Table of contents

Electron-phonon scattering leads naturally to cubic and exponential exp(- theta /T) temperature dependences of quasiparticle scattering rates at low temperatures. The authors describe a mechanism by which a quadratic dependence can arise in orbital averages without necessarily appearing in point rates anywhere on the orbit. The estimated magnitude agrees with recent radio-frequency size effect data in Cd. They predict that cubic or exponential behaviour should be recovered at temperatures below about 1K.

Precise measurements of the electrical conductivity sigma (T) are presented for Cu_100-xTi_x (x=37, 40, 50, 56, 59) metallic glasses between 1.5 and 25K. The temperature and concentration dependences can be essentially accounted for by quantum interference and electron-electron interaction corrections. Characteristic magnetic fields describing the strength of inelastic and spin-orbit scattering as well as the screening factors are consistent with results from the magnetoresistivity.

The Hall effect of UPt₃ has been measured between 1.9 and 300K. The magnitude of the Hall coefficient decreases by a factor of ten as the temperature is lowered from 30 to 4.2K whether the specimens become superconducting at T_c=0.48K or not. Unlike other heavy fermion materials for which the Hall effect has been measured, the Hall coefficient is negative above 5K. The large drop below 30K is attributed to a correlation in scattering on the Fermi surface that appears at a somewhat higher temperature than that at which the heavy fermion state is clearly established.

Mossbauer spectra with ¹⁵⁵Gd were taken of a Gd₂Fe₁₄B single-crystal absorber. These spectra, which ave a more pronounced structure than those of powder absorber, lead to considerably improved precision in the hyperfine interaction parameter values and in estimates for the second-order crystal field parameters. A comparison between spectra taken in zero field and in a field 8 T shows the magnetic hyperfine field at Gd nuclei to be positive at both lattice sites.

The authors report on the first measurements of the change in electrical resistivity of potassium during low-temperature irradiation with 3 MeV electrons or fast neutrons, and during subsequent isochronal annealing. An analysis of defect production in pure and rubidium-doped samples shows that one of the defects produced-the interstitial-is already mobile at the irradiation temperature T_irr<or=6K. The main feature of the annealing spectrum is a dominant peak at about 16K. Its position is shifted to lower temperatures if the total annealing in this stage increases. A detailed analysis of this behaviour based on a rate-equation approach yields an activation energy of E_M=0.038+or-0.004 eV for the defect-identified as the vacancy-to migrate in this stage. This value of E_M is compared with recent theoretical and experimental results.

The lattice statics method has been used for relaxation calculations in lead due to a vacancy created in the lattice. Using the measured phonon dispersion relations and the elastic constants as constraints the dynamical matrix is derived considering progressively the interactions up to seventh neighbour. The calculation shows that the first twelve neighbours relax towards the vacancy whereas the second six relax outwards. In general, the displacements are found to be larger in the close-packed directions. The divacancy binding energy is found to be in accordance with other calculations and the bond model. The formation energy of the monovacancy is also found to be very close to the experimental value.

By means of multiple beam interferometry in evaporated wedge-shaped metal films and movable beam of coherent pulsed phonons it is possible to determine the attenuation and phase velocity of longitudinal and transverse hypersound in these films. With a pulse echo apparatus for 35 GHz the authors measured the phase velocity and the amplitude damping coefficient for longitudinal hypersound in evaporated silver films.

Charge transfers in liquid Cs-Na, Na-Pb and Na-Sn alloys are estimated using partial structure factors in the long-wavelength limit and Thomas-Fermi screening theory. It is verified that the results are consistent with the magnetic susceptibility data for liquid Na-Pb and Na-Sn previously measured by the present authors.

