Table of contents

Computational techniques predict that several microporous zeolitic structures have unusual negative coefficients of thermal expansion. High-resolution powder diffraction techniques confirm the qualitative aspects of these simulations.

A simple one-dimensional model to simulate the establishment of the Bean critical state is introduced. It is shown that the dynamics of the flux lines as they enter the superconductor are dominated by 'avalanches'. The distribution of distances moved by vortices in the avalanches obeys a power law with an exponent of -1. This suggests that the Bean state is a self-organized critical state. The density of flux lines is parabolic.

An analysis of low-energy electron diffraction (LEED) I-V spectra from the clean Pt₃Ti(111) surface was performed by comparing measured intensities with data calculated using an automated tensor LEED program, which employs a directed search optimization procedure. It was found that the topmost layer is pure Pt and that the other layers have the bulk composition. The first and second interlayer spacings are 2.23+or-0.03 AA and 2.21+or-0.03 AA respectively, corresponding to a contraction of 0.9% and 1.8% of the bulk value. The perpendicular buckling is 0.04 AA +or-0.05 AA in the top layer and 0.15 AA +or-0.04 AA in the second layer. The results are in full accordance with previous investigations of the physical and chemical properties of this surface.

A new method for calculating the stored elastic energy of epitaxial strained layers containing misfit interfacial dislocations is presented. The method differs from previous approaches in that it explicitly takes into account all dislocation-dislocation interactions and interactions between dislocations and the mismatch stress. The method is unique in that it can be used to calculate the energy of finite and/or irregularly spaced systems of dislocations. Examples are given comparing the energy per unit area of finite and non-uniform systems of dislocations calculated using the present approach with that calculated for the infinite uniform case.

The authors have measured the work function changes induced by the coadsorption of oxygen or carbon monoxide with sodium as a function of alkali precoverage for the systems O₂/Na/Cu(111), O₂/Na/Ni(111), O₂/Na/Cu(110), and CO/Na/Ni(111). In agreement with results for many other related systems, they observe anomalous work function decreases for initial doses of both O₂ and CO for high alkali precoverages. Their results are inconsistent with the standard explanations. They propose a new model in which the work function changes can be accounted for by a competition between two elements; a work function increase due to the coadsorbate and a decrease due to coadsorbate-induced undepolarization of the alkali overlayer.

The connection between the microstructure of sputter-deposited Ag/Ni superlattices and their elastic response has been investigated in order to give evidence of the role of interface effects in determining the physical properties of this kind of structure. X-ray and transmission electron microscopy (TEM) experiments have shown that the Ag/Ni interfaces are semicoherent even at very low periods; in addition, there are an appreciable number of structural imperfections, such as stacking faults, twins and atomic disorder at the interfaces. The softening of the shear effective elastic constant c₄₄, observed at low periods by means of Brillouin scattering experiments, can be simulated by assuming the presence of a modified layer at each interface, 2/2 atomic planes thick, characterized by a reduced value of the shear effective elastic constant. A comparison between the set of experimental data relating to Ag/Ni superlattices and existing theoretical models suggests that elastic anomalies can be attributed to the presence of atomic disorder at the interfaces, rather than to coherency strains, surface tension stresses or electronic effects.

The bulk and surface magnetoplasmon modes of a Cantor-type superlattice, in which the cells consist of two alternating materials with their thicknesses following the Cantor sequence, are studied in the static approximation. The dispersion relations of the modes are presented concisely in terms of the transfer matrix elements. The authors find that the modes have fractal structure due to the special geometry of the superlattice. As a result of lowering the symmetry by applying an external magnetic field, the propagations of the surface magnetoplasmon become non-reciprocal. By plotting the profiles of the scalar potential amplitudes, they have also investigated the localization properties of the modes.

The spatial distribution of the electric field E between two current terminals along the quasi-one-dimensional TaS₃ conductor has been studied using a multicontact configuration. In the low-temperature range (30<T<50 K) for a DC voltage V applied between current terminals less than the threshold value V_T, the authors have measured a strong increase of E near the negative terminal with respect to its value in the middle of the sample. In contrast, E is reduced near the positive terminal. The distribution of E revealed by these experiments is supposed to be related to the formation of two Schottky barriers on the metal-Peierls semiconductor contacts which are biased one in the reverse and the other in the forward direction. When V is increased, both the size of this region L and the appropriate band beading in the Peierls semiconductor are enlarged. In the range V>V_T the spatial distribution of E is governed by an additional voltage developed near both current terminals to maintain the conversion of the current of electrons in the sliding charge-density-wave condensate.

