Table of contents

A method is developed to obtain values for the free enthalpy of binding of a vacancy to an impurity in very dilute FCC alloys. On the basis of the five-jump frequency model a computer simulation of the diffusion process is carried out using Monte-Carlo methods. This simulation method is then applied to some copper alloys from which experimental values are well known.

Angle-resolved photoemission spectra of Ag (100) indicate that surface resonances exist near the top of the bulk d bands around the M symmetry point and in the midst of the d bands around the Gamma point. However, they exhibit no separate surface state feature in the sd gap at M in sharp contrast to the observations for Cu (100) and Au (100) and to the prediction of self-consistent surface electronic structure calculations. A LEED type photoemission calculation is carried out for interpreting the observed features in the spectra.

Amorphous Eu₈₀Au₂₀, Gd₈₀Au₂₀ and Eu₆₀Yb₂₀Au₂₀ alloys have been investigated by ¹⁵¹Eu Mossbauer spectroscopy over a range of temperatures (1.5 to 245K) and by ¹⁵⁵Gd and ¹⁹⁷Au Mossbauer spectroscopy at 4.2K (ordered state). In all cases, electric-field gradients (EFG) have a well defined sign; neither isomer shifts nor EFG present significant distributions. This indicates strict compositional and structural short-range order around both rare-earth and Au atoms. Previous magnetisation measurements have revealed dominant ferromagnetic interactions with a small reduction of the spontaneous moment due to asperomagnetism. Consistent with the absence of single-ion anisotropy and with the amorphous arrangement, the principal direction of the EFG tensor is random with respect to the hyperfine field at all atomic components. The distribution of the Eu²⁺ and Gd³⁺ magnetic hyperfine field is interpreted in terms of anisotropic conduction electron polarisation, correlated with spin misalignment at these sites, since the transferred field at Au presents no distribution.

The effect of local phonon modes, due to light impurities in a heavy host, on the electrical resistivity, is calculated. The theory of Kagan and Zhernov (1966) is extended to include inelastic impurity scattering as well as the energy dependence of the electronic distribution function. An analysis of the various effects causing the deviations from Matthiessen's rule (DMR) shows that, although a distinctive signature of the local modes exists, it will usually be overwhelmed by other effects causing DMR.

The EMF of the liquid Na-Hg system has been measured as a function of temperature and from these measurements we have obtained the thermodynamic quantities of mixing: Delta G, Delta S and Delta H. The temperature variations of Delta H were observed. The authors then estimated Delta C_p, the additional specific heat by alloying. In connection with this, the fluctuations in the long-wavelength limit S_cc(0), S_NN(0), S_Nc(0) and the partial structure factors a_ij(0) of Faber-Ziman type were estimated. They also estimated the local ordering parameter from these partial structure factors.

Magnetic after-effect measurements of H, D and T in both dilute Ni alloys and in a face-centred cubic Co alloy at low temperatures have revealed significant isotope effects of the jump frequencies nu , the pre-exponential factors nu ₀ and the activation energies Q. The isotope effects are much more pronounced between H and D than between D and T. Identical jump frequencies and activation parameters ( nu ₀, Q) were determined for D and T in the Co alloy. In Ni and its alloys between 120K and 160K, the order of jump frequencies was found to be nu ^H>> nu ^T> nu ^D, i.e. T was found to be more mobile than D. A comparison with the results of H, D and T diffusion experiments in Ni at higher temperatures shows that for H and D high- and low-temperature ranges must be distinguished. The transition from high- to low-temperature diffusion is a gradual one and begins for H below room temperature and for D below 170K. No transition is observed for T. The diffusion of H isotopes is interpreted as being a competitive process of over-the-barrier jumps and tunnelling transitions. For the diffusion of H, D and T in Ni, from 120K to 160K, the following effective activation parameters were determined: Q^H approximately=0.350 eV, Q^D approximately=0.395 eV, Q^T approximately=0.400 eV, nu ₀^H approximately=8.9*10¹⁰ s^-1, nu ₀^D approximately=7.1*10¹¹ s^-1, nu ₀^T approximately=1.4*10¹² s^-1.

