Table of contents

The total structure factors obtained by neutron diffraction of pure liquid Li and liquid Li₉Ag look very different. It will be shown that the difference is due mainly to chemical short-range order which is quite pronounced in the alloy at distances smaller then about 4 AA. At larger distances the atoms are distributed almost at random. The global structure of the alloy and its temperature variation are similar to those of pure liquid lithium.

The first use of the neutron diffraction double-null isotopic substitution technique is reported to study melt-spun amorphous Dy₇Ni₃. The appropriate scattering equations are discussed, and measured diffraction patterns presented to illustrate that the technique is indeed a practical experimental method. The ability to set the neutron scattering length of either or both of the constituent elements to zero allows a range of elegant neutron diffraction and inelastic scattering experiments to be performed.

The resistivity rho of some intermetallic compounds such as Nb₃Sn, V₃Si, UAl₂, etc. has an anomaly at a temperature T₀ near 100K. The resistivity rises steeply with T below T₀ and less steeply above it. Here the authors account for this anomalous temperature dependence using the 'classical' s-d model. Below T₀ the s-like and d-like states are hybridised by a small electronic matrix element J_sd (J_sd approximately=10-50 meV) and the electron-phonon coupling constant lambda is large. Above T₀ the electron-phonon scattering rate h(cross)/ tau exceeds J_sd, and the s-like and d-like states are decoupled. The electrical conductivity is then due mainly to the s-like states which possess a smaller electron-phonon coupling constant lambda , giving rise to a smaller value of d rho /dT. The anomaly occurs when the mean free path l is about an order of magnitude larger than the interatomic spacing d.

Doppler-broadened lineshapes of positron annihilation in pure alpha -Ti were measured in the temperature range between 293 and 1143K. The peak parameter S indicated significant effects from positron interactions with equilibrium vacancies above 970K. A vacancy formation enthalpy of 1.27+or-0.05 eV was obtained from the temperature where the curve of S against temperature deviates from the linear increase due to thermal expansion. A long exposure of the specimen to high temperatures, however, yielded a predominantly linear increase of S with temperature, as reported by Hood et al. (1976, 1978, 1979), and effects due to thermally created vacancies were smeared.

X-ray ( lambda _Mo) and neutron ( lambda =0.5 and 0.7 AA) diffraction measurements were performed on an amorphous alloy (Tb_0.22Cu_0.78)_0.93H_0.04O_0.015Ar_0.015. Interatomic distances and coordination numbers were determined separately from the X-ray and neutron data. These results are compared and discussed in relation to the measurement methods and data treatment. A supplementary interatomic distance obtained from the neutron measurements is discussed. The analysis of the neutron data at two wavelengths enables the authors to establish the alloy concentration, the hydrogen percentage and its effective mass. Evidence was obtained of a chemical order, whose nature is discussed.

It is by now well established that the structures of the elements in the liquid state follow characteristic trends reminiscent of the trends among the crystal structures of the elements. The crystal structures have been discussed very recently by Hafner and Heine (1983) in terms of the systematic variations of the effective interatomic pair potentials with electron density and pseudopotential. The present analysis of the liquid structures is based on the same set of interatomic potentials and on the optimised random-phase approximation (ORPA) for the calculation of the liquid structure. The authors show that the structural trends in the liquid and crystalline states have the same physical origin. The complex structures of the light polyvalent liquid metals (Ga, Si, Ge) are shown to arise from the interplay of two characteristic distances: the effective diameter of the hard repulsive core expressing the geometrical requirements of sphere packing and the Friedel wavelength of the oscillatory part of the potentials characterising the electronic effects in the metallic bonding. The return to hard-sphere-like structures for the heavy elements stems from the disappearance of the leading term of the Friedel oscillations as the core radius R_c approaches a value for which 2k_FR_c=¹/₂ pi .

A model presented recently by the authors for the structure and thermodynamics of liquid alloys with strong chemical interactions is applied to Na-Pb and Na-Sn. They show that the thermodynamic 'anomalies' observed in these systems arise from the interplay of ordering effects induced by a preferred interaction between unlike atoms and packing effects. They demonstrate that the structural implications of the model are supported by a diffraction experiment.

