Table of contents

Experiments on potassium and K-0.077 at.% Rb strained at 60K indicate the presence of a new temperature-dependent scattering mechanism at T approximately 0.25K. The authors ascribe this to inelastic scattering of vibrating dislocations. At the same time the suppression of phonon drag at higher temperatures results in the emergence of the normal electron-phonon term T⁵. Both these effects obscure the effect of dislocations on the electron-electron scattering term AT².

It is shown that there is excellent agreement between the theory and the recent data of Zwart et al. (1983) for the strain dependence of the electron-electron scattering resistivity for dilute CuAg alloys.

The temperature and pressure dependence of the electrical resistivity of a Cr+0.92 at.% Ga alloy are reported. At atmospheric pressure the resistivity behaves anomalously at the Neel temperature but not at the incommensurate-commensurate SDW transition point. On the application of hydrostatic pressure there appears a second resistivity anomaly, not seen before, which accompanies the incommensurate to commensurate transition. This anomaly increases with increasing pressure.

A theory is presented for the linear increase in resistivity, in simple metals, with large magnetic field. In calculating the relaxation time the author shows that the magnetic field introduces symmetry breaking between the out- and backscattering of the electrons by impurities and phonons.

Doppler-broadening and positron lifetime data have been taken for well annealed ultra-pure Al(110) crystals from 85 mK to 300K. The Doppler-broadening data are independent of specimen temperature from 85 mK to 77K, while they show an increase between 77K and 300K consistent with the linear slope observed previously by Hood and Schultz (1979). Over the whole temperature range studied, the Doppler data scale to the static thermal expansion coefficient. The lifetime data also show no obvious dependence on temperature over the entire range studied, although they were statistically limited by the cryostat geometry. These results, which represent the lowest temperatures studied by any positron annihilation technique, provide no evidence for positron localisation at low temperatures in very well annealed single crystals.

The neutron diffuse elastic cross sections of dilute Fe_1-cSb_c alloys have been deduced from polarised and unpolarised neutron measurements. The magnetic cross sections are proportional to the concentration whereas the nuclear and the nuclear-magnetic cross sections present an abnormal concentration dependence. The data are analysed in terms of magnetic perturbation and static lattice distortions, the short-range order contribution being negligible. The magnetic perturbation corresponds to a model of 'isolated' impurities where the magnetic moment on antimony is small and negative (-0.2+or-0.3 mu _B) and the average magnetic moment on iron increases with Sb concentration with a slope d mu _Fe/dc=1.2 mu _B atom^-1. The results concerning the nuclear terms can be qualitatively explained by the contribution of large static lattice distortions around Sb atoms described in a Kanzaki force model. The unusual concentration variation is mainly attributed to the presence of direct defect-defect interactions.

Neutron diffraction measurements of liquid Na-Sn alloys provide evidence for appreciable ordering in the liquid. At small Sn concentrations (about 20 at.%) the liquid is similar to a simple ionic mixture with maximum distance between the negatively charged Sn atoms. At higher Sn concentrations, the Sn atoms tend to form aggregates, perhaps in the form of tetrahedra of covalently bonded Sn atoms. Elements of the solid structure seem to be preserved in the liquid state.

The authors reconsider the trends among the crystal structures of the elements in terms of the real-space pair interaction. Experience with the on-Fermi-sphere approximation suggests that a simple empty-core pseudopotential is sufficient to explain the trends and indeed they find it so. Calculations of the interatomic potentials have been made for different core radius R_c, atomic radius T_a and valence Z. These are presented in a form to demonstrate the trends within groups I to V and within periods of the periodic table. They show that the structural trends arise from a characteristic variation of the real-space interatomic potential with electron density and pseudopotential. This more critical and detailed discussion now justifies and supplements the qualitative ideas advanced in the 1960s. An important effect not previously discussed stems from the disappearance of the Friedel oscillations as 2k_FR_c approaches ¹/₂ pi .

The properties of self-interstitials and vacancies in the noble metals Cu, Ag and Au were calculated using interatomic potentials derived entirely from first principles. Self-interstitials were found to be stable in the (100) dumbbell (or split) configuration. Their formation and migration energies were 2.61, 2.20, 3.81 and 0.11, 0.07 and 0.11 eV for Cu, Ag and Au, respectively. On the other hand, the respective energies of vacancy formation and migration were calculated to be 1.42, 1.37, 2.75, and 0.82, 0.54 and 0.71 eV. Good agreement between the present theoretical predictions and experimental measurements was obtained for Cu and Ag. In the case of Au, only the calculated vacancy migration energy was in good accord with experiment; the other quantities showed various inconsistencies with regard to experimental observations. For example, the results for self-interstitials could not explain the abnormality of the post-irradiation recovery stage I and interstitial clustering at about 5K observed experimentally in Au. Furthermore, the energy of vacancy formation was three times larger than the measured value. These discrepancies are ascribed mainly to the less reliable nature of the Au potential, relative to those for Cu and Ag, which results from the decrease in the justification for the applicability of perturbation theory. Calculations of the defect properties in Al, another face-centred cubic metal, for which the first-principles interatomic potential has been available, are also included for the purpose of comparison.

