Table of contents

Small silver clusters have been investigated by electron microscopy. A bimodal size distribution is formed containing particles of diameter 1-2 nm and 4-8 nm. Among the largest particles icosahedra and cuboctahedra appear with equal frequency in HREM. 1 nm size icosahedra, however, could not be found, while cuboctahedra of this size could be resolved. Computer simulations of high resolution images for both structures reveal that icosahedra are more difficult to resolve under the experimental conditions. A dilation of approximately 5% of the lattice parameter could be measured for 1 nm particles. For 2 nm particles, this dilation drops to 2.5% compared with the bulk value of the particles larger than 3 nm.

The second moment approximation of tight-binding theory gives an electronic band energy per atom which has the form of the square root of a pairwise sum of squared hopping integrals. This functional form for the dependence of the energy on local coordination has been the basis of empirical non-pairwise force models for transition and noble metals. Hitherto, it appeared to be valid only for a half-filled band. By introducing the assumption that the atoms keep the same charge in all environments, the authors show that the square root function is appropriate for all band fillings.

The characteristics of the aging internal friction and modulus associated with the non-equilibrium phase transformation were investigated in an as-quenched eutectoid Al-Zn alloy. It was observed that the loss of stability in the lattice during spinodal pretransition did not induce any localised soft modulus effect and there were two aging internal friction peaks during the phase transition. These results are discussed.

The effect of the partial replacement of Bi by Eu, Gd, K and Pb, and Cu by Ni, on the superconducting transition temperature (T_c) of Bi₄Ca₃Sr₃Cu₄O_y (y=undetermined) is investigated by DC resistivity (R) and AC susceptibility ( chi ) measurements. T_c is not significantly altered by Eu and Gd substituents (T_c close to 65 K); however, K and Pb substituents tend to enhance the fraction of a phase superconducting around 110 K. Interestingly, for nickel specimens, T_c of a large fraction, as suggested by AC chi data, initially decreases to 40 K with increasing concentration of nickel, followed by an increase, whereas R data shows no appreciable variation of T_c.

The cluster field method (CFM) described earlier is applied to the calculation of the correlation function K_ij=K(r_i-r_j) and short-range order parameters a_ij in alloys. The method generalises the previous approaches of Krivoglaz (1957) and Clapp and Moss (1966, 1968) to the case of realistic values of the interaction constants V_ij, which are not small, as compared with the temperature T. The cluster expansions for the matrix elements S_ij=(K^-1)_ij are presented, which converge rapidly at normal values of the 'virial' parameter c(1-c)f_ij where c is the mean site occupation number and f_ij=exp(-V_ij/T)-1 is the Mayer function. The sum rule a_ij=1 is usually fulfilled in the CFM with good accuracy, unlike in the previous approximate methods. The method enables the authors to estimate quantitatively the interaction constants V_ij from the diffuse scattering data. This is illustrated by using the data of Lefebvre et al. (1981) for Ni_0.765Fe_0.235. The results show that the conventional Clapp-Moss approximation underestimates V_ij by 30-40%. Applications to interstitial alloys are illustrated by calculating K_ij for the NbH_x-type alloys.

Starting with the real metallic potentials constructed from a full non-local model pseudopotential theory, the author carries out two computer simulation studies for yttrium and zirconium monatomic glasses. By comparing the Wendt-Abraham parameter, R_LG, at the liquid-glass transition point obtained with previous results for chromium and various other, seemingly different monatomic glasses such as the Lennard-Jones, the soft core, etc., the author finds that the Wendt-Abraham criterion satisfies R_LG=0.1+or-0.02 and may well be a universal constant. The physical indication of this R_LG value is interpreted and various ubiquitous structural features related to liquid-glass transition are also presented.

The temperature dependences of the liquid structure factors of caesium metal have been determined using a Monte Carlo simulation technique. The calculation adopts the generalised non-local model pseudopotential theory developed recently by Li et al. (1986) to construct the interionic pair potential. The author finds that the latter pair potentials are rather accurate and, with proper choice of the asymptotic pair correlation function, are capable of reproducing highly reliable liquid structure factors at various temperatures.

The metallic glass Cu₆₄Ti₃₆ was irradiated at 4.6 K with 3 MeV electrons up to fluences of 2.8*10¹⁹ cm^-2. In the whole fluence region the electrical resistivity decreases almost linearly without a tendency to saturate. The subsequent isochronal annealing treatment up to 723 K reveals the existence of two contributions of opposite sign to the electrical resistivity. Topological defects are assigned to an increase and compositional defects to a decrease of the electrical resistivity. Both types of defects are totally annealed far below the crystallisation temperature.

The migration energies of tri- and tetravacancies in copper are calculated to be 0.56 eV and 0.38 eV, respectively, using a new energy minimisation scheme and a first-principles interatomic potential. Comparison with the previously calculated single and divacancy migration energies of 0.82 and 0.55 eV, respectively, indicates that the tri- and tetravacancies are very mobile in copper. The calculated binding and migration energies were used in a rate equation model which predicts the concentrations of vacancy defects during quenching and electron-irradiation experiments. The predicted activation energy of the model follows the same general behaviour as that measured in annealing experiments. The maximum concentrations of tri- and tetravacancies predicted by the model are very low (less than 0.001 p.p.m.), which suggests that these defects many not be observable using experimental techniques.