Structural, magnetoresistance (MR), and magnetic characteristics were studied on FeCo-Ag grain-type alloy thin films annealed under various conditions. Internal diffusion and interfacial diffusion of the Ag atoms in the ferromagnetic FeCo grains possessed negative and positive contributions to obtaining a large MR ratio, respectively. The magnetic anisotropy field (H_k) associated with the MR characteristics decreased in accordance with the decrease in resistivity as the annealing temperature and/or annealing time increased. This suggests that the Ruderman-Kittel-Kasuya-Yosida interaction between the ferromagnetic grains dominates the H_k in such systems.

The development of thin-film electronic and magnetic properties with film size are addressed in a study of Fe on Au(001), up to 10 Fe layers thick and including the Au overlayer formed during normal growth conditions. Enhanced moments at the Fe/Au interface are observed, decaying to bulk over 3 Fe layers, and the weak Fe-Au interaction is reflected in moments similar to those of the Fe(001) surface and in a very small asymmetry in the magnetic properties of the Fe film. Calculated hyperfine fields are found to exhibit a strong dependence upon film thickness and no simple relationship with the local spin moments. Densities of states and core-level shifts reflect changes in nearest-neighbour species and give a consistent picture of the interface bonding. The development of zone-centre states with film thickness is followed, and bulk-band overlap and interfacial bonding are found to influence the distribution of quantum well states strongly.

A method is described for the computation of the back-scattered intensities of atomic beams, diffracted from the evanescent field generated outside an optical plate by internal counter-propagating laser beams. The method derives from a procedure developed for the similar but importantly differing problem of low-energy electron diffraction (LEED). Modifications to that theory required for the present problem bring the equations closer to the reflection high-energy electron diffraction (RHEED) solution proposed by Ichimiya (1983). Preliminary calculations for the strong-field case indicate that diffracted beams occur with intensities of 7% and 8% of the incident beam which if correct should lead to detectable diffraction effects.

The influence of preadsorbed sodium on the early stages of oxidation of Al(111) has been investigated using synchrotron radiation photoemission measurements of the Al and Ha 2p states. There is a strong promotion (by a factor of up to 100 or more) of the rate of oxygen dissociation and formation of the three-dimensional oxide induced by the presence of the sodium. At the highest Na coverage of 0.33 ML (associated with the Al(111) ( square root 3* square root 3)R30 degrees -Na phase), the Al 2p chemically shifted states characteristic of the chemisorption precursor to oxidation seen on the N-free surface are no longer observed. However, a state of intermediate chemical shift is seen in the Na 2p photoemission peak, at much lower oxygen exposures and appears to be a characteristic of a new oxidation precursor.

The authors have investigated by means of both scanning tunnelling microscopy (STM) and low-energy electron diffraction (LEED) Ge(111) vicinal surfaces misoriented by 1 degrees and 2 degrees toward the (112) and (112) high-symmetry directions. Surfaces of moderately Ga-doped (dopant concentration (DC) approximately=1*10¹⁹ cm^-3) Ge reconstruct c(2*8). No straight step edges but meandering kinks between terraces are observed with the STM. For highly Ga-doped (DC approximately=2*10²⁰ cm^-3) Ge crystals rich in dislocations and Ga inclusions, the vicinal surfaces mentioned exhibit at room temperature the previously reported LEED pattern for a phase of the well oriented clean Ge(111) surface appearing at approximately 300 degrees C. STM images at room temperature clearly display the incommensurate honeycomb-domain reconstruction, I(2*2), proposed by Phaneuf et al. (1985) in order to explain the LEED pattern. High dislocation density and misorientation appear to be essential in order to observe this reconstruction. This is, to the best of their knowledge the first real-space observation of such a structure.

The author presents a detailed investigation of the atomic geometry, electronic states and bonding of an ordered monolayer of Sb on the GaP (110) surface by using a first-principles pseudopotential method. It is found that the epitaxially continued layer structure is decisively more stable than the relaxed p³, epitaxial on top, and epitaxial overlapping chain structures. The equilibrium geometry is characterized by a very small vertical shear between the Sb atoms and agrees well with measurements reported from the X-ray standing wave and surface extended X-ray-absorption fine-structure techniques. The calculated energy location and dispersion of the highest-lying occupied interface state compare very well with reported angle resolved photoemission studies. A thorough analysis of the nature of bonding between the overlayer and the substrate is provided. Finally the location and nature of the interface Fermi level pinning is discussed.

A fully relativistic Green function formalism is applied to calculate the layer-, k/sub ///- and double group symmetry-resolved densities of states (DOS) and the spin-resolved normal photoemission intensities produced by normally incident circularly polarized light for the (001) surface of Cu₃Au in its ordered phase. In the calculated photoemission spectra, a number of features can be ascribed to interband transitions in the authors simultaneously calculated bulk band structure, with a real self-energy correction of -0.45 eV for the lower and of 2.5 eV for the upper states. In addition, the authors identify seven features as due to surface states or surface resonances. A comparison with spin-resolved experimental data shows reasonable agreement with regard to existence, energy and preferential spin polarization (i.e. symmetry type of initial states) for most features.