The accuracy of the various liquid theory methods for calculating the thermodynamic properties of liquid metals is investigated. The calculations have been made for liquid Na, Rb, Mg and Al over wide ranges of temperature T and density n using (i) the variational method and hard-sphere reference system (the HSV method), (ii) the Barker-Henderson (BH) method, (iii) the Weeks-Chandler-Andersen (WCA) method and (iv) the expressions for energy and pressure in terms of the radial distribution function found in the Percus-Yevick approximation (the PYEP method). The results of the calculations are compared with each other and with experiment. The pseudopotential model is used which describes all the atomic properties of the alkali metals to a high precision, those of Mg rather well, and of Al fairly well. It has been shown that the simple HSV approximation suitable for a semi-quantitative description of the liquid metal properties is insufficient for quantitative calculations of differential characteristics, such as melting temperatures, entropies of fusion, entropies of liquid alloy mixing, etc. The PYEP and BH methods allow the authors to calculate some thermodynamic potentials rather precisely but the computational difficulties prevent them from being used for calculating other thermodynamic properties. The WCA method has been found to combine high accuracy with sufficient simplicity of calculations and seems, therefore, to be the most promising one for quantitative calculations of liquid metal thermodynamics. They also discuss the problems of applicability of the methods considered at various T and n (particularly in the vicinity of the critical point), of the dependence of their accuracy on the metal valency and the sensitivity of the results to the type of pseudopotential model used.

The authors study the electronic structure and the stability of A15 transition metals and alloys at 0K, using a tight-binding approximation and the recursion method. The band contribution to the energy is shown to favour the A15 structure for pure metals when the number of d electrons is about four. This is in agreement with some experimental observations concerning W, Nb, and Cr. Then the energy of formation and the ordering energy of A15 compounds are calculated. They are shown to be quite similar to those of other comparable ordered structures such as the Cu₃Au (L1₂) structure. The relative stability of the A15 and L1₂ structures is studied in some detail. A15 and L1₂ are found to be stable for small and large values of the mean number of d electrons respectively, as is observed experimentally. Finally a general theoretical structural map is presented which accounts for many observed A₃B transition compounds.

Electronic energy bands of the FCC transition metals Ir and Pt have been calculated by the KKR method for a large number of k points inside the Brillouin zone in the energy range 0.-35 eV. The method of Segall and Yang (1980) was used for efficient calculation of the structure constants. Gaspar's (1954) universal potential and l-dependent muffin-tin constants have been used to obtain the correct positioning of the d bands. From the band structure so obtained, the density of states and epsilon ₂ spectra were calculated for both metals. The rich variety of structures found in the spectra was compared with the latest available experimental results. The agreement was found to be much better than other previous theoretical calculations. The structures have been successfully interpreted in terms of various band-to-band transitions inside the Brillouin zone.

Measurements of quenched-in resistivity ( Delta rho _obs)_Q as a function of temperature can yield values of the effective enthalpy for vacancy defect formation H_eff^F. The samples are usually made thin to achieve high quenching rates and pure to avoid vacancy-impurity clustering during the quench. The higher the sample purity and the thinner the sample, the more the sample resistance is affected by scattering of electrons from the sample surface. Inadequate correction for this size effect can result in errors in the values of Delta rho _Q and thus H_eff^F. The magnitude of the errors using Fuchs-Sondheimer (1938, 1952) and Dingle (1950) theories is discussed. A possible method of obtaining reliable size-effect corrections is suggested by an application of Soffer theory (1967).

A quenching apparatus was built in which the gaseous environment could be controlled to an oxygen partial pressure of less than 10^-14 Pa, in order to examine the effect of oxygen on the quenched-in resistivity, Delta rho _Q, of thin, high-purity, copper samples. Oxygen was found to significantly affect measurements of Delta rho _Q through internal oxidation of magnetic impurities, through precipitate formation and through changes of the size-effect parameters (surface specularity and bulk mean free path of the electron). The complex interaction of these effects of oxygen on the quenched-in resistivity of copper samples is discussed.

Between 0.4 and 1K a T² temperature variation of electron-electron scattering is observed. Below 0.4K there are departures from T² in the opposite direction to those observed in K, K-Rb, Ag and Au. The change in the electron-electron scattering term upon straining is in qualitative agreement with the theory of Kaveh and Wiser (1982, 1983) but metallurgical problems prevent a definitive test.

