomas-Fermi approximation for this model is discussed.

Recent calculations of the deviations from Matthiessen's Rule (DMR) using the scattering time formalism are shown to be equivalent to the generalized two band model of MacDonald (1956). This provides a unified physical model which explains in a qualitative way the similarities in the DMR observed in a wide variety of metals with widely varying band structures. Some simple calculations for dilute alloys of silver are presented and shown to have some of the essential features of the experimental results for the DMR.

The authors present the results of a calculation of the low temperature ideal electrical resistivity of potassium which includes a consistent treatment of phonon drag. These results differ significantly from those of recent calculations which include the effect of phonon drag via the variational approach using the simplest trial functions for the electron and phonon distributions. It is shown that electron-phonon scattering does not, by itself, limit phonon drag to being a low temperature effect. Finally, it is pointed out that, as phonon drag significantly increases the relative contribution of the Umklapp process to the low temperature electrical resistivity, all recent investigations of the effect on the electrical resistivity of anisotropic relaxation of the electron distribution should be re-examined in the phonon drag regime.

The authors calculate the reduced electronic thermal conductivity of a superconducting alloy containing paramagnetic impurities with local states within the gap. The Green function given by Shiba or Rusinov (1968-69) (for s wave scattering) is used which is appropriate for the case when the exchange scattering of a conduction electron from a magnetic impurity is strong. The temperature dependence of (K_s/K_n) for various impurity concentrations is shown and for a typical value of epsilon ₀ (=0.6; epsilon ₀ is a parameter which determines the strength of the conduction electron-impurity interaction). Significant deviations are found, from the theoretical results based on Abrikosov-Gor'kov theory ( epsilon ₀=1.0).

The electrical resistivity of nickel alloys, if extrapolated to B=0, depends on the angle between the current and the magnetization. The authors have measured, at 4.2K, this resistivity anisotropy for ternary alloys of general composition Ni₉₇A_3-xB_x, where A and B denote different metallic elements. These alloys include Ni(AlCr), Ni(AuCo), Ni(AuRh), Ni(CoRh), Ni(CoRu), Ni(CrFe), Ni(CrMn), Ni(CrSn), Ni(CrTi), Ni(CuRh), Ni(CuRu), Ni(CuV), Ni(FePt), Ni(FeRe), Ni(RhRu) and Ni(RuTi). The experimental results have been analyzed in terms of Mott's two current conduction model. In this model it is assumed that the conduction electrons of spin direction parallel or antiparallel to the magnetization contribute individually to the current. Each impurity element in nickel has got two residual resistivities, one for each spin current. These residual resistivities are derived for the ferromagnetic moment parallel and perpendicular to the current. It is found that the resistivity anisotropy is large for one of the spin currents (about +10%) while that for the other varies around -2%.

The authors have measured the variation with magnetic field and temperature of the amplitude of giant quantum oscillations in the thermopower of aluminum from 1.7 to 4.8K in fields up to 21 kG. The field and temperature variations are dominated by a term exponential in the ratio m*(T+T_D)/H. Analysis using the de Haas-van Alphen procedure for determining effective masses and Dingle temperatures yields m*=(0.093 +or-0.01)m_e and T_D approximately=1-4K. This value for m* agrees with the value obtained from de Haas-van Alphen measurements for the same electron orbit, and corresponds to a phonon-enhancement of approximately 40%.

The radio frequency size effect with a magnetic field parallel to the surface of the specimen has been used to determine mean free paths in indium. The phonon limited mean free path l_p for carriers on the second zone surface was determined from the temperature dependence of the amplitude A_T of the size effect. A cubic dependence of l_p^-1 on temperature was found. However, in some cases deviations from the expected behaviour of the amplitude in A_T=a-bT^-3, where a and b are constants, have been recorded. From the field position of the extremum of the peak near zero field, values of the carrier mean free path have been extracted. The preliminary results of both methods agree fairly well.

The X-ray photoelectron spectra of Cu_0.9Zn_0.1, Cu_0.5Zn_0.5 and Cu_0.2Zn_0.8 have been observed using Mg K alpha X-rays. The valence band spectra show that the Cu 3d band narrows and shifts to increasing binding energy relative to the Fermi energy as the Cu is diluted by Zn. The Zn 3d state width is also reduced as the Zn becomes more dilute. These results are suggested to be due to reduction of nearest neighbour 3d overlap.

The transport of magnetic flux in type 1 and type 2 superconductors under the influence of a transport current gives rise to a noise voltage with power spectrum W_f superposed on the d.c. voltage (V). The authors report measurements of W_f for cold-rolled strips and discs of type 1 and type 2 InPb alloys. For the type 2 strips a new phenomenon has been observed: peaks in W_f which are very sensitive to current density and magnetic field. The general features of the W_f measurements are compared, where possible, with various model of flux motion. For the type 1 strips with leads of different separation a similar dependence of W_f on frequency f is found, a result which cannot be interpreted according to any of the simple random pulse models.

