Table of contents

The metal-insulator transition in the four-dimensional Anderson model is studied numerically using the finite-size-scaling method. Critical disorder is found and the spectrum of Lyapunov exponents of the transfer matrix is shown to change from linear to the third-root shape when the system undergoes the metal-insulator transition. On the basis of our results, we conjecture the dimension dependence of the critical disorder and of the spectrum of Lyapunov exponents for dimension d>2.

The compound Y₃(Fe,Ti)₂₉ has been successfully prepared with a starting composition of Y₉Fe₈₆Ti₅ via annealing at 960 degrees C for three days. Thermo-gravimetric analysis (TGA) gives a magnetic ordering temperature of 386(3) K. Refinement of powder X-ray diffraction data with the newly discovered monoclinic Nd₃(Fe,Ti)₂₉ structure model gives the lattice parameters a=10.6068(9) AA, b=8.5224(6) AA, c=9.6926(8) AA and beta =97.153(7) degrees . ⁵⁷Fe Mossbauer measurements yield an average hyperfine field of 20.2(2) T at 293 K for the new Y₃(Fe,Ti)₂₉ compound, which corresponds to an average Fe magnetic moment of about 1.31 mu _B.

Raman spectra of CuInS_2xSe_(1-x) mixed crystals are reported. The dependence of the mode frequencies on the composition is studied. Two-mode behaviour is observed for the higher-frequency phonons. One-mode behaviour is found for the low-frequency modes corresponding to the zone-edge acoustic phonons of zincblende.

Data are presented for the effect of U substitutions for Ce on the thermopower of the so-called 'Kondo insulator' CeNiSn. These suggest that the gap that begins to open below 5 K is destroyed by 10% U, with the appearance of metallic Kondo character. This is in turn attenuated by further substitutions of U.

Fe/Pd multilayers with different thicknesses of Pd layers prepared by e-beam evaporation have been studied by ⁵⁷Fe conversion electron Mossbauer spectroscopy (CEMS) and X-ray diffraction (XRD). A structural transition of Fe layers from BCC to FCC in Fe/Pd multilayers was found with the thickness of Pd layers d_Pd>or=36 AA. The CEMS measurements indicated that the hyperfine field H_hf of the inner parts of the Fe layers was almost the same as that of a single-layer Fe film and H_hf of the interface Fe was enhanced slightly from that of a single-layer Fe film. The moments of Fe layers were independent of the thickness of Pd layers and the structure of Fe layers, which suggested that the FCC Fe had the same magnetic moment as BCC Fe in Fe/Pd multilayers. The isomer shifts and quadrupole splitting increased abruptly for d_Pd>or=36 AA. This would correlate with the structural transition of the Fe layers in Fe/Pd multilayers.

Theoretical description of Brillouin scattering from shear horizontal (SH) acoustic surface phonons in a generic multilayer structure with arbitrarily smooth interfaces is given. To evaluate the p-s Brillouin cross-section, the SH phonon spectrum, the transmitted zeroth-order field and the fluctuating polarization vector radiating the Brillouin light are computed numerically by means of a method that can take into account any depth profiles of elastic, elasto-optic and dielectric properties of the medium including electromagnetic absorption. As an application the case of a silicon on insulator structure with both sharp and diffuse interfaces is illustrated.

In the application of X-ray standing wave (XSW) methods to the determination of surface structures, the experiment provides two structural parameters: the coherent position and the coherent fraction. For simple, single-high-symmetry-site adsorption systems, the interpretation of these parameters in terms of a structural model is trivial, but in the case of lower-symmetry adsorption sites, or multiple adsorption sites (including those associated with coincidence lattice structures), these parameters are related to spatial distribution functions through a Fourier integral. A particularly simple way of viewing this result, in terms of vector (Argand) diagrams, allows many simple cases and general theorems concerning the interconnection of the structure and the XSW fitting parameters to be visualized. The application of this approach is illustrated with particular reference to recent studies of adsorption on FCC (111) metal surfaces, but some generalization to other surfaces is included.

A regular method for finding solutions of the equation for the depth distribution of momentum deposited by atomic particles in elastic collisions is developed, using the model of an infinite medium and power cross section. It is shown that in neglecting threshold energy the distribution has a specific singularity at the target surface; its physical origin and the respective asymptotes are investigated. The calculated distributions and other functions connected with the problem are given.

The inner valence region of CO/Pd(100) p(2 square root 2* square root 2)R45 degrees has been studied by angular resolved photoemission at the Pd 4d Cooper minimum, and with resonant Auger spectroscopy at photon energies corresponding to the C 1s and O 1s X-ray absorption (XA) maxima of the unoccupied parts of the 2 pi *-Pd 4d hybrid (2 pi _un). Previously unobserved inner valence states are revealed in the direct photoemission and are compared with resonant Auger results. The interpretation and assignment of the different spectral features to different main final state configurations are based on energy, symmetry and intensity arguments, as well as comparisons with previous results.