Bulk copper and nickel samples containing 3 and 5 at. % krypton respectively have been studied by conventional positron annihilation techniques. Most of the emphasis has been placed on the changes in positron lifetime and angular correlation parameters during isochronal annealing from ambient up to near the metal melting points. The study was complemented by transmission and scanning electron microscopy, together with macroscopic measurements of weight and dimensions. These techniques, combined with previous studies, provide a fairly detailed picture of both the as-prepared materials, where the krypton is present as a high density of solid phase precipitates (solid bubbles), and the subsequent marked response of the substructure to annealing. A number of features are of particular interest. The onset temperature for bubble coalescence events is correlated with the melting temperature of the solid Kr inside the bubbles. At higher temperatures extensive swelling is observed prior to and simultaneously with the release of the majority of the krypton. The structure after the Kr release contains micrometre-sized pores and approximately 10 nm bubbles. This structure partly recovers at higher temperatures, but sufficient krypton is retained in the pores to maintain a large swelling up to close to the metal melting points. Positrons are found to become trapped at the Kr-metal interface in Kr bubbles, and there is clear evidence for a quantitative relation between lifetime and Kr density. Both these features are in agreement with the recent theory of Jensen and Nieminen (1987). Finally, positron trapping into dislocations is observed to occur at a rate much lower than that predicted from published specific trapping rates.

The authors have analysed the acoustic emission signals generated during the thermoelastic martensitic transformation of a Cu-Zn-Al alloy using a broad-band detection system. The analysis has been restricted to rise times which contain information related with the dynamics of the transformation mechanism. The results show the existence of two well separated rise-time signal populations. The average values of each class are 0.2+or-0.1 mu m and 2.2+or-0.5 mu m. The presence of signals with a characteristic 'fine structure' superimposed on the mean long rise-time wave has been attributed to a jerky character of the martensitic growth process.

The anharmonic shape of the interstitial Pd potential well is analysed within first-order perturbation theory. The anharmonic part of the octahedral potential raises the ground-state energy of H in Pd by about 18%, which means that anharmonicity cannot be neglected when studying hydrogenous vibrations. The influence of anharmonicity on the Pd force constants is evaluated in a semi-empirical way. The transversal force constants are found to be negative and of similar importance to the longitudinal force constants. The negative transversal force constant weakens the total force constant considerably. It is argued that, due to a weak effective force constant, resonant vibrations are possible for hydrogen isotopes in Pd.

The contribution of resonant modes (band modes) to the electron scattering in dilute interstitial Pd-hydrogen alloys is investigated computationally. The interaction of Bloch electrons with the vibrating defect is treated in the adiabatic approximation. Electron scattering from a frozen defect is described by a KKR-Green function method which accounts for all defect-induced multiple scattering effects. It appears that the combined oscillations of a hydrogen atom together with surrounding Pd atoms give rise to a considerable enhancement of the scattering quantities. It is conjectured that the measured isotope effects in the de Haas-van Alphen quantities of Pd(H), Pd(D) and Pd(T) can be attributed to differences in the resonant modes of the isotopes. The occurrence of a resonant mode depends critically on the Pd-hydrogen force constants, which are, due to anharmonicity, different for the three isotopes.

Attempts to explain the stoichiometrics of intermetallic compounds in binary phase diagrams are described. The model presented yields compound compositions which are simple rational numbers and is expected to be useful for work on intermetallic phase diagrams. The occurrence of compounds in 10 binary systems (Si-Hf, Co-Y, Hg-Sr, Rh-Gd, Au-Sm, Hg-Ca, Pt-In, Au-In, Fe-Al, and Pd-Ti) is compared with the results of the calculations. It is concluded that a considerable number of compounds occurring in these systems is justified by the proposed hard-spheres packing condition. The dependence of the efficiency of the method on the radius ratio and on the compound components is indicated.

The solute partitioning and the unmixing kinetics of a Cu_0.43Ni_0.42Fe_0.15 alloy have been studied by anomalous small-angle X-ray scattering (ASAXS). From ASAXS data, partial structure factors (PSF) have been extracted: the homoatomic PSF curves cannot account for a pseudobinary system and one has to consider a Ni (or Fe) segregation at the interfaces between the Cu-rich phase and the Ni-Fe phase. The main contribution to the PSF comes nevertheless from the two-phase term which behaves as in the late stage of deep quenches of concentrated binary systems: PSF present a hump which increases and shifts toward low q values as power laws of the aging time: q_m^{varies as} t^-0.25, S_m^{varies as} t^0.7.

