Table of contents

The authors present a simple model for the calculation of the entropy of mixing of liquid transition-metal-based alloys with strong chemical interactions. They show that the thermodynamic 'anomalies' observed in these systems arise from the interplay of electronic and packing effects.

The author suggests that a strongly exchange enhanced paramagnet in a magnetic field will undergo a sequence of metamagnetic instabilities, due to the Landau level structure in its density of states. The sizes of the discontinuities in the magnetisation are calculated, as well as the low-temperature thermal properties. The phenomenon may be observable in the millikelvin temperature range in materials such as TiBe₂ or Ni₃Al.

The compound Gd₂Fe₁₄B was investigated by means of ¹⁵⁵Gd Mossbauer spectroscopy. From the value of the quadrupole splitting found in this compound and in the compounds GdCo₅ and Gd₂Co₁₇ investigated earlier the authors determined experimental values for the corresponding second-order crystal-field parameters V₂⁰, which mainly govern the magnetocrystalline anisotropy in related permanent magnet materials such as Nd₂Fe₁₄B, SmCo₅ and Sm₂Co₁₇.

The structure of the boron-rich metallic glass Ni₆₄B₃₆ has been investigated by neutron diffraction under conditions of high spatial resolution. Three specimens with different isotopic abundances have been examined and the total radial distribution functions (RDF) and the more fundamental pair correlation functions have been obtained. A comparison between these data suggests that caution is needed in the interpretation of the total RDF of metallic glasses at high resolution. Even when the first-neighbour peak is split into distinct components, it may be impossible to make an unequivocal identification of the atomic correlations involved, as might be the case, say, for a locally ordered conventional glass. This suggests, for metallic glasses at least, that measurements at high resolution cannot be a true substitute for the proper derivation of pair correlation functions in the way that has sometimes been suggested in the past.

The diffuse X-ray scattering from the displacement field around N and H in single crystals of Nb was studied at various temperatures. The observed increase of the scattering intensity with decreasing temperature can be explained by the trapping of H by N. A trap model can describe the observed changes. The displacement fields of H and N do not add linearly. On the assumption that the force dipole tensor P=(^AB_B) describing the displacement fields does not change for N, the trace (A+2B) of the tensor for H is lowered by 10-20% and a tetragonality mod A-B mod /(A+2B)=0.08+or-0.02 is observed for the trapped state.

Calculations were carried out for the properties of multiple vacancies and interstitials in Cu and Ag using the molecular dynamics technique in conjunction with interatomic potentials derived from first principles. The most stable configurations and energies of formation and binding of defect clusters as well as the mechanisms and energies of migration for divacancies and di-interstitials are reported. The theoretical results are compared with the available experimental observations and with previous defect calculations based on short-range empirical potentials.

The thermal variation of atomic beam diffraction intensities on (1, 1, 2n + 1) faces of Cu is discussed for large n. the dramatic decrease of intensity at about 600K is attributed to a roughening transition, above which the steps are fuzzy. The steps are assumed to repel themselves with a repulsion energy proportional to (2n+1)^-2, of elastic or electrostatic nature or both. The main parameter of the theory is the energy W₀ of a kink on a step. Experimental data are consistent with a value of W₀ between 0.2 and 0.3 eV, in agreement with very rough theoretical evaluations. Predictions are presented concerning the lineshape of diffraction spectra in the high-temperature rough phase, where intensities cannot be written as delta functions. The Bragg peak intensity vanishes just below the roughening transition temperature, T_R, as exp(-constant*(T_R-T)^-12/), where the constant depends on the particular Bragg peak.

Based on previous work on the magnetic-moment distribution, the authors demonstrate that it is possible to obtain fairly detailed information about the local atomic configuration in an artificial superstructure film of the Fe-V system by analysing experimental data on the hyperfine field distributions. A new method is used to treat the atomic short-range order in systems with a concentration modulation from layer to layer. As examples, detailed analyses are made for some of the samples prepared by the ultra-high-vacuum deposition technique.

The surface and bulk densities of states of a solid described by the stacking of principal layers are obtained by means of an iterative procedure which allows (i) the inclusion of 2ⁿ layers after n iterations, (ii) the simultaneous calculation of the Green functions for both the 'right' and 'left' surfaces as well as for the bulk (or central) principal layer, and (iii) the use of imaginary parts eta as small as one wishes in the energy without any large increase in computing time, so that the limit eta to 0 can really be obtained. As a by-product the authors obtain (i) the 'right' and 'left' transfer matrices of the 'effective field' or continuous fraction approach and (ii) a factorisation theorem which relates the Green functions of both surfaces to the Green functions of both surfaces to the Green functions of the bulk and the free metal atom.