The behaviour of positive muons ( mu ⁺) has been studied in Cu, V, Nb and Ta single crystals, using the muon spin rotation ( mu SR) technique. The measurements were performed in the temperature regime 5-365K, in transverse external magnetic fields of 50-5500 G. In FCC Cu the mu ⁺ diffusional motion can be described in terms of quantum mechanical jump processes of a small polaron among octahedral interstitial sites. An orientation average leads to an activation energy of E=79.2+or-3.1 meV, a pre-exponential factor of D(0)=(1.66+or-0.33)*10^-8 cm² s^-1 and a tunnelling matrix element of J=(18.3+or-1.8)*10^-6 eV. The mu ⁺ total volume expansion was found to be Delta V=3.91 AA³. Furthermore, the electric field gradient (EFG) at the mu ⁺ nearest-neighbour nuclei has been determined as q=0.30 AA^-3. In BCC V, Nb and Ta the mu ⁺ diffusion is trap-limited. As a function of the temperature the mu ⁺ probe different classes of trapping sites associated with host metal impurities. At low temperature, before decaying, the mu ⁺ always end up in a trapped state which shows that the intrinsic mu ⁺ motion between traps is fast and seems to reflect properties of a band diffusion.

Analytic expressions representing the local density of states for a binary chemically disordered lattice are derived and used to investigate electronic and cohesive properties of palladium hydride. This application demonstrates the usefulness of the model to account for the electronic properties of PdH_x (x<or=1) in a way that retains the essential chemical features of the system yet does not require elaborate computation.

The vibrational densities of states of the first-stage alkali-metal-graphite intercalation compounds C₈K, C₈Rb and C₈Cs have been calculated by the recursion method. It is shown that these vibrational densities of states have two peaks in the low-frequency region caused by the alkali metals.

Electronic energy bands of ScH₂ in the fluorite structure have been calculated by the composite-wave variational version of the APW method. The crystal potential was constructed by superposition of the atomic potential of Herman and Skillman (1963). From the band structure data the density of states (DOS), the joint DOS (JDOS) and the Fermi surface have been worked out. The results of the JDOS calculation have been compared with the theoretical and experimental results of other workers. Although the theoretical results agree, the experimental comparison is not very good. A possible explanation for this discrepancy has been attempted.

A fully relativistic first-principles band structure calculation on iridium is presented. The energy range covers both the occupied electron bands which originate in the free atomic 5d and 6s levels, and the excited bands up to 30 eV above the Fermi level E_F. The calculation is based on the Hedin-Lundqvist form of the exchange-correlation potential and was carried to self-consistency including 525 k points in the irreducible wedge of the Brillouin zone. Complete relaxation of all core states was allowed for. The results are compared with a recent angle-resolved photoemission study by van der Veen et al. (1980). The experimental values for critical points in the band structure of both valence and final states agree with the respective special points of the authors' bands within experimental error. In addition, most of the rich structure seen in the photoemission spectra can be quantitatively interpreted in terms of the energy bands.

A theory is presented of the sample dependence of the electron-electron scattering contribution to the electrical resistivity, rho _ee(T)=A_eeT². A central role in the analysis is played by anisotropic electron scattering processes such as those caused by dislocations, stacking faults, grain boundaries, etc. The main idea is that such anisotropic scattering centres lead to anisotropy in the electron scattering relaxation time, which in turn leads to an enhancement of A_ee. As a result, samples which are strained or unannealed, and thus contain a relatively high density of anisotropic scattering centres, will exhibit a larger value for A_ee than samples which are annealed; this implies a sample dependence for A_ee. It is shown for the alkali metals that the enhancement of A_ee can be as large as an order of magnitude, whereas for the noble and polyvalent metals the enhancement is much smaller. These results are in accord with all the available data. A model is introduced for the anisotropy of the electron scattering relaxation time for the alkali metals, and an explicit calculation is carried out for the sample dependence of A_ee. The results of the calculation are applied to potassium and yield excellent agreement with experiment. Finally, the data of A_ee for lithium are also explained.

The authors have measured the electrical resistivity, rho , the thermopower, S, d rho /dT and dS/dT for the liquid substitutional alloy Sr-Ba across the entire phase diagram. d rho /dT is negative for all concentrations and shows a minimum of about -0.17 mu Omega cm degrees C^-1 at about Sr_0.2Ba_0.8. The concentration dependences of rho and S are analysed by means of several theoretical models, and provide a good test of these models as applied to strong-scattering liquid metals.

Starting from the hypothesis that conduction electrons are in a static charge-density-wave (CDW) state, the authors evaluate the relaxation times and the anisotropy in the residual resistivity of potassium by means of an iterative procedure which allows for a convergent and accurate solution of the linearised Boltzmann equation. Numerical results, obtained for values of the parameters within the range characteristic of potassium, show that the CDW hypothesis cannot by itself satisfactorily explain the anomalies in the induced-torque experiments.

