Table of contents

A shock-induced transformation which destroys de Haas-van Alphen signals at 1K has been observed in free-standing single crystals of rubidium. The change is almost completely reversed by annealing at room temperature. Surface markings on the samples suggest that a martensitic transformation may be responsible for the effect.

A calculation and system of binding energies of metalloid atoms embedded in the host lattice of a transition metal are reported. The metalloid binding energy is determined within the framework of a theory containing a number of parameters which are known for the transition metals. The metalloid binding energy is found to be negative and to decrease numerically as the atomic number of the transition metal increases.

The ⁷Li Knight shift in liquid lithium-germanium alloys has been measured through the main part of the composition range. The results, particularly those obtained close to the liquidus, indicate both charge transfer and the formation of homopolar compounds in the liquid.

A comparison has been made between two recent, independent X-ray determinations of the structure of Fe-B metallic glasses which are in disagreement. The structure factors for two similar Fe-B glasses obtained in these two studies have been critical compared, and their differences discussed. The effect which these differences in S(Q) have on the reduced radial distribution function G(r) has then been examined in a systematic way. It is suggested that some of the disagreements in the S(Q) curves can arise through structural differences in the samples, which are related to their state of quench. In addition it is shown that errors in the experimental method and systematic variations in data analysis can also influence the form of S(Q) in a similar way, and hence affect G(r). A conclusion reached from this comparison is that the second peak splitting in G(r) appears to be too susceptible to experimental uncertainty to be considered an important characteristic of metallic glass structure. Finally it is suggested that an adequate treatment of chemical as well as topological order is needed in order that reliable structural data may be obtained for these transition metal-metalloid glasses.

The behaviour of deuterium implanted in Al was studied by the D(³He,p)⁴He and the D(d,p)T nuclear reactions. Changes of the depth profiles on the deuterium after heat treatments indicated that the implanted deuterium was trapped by the defect produced during the deuterium implantation and the release probability of the trapped deuterium increased as the specimen temperature was raised. Assuming a thermal equilibrium locally in the region of high defect concentration, the trapping energy of deuterium in Al was determined to be 0.12 eV. Since the release probability for the single crystal was considerably larger than that for the polycrystal specimens, the deuterium was considered to be strongly trapped in the grain boundaries. Distributions of displaced Al atoms and the recovery of the lattice damage by annealing were measured by the channelling technique.

The first part of this paper is devoted to the self-consistent calculation of the phonon spectra for metals whose band structure can be represented by the tight-binding approximation. Firstly the authors take into account the displacement of the tight-binding basis orbitals and obtain self-consistent expressions for the electronic density and potential changes induced by the atomic displacements: they discuss the relation between the Frolich and the Bloch approaches and show that a reasonable approximation allows the Frolich result to be obtained from the Bloch method. Finally the authors obtain simple self-consistent expressions for the electron-phonon matrix elements and for the dynamical matrix. In the second part they apply the previous formal theory to the transition metals. The authors assumed that the band structure can be approximated by a tight-binding nearly free electron model and show-as previously stated by other authors-that the main role of the 'sp' electrons is to screen the interaction between 'd' electrons. The physical meaning of some approximations which must be used in a practical calculation is discussed and the self-consistent expressions of the elastic constants are derived.

The incommensurate-commensurate phase transition of CrMn single crystals is shown to be of first order. Coexistence of the two phases reported in the past is shown conclusively to be due to inhomogeneities in the Mn concentration of large samples. The observed relative length changes and latent heat effects at the transition correlate well with the known pressure dependence of the transition temperature. At this temperature very small resistivity anomalies are present as a result of a different temperature dependence of the resistivity of the incommensurate and commensurate phases.

The EMF of Na in liquid Na-Sn has been measured as a function of temperature using a concentration cell. The associated thermodynamic quantities of mixing such as free energy Delta G, entropy Delta S and enthalpy Delta H are obtained. A breakdown of Kopp's law, which expresses the additivity of specific heats, has been found. From the thermodynamic quantities obtained, the authors find that liquid Na-Sn alloys have two kinds of compound formation, NaSn and Na₃Sn. The curve for the concentration-concentration fluctuation S_cc(0) in the long-wavelength limit obtained from Delta G also supports the coexistence of these two compounds.

Reports detailed data on the chemical potential of Na in molten Na-Bi and Na-Cs alloys. The measurements employed an electrochemical cell with beta alumina as the electrolyte. The concentration range was from 0 to 50 at.% Na in Bi and from 20 to 30 at.% Na in Cs. The temperature range was from the liquidus to 600 degrees C in the Bi system and from the liquidus to 150 degrees C in the Cs alloy. The latter measurements were extended very close to the eutectic at -29 degrees C. The Na-Bi system exhibits strong deviations from ideality at all but the most dilute Na concentrations while the Na-Cs system is very close to ideal throughout the range studied.

