Table of contents

Charge distributions in Ni_nH clusters with n=6, 14 and 38 are calculated by a semi-empirical INDO (intermediate neglect of differential overlap) Hartree-Fock SCF (self-consistent field) method. The computations in this formalism allow also for the simulation of small H concentrations (e.g. n=38) in the Ni lattice. The magnitude of the charge transfer is studied as a function of n (i.e. size of the Ni cluster) and the lattice parameter. The influence of intra-atomic correlations on the magnitude of the charge transfer is estimated by an 'atom in molecule' type of approach. The semi-empirical INDO results are compared with local density functional calculations using the LMTO formalism and calculations of a proton in jellium. It is concluded that the proton in Ni can be interpreted as an overscreened particle and that the electronic redistribution caused by H in Ni is localised within the first shell surrounding the proton site.

The feasibility of a combined neutron and X-ray scattering experiment in a liquid metal for determining the static dielectric function is discussed. It is shown that a reasonable response function is obtainable from available experimental data in the case of aluminium. The sources of errors present in the proposed approach are also discussed.

The self-consistent linearised augmented plane-wave method with a general potential has been applied to study the electronic structure of the heavy fermion superconductor UPt₃ within the local density approximation. Comparison with experimental data supports a strongly frequency-dependent self-energy correction which diminishes rapidly away from the Fermi energy E_F but can be characterised by a large mass enhancement factor of 21 at E_F. Calculation of the conventional electron-phonon interaction strength leads to the rather small value of lambda _ep=0.28. It is argued that conventional superconductivity is extremely unlikely in UPt₃ because of the interactions responsible for the large mass enhancement.

Theoretical calculations of momentum distributions of positronium (Ps) atoms ejected from clean metal surfaces are presented and compared with recent experimental results. The authors find that the momentum dependence of the Ps-forming interaction significantly affects the shape of the spectra. They also show that, within the model, the energy distribution of Ps atoms does not vary directly with the density of electronic states just outside the surface.

For pt.I see ibid., vol.16, p.421 (1986). The decomposition of CuNiFe alloys containing 46-48 at.% ⁶²Ni and 2 at.%, 4 at.% and 8 at.% Fe is studied by means of small angle neutron scattering (SANS). The data are evaluated by fitting model calculations to the scattering curves, by scaling analysis and by Fourier transformation. It turns out that consideration of the periodic cluster arrangement in the interparticle interference function is necessary to describe the structure function in an appropriate way. Further, the authors find that the SANS curves can successfully be scaled except for the very first decomposition stages, in spite of a change in morphology which occurs during aging. A comparison of the results of the different analysis procedures is given, and kinetics and temperature dependence of the relevant scaling parameters are outlined and discussed.

The structure and thermal stability of rapidly solidified Al_86-xFe₁₄Si_x and Al_86-xMn₁₄Si_x (0<or=x<or=16) alloys have been investigated. The Al-Fe-Si alloys show predominantly a quasicrystalline icosahedral phase for x<or=6 and an amorphous phase for larger x. X-ray and thermal analysis measurements indicate that, for a region of compositions near x=6, one phase grows at the expense of the other. Compositions with x near six show a mixture of these two phases but these two phases remain distinct. The alloys Al-Mn-Si show similar behaviour but here the transition from icosahedral to amorphous occurs near x=13.

The variations of the activation energy, the Curie temperature and the Mossbauer parameters of metallic glasses containing bismuth have been investigated by means of DTA and magnetic measurements. It is found that there are two distinguishable relaxation stages in the thermal relaxation processes. The mechanism of formation of these two relaxation stages has been analysed in terms of short-range localised atomic ordering and medium or long-range atomic ordering. The variations of the Mossbauer parameters during the thermal relaxation processes are explained in terms of TSRO and CSRO relaxation processes. Moreover, isothermal relaxation processes of metallic glasses containing bismuth have also been discussed.