New results of resistivities and NMR at ^121-123Sb, ⁵⁵Mn and ⁵¹V nuclei in ternary FeSbM alloys (for M=Co, Mn, V and Ti) are presented and analysed in terms of isolated and paired impurities. For M=Mn, V and Ti the alloys are short-range ordered whereas for M=Co they are random. The dependence of Sb-M interactions through the valence of 3d elements from Ti and Ni is discussed. The Fe-M pair energies appear to play a quite important role in the pair interactions of these ternary alloys.

The authors have performed neutron diffraction experiments on CeMg₃, NdMg₃ and CeInAg₂ intermetallic compounds. The crystallographic structure of CeInAg₂ is shown to be of the Heusler (L2₁) type. CeMg₃ and CeInAg₂ order antiferromagnetically at 4 and 2.5K respectively; the magnetic ordering of the face-centred cubic cerium array is the first antiferromagnetic type. NdMg₃ orders at 6K, the antiferromagnetic structure is the second type like TbMg₃. The ordered moments for CeMg₃ (0.59 mu _B at 2K) and NdMg₃ (1.30 mu _B at 2K) are smaller than those calculated from a simple Hamiltonian including crystal field and isotropic exchange. This reduction may arise from a contribution of the conduction band, which in the cases of CeMg₃ and CeInAg₂ leads to the appearance of a Kondo effect revealed by the behaviour of electrical resistivities.

The magnetoresistance of (Au_1-xCu_x) Mn spin glass alloys has been studied for large variations of copper and manganese atom concentrations at temperatures above the freezing temperature. The magnetoresistance has been observed to depend strongly on the concentration of both the magnetic (Mn) and non-magnetic (Cu) impurities. The variation in the Cu concentration for a constant value of the Mn concentration has enabled the authors to study the dependence of magnetoresistance on the strength of the RKKY interaction.

The effect of the RKKY interaction on the approach to saturation of the magnetisation of a dilute magnetic alloy is discussed within the nearest neighbour pair approximation. For very strong fields the well known asymptotic law in H^-1 is recovered, but for intermediate field strengths another regime of linear behaviour is found around an inflection point in the plot m versus H^-1. The coefficient of this transitory regime quantitatively follows the evolution with temperature observed for many typical RKKY spin glasses. Data on the AuMn and CuMn systems are analysed.

The pseudobinary system YMn_2-2cCo_2c is shown to develop ferromagnetic order over a limited range of composition. This is explained on the basis of ferromagnetic Mn-Mn coupling in the large c limit, using a local model for exchange. Behaviour reminiscent of a spin glass develops at the Mn-rich side of the homogeneous range. A weak intrinsic coercivity exists at 4.2K, but no mictomagnetic asymmetry of the hysteresis is observed.

Hyperfine fields have been measured at various nuclear sites in a number of materials of the REMn₂X₂ type in their ferromagnetic state LaMn₂Si₂ seems to have nearly ideal crystallographic order. In NdMn₂Ge₂, the hyperfine field at the Nd nucleus is significantly less than expected and this is attributed to crystal field effects on the nd³⁺ ion.

Formulae for the anisotropic magnetoelastic coupling constants B₁ and B₂ related to shear strains in itinerant electron ferromagnets with cubic symmetry are derived by the perturbation theory where a tight-binding parametrisation scheme for the electron-ion and spin-orbit couplings is used. Numerical computations of B₁ and B₂ are carried out for iron. The results are compared with the empirical values estimated by the phenomenological theory and the agreement between them is satisfactory for B₁ but not for B₂. The discrepancy is discussed.

Electron spin resonance (ESR) is shown here to constitute a convenient probe for studying metal-oxide-metal junctions. The experiments have been carried out on lithium-alumina-copper layers. From the observed Li ESR linewidths, the authors deduce the transmission coefficient (t) from Li to Cu, for alumina thickness ranging from 10 to 30 AA. A simple model involving direct tunnelling through the oxide barrier supplies theoretical t values in accordance with the experimental results.

The cold-work induced substructure of metal single-crystal specimens has been studied by means of polarised neutron diffraction, optical microscopy and scanning electron microscopy. A cold-working procedure has been applied to the samples with the aim of decreasing extinction effects in neutron diffraction experiments. A practical method of treating 'as-grown' single crystals, applicable to non-brittle materials like metals and alloys, is described. This method enables one to approach the zero-extinction limit for relatively thick specimens in neutron diffraction. A relation between the actual substructure and the primary extinction effects is given.