Earlier work in this laboratory obtained values of the changes in the thermoelectric EMF (TEMF) per unit tensile stress per unit temperature gradient (i.e. the piezothermoelectric coefficient, Pi _L) for Cu, Au and Ag. This extension to the present work studies four more metals (Nb, Ta, Co and Ni). More significantly, Pi _L has been used to calculate the relative changes in the Fermi surface areas of the noble metals and these calculated values have been compared to those obtained by other workers using other techniques, notably the de Haas-von Alphen (DHVA) effect.

By solving the Schrodinger equation exactly for representative clusters of nonoverlapping spherically symmetric potential wells, designed to describe one-electron potentials in metallic glasses, the authors have studied the electronic structure of Pd_cZr_1-c, Fe_cZr_1-c, Co_cZr_1-c, Ni_cZr_1-c and Cu_cZr_1-c with c approximately 0.3. The authors' clusters were selected to display the significance of both positional and chemical order. Their results suggest that the latter is the more important effect as far as the electronic states are concerned. They found that the nonselfconsistent potentials, constructed by overlapping atomic charge densities and taking the exchange interaction into account by using the Slater approximation ( alpha =1), gave rise to a 'band structure' consistent with the measured photoemission and soft X-ray spectra.

The local and non-local spin susceptibilities with regard to lattice sites and orbitals, chi _nL,n'L', are calculated for transition metals using the muffin-tin potential model. Their general properties are discussed. It is found that chi _nL,n'L' provides an insight from the real space point of view into the qualitative aspect of the magnetism of transition metals. A discussion is also given of an induced magnetic moment at an impurity site in a host metal with a large exchange enhancement.

A simple model of damping of excited electrons in crystals is investigated, and the consequent broadening of angular resolved photoemission spectra is discussed. For a photon energy of 11.7 eV the exit angle dependence of the peak width of photoelectrons emanating from the sp band of Cu(111) is estimated from lifetime and band structure data and compared with the authors previously published spectra. A predicted sharp linewidth enhancement and maximum for k/sub /// approximately=0.83 pi /a along Gamma M is seen experimentally by a more gradual rise saturating at k/sub /// approximately=0.75 pi /a. Finally, angular tuning is used to obtain electron inverse lifetimes in copper of 2.7+or-0.4 eV and 3.5+or-0.4 eV at energies of 16.6 and 20.7 eV above the Fermi level, respectively.

Describes a procedure for calculating the residual resistivity of impurities in metals taking both the Bloch-wave character of the electronic states and the multiple scattering by the host lattice into account. The Boltzmann equation is solved for an anisotropic Fermi surface. For localised impurities the Boltzmann equation has a degenerate kernel and can be transformed into an inhomogeneous linear algebraic set of finite rank.

A generalised pseudopotential formalism for calculating the change in electronic structure Delta N(E) of transitional impurity systems has been employed to recast the Green function analysis of Riedinger (1971) for dilute transitional impurity systems into a pseudopotential framework. The required pseudo Green's functions are evaluated from a model Hamiltonian representation of the host metal band structure. For the difference pseudopotential at the impurity site the authors employ a parametrised, optimised model potential which is made partially self-consistent through the application of the Friedel sum rule. This pseudopotential method is then applied to a variety of dilute transitional impurity systems in an endeavour to assess its overall validity. The first systems considered are the dilute noble metal-palladium alloys CuPd, AgPd and AuPd and the resulting Delta N(E) are found to yield Pd virtual bound state (VBS) resonance peaks whose positions and widths are in fairly good agreement with experiment. The method is then applied to the dilute PdAu alloy system and produces a Au VBS resonance near the bottom of the Pd d band as supported by photoemission studies. Finally the authors consider the dilute noble alloys for which the difference in chemical valence of the two components is zero. Calculations for both AgAu and AuAg confirm that there is an effective charge transfer in these systems of around 0.3 e.

The authors present experimental results for the resistivities, the temperature and magnetic field dependencies of the resistivity, and the thermopower of a number of Cu-Zr, Cu-Ti, Ni-Zr and Ni-Nb alloys together with a Cu₅₀Zr₄₀Fe₁₀ alloy in the temperature range 4.2-270K. They show how the observed temperature dependence of the thermopowers may be understood in terms of electron-phonon enhancement. The resistivities and thermopowers of these systems are then calculated in the Ziman and Mott models and the authors show that neither model gives the correct resistivities, but that the Mott model gives reasonable agreement with the thermopowers. Surveying the published thermopower results they then find that none of them seem in conflict with Mott's s-d scattering model. The paper concluded with a discussion of the effects of hybridisation and of d band conduction in which the authors indicate how this agreement with the Mott formula may be retained in a more physically sound theory. In an appendix the general features of the Ziman and Mott models are illustrated by rough calculations of the resistivities and thermopowers of the liquid 3d transition metals.