The behaviours of transition adatoms on two stepped surfaces (m(110)*(110)) and (m(110)*(001)) of body-centred cubic transition metals are studied and compared using the tight-binding approximation. One finds that the most stable adsorption sites are always close to the ledge as expected. The presence of the step very slightly modifies the diffusion rate in a direction parallel to the ledge except for the path close to it in the lower terrace when the protruding edge is not close packed. On the other hand an adatom diffusing towards the edge of a terrace is reflected by a potential barrier lying near the outer edge. This study confirms the main trends found in a previous work devoted to the (m(110)*(011)) step; it also gives evidence of the role played by the specific geometry of each step edge, for instance the adatoms at the most stable adsorption sites are less strongly bound for the (m(110)*(011)) and (m(110)*(110)) steps than in the case of the more open (m(110)*(001)) ones.

LCAO-CPA theory is applied to substitutionally disordered ferromagnetic BCC FeCo alloys. Calculations are based on charge and spin self-consistent LCAO parameters determined for the pure component metals within the local spin density functional theory. Effects of charge and spin self-consistency on the alloy magnetisation are treated in an ad hoc manner. Mechanisms determining the concentration dependence of the local magnetic moments are explained. The correlation of magnetic moment formation and maintenance of local charge neutrality is discussed. At Fe sites the exchange splitting is enhanced by a hybridisation-driven local reoccupation within the spin sub-bands. This causes the observed increase of the Fe moment by about 30%. Numerical results are presented for the saturation magnetisation, local magnetic moments, d-band population and specific heat coefficient as functions of concentration. Qualitative agreement was obtained with available experimental data. Numerical results for the concentration dependence of the total and component density of states are shown. Quite different degrees of disorder are found for both spin states. The majority spin states are typical rigid-band-like, whereas a strong influence of alloy disorder on the minority spin states is observed. This causes a strong broadening of the minority spin Bloch spectral functions and Fermi surface.

The authors note that the current technique for studying the electronic structure of disordered alloys with the relativistic Korringa-Kohn-Rostoker coherent potential approximation (RKKR-CPA) method is difficult to use and inappropriate for future developments including charge self-consistency, total energy evaluations and studies of order-disorder transitions. They detail here an improved method which is suitable for such applications and easy to use. As an illustration a study of a disordered Cu_0.75Au_0.25 alloy is presented.

The electronic structure of Mg-reinforced nickel-based superalloys has been studied. By setting up models for the mechanism of Mg in nickel-based superalloys, the local density of states, charge transfer, structural energy, and Fermi level of the systems have been calculated with Haydock's recursion method (1972, 1975). The difference in structural energy of clusters including Mg and its neighbours in between systems in model I and model II shows that Mg prefers and stabilises gamma ' phases. The amount of charge transfer of Mg in the gamma ' phase of a nickel-based superalloy indicates that Mg tends to enter a gamma ' phase.

Changes in the electronic density of states (DOS) caused by plastic deformation and hydrostatic pressure in dilute Al-Mg-Si alloys are examined. Significant changes in the DOS were observed which depended on the mechanical treatment of the alloy. The observations are compared with the DOS of the same alloy in metastable states.

The electron-electron scattering contribution to the thermal resistivity has been calculated for the alkali metals Li, Na and K from 0 kbar to 100, 300 and 120 kbar, respectively. The electrons are described using a Fermi liquid model, which has been modified in order to account for the crystal. It turned out that the main part of the implicit pressure dependence of the thermal resistivity is due to the band mass, which has been calculated using the LMTO method.

Electrical resistivity and magnetoresistivity measurements were performed on MnRhAs between 4.2 K and 300 K. The reveal an anisotropic behaviour and strong anomalies giving evidence for changes in the magnetic ordering. The anomalies are most pronounced in measurements along the hexagonal easy axis. The resistivity curve recorded in the ferromagnetic state (160<T<200 K) coincides well with that measured at lower temperatures, when the applied magnetic field is large enough to flip the initial AF structure. It can be concluded that the resistivity of the two phases (F and AF) would be equal near 78 K. When the temperature is increased above the magnetic ordering temperature (T_N=238 K), rho (T) reaches an almost constant value. This could result from the superposition of a weak electron-phonon interaction term and the contribution of the paramagnetic scattering of the conduction electrons.

The authors present a study of the electron momentum density of BCC iron in both its ferro- and paramagnetic phases and interpret the results in terms of the appropriate electronic structure. The paramagnetic state is modelled by the disordered local moment (DLM) theory of magnetism and a 'first principles' multiple-scattering formalism is used. They make a suggestion for a novel, single-crystal Compton scattering experiment on iron at varying temperatures which they predict will measure directionally dependent changes of the Compton profile between the two phases. It may thus be used to test the basic picture of the thermally induced magnetic fluctuations considered in the DLM theory.

The EPR spectrum of Gd³⁺ ions diluted (0.5% and 0.1%) in powdered samples of the high-T_c(T_c-92 K) superconductor Y₁Ba₂Cu₃O_x has been measured at X-band in the temperature range 4.2-300 K. It was shown that the EPR spectrum exhibits significant changes in g-shift, thermal broadening and intensity well below T_c. The results are analysed in terms of existing theoretical models.