Liquid sodium has been studied at small momentum transfers by the neutron scattering method. The temperature was near the melting point. Two different neutron energies were used (3.34 and 47.7 meV with approximate resolutions of 0.16 and 1.5 meV respectively) to measure the dynamic scattering function for momentum transfer from 0.3 to 1.8 AA^-1. The observed region of energy transfers covers a range from 4 to 18 meV. The zeroth and second moments of the scattering function are determined and are compared with an independent diffraction experiment and an exact theoretical value, respectively. A comparison is made with the results from an approximate molecular dynamics simulation solely made to create an input kernel for the multiple scattering correction. The width of the scattering function is also obtained and its general feature is compared with theoretical results on other substances. An effective frequency distribution of the velocity autocorrelation function is determined and discussed.

By introducing a temperature-dependent interchange energy, omega , the entropy of mixing of NaCs liquid alloy is studied. It is found that the contribution to the entropy of mixing arising from widely differing sizes of atoms is almost compensated by the temperature-dependent interchange energy term, the NaCs alloy thus exhibiting a nearly ideal entropy of mixing in agreement with experiment. The temperature dependence of concentration fluctuations is also briefly discussed.

Molecular dynamics calculations of the shear and bulk viscosity and the self-diffusion constant were made for several states of liquid Na and K along the saturated vapour-pressure curve up to the temperature 1800K. The authors used the effective pair potential between ions developed by Price et al. (1970) with the Ashcroft pseudopotential and the dielectric screening function of Geldart and Vosko (1966). Over the whole temperature range the difference between the values obtained for the transport coefficients and the experimental ones is not more than 30%.

The positron annihilation peak coincidence rate has been measured in thorium from 20 to 1460 degrees C. A gradual S-shaped rise due to positron trapping at vacancies has been observed above 700 degrees C. The peak rate remained unchanged across the FCC to BCC transition at approximately=1345 degrees C, which indicates that positron trapping at vacancies is saturated at this temperature. A trapping-model analysis of the data yields a vacancy formation energy in alpha (FCC)-thorium of (1.28+or-0.23) eV.

The quantum diffusion of H, D, T and eta mu ⁺ in copper is studied. The impurity-host configurational interaction is evaluated according to the prescriptions of Teichler. The impurity is represented as a Gaussian wave packet and the total transition probability is evaluated explicitly at low, high and intermediate and temperatures across a piecewise quadratic potential barrier. It is found that the tunnelling transitions dominate at low temperatures while over-barrier transitions dominate at high temperatures. The characteristic transition temperature for mu ⁺ in Cu is found to be in good agreement with the experimental value. The O-T-O diffusion path is adopted for H, D and T and the O-O path for mu ⁺ diffusion. The contribution of excited states in the respective potential wells is also included. The calculated pre-exponential factors D₀ and migration energies E_m are found to be in good agreement with the experimental values wherever available. an inverse isotope dependence in D₀ and E_m is found above T=250K while it is absent at low temperatures, D₀ and E_m are also reported for the intermediate temperature range.

The electronic structure of Hg_{3- delta} AsF₆ has been considered in a model which takes the incommensurability of the Hg linear chains with the AsF₆ anion lattice into account. Despite the strength of the potential, the authors find that symmetry properties and selection rules play a dominant role in causing some matrix elements connecting states on the Fermi surface to vanish, in analogy with the conservation of crystal momentum in commensurate systems. By considering the effects of incommensurability, very good agreement is found with experimental de Haas-van Alphen data. They find that, paradoxically, some orbits in these materials exist in the absence of long-range order only.

The self-energy effects in normal-metal tunnelling were studied experimentally. Using the data for Pb, Bi and Pb₇₀Bi₃₀ the electron-phonon interaction functions were obtained.

A transport theory of itinerant ferromagnets at high temperatures is formulated. It is consistent with the modern picture of these materials in which the bands remain spin-split even above the Curie temperature. Interband scattering caused by magnetisation fluctuations is identified as the most important source of dissipation. The temperature dependence and overall magnitude of the electrical resistance of Ni and Fe are computed and found to be in good agreement with experiment.

Self-consistent band-structure calculations using the LAPW method are performed for layered structures made of Fe and V in order to study the distribution of magnetic moments near the (001) and (110) interfaces. It is found that the atomic magnetic moment of Fe at the interface is appreciably reduced from its bulk value of 2.2 mu _B and a small negative magnetic moment (antiparallel to the Fe moment) is induced on V at the interface. The result is analysed in terms of the local-environment effect and is compared with previous work on random Fe-V alloys.

