Table of contents

EMF measurements have been performed on Na-NH₃ solution above its well established liquid-liquid critical point. The existence of a new phase transition reported earlier by Steinberg et al. (1980) is confirmed and its coordinates are traced throughout the interval of concentration from 4.8 to 8.5 mol% Na and for temperatures from 238 to 285K. The difference in partial molar entropy of the phases has been estimated to be of the order of the gas constant.

The authors show that the reply of Rao et al. (1980) to the letter of Brosens et al. (1980) regarding the existence of singularities in the local field factor G(k, omega ) is in error. The function G indeed has logarithmic singularities at omega _s= mod (k²/2)+or-k mod as pointed out by us earlier.

The electronic structure of interstitial H in Cu is calculated by embedding a self-consistent cluster calculation of H+4 Cu atoms into bulk Cu. There is a peak in the H density of states at the bottom of the Cu s-p band, and the d levels on the neighbouring Cu atoms are pulled down relative to the substrate Cu by the H potential.

Presents a general method for calculating the stacking-fault energy in simple metals, and applies this to the (112) faults in body-centered cubic (BCC) metals. The method contains no approximations for a given wavenumber characteristic (or equivalently the pair potential). The results show that metastable faults do indeed exist in the simple BCC metals (Li, Na, K, Rb, Cs, Ca, Sr, Ba), but the currently available potentials do not yield sufficiently accurate stacking-fault energies because they do not predict BCC as the lowest energy phase.

The effective charge on impurities is calculated considering the bound state, size effect and ionic configuration of the copper host matrix. It is found that both the impurity-interstitial capture radii and the vacancy-impurity binding energy are proportional to the magnitude of the effective charge on the impurities. The occurrence of mixed dumbbells is predicted on the grounds of the attractive interaction of each atom of the pair with the pseudo-vacancy left in between.

Positron lifetime and annihilation lineshape measurements have been performed to study the behaviour of vacancies during GP zone formation and growth in various Al-Zn alloys quenched to 200K. The results suggest that vacancies in these alloys are surrounded by Zn atoms. Vacancy recovery is found only above 220K when GP zones already exist in the lattice. Above room temperature the well-developed zones are free of vacancy-type defects.

The resistivity ( rho ) and thermoelectric power (S) have been measured as functions of concentration and temperature for the liquid alloy system Ag-Pd. Comparison is made between the experimental results and the theoretical predictions of a nearly free electron model adapted for liquid transition metals. It is concluded that such a theory is unable to reproduce the gross features of the concentration dependence of rho and S even when concentration-dependent effective valencies are introduced. The close similarity between the concentration dependence of the resistivity and thermoelectric power of solid Ag-Pd alloys at high temperatures and that of the liquid alloys lends support to the alternative s-d scattering model proposed by Mott (1972).

The low-field magnetoresistance of single crystals and polycrystals was measured for some dilute alloys of K(Rb) at 4.2K. Measurements on the polycrystals showed that Kohler's rule is obeyed; at the lowest fields, the transverse magnetoresistance is given by Delta rho / rho ₀=4.0*10^-21 (B/ rho ₀)² with B in teslas and rho ₀ in Omega m. The order of magnitude of the results can be explained by the variation of the electron-impurity scattering rate over the Fermi surface of potassium as revealed by de Haas-van Alphen measurements.

A calculation is presented of the temperature dependence of the electron-phonon scattering contribution to the electrical resistivity of the alkali metals as a function of the dislocation density present in the sample. Explicit account is taken of the partial quenching of phonon drag by phonon-dislocation scattering. The calculated results are compared with the recent low-temperature resistivity data for a strained sample of potassium. Excellent agreement is found between the calculated values and the measured data.

AuCo solid solutions have been prepared by co-evaporation of both components onto a liquid-helium-cooled substrate. The electrical resistivity of these solid solutions was then measured in situ after different annealing treatments (T_a=80K, 150K, 290K, 360K) as a function of temperature and Co concentration (c=0.25-23 at.%). The residual resistivity increases linearly with Co concentration up to c=8 at.%. The negative slope of the resistivity d rho /dT taken at 4K is proportional to the square of the Co concentration (up to c=1.8 at.%). The c² law, valid for the random solid solution (T_a=80K), means that the temperature-dependent cross section for the electrons is due to interactions between the Co atoms. Annealing the films to T_a=360K allows clustering and deviation from the random distribution of the Co atoms in Au. For this case the negative slope d rho /sup //dT is proportional to c³/² in agreement with the bulk sample results of Loram et al. (1970).

The authors extend a recent theory of superconductivity in pure anisotropic metals, with electron-phonon interaction of the order of or even smaller than the Coulomb repulsion mu *, to include non-magnetic impurities. It is shown that in the regime when superconductivity exists only by virtue of anisotropy, the critical temperature T_c drops much faster than in the case mu *=0, but that T_c remains non-zero. As long as the mean square anisotropy (a²) is non-vanishing, superconductivity exists and, for attainable sample purities, T_c should be measurable in this regime, at least within the present theory in which the anisotropy is incorporated via a separable model.

The kinetic equations for non-equilibrium superconductivity are solved numerically using two techniques. The first, based on a 'histogram' quadrature, is used to find the excess distribution functions at various temperatures for a BCS superconductor perturbed by NIS tunnelling, phonon injection and microwave irradiation and these solutions are then used to predict the change to first order in the applied field of the gap and the nuclear spin relaxation rate. The second method uses a variational approach and compares favourably with the first. The possibility of using non-equilibrium superconductivity as a probe of the deviations from a BCS density of states is briefly discussed.