The superconducting properties and normal electrical resistivity were investigated for series of Zr-Si and Zr-Ge alloys in metastable BCC and amorphous phases, produced by the melt-spinning technique. The formation of each metastable phase was limited to the composition ranges: 8-11 at.% Si or 9-12 at.% Ge for the BCC phase and 12-24 at.% Si or 13-21 at.% Ge for the amorphous phase. Two significant features were observed. Firstly, both the superconducting properties, T_c and -(dH_c2/dT)_Tc, and electrical properties, rho _n and (1/ rho _RT)(d rho /dT) varied monotonically with silicon or germanium content without any drastic discontinuity at the BCC and amorphous phase boundaries. Secondly, a strong correlation existed between T_c, -(dH_c2/dT)_Tc, rho _n and (1/ rho _RT)(d rho /dT). The strong correlation between T_c, rho _n and (1/ rho _RT)(d rho /dT) was suggested to arise from the increase in N*(E_F) and D which led to a decrease in rho _n and increases in T_c and (1/ rho _RT)(d rho /dT) with decreasing silicon or germanium content.

The authors present the results of extensive new calculations of the thermodynamics of strong coupling superconductors with structures in the electronic density of states (EDOS) around the Fermi surface. For the first time they treat Lorentzian peaks with Fermi energy off-centre as well as centred and also consider valleys. It is found that in some cases energy dependence in the EDOS can lead to large differences in thermodynamic properties for the same electron-phonon spectral shape and critical temperature.

The spin-glass behaviour exhibited by certain pseudobinary Laves phase intermetallic compounds is discussed qualitatively in terms of the electronic structure. Following this picture, a simple model based on a two-Gaussian distribution for the exchange interaction is developed. Phase diagrams are obtained and compared to experiment.

Bremsstrahlung isochromat spectroscopy (BIS) measurements performed on La_1-xGd_xAl₂ showed a linear change of the density of states at E_F with Gd concentration from 0.5 (eV at spin)^-1 for LaAl₂ to about 1.1 (eV at spin)^-1 for GdAl₂. A theoretical analysis of the resistivity and T_C data was performed using the values the rho (E_F) obtained for various concentrations. A general formula for T_C for binary alloys with U_A not=U_B and rho _A not= rho _B was given. Good agreement was found between the theoretical and experimental results.

Microscopic calculations are given of the wavevector- and temperature-dependent spin susceptibility, chi (q), of itinerant-electron magnets by generalising the approach initiated by Hubbard and Hasegawa (1979-1983). The expressions for chi (q) in the limit of q=0 and q=Q (the antiferromagnetic wavevector) are shown to reduce to the results obtained previously by the present author. Detailed calculations of chi (q) of paramagnetic iron are made using the density of states generated by the recursion method. The amplitude of local moments evaluated by summing chi (q) over the BCC Brillouin zone using the classical fluctuation-dissipation theorem is nearly in agreement with effective moments deduced from the susceptibility chi (0). The equal-time and spatial spin-spin correlation functions are also calculated, and discussed with reference to recent neutron diffraction experiments.

Densities of states of disordered and ordered ferromagnetic Fe₃Pt alloys are calculated in the tight-binding d-band model with intra-atomic Coulomb interaction by making use of the coherent potential approximation for disordered alloys and the recursion method for ordered alloys. Magnetic properties such as the magnetisation, local magnetic moments low-temperature specific heat coefficient and spontaneous volume striction are calculated and compared with experiment. It is found that both the ordered and disordered Fe₃Pt alloys are strong ferromagnets and that the ground-state properties can be explained by the Hartree-Fock approximation.

The magnetic susceptibility was measured in the temperature range 77-1200K for a number of (Ag_0.5Cu_0.5)_100-xGe_x alloys in both amorphous and liquid states. The measured magnetic susceptibility chi _exp is diamagnetic and is of the order of 10^-5 EMU mol^-1, being typical of non-magnetic simple metals. In its composition dependence just above the melting point, a minimum is observed in the range x=20-30. The temperature derivative delta chi _exp/ delta T, in the liquid state also exhibits a sharp maximum. The amorphous phase is easily formed in this particular Ge composition range, where both chi _exp and delta chi _exp/ delta T show anomalies. The magnetic susceptibility of the amorphous state is further diamagnetic and its magnitude is close to the extrapolation of the chi _exp value of the liquid state to low temperatures. Apart from the anomalies, the chi _exp value in the liquid state can be well explained in terms of the free electron model. The anomalous behaviour at x=20-30 is ascribed either to the formation of pseudomolecules or to the coincidence of the Fermi wavenumber 2k_F with the first peak of the structure factor.