The temperature dependence of the electrical resistivity of Li at high temperatures is calculated, both at constant volume and at constant pressure. The calculation includes the effects of the non-sphericity of the Fermi surface, multiple plane wave pseudowavefunctions, the anisotropy of the electron-phonon scattering relaxation time and the anharmonic effects. Good agreement with experiment is obtained. The resistivity of the ⁶Li isotope is also calculated and here too good agreement is found with the experimental data.

The results of several years of work devoted to developing an understanding of the process by which the resistive contribution of transition metal impurities, primarily iron, in copper is removed by internal oxidation are presented. The majority of the investigations were made on a CuFe alloy series of precisely determined composition from 1 to 100 atomic parts per million (at PPM) Fe. Electrical resistance and magnetic susceptibility measurements at room temperature and at 4K on both unoxidised and oxidised samples are reported. These measurements, supported by scanning electron microscopy (SEM) and Curie temperature determinations, suggest a two-step internal oxidation process at even the lowest impurity levels.

The authors have derived a model theory for the superconductivity of A15 compounds, with normal and paramagnetic impurities, which is based on a generalised version of the Eliashberg equation that includes the possibility of energy dependence in the electron density of states. This variation is simulated by a Lorentzian distribution symmetric about the chemical potential and superimposed on a constant background density. To be specific, parameters similar to those characteristic of Nb₃Sn are used but with a high and narrow peak in the density of states.

A model previously used for treating thermodynamic properties is developed further to discuss the wave-vector-dependent susceptibility chi (q) and the associated static correlation functions. A consistent theory is given of the temperature dependence of the uniform susceptibility and of the wave-vector and temperature dependence of paramagnetic diffuse neutron scattering. The theory is applied quantitatively to CeFe₂, Fe and MnSi above their magnetic ordering temperatures and to FeSi at high temperatures. In CeFe₂ and MnSi the enhancement of the Curie constant and the forward scattering peak in the neutron data are both attributed to long-wavelength 'longitudinal' fluctuations of the magnetic moment superposed on strong transverse spin disorder. The situation is predicted to be similar in BCC Fe and this differs sharply from the strong short-range order picture of Prange and Korenman (1978), Capellmann (1979) and Sokoloff (1975).

The magnetic free energy in the so-called s-d model where the itinerant d electrons and localised spins coexist is expanded in a power series of the total magnetisation M. The first-order transition observed for DyCo₂, HoCo₂ and ErCo₂ is explained from the negative values of the coefficient of the M⁴ term at T_C where the paramagnetic susceptibility diverges in the expansion of the free energy. Volume dependence of the first-order transition temperature T_C' is studied by using the expansion of the free energy up to the M⁶ term. It is shown that large experimental values of d ln T_C'/d ln V for DyCo₂, HoCo₂ and ErCo₂ can be explained by this model.

Unpolarised neutron diffraction measurements on weakly ferromagnetic PtCo and PtFe alloys give elastic diffuse magnetic cross sections which exhibit marked forward peaks confirming the inhomogeneous distribution of magnetisation in these alloys. The local moments on the solute atoms are 2.00+or-0.06 mu _B and 3.16+or-0.04 mu _B for Co and Fe respectively. These are approximately independent of composition, but it is not possible to obtain a characteristic polarisation range for the Pt matrix algorithm indications are that this should be of the nearest-neighbour distance only. The forward cross sections are much larger than would reasonably be expected from the bulk magnetisation values of (d mu _sat/dc), where mu _sat is the saturation magnetisation. It is therefore argued that the Marshall model is inapplicable to unpolarised neutron data for weakly ferromagnetic alloys. Consequently an alternative model has been invoked wherein the cross sections have been interpreted as arising from the critical scattering of neutrons by concentration-driven critical fluctuations of the spontaneous magnetisation which occur near the percolation concentration for the onset of long-range ferromagnetic order.

The forced magnetostriction arising from the magnetic field dependence of the volume and linear magnetostriction of amorphous alloys is calculated within the random anisotropy model. It is shown that forced magnetostriction in amorphous materials is determined by several mechanisms with different origins.

The anomalous skin effect is treated classically using a convective derivative to describe non-local effects. This approach yields a simple and accurate means by which to evaluate the optical properties of a metal over the entire frequency range in which the anomalous skin effect is significant. Furthermore, it can cater for any mixture of specular and diffuse surface scattering. In the infrared and visible regions, where computational difficulties are encountered with the exact theories, the model provides a more complete description of the optical properties than the approximations presently in use.