The hydrostatic pressure dependence of cyclotron effective mass of potassium up to 4.4 kbar in solid helium was obtained from de Haas-van Alphen measurements using the field modulation technique. The logarithmic pressure derivative of the cyclotron effective mass was found to be -(0.71+or-0.12)*10^-2 kbar^-1 and the zero-pressure mass was (1.211+or-0.005)m₀, where m₀ is the free-electron mass. The decrease of the quasiparticle mass with increasing pressure is compared with theoretical predictions of the changes of band mass, electron-electron and electron-phonon mass enhancement parameters with pressure.

The Fermi surface of ordered Cd₃Mg has been studied by using the DHVA effect. Eighteen separate DHVA frequency branches were observed and the frequencies ranged from 2*10⁵ G to 4*10⁷ G. The abundant branches observed are associated with the superlattice structure of Cd₃Mg. The overall features of the Fermi surface of Cd₃Mg are relatively simple and resemble those of the Fermi surface of Mg rather than Cd; the third-zone electron surface centred at the K point, the third- and fourth-zone electron surfaces around the L point and the so called monster waist all exist in Cd₃Mg. However, the dimensions of these Fermi surfaces are about half those in Mg.

The temperature dependence of the electron-phonon mass enhancement parameter through a second-order structural phase transition is calculated and is shown to agree with the observed temperature dependence of the effective mass in samples of Pb_1-xSn_xTe through the structural phase transition.

Sondheimer's exact solution of the Boltzmann equation with a degenerate kernel takes a particularly simple form when the scattering can be described by a small number of phase shifts. The authors discuss the practical implementation of this form, stressing its economy (which is comparable with commonly used non-trivial approximations) and its generality (independence of the formalism used to compute the phase shifts). Various approximations are compared with the exact resistivity as functions of the phase shifts, using aluminium as the host metal. This comparison helps distinguish several 'band-structure' effects as compared with free-electron predictions. The most dramatic band-structure effect (in the example of aluminium) is the vanishing of the resistivity as all phase shifts except for the s wave approach zero. They discuss the potential usefulness of these results in free-electron-based calculations as well as those which include host band structure from the beginning.

Measurements of the resistivities of liquid K-Pb, K-In, Li-Pb and Na-In alloys are presented. The resistivities of the lead alloys exhibit a sharp peak at c_Pb=21 at.% for Li-Pb and at c_Pb=49 at.% for K-Pb. This is in agreement with the transition from simple ionic behaviour in Li-Pb to poly-anion formation in K-Pb. The maximum resistivity of K-Pb distinctly exceeds the range of metallic conduction. In the indium alloys the resistivity maximum is less pronounced and the theoretical description is more complicated.

The authors present experimental results for the thermopowers of a number of amorphous transition-metal alloys in the temperature range 2-300K. They show that the observed temperature dependences arise from the temperature-dependent electron-phonon enhancement factor lambda (T). The form of lambda (T) obtained implies that the electron-phonon coupling factor, alpha ², does not vary as omega ^-1 as is the case for amorphous simple metals but in fact increases as omega increases. The magnitudes of the observed enhancements are generally in good agreement with those obtained from superconductivity data and the inclusion of spin-fluctuation and electron-electron interaction effects does not alter this agreement. They do, however, find significant disagreements for a number of alloys, which may be an indication of further renormalisation contributions. They also briefly discuss the observed signs of the thermopowers, considering the predictions of the Mott s-d scattering model and the possibility of significant d-band conductivity.

The thermopower of five amorphous CuTi films of varying composition has been measured at temperatures from 4.2 to 300K. An enhancement effect is seen as the temperature decreases (as in metallic glass ribbons) which is consistent with that expected from the electron-phonon interaction.

The total energy, pressure and spin susceptibility of paramagnetic Cr and Fe are calculated as functions of a large tetragonal deformation using the local density and local spin density approximations. The electronic structures, structural stabilities, elastic constants and their volume derivatives, theoretical tensile strength and strain dependence of the spin susceptibility are discussed on the basis of the results. Although the muffin-tin approximation is used for the charge and potential distributions, the calculated results agree essentially with the elastic properties observed.

A reanalysis of the unpolarised neutron magnetic diffuse scattering data for weakly ferromagnetic PdFe and PdCo alloys shows that it is the geometrical critical scattering model rather than the Marshall model (used by Low and his co-workers 1963, 1966) that provides a self-consistent interpretation of these data. Consequently the earlier deductions from these data of a large polarisation range ( approximately 10 AA) of an Fe or Co atom in Pd and the relatively low solute concentration at which the Pd atoms become magnetically saturated are incorrect. Instead it is argued that the polarisation range in Pd is probably not much larger than the nearest-neighbour distance and that ferromagnetism in Pd alloys arises through a percolation process involving ferromagnetic correlations between otherwise isolated polarisation clouds rather than through the direct overlaps of very spatially extended polarisation clouds.