The author compares the thermodynamic excess quantities for nine liquid metals which are calculated from the pseudopotential-induced interionic potential and the thermodynamic perturbation methods from the hard-sphere fluid. He considers two useful and widely used methods: the variational approach (VA) and the optimised random phase approximation (ORPA). The latter method allows the determination of the deviation of the liquid metal structure from the hard-sphere one, but not the VA. He shows that the corresponding change of the structure energy is not well described by the variational procedure. In the alkalis, the difference between the VA and the ORPA comes form the estimation of the short-ranged repulsive forces, as in the case of the Lennard-Jones fluid, but in the polyvalents metals the Friedel oscillatory forces also make a significant contribution. He points out the usefulness of the ORPA for the analysis of the thermodynamic stability of the liquid structure and he shows that the distortions of the liquid metal structure discussed recently by J. Hafner and G. Kahl (1984) are energetically favourable.

The authors' recently suggested approach (1984) for calculating the Helmholtz free energy, including the usual high-order pseudopotential perturbation corrections for the nearly-free-electron-like liquid metals, is critically examined and is now improved so as to be applicable in principle to any sp-type disordered metal in its liquid phase. It appears that the mean atomic volume, the packing density parameter and the related thermal quantities can now be predicted from first principles with reasonable accuracy by the presently described variational thermodynamic calculation for nearly all the sp-type liquid metals.

Calculations of the densities of states of amorphous Zr_cT_1-c alloys where T=Fe, Co, Ni or Cu, are presented. The method of calculation is an adaptation of the method used by the authors to analyse the virtual bound states on impurities in Ag and Cu alloys (1985). The values of the calculated densities of states at the Fermi level are used to draw conclusions concerning the magnetic susceptibilities chi and the specific heat coefficients gamma which are discussed as a function of Zr concentration, c. Agreement with some experimental trends is obtained and, in particular, a decrease in chi of gamma with increasing c in Zr_cCo_1-c and Zr_cFe_1-c alloys can be understood. This behaviour is attributed to the close proximity of the Co or Fe d band to the Zr d band.

Fully relativistic self-consistent LDA band-structure calculations for CeSn₃ are reported using the LMTO method. Fermi surface areas are in agreement with experimental but band masses, density of states and electronic pressure are in wide disagreement. A one-parameter renormalisation prescription based on the large-N limit of the infinite-U Anderson model is introduced and the band structure recalculated. The 4f band then splits into distinct j=5/2 and j=7/2 bands. Using the renormalised band structure and band masses, density of state and electronic pressure are found to be in agreement with the experimental data.

Magnetic properties of metallic surfaces at finite temperatures are discussed by using the single-site spin-fluctuation theory in which the effects of spin fluctuations both at the surface and in the bulk are taken into account on the same footing by means of the static functional-integral method. The temperature dependence of the layer-by-layer magnetic moments on the first time layers is calculated numerically for the semi-infinite simple cubic lattice with the (100) surface. Model calculations using the parameters appropriate to Ni show that the magnetic order at the 'clean' surface persists above the bulk Curie temperature, T_Cb, and that the amplitudes of local spin fluctuations on the first few layers are significantly enhanced near T_Cb compared with that in the bulk. Calculated results are compared with those obtained in the mean-field theory for the semi-infinite simple cubic Ising model and they are discussed with reference to relevant experimental data.

Detailed AC susceptibility measurements have been performed on amorphous Fe_xNi_80-xB₁₉Si₁, x=10, 13, 16, and Fe₉₁Zr₉ alloys in the critical region. In accordance with the predictions of the renormalisation group calculations and the Harris criterion, the critical exponent gamma of the pure (ordered) system is found to remain unaffected (within the error limits) in the presence of a quenched disorder which is not too strong.

The lattice expansion in 99.999% pure nickel has been studied to within epsilon = mod (T-T_c)/T_c mod =10^-3. It is observed that during the ferromagnetic-to-paramagnetic transition, nickel undergoes anomalous expansion leading to a lambda -type anomaly in the thermal expansion coefficient, beta (T). From beta (T), the critical part of the thermal expansion coefficient beta _c(T) was separated and fitted to the power-law equation beta _c^+or-=(A^+or-/ alpha ^+or-) epsilon ^{- alpha +or-}(1+E^+or- epsilon ^x+or-)+B^+or-. The parameters thus obtained are in good agreement with the values predicted by the renormalisation group theory and those derived from specific heat measurements.