Model systems evolving in time to attain equilibrium, with spin exchange dynamics order by forming domains after a temperature quench within the ordered region. They are described in terms of time-dependent diffusion coefficients and a distribution of activation energies. This description is analogous to the relaxation in disordered systems such as a:Si-H (amorphous hydrogenated Si).

The lattice parameters and density of the polydimethylsilane crystal, and the stiffness constants of the film prepared by the evaporation of the source polydimethylsilane powder, have been estimated using the X-ray diffraction technique and the surface acoustic wave, respectively. In the former case, the experimental X-ray diffraction patterns are compared with the theoretical patterns. As a result, it is found that the crystallographic structure of the polydimethylsilane crystal is monoclinic, with the lattice parameters a=0.745+or-0.001 nm, b=0.724+or-0.001 nm, c=0.389+or-0.001 nm and gamma =67.1+or-0.1 degrees . In the latter case, the stiffness constants are estimated from the velocity of the surface acoustic wave propagating along the evaporated film/piezoelectric substrate structure. The values obtained on the assumption that the film is isotropic are c₁₁=(5.9+or-0.9)*10¹⁰ N m^-2, c₁₂=(3.5+or-0.9)*10¹⁰ N m^-2 and c₄₄=(1.2+or-0.3)*10¹⁰ N m^-2.

Neutron scattering and specific heat measurements on pressure-quenched AlGe and AlSi alloys are reported. The neutron results corroborate earlier findings of a pronounced softening of the transverse zone boundary modes in these polycrystalline nonequilibrium alloys, reminiscent of a similar softening observed in metallic glasses. At lower frequencies, however, no pronounced glassy anomaly is found. The additional excitations seen around 9 meV in AlGe can be explained in terms of resonant modes of the heavier Ge atoms. At still lower frequencies, neutrons see only sound waves. The more sensitive specific heat technique shows the corresponding lattice T³ term and an electronic contribution which follows the classical BCS behaviour at the transition to superconductivity. Below 0.5 K, there remains a linear term about a factor of ten smaller than in typical glasses.

A random central force network model of the erythrocyte membrane skeleton is constructed by the relaxation of a bond diluted triangular lattice of Hooke's law springs under tension. The fracture of such a network is studied numerically by including a maximum value, L_b, for the extension of any spring before irreversible breakage. The model shows a mechanical instability for values of L_b less than a well defined critical value at any bond fraction above the percolation threshold. For undeformed networks that are mechanically stable, the fracture characteristics under an applied strain can be quantified as a function of L_b, the bond fraction and the type of sample deformation.

The optical activity, birefringence and the indicatrix rotation of a (001) sample of (N(CH₃)₄)₂ZnCl₄ are measured in the high-temperature, incommensurate and partly lock-in phases with a new high-accuracy null polarimeter. The optical properties are shown to be sensitive to the annealing and X-irradiation of a sample via the defect concentration changes. The results can be explained by consideration of a modulated gyration tensor in the incommensurate phase, using the idea of modulation wave distortions imposed by defects.

Nuclear magnetic resonance (NMR) measurements of the proton spin-lattice relaxation times in the C15-type compounds ZrCr₂H_x (x=0.2 and 0.5) have been performed over the temperature range 11-440 K. The experimental results are analysed to obtain the electronic (Korringa) contributions to the relaxation rates and the parameters of hydrogen motion. For both samples hydrogen is found to retain high mobility on the NMR frequency scale down to 78 K. The average activation energies for hydrogen diffusion in ZrCr₂H_x (81 and 84 meV for x=0.2 and 0.5, respectively) are lower than in the other studied hydrides of intermetallic compounds.

Local correlations between kinetic events of cation/vacancy ordering in framework structures lead to 'partial conservation' of the order parameter. Mixing non-conserved and conserved kinetic processes leads to a bifurcation behaviour with uniform states for mainly non-conserved order parameter and periodic pattern formation if more than 1 out of 12 steps is conserved. Possible correlations with experimental observations are discussed.

The electronic structures of metallo-carbohedrenes M_mC_n, a new class of molecular clusters, have been investigated using the universal parameter tight-binding method. The authors have studied the cases of different transition metals M (M=Ti, V, Zr) with single-cage structure and unusual double-cage structure and triple-cage structures.