The lattice expansion of Pd during the absorption of hydrogen has been investigated with a combined technique in which the length changes and hydrogen concentration were measured simultaneously and on the same sample in the single-phase region of the PdH_c phase diagram above the critical point. For atomic ratios up to c=0.65, the hydrogen induced relative volume changes are found to be Delta V/V₀=0.200 c-0.0493c². This concentration dependence of the lattice expansion, as well as that of the partial molar enthalpy Delta H_H of hydrogen in Pd, can be described successfully by taking into account the volume and concentration dependences of the characteristic band-structure energy (E_F-E_s) introduced in the semi-empirical heat of formation model of Griessen and Driessen. Using this model the thermally broadened density of states in PdH_c, as a function of band fillings above the Fermi level in pure Pd, is deduced from the experimental data and it is shown that the strong concentration dependence of Delta V/V₀ in PdH_c originates from the difference between the volume dependences of the almost completely filled d band at low hydrogen concentrations and the occupied s-p band at higher hydrogen concentrations. In NbH_c where such a d band to s-p band 'transition' does not occur the concentration dependences of Delta H_H and Delta V/V₀ are dominated by the volume dependence of the partially filled d band.

The absorption of hydrogen in thin Pd films with thicknesses of 300 nm and 50 nm is studied experimentally by measuring pressure-composition isotherms with a microbalance at temperatures up to 590K and hydrogen pressures up to 50 bar. In contrast to earlier findings obtained with the vibrating quartz technique, the 300 nm isotherms do not deviate significantly from those of bulk PdH_c. For the 50 nm films dissociation plateaux are narrower than for bulk PdH_c and the critical temperature is reduced by approximately 26K. The values of Delta H_H and Delta S_H obtained from the plateau pressures are practically the same as those for bulk PdH_c. The influence of volume restrictions (e.g. clamping) on the effective H-H interaction is described with a recently developed model based on the semi-empirical heat of formation model of Griessen and Driessen. It is argued that apart from clamping, the narrowing of the isotherms may also be related to an enhancement of the surface to volume ratio due to the crystallite formation in thin films and a distribution of site energies caused by intrinsic stresses in films that adhere to substrates.

The absorption of hydrogen in several amorphous Pd-Si alloys as well as in crystallised Pd₈₀Si₂₀ is studied experimentally for hydrogen pressures up to 50 bar and temperatures between 333 and 493K. The absorption of hydrogen is detected in situ with a combined dilatometric/resistometric apparatus and with a microbalance. From a comparison of these measurements useful relations between lattice expansion, resistance increase and hydrogen content are determined. The pressure-composition isotherms of the structurally disordered host metals are described theoretically by a gaussian distribution of site energies, in combination with a nonlocal effective H-H interaction that arises from elastic and electronic changes in the host metal. From the concentration dependence of the interaction energy, it is deduced that the Fermi level in the glass-forming Pd-Si alloys is located at a minimum of the electronic density of states.

Elastic constants for amorphous iron are calculated by making use of a simulated amorphous structure. The bulk modulus is 4% and the shear modulus 30% lower than for the crystalline material, in agreement with a number of experimental data. The atomic rearrangement contribution is of crucial importance in obtaining this agreement. After a simulated mechanical annealing, the bulk modulus changes markedly, due to the rearrangement term. Accurate and simple expressions for the rearrangement contributions are derived, which are at the same time illustrative of the origin of these terms; they can be calculated without simulating the deformation itself, using force quantities f (with different components) which measure the local state of relaxation around an atom. Using these f quantities a plausible explanation can be given for a number of observations. The f can be interesting tools for future research on the relaxation of amorphous materials.