The magnetic structure of the hexagonal polymorph of FeGe has been investigated by means of neutron diffraction on single crystals at low temperature and for magnetic fields applied perpendicular to the c axis. Between 410 and approximately 55K the magnetic structure is collinear c-axis antiferromagnetic. Below approximately 55K the structure changes to c-axis double-cone antiferromagnetic with an inter-layer turn angle for the basal-plane moment component of 194.4 degrees , independent of temperature and applied field. The cone half-angle increases with decreasing temperature to approximately 14 degrees at 4.2K, but its temperature dependence shows a pronounced kink at 30K, indicating a phase change at this temperature. At 4.2K the authors observe an anomalous decrease of the basal-plane moment component at a critical field (B perpendicular to c) of 1.4 T. As the temperature is increased the critical field decreases and the anomaly becomes less pronounced. The cone structure is found to persist up at least 3.9 T, which was the upper limit of the applied field. The neutron diffraction data are discussed and related to previously published macroscopic data (e.g. magnetic susceptibility, torsion and Mossbauer data) as well as to earlier neutron diffraction results.

The pressure derivative of the spin-density-wave Q vector in antiferromagnetic Cr, indicating an irreversible shift of the Q vector relative to the lattice on initial application of change in the vector Q in the unpressurised or 'virgin' sample is large and irreversible while the value for the derivative d ln Q/dp in a previously pressurised or 'hardened' sample, is essentially zero and reversible, apart from a small hysteresis at higher pressures, thus indicating an irreversible shift of the vector Q relative to the lattice on initial application of hydrostatic pressure, with subsequent locking. The values are independent of sample quality and origin and are reproducible within experimental error. This, as reported previously, is in essence the explanation for the historical discrepancy between the two classes of experimentally determined values for the derivative d ln Q/dp. The virgin state of the sample can be recovered by simply warming the sample to about 80K. The width of the hysteresis loops, expressed in units of external hydrostatic pressure, shows an initial exponential increase with the pressure p_h up to which the sample has been hardened. It decreases linearly with magnetic field and is found to relax to its final value after initial application of p_h with a time constant of approximately one hour. No evidence was found to support the prediction of Fenton (1980) that the loop width would be a function of specimen dimensions. An explanation of the hysteresis is proposed in terms of a progressive breakdown of the hardened condition.

The longitudinal and transverse magnetostriction of polycrystalline Pd, Pd₉₉Rh₁, Pd₉₇Rh₃ and Pd₉₅Rh₅ is measured at 4.2K before and after charging with about 1 at.% H. The alloys are believed to be in the alpha phase and the systematic decrease in the volume magnetostriction is explained in a semi-rigid band model as being due to filling the 4d band at a rate of 0.3+or-0.2 states per H atom. The shape magnetostriction is unchanged by hydrogenation.

Nuclear orientation and NMR measurements on ¹³¹I nuclei implanted in Fe single crystals have been compared with a simplified theory of the magnetisation behavior of the single crystals. It is shown that both the angular distribution and the resonance frequency as a function of an external magnetic field can be described reasonably with this model. Additionally, a clear difference between the zero-field resonance frequency in single crystals and that in polycrystalline samples is observed.

The results of earlier charge self-consistent KKR CPA calculations on AgPd and CuPd have been used to calculate the nuclear spin lattice relaxation rate (T₁T)^-1 of Ag and Cu in these alloys. For Ag in AgPd the contact contribution to (T₁T)^-1 is found to be dominant for the whole range of concentrations and is primarily responsible for the concentration dependence of (T₁T)^-1. It was necessary to use the relativistic expression for the contact hyperfine field to bring the numerical results into close agreement with experiment. For Cu in CuPd the contact contribution is also dominant on the Cu-rich side. On the Pd-rich side, on the other hand, this part of (T₁T)^-1 is exceeded by the d-dipolar and orbital contributions. In spite of the wide spread in experimental data for CuPd, it can be concluded that CuPd does not behave like a rigid band. Thus is in full agreement with calculations.

The authors determine the one-particle and two-particle spectra of the degenerate Hubbard model with partially filled bands within the self-consistent T-matrix approximation and the ladder approximation, respectively. They have found that the effect of self-consistency is important even for metals with a small particle concentration, e.g., Ni. The results are compared with the theory of Treglia et al. (1980, 1981) as well as with the theory of Cini (1976, 1979), which represents basically the non-self-consistent version of the approach.