Applied fields of up to 5 T in the b direction give anomalies in magnetisation curves, the number and character of which change with temperature. Only that at highest field ( approximately 3 T) in this range is visible above the temperature (8.3K) of the susceptibility maximum and up to 14K; the field for this is a maximum at about 8K, falling slightly at lower and higher temperatures. A second, more hysteretic and more temperature-dependent anomaly is seen below 8.3K, falling from 2.9 T there to 2.4 T at 4.2K. Two low-field transitions are visible below 7K, the lower and more hysteretic one dominating below approximately 5K; both rise in field (to approximately 0.8 and 1.0 T at 4.2K) with falling temperature. Some correlations can be made with other field-dependent properties. For fields along the c axis only one major effect is found-a regime of enhanced dM/dB around 0.9 T visible below 8.3K and becoming more marked on cooling to 4.2K.

Gd_xCo_100-x (50<or=x<or=75) amorphous alloys have been prepared by melt quenching and high-rate sputtering in order to investigate the Curie temperature and its pressure effect. The concentration dependence of the Curie temperature does not show a steep variation in the composition range from 75-50% Gd, in contrast to that in the range of low Gd concentrations. The Curie temperature of the crystalline compound Gd₄Co₃ is almost the same as that of its amorphous counterpart and the pressure effect on the Curie temperature is also very similar in both cases. The Curie temperature of amorphous Gd₇₅Co₂₅ is 145K, although the crystalline state is antiferromagnetic at the same composition. The shift of the Curie temperature T_c under hydrostatic pressure P for alloys with about 50% Gd is very significant, being comparable with that of crystalline Invar alloys, and the magnitude of dT_c/dP decreases as the Gd content increases, reaching almost zero at 75% Gd. The plots of dT_c/dP against 1/T_c show a positive slope, in contrast to Co-based crystalline compounds. On the other hand, the slope of plots of dT_c/dP against T_c for Gd-Co amorphous alloys is about twice as large as that of Fe-based amorphous alloys containing metalloids with a low Curie temperature.

The basal-plane anisotropy in a hexagonal easy-plane ferrimagnetic system yields discontinuities in the magnetisation curve for external fields applied in the easy plane. At the discontinuities first-order moment reorientation (FOMR) transitions occur. The lowest transition field depends strongly on the difference between the values of the sublattice magnetisations and is very sensitive to the intersublattice coupling. The transition field provides a direct method for evaluating the exchange coupling between the 3d and 4f spins. The lowest transition field in the series of the easy-plane R₂T₁₇ compounds (where R=rare-earth metal and T=Fe, Co) occurs for Ho₂Co₁₇ and amounts to 21 T.

The method of electron paramagnetic resonance (EPR) is show to make possible the study of orbit-lattice coupling in bulk metal crystals. The magnetic resonance of dilute Er impurities in copper and silver single crystals under uniaxial stress has been investigated experimentally. The authors have also measured the magnetisation of CuEr crystals in magnetic fields up to 90 kOe. They have used the angular variation of the g factor and the field dependence of the magnetisation to obtain the orbit-lattice coefficients V_r⁽²⁾. Theoretical calculations of the parameters V_r⁽²⁾ have been carried out in the framework of the 5d virtual states model. It has been established that the covalency effects between 5d states of erbium and 3d states of copper or 4d states of silver are of importance in the formation of orbit-lattice coupling.

New theoretical developments in impurity lattice dynamics are presented, with the emphasis on phonon broadening of core levels in metals. Detailed calculations have been made for Li and Na. The effects of non-linear screening are evaluated for the first time and are shown to increase the phonon broadening by about 70% for Li compared with the linear-screening value. The present a priori estimate of the phonon broadening of the Li 1s level is the first which agrees with experiment. The effects of core-hole-lattice couplings of second order in ion displacements are also evaluated and shown to be unimportant for Li and Na. The usual linear-screening theory for lattice dynamics and for impurity-lattice couplings is analysed critically with regard to its sensitivity to the approximations for the pseudopotential and for the screening function. Both optimised model potentials and norm-conserving a priori potentials are considered. Finally, an indirect phonon broadening mechanism proposed previously is analysed and shown to give a negligible effect.

The semi-empirical model of Griessen and Driessen (1984) which relates the heat of formation Delta H of a metal hydride to the difference of the Fermi energy E_F and the centre E_s of the s band of the host metal, is used to calculate the volume expansion accompanying hydrogen uptake by a metal. The values obtained from the model are in good agreement with experimental data.