Be and Ca substitutional impurities in bulk MgO are studied using the periodic ab initio Hartree-Fock method and the considerably simpler ionic model based on two-body forces and the dipole shell model. In both cases a supercell approach was used; in the Hartree-Fock calculations the largest unit cell contained 32 atoms whilst in the case of the ionic model supercells of up to 250 atoms were employed. The Hartree-Fock results for the Be and Ca defect formation energies are -4.10 and 6.25 eV respectively, the corresponding results from the ionic model being -3.88 and 6.37 eV, these data being defined with respect to ionic, rather than atomic, substitutions. The effect of such defects on the Hartree-Fock charge distribution of the host is examined in terms of nuclear displacements (these being compared with the corresponding results of the ionic model) and electronic redistribution, the latter being analysed in terms of charge density maps and induced atomic multipoles. The convergence pattern of the Hartree-Fock defect formation energies is discussed using data from cells containing 8, 16 and 32 atoms. The larger unit cells employed with the ionic model allow a more careful study of the convergence rate of the supercell method.

A systematic study has been carried out on the electronic structure of YH₃. The attention was focused on the conditions for finding an energy gap, in accordance with experimental evidence. The calculations are made with a localized spherical wave (LSW) program, which is based on a first-principles tight-binding version of the augmented spherical wave method. No gap was found for the FCC structure in spite of the use of a great variety of model parameters. However, a non-self-consistent run of the program for this structure reproduced the gap of 2 eV found by Switendick. No gap was found for several versions of the HCP structure either, including the rather complicated HoD₃ structure. The separation of the bands around the Fermi level as found for the LaF₃ structure, could explain the light transmission measurements.

The interaction of quasi-one-dimensional electrons and longitudinal optical (LO) phonons placed in a perpendicular magnetic field is calculated. Results are presented for the polaron correction to the Landau levels (N) and the polaron cyclotron mass. It is shown that under the condition omega _L>N Omega (where omega _L is the LO phonon frequency, and Omega is the frequency of the parabolic confinement potential) level crossing occurs, resulting in a resonant magnetopolaron in the vicinity of the cyclotron frequency omega _C=(( omega _L/N)²- Omega ²)¹2/. The polaron cyclotron mass increases in the vicinity of this omega _C. It is shown that in a cyclotron resonance experiment it is only possible to measure the polaron cyclotron mass if the cyclotron frequency is larger than a critical frequency, determined by the confinement potential.

Extensive measurements of the temperature and of the Al volume fraction dependence of the resistivity of granular Al-Ge have been made near the percolation threshold phi _c. The results at 295 K are analysed using the percolation equations, as modified by Efros and Shklovskii, and by Straley, for systems where the two components have finite conductivity ratios, and by fitting the results to the general effective media (GEM) equation, which also takes into account the finite conductivities of both components. The parameters of these equations are the conductivities (resistivities) of the two components, the critical conductivity exponents s and t, and the critical (percolation) volume fraction phi _c. The experimental value of phi _c, obtained from resistivity and magnetoresistivity measurements at and below the superconducting transition temperature for Al, agrees remarkably well with the values obtained from the percolation and GEM equations. The observed exponents are found to be high, and the width of the critical region surprisingly large. Attempts to extend this type of analysis to lower temperatures proved unsuccessful, and it is concluded that the resistivity of the more insulating component, namely of the amorphous Al-doped Ge, depends on the total Al content of the sample. It is shown that phi _c cannot be identified from the resistivity versus temperature curves between 5 and 295 K, nor from temperature derivatives of these curves. Graphs of the resistivity versus temperature of the amorphous Al-doped Ge for individual samples are extracted using the GEM equation.

Two Luttinger chains coupled by an attractive interaction are investigated. Because of the attractive interaction the charge-1e excitations have a gap, in contrast to a holon with charge 2e which is massless. A superconductor order parameter for the two charges is introduced. The pair-pair correlation is computed. The authors find a power-law behaviour controlled by the 2e Luttinger excitations. They comment on the relevance of this work to the recent suggestions for triplet pairing between planes and s waves in the plane. They mention the possibility of a persistent current with a period of Phi ₀2/=hc/2e in coupled rings.

The authors determine the ground-state phase diagram of the Hubbard model on the square lattice allowing for homogeneous spiral, antiferromagnetic, ferromagnetic and paramagnetic phases in the overall parameter range. This is obtained from a saddle-point approximation of a spin-rotation-invariant form of the slave-boson representation introduced by Kotliar and Ruckenstein (1986). In addition they determine analytically the transition point for infinite repulsive interaction strength between the Nagaoka state and the paramagnetic state on any bipartite lattice.

The first-order Raman spectrum for the A_1g mode of an MgF₂ single crystal was measured at various temperatures from 301 to 973 K. The frequency shift and linewidth change significantly with increasing temperature. In order to analyse the temperature dependence of the frequency shift, first the authors calculated the anharmonic force constants of an MgF₂ crystal. Then applying the lattice-dynamical perturbative treatment to the third- and fourth-order proper self-energy, they evaluated each contribution from the cubic and quartic anharmonic terms to the whole shift and could explain the frequency shift throughout the whole range of temperatures. They found that the cubic as well as quartic anharmonic terms play an important role in the temperature dependence of the Raman frequency shift for the A_1g mode.