The generalised model potential theory, developed recently for liquid sp-type metals, is extended to the variational calculation of the Helmholtz free energy for the alloys of these metals. Hence, model potentials, which include the usual high-order pseudopotential perturbation (HOPP) effects, are obtained for use in alloy calculations. Thesese model potentials are applied to first-principles calculations of thermodynamic quantities and electrical resistivities for some typical liquid metallic binary alloys. The results thus obtained are found to be in reasonable agreement with observation. When compared with the results obtained using the usual low-order pseudopotential perturbation theory, they indicate that HOPP effects become extremely important in the case of the heat of mixing for any metallic binary alloy, whereas for some other properties these effects are essential only for those metallic alloys in which charge transfer is significant.

The self-consistent symmetrised relativistic APW method has been employed to calculate the band structure of Pb at normal and high pressure. A one-electron potential is constructed on the basis of the local density approximation. The calculated electronic density of states at normal pressure gives a satisfactory agreement with both X-ray photoemission and specific heat data. Calculated de Haas-van Alphen frequencies for the Fermi surface at normal pressure agree well with experimental results. The pressure derivatives of the Fermi surface cross sections are also calculated and they agree satisfactorily with experimental results.

The de Haas-van Alphen effect has been used to investigate the exchange splitting on the X-ellipsoid sheet of the Fermi surface of Pd and dilute Pd(Fe) and Pd(Ni) alloys. In pure Pd the g-factor shows appreciable anisotropy which can be explained in terms of the spin-orbit interaction and an isotropic Stoner enhancement. In the alloys an extra ferromagnetic exchange splitting is found. In Pd(Fe), at 2K, this is saturated at fields of order 3 T while in Pd(Ni) it varies linearly with field and decreases with increasing temperature between 1 and 2K. A striking but unexplained feature is the complete absence of spin-dependent scattering in Pd(Fe).

De-Haas-van Alphen (DHVA) frequencies corresponding to the neck orbits of a majority-spin Fermi surface have been observed over a wide range of orientations in the ferromagnetic FCC alloy Co_0.92Fe_0.08. Self-consistent band-structure calculations can largely explain the experimental results, albeit that the predicted neck diameter is 20% too small.

Transverse electron focusing (TEF) of the second-zone hole surface orbits in aluminium single crystals has been observed and used to investigate the probability (p) of specular reflection of conduction electrons at sample surfaces for the three principal crystallographic planes-(100), (110) and (111). The p-values obtained range from 0 to 0.61+or-0.05 depending on both the crystallographic plane and the geometry of the orbit when the surfaces are electrochemically polished like a mirror. Even on the chemically etched dull surface, up to a quarter of conduction electrons are specularly reflected. For specific orbits, the signal shapes are compared with those computed from the four-OPW Fermi surface model.

The matrix elements of the electron-phonon interaction are calculated by the APW method for copper and silver. The matrix elements of the electron-impurity interaction are also evaluated by the KKR method. On the basis of these matrix elements the authors have discussed the low-temperature electrical resistivity of the nominally pure samples of copper and silver and their dilute alloys. The Boltzmann equation is solved by the method of the vector mean free path combined with Ziman's variational principle. The theoretical results are compared with the available experiments.

A new description of the DC conductivity in polycrystalline bulk metal and thin metal films is proposed in which isotropic background scattering and scattering due to grain boundaries are considered simultaneously. The general formula for the relaxation time for partly diffuse and partly specular scattering grain boundaries is obtained. The well known Mayadas and Shatzkes and Pichard-Tellier and Tosser models represent special cases of the author's description. For appropriately high specular reflection and small transmission coefficients of electrons through the grain boundaries a negative value of the temperature coefficient of the resistivity is obtained.

