Table of contents

Thermal atom diffraction from a densely packed metal surface has been observed for the first time, by using a supersonic beam and a low temperature Ag (111) crystal face. For He-Ag, the intensity of the (1,0) diffraction peak is about 2*10^-3 times that of the specular beam, while for H₂-Ag it is about 3*10^-2 times that of the specular beam.

The atomic volumes, the bulk moduli, the magnetizations, the gain susceptibilities and the derivatives of these quantities with respect to pressure have been obtained from first principles for Fe, Ni, Rh, Pd, Ir and Pt at 0K using canonical band theory and the local spin-density approximation for exchange and correlation. The agreement with experiment is surprisingly good. For Fe the calculated ferromagnetic contribution to the pressure is nearly large enough to account for the relative softness and large volume in the BCC phase and it creates an instability in the FCC phase.

Evidence for the possible existence of a phase change, at approximately 50K, in ZrZn₂ is presented. It is believed to be nonmagnetic in origin.

Using a perturbation expansion, a moment method of calculating the hopping conductivity in the tight-binding representation is described.

The authors report high-resolution photoemission measurements from the (100), (110) and (111) faces of Ni at photon energies of 21.22, 16.85 and 11.83 eV, the photoelectrons being emitted normal to the surface. The results from Ni (100) and Ni (111) can be explained by recent bandstructure calculations. The exchange splitting of two bands of opposite spin is thought to have been observed. The emission from Ni (110) is attributed to surface effects.

Low energy photoemission measurements (21.2 eV) of liquid bismuth are reported. The results show two distinct bands, corresponding to the occupied 6s and 6p states, in contradiction to the weak scattering description of liquid metals.

A nearest-neighbour binding model is taken; the only interaction between defects is a binding energy Delta g for point defects on a nearest-neighbour site to an impurity atom of the host crystal. This model enables almost exact expressions for the equilibrium concentrations of both free and bound point defects to be obtained. It is found that free vacancies obey Boltzmann statistics, while bound vacancies obey Fermi-Dirac statistics. Both free and bound interstitial impurities obey Fermi-Dirac statistics. The distribution of point defects of either type in this model may be far from random. Under certain conditions the enhanced Boltzmann population factor adequately describes the deviation from randomness and the exact solutions are not required.

In irradiated alloys of Al containing approximately 0.1 at.% Ag, Al-Ag (100) dumbbells are created by the trapping of Al interstitial atoms at Ag solute atoms. The results demonstrate that the addition of 0.2% at.% Mg retards not only the formation of Al-Ag dumbbells during annealing from 30 to 100K but also their annihilation during annealing from 180 to 240K. Al interstitials are released from Mg traps at 100-160K, causing further trapping at Ag atoms. Approximately 70% of the Ag atoms return to lattice sites at approximately 200K (stage III) and the remainder at approximately 260K. These results favour migration of Al-Ag dumbbells rather than vacancies during stage III annealing.

The electronic bandstructure of zirconium, which has a HCP lattice structure, has been calculated using crystal potentials which include the effect of correlation as well as exchange. Two methods, one due to Kohn and Sham (1965) and the other to Overhauser (1971) are used for calculating the exchange and correlation contributions to the crystal potential. With the inclusion of correlation the levels around the symmetry point K become well separated, and the entire set of bands is shifted by about 0.25 Ryd. The density of states has been obtained using one of the potentials. The bandstructure results are used to calculate the momentum density distribution of positron annihilation along the c axis.

A simple method which gives the main effect of the ions on the surface potential of metals is developed. This method is applied to studying self consistently the surface potential and the relaxation of ionic planes. The surface potential of the Na (100) face as well as the ionic relaxation of Al (111) and (110) faces have been obtained and compared with experimental data and other theoretical results.

Using a Gaussian free-energy functional with randomized local T_c, the author has calculated the fluctuation specific heat of an inhomogeneous superconducting film. The specific heat is found analytically to diverge as (T-T_c*)^-1/2 where T_c*>T_c. The crossover from the Gaussian to the random Gaussian regime is studied.

The density of bound states in a thick superconductor coated by a thin normal magnetic layer is calculated. A model pair potential: Delta (T,z)= Delta _infinity (T)tanh(z/ square root 2 zeta (T)) is used for the calculation. The thermal conductivity is calculated under the assumption of an effective quasi-particle velocity. Quantitative agreement with experiment is obtained above 0.6T_c. The discrepancy at lower temperatures could be explained by a non-zero pair potential at the interface (z=0) or by boundary scattering.

The usual theory of scattering by non-magnetic impurities in superconductors is extended to cases where the scattering is not isotropic and the mean free path, l, is not very short. The starting point is the equation for the vertex function from which the electromagnetic response of the superconductor is derived. The anomalous limit, ql>>1 where q is a typical wavevector of the field, is studied in some detail and, in particular, the dependence of penetration depth on mean free path is derived. The results show that anisotropic scattering gives rise to measurable deviations from the theory of Mattis and Bardeen (1958).

The proximity effect in the Pb-Sn system in zero magnetic field has been studied by measuring the surface impedance at 3 GHz of a thin film of tin evaporated on to a bulk lead substrate. The results are compared with the predictions of theories of the proximity effect. It is found that good agreement can be obtained by using a theory due to Hook and Waldram (1973) of the spatial variation of the superconducting order parameter Delta inside each metal together with suitable boundary conditions on Delta at the interface between the metals.

An extension of the CPA technique is presented to account for the long-range order of atoms and/or antiferromagnetic solutions for simple cubic or body centred cubic crystallographic structures. Numerical calculations indicate that the influence of the ordering of atoms on the density of states is small for Fe_1-xNi_x alloys. The electronic energy shows that the non-ordered state is favourable. For Cr in antiferromagnetic state is obtained with energy close to the energy of the paramagnetic state. For Cr_0.5Fe_0.5 alloy the minimum of the free energy corresponds to the fully ordered state, which is a non-magnetic one.

A few years ago the Bloch-Hasagawa equations were derived from first principles for the EPR of ions in metals. These results have been extended to include the Kondo effect for high temperatures (T>>T_K). The recent generalization of Sy (1975) and the original work of Langreth and Wilkins (1972) are shown to be incorrect in this aspect. Various possible experimental regimes and the structure of a magnetic impurity in a Kondo system are briefly discussed.

Previous work is briefly summarized to present the theoretical ground for a study of the number of conduction electrons. Important objections are raised on the theoretical handling of such a quantity in the vicinity of the Neel temperature. The number of conduction electrons is written in terms of an appropriate self energy. This improves upon a too crude mean field treatment. The critical exponent of dn/dT is found to be alpha + gamma -1. The implication for the temperature behaviour of the electrical resistivity in the vicinity of the Neel temperature is outlined.

New measurements are reported which show that: (i) the non-hyperfine specific heat of Ho is smaller than previously reported, and the electronic specific heat coefficient gamma =6+or-1 mJ K^-2 mole^-1, (ii) the hyperfine specific heat of HoAg is substantially smaller than that of Ho, and is consistent with a magnetic moment of 8.6 mu _B, independent of specimen and heat treatment, (iii) LaAg is a superconductor with a zero-field transition temperature of 1.08K.