The magnetic and structural properties have been examined for pseudobinary compounds for many stoichiometries and solubility ranges in the phase diagrams defined by Y(Fe, Co) and Y(Co, Ni). Attempts have been made to account for the disappearance of certain phases in terms of atomic substitution and changes in the electron concentration. The magnetic behaviour is related to that of the alloys of the pure metals iron, cobalt and nickel; however a deep minimum is found to occur in the molecular moment in several stoichiometries near to the nickel compound. There is also a corresponding minimum in the exchange interaction strength. The observations may be understood in terms of an analysis involving the form of the density of states curves in these materials.

Magnetization measurements are reported for homogeneous materials Sm_0.167Co_0.833-xAl_x both in bulk and 'oriented powder' form over a temperature range of 1.6-280K with fields up to 45 kOe. At cryogenic temperatures (4.2K) outstanding magnetic hardness is found in both the Al rich powder and the bulk specimen, with coercive forces, H_c, of order 40-50 kOe. This magnetic hardness decays with temperature according to a Curie type law. It is inferred from these findings that the domain wall energy increases strongly on Al substitution. Moreover, when Al is substituted for Co, the predominant mode of domain reversal changes at cryogenic temperatures from a mechanism involving domain wall nucleation to one of domain wall pinning. The mechanism of magnetization in highly energetic domain wall materials is shown to reflect in a delicate way details of structure on an atomic scale. Some preliminary results are presented for alloys Sm₂Co₁₇ and Sm₂Co₇ with Al substitution for Co. An exploratory effort at coating SmCo₅ powders with its Al substituted derivatives has been carried out via sintering with either Al or Sm-Al compounds. The potential for applications of Al derivatives of SmCo₅ in the fabrication process of permanent magnets is discussed.

An earlier calculation based on the generalized Landau theory of phase transitions is extended to cover the ferromagnetic regime where the Landau coefficient B has to be included. The theory is applied to dilute Pd-Ni alloys. Contrary to the earlier calculation, nearest neighbour pairs of nickel atoms in a matrix consisting of palladium and single nickel impurities are considered. The Landau equation is solved to give the dependence of the magnetization M on external field (Arrott plots) and on the spatial position relative to the impurities. The agreement of the Arrott plots with those observed is good.

The Eu Mossbauer effect was studied on the compounds EuZn₁₃, EuZn₅ and EuZn₂ at various temperatures. The hyperfine fields and the isomer shifts were determined. For these compounds the lattice constants and the magnetic properties were also studied: EuZn₁₃ is ferromagnetic, EuZn₅ antiferromagnetic and EuZn₂ metamagnetic. The ordering temperatures, asymptotic Curie temperatures and effective moments were determined. From all these results it was established that the Eu in these compounds is almost completely in the divalent state, isoelectronic with Gd. The data available for EuZn₂ were used to analyse the strength of the magnetic coupling between the localized moments by means of the RKKY model. The sign and magnitude of the s-f exchange integral were determined.

The author presents a model calculation for the determination of the unquenched local moment and the electrical field gradient on dilute heavy transitional impurities in iron hosts. Selfconsistent Hartree-Fock equations taking full account of the spin-orbit terms are written for the localized matrix elements of the perturbing potential on the impurity site. Truncation of the Hartree-Fock equations and Friedel's rule are used in order to determine the impurity potential. The band structure of pure iron is described by a hybridized tight binding nearly free electron (TBNFE) model and the calculation is performed for dilute impurities of the third transition series. The results are compared with hyperfine anomaly, Mossbauer data and nuclear magnetic resonance on Ir impurities in iron.

The core enhancement factor, alpha (k_F), of the electric field gradient has been calculated for k_F along (110), (100), (211) and (111). The convergence of alpha has been studied using the model potentials of both Ashcroft (1966) and of Heine and Abarenkov (1964). Calculations have also been made using the semi-empirical pseudopotential band structure of Anderson and Lane (1970). The results show a small anisotropy ( approximately 20%) and are relatively close to the single OPW estimate. Comparisons have been made with recent theoretical calculations based on first order perturbation theory and quantitative differences are observed. Various approximations in the existing theory for the electric field gradient in alloys and their relevance to experiment are also briefly discussed.

Mossbauer studies on a disordered Ni₄₈Fe₅₂ alloy are presented in the temperature range 4.2-750K. The temperature dependence of the hyperfine field at an iron site is interpreted in terms of a distribution of exchange fields due to iron atoms being present in different near neighbour environments in the disordered alloy.

The electronic band structure of beta -Ga has been investigated using the augmented plane wave method. The E_k relation along symmetry directions, the electronic density of states and the Fermi surface have been calculated. In addition, electron wavefunctions have been evaluated at points on the Fermi surface and used to estimate the Knight shift. The band structure was found to deviate little from free electron behaviour; in particular the band width and the density of states at the Fermi level were within 4% of the corresponding free electron values. The calculated Knight shift was 0.40% in good agreement with the observed value, extrapolated to 4K, of 0.35%. The observed differences between the Knight shifts of alpha -Ga and beta -Ga was investigated and found to be largely due to a difference in the density of states and to a lesser extent to differences in the wavefunction character at the nucleus.