The authors present magnetoresistance measurements at low temperature on a series of metallic glasses: Mg₈₀Cu₂₀, Cu₅₀Lu₅₀, Pd₈₀Si₂₀, Cu₅₀Y₅₀, Y₈₀Si₂₀, Cu₅₇Zr₄₃, CuYGd and CuLuGd. The magnetoresistance clearly exhibits three-dimensional localisation effects with a strong influence of the spin-orbit scattering: the magnetoresistance is negative for weak spin-orbit scattering (Mg₈₀Cu₂₀) and positive for strong spin-orbit scattering (Cu₅₀Lu₅₀, Pd₈₀Si₂₀), and changes from positive at low field to negative at high field for intermediate spin-orbit scattering (Cu₅₀Y₅₀, Cu₅₇Zr₄₃). The magnetoresistance curves can be accounted for by only localisation terms for most alloys and interaction effects are significant only for Y₈₀Si₂₀. An additional contribution from superconducting fluctuations is observed for the superconducting alloy Cu₅₇Zr₄₃. The analysis of the magnetoresistance yields the spin-orbit and inelastic scattering rates. The inelastic scattering rate of the alloys varies as T² at high enough temperature and departs from T² below about 5K. The measurements in a dilution refrigerator on Cu₅₀Y₅₀ and Y₈₀Si₂₀ have shown a variation as T^1/2 below 1K. They discuss the possible origin of such a variation. They finally present measurements on Cu₅₀Y₅₀ and Y₈₀Si₂₀ doped with Gd. Surprisingly the addition of magnetic Gd impurities does not depress strongly the localisation effects, as would be expected from the existing theories.

Soffer's angular-dependent model of the scattering of electrons at a metal surface is applied to both the longitudinal and the transverse magnetoresistance of thin foils and films. In the transverse case the Hall effect is also evaluated. The results of these calculations are presented in graphical form. For the case of a longitudinal field the results are very similar to those obtained by Way and Kao using the Fuchs model of surface scattering. In the case of the field transverse to the current and normal to the plane of the sample, oscillations in conductivity as a function of magnetic field are found which, although similar to Sondheimer's results based on the Fuchs model, differ in the interesting regime of thin, rough samples at low magnetic fields. Furthermore, while the Hall coefficient shows the Sondheimer plateau behaviour at low fields, its variation at higher fields is significantly different.

Co₇₃Fe₆Mo₃B₁₈ amorphous thin films have been prepared by using a RF sputtering technique. The optimum sputtering parameters are determined for this case. The high-frequency impedance of the films has been studied experimentally. The frequency relation for the impedance is described very well by the relation Z=(a₀+ia₁ omega )/(b₀+ib₁ omega ). The calculated value of a₀/b₀ is consistent with that extrapolated from the resistance-frequency curve.

The properties of the heavy-fermion material CePb₃ have been examined within the framework of density functional theory. The LMTO method has been used to study the electronic structure in both the paramagnetic and antiferromagnetic phases. Comparison of the bare susceptibility and specific heat with experiment enables the authors to estimate the enhancement factors which turn out to be very large. The details of the magnetic properties are discussed. They have also calculated the superconducting transition temperature T_c and show that the electron-phonon interaction is a feasible mechanism for superconductivity. The results are compared with an analogous calculation for CeSn₃. The implications of this work for current theories of heavy fermions are discussed.

The energy spectrum of a rare-earth impurity in a normal metal is considered. It is possible to achieve the degeneration of the magnetic energy sublevels of the impurity under an external magnetic field. This degeneration gives the anomalous Kondo scattering of conduction electrons. In the region of weak coupling, the amplitude of scattering and the contribution of this mechanism to the electroresistance of the metal are calculated. The effect of conduction-electron scattering on the impurity couples is considered. The influence of Landau quantisation on the effect is studied. It is shown that the investigation of Kondo anomalies can be used as a tool to obtain information about crystal fields in metals.

The magnetic susceptibility of magnetron-sputtered B1 phase MoN films has been determined by magnetisation measurements with a Faraday balance. The electronic DOS at the Fermi level N(0) estimated from the susceptibility data is much lower than the band-structure value for an ordered compound. The discrepancy is explained by structural defects in the experimentally prepared films. Two models for defect structures appear to be consistent with the results of annealing experiments. One model suggests that nitrogen atoms are located on Mo sites accompanied by distortions of the Mo sublattice. The other model postulates structural instabilities when nitrogen is incorporated on the octahedral N sites beyond the gamma -Mo₂N composition.