For pt.I see ibid., vol.11, no.12, p.2595 (1981). Using a tailored quadrature method to solve the kinetic equation for a non-equilibrium superconductor perturbed by SIS tunnelling, the possibility of multiple gap states is explored. It is found that above a threshold bias voltage two locally stable gaps appear, the competition between which can give rise to a first-order phase transition. As the voltage is increased still further the normal state becomes the stable state. The results show a slight disagreement with the earlier work of Schon and Tremblay (1979).

A magnetic environment model of the moment distribution in ferromagnetic alloys is applied to the Ni-Mn alloy system. The variation of spontaneous moment with composition is fitted and, from the parameters obtained, neutron diffuse scattering cross sections calculated with the further assumption of dominant first-neighbour interactions. These compare well with previous neutron measurements and much better than calculations based on any of the simple Mn spin-reversal models. The internal ratios between the three first-neighbour interactions obtained from the model also compare well with two experimental determinations from the spin-wave stiffness constant.

The authors consider whether a state of disordered local moments (DLM) can exist in an itinerant magnetic metal above T_c. They develop a criterion Il(E_F)>1 for the non-magnetic state to be unstable with respect to the formation of DLM. This criterion is more easily satisfied than the Stoner criterion for ferromagnetism. In(E_F)>1, near the centre of a tight-binding band and less easily satisfied near the band edges. They report calculations on tight-binding d bands of 3d transition metals by means of the recursion method. These suggest that BCC Mn and Fe do possess a DLM state, whilst Cr, FCC Mn and Ni do not. They present a method of calculation of the real part of the Green function by the recursion method.

Low-temperature (1-20K) low-field (¹/₂-4 kG) magnetisation and low-temperature (1-20K) zero-field electrical resistivity measurements have been made on a series of (Pd₉₅Rh₅)_1-xNi_x alloys (x<or=0.03). At 4K the magnetisation measurements were extended to 56 kG and in the case of the more dilute alloys the resistivity measurements were extended to 300K. Curie temperatures and spontaneous magnetisation obtained by an Arrott plot analysis of the magnetisation data yield a critical concentration for the onset of ferromagnetism of 1.9+or-0.05 at.% Ni. In the paramagnetic region the magnetisation, susceptibility (dM/dH) and resistivity are well fitted by a model which considers local spin fluctuations to be associated with two types of magnetic centre: a pair of interacting Ni atoms and a single non-interacting Ni atom. The local enhancement associated with a Ni pair is found to be roughly three to four times as large as that of a non-interacting Ni atom.

The crystallisation of the metallic glass Y_0.66Fe_0.34 has been investigated using ⁵⁷Fe Mossbauer effect spectroscopy, differential scanning calorimetry, X-ray diffraction, and high-temperature resistivity measurements. The emphasis of this work is on the microscopic changes which occur in the local environments of the Fe atoms during the crystallisation process. The crystallisation of this glass occurs in two distinct steps associated with the formation of HCP Y and the Laves phase compound YFe₂. The activation energies for these two steps were found to be 2.2 eV and 3.6 eV respectively.

The magnetic properties of amorphous yttrium-iron alloys Y_1-xFe_x have been studied over a wide concentration range 0.32<or=x<or=0.88. Appearance of a moment on the iron in the region x=0.2-0.4 is associated with band metamagnetism, just as for cobalt or nickel alloys. An yttrium-rich alloy (x=0.32) shows exchange-enhanced Pauli paramagnetism in large applied fields. In the other samples, the temperature variation of the susceptibility is not reversible below a temperature T_H, but the susceptibility maximum occurs at a lower temperature, T_M. These properties are characteristic of a spin glass. Competing exchange interactions of both signs, which stabilise helimagnetic structures in some crystalline R₂F₁₇ alloys, lead in the amorphous state to spin-glass behaviour and asperomagnetic order. The dominant positive interactions produce short-range ferromagnetic correlations which persist up to room temperature. However magnetic saturation cannot be achieved for any of the alloys in applied fields of up to 180 kOe, indicating that strong negative interactions are also present.

For pt.I see ibid., vol.11, no.12, p.2727 (1981). Mossbauer data are presented for eight Y_1-xFe_x alloys with 0.32<or=x<or=0.88 in the temperature range 1.5-300K. In some cases external magnetic fields of up to 140 kOe were applied. Iron in the x=0.32 sample has essentially no magnetic moment. Magnetism in zero field appears near x=0.4 and the fraction of non-magnetic iron corresponds to the fraction possessing seven or more yttrium neighbours. The iron moment increases with the number of iron neighbours beyond six. An ordered state with a random non-collinear structure is found for all compositions except x=0.32. Data in an applied field of 50 kOe are interpreted in terms of idealised asperomagnetic structures where the iron moments are oriented at random within a cone of half-angle psi . psi decreases with increasing x, but collinear saturation cannot be achieved in any laboratory field. The average hyperfine field of iron in Y_0.29Fe_0.71 varies as (H_hf(T))=(H_hf(0))-cTⁿ below 40K where n=1.1+or-0.1. An explanation is suggested in terms of single atom excitations where iron atoms in environments barely suitable for the existence of a moment are excited to low-lying non-magnetic states.