An exact generalization of the classical Bloch-Jensen hydrodynamical model for the homogeneous free-electron gas is obtained for a two-layer spherical metal system. The dispersion relation of surface plasmons together with the analytical expression of the self-consistent electric field are calculated exactly. It is found that small and large systems exhibit notably different behaviours. In particular, the surface-plasmon frequencies strongly depend on the layer thickness. As is the case for the metal-vacuum interface, the electric field is shown to present a sharp maximum at the interface between the two metals. A simple physical interpretation of this effect is provided. The present two-layer system also gives interesting indications about the part played in collective surface modes by the ion-density profile near the surface of small metal clusters, which may be helpful in interpreting recent puzzling experimental data.

In recent years advances in materials science in the area of molecular beam epitaxy have made possible the synthesis of novel materials and novel structures such as superlattices and multilayers. In the face of the expanding experimental interest in superlattices there is a need for theoretical work to be done. In addition there has been a lot of attention focused on magnetic semiconductors (MS), materials whose magnetic and semiconducting properties interact with each other. In this paper a MS superlattice is investigated using a Green function method applied to the s-d interaction model. The spectrum of spin-wave excitations is calculated and the spectral intensity of certain spin-wave modes is derived from a consideration of the imaginary part of the Green function.

A calculation is presented via the standard mean-field theory for the magnetic properties (transition temperature, compensation temperature, and magnetization curve) of a ferrimagnetic multilayer system, consisting periodically of two layers of spin- 1/2 A atoms, two layers of spin-3/2 B atoms, and a disordered interface in between that is characterized by a random arrangement of A and B atoms of A_pB_1-p type and a negative A-B coupling. The consequences of the disordered interfaces and different anisotropies between the bulk and interface for the magnetic properties are examined. A number of characteristic behaviours, such as the possibility of two compensation points, are found for the magnetic properties.

This paper describes highly oriented films prepared by the evaporation of poly(dimethyl silane). The specimens are prepared at various evaporation speeds and substrate temperatures. The polymer is divided into species of about 2 nm length on the evaporation heater, and the species recrystallize on the substrate. The structure and orientation of the films are investigated using wide-angle X-ray diffraction and ultraviolet absorption methods. The substrates used are single-crystal silicon (100) for the X-ray diffraction measurement and fused silica for the ultraviolet absorption measurement. The polymer exhibits the same orientation on both substrates. It is found that most of the silicon chains are perpendicular to the substrate surface in the film prepared under the condition whose growth rate is lower than 20 nm min^-1. On increasing the substrate temperature, the number of silicon chains whose orientation is normal to the substrate surface increases.

Using Brillouin spectroscopy we have studied the dynamical properties of epitaxial CaF₂ films, with thicknesses in the range h=61-300 nm, grown on Si(001) substrates. Theoretical calculations for the velocity values, based on a simple elastic continuum model, give results that agree quite well with experiment. A correlation between the dynamical properties of CaF₂/Si(001) systems and their structure is found and discussed.

The atomic structure of the perfect AlPdMn icosahedral phase has been studied on a single grain by anomalous X-ray diffraction close to the Pd edge. A method is proposed for relative normalization of the data, in order to extract the partial structure factor F_Pd of the Pd atoms. The Pd is found to be very ordered in the six-dimensional description of the quasicrystal, as was previously inferred from X-ray and neutron single-crystal diffraction data. However, a phason Debye-Waller-like term has to be introduced to account for the data. Comparison of the present data set with low-resolution data indicates the existence of diffuse scattering located close to the Bragg reflections whose intensity scales as (1-exp(-B_perpQ²_perp)) as expected from random fluctuations of the atomic surfaces. The average fluctuation is found to be of the order of 1 AA. Moreover the comparison of the data with a spherical model shows that the atomic surface describing Pd atoms is likely to be made of several disjointed pieces.

Substitution of Fe for Mn in i-Al₇₀Pd₂₀Fe_10-xMn_x was found to decrease the thermal stability of the alloys. All Mn-containing alloys exhibit paramagnetic behaviour and have an effective magnetic moment per Mn atom in the range (1.39-1.86) mu _B. Measurements of the temperature dependence of the resistivity of this series of alloys reveal that the Al₇₀Pd₂₀Fe₁₀ alloy has the highest resistivities, 5600 mu Omega cm and 5800 mu Omega cm, at room temperature and at 5 K, respectively, and the Kondo effect is observed in Mn-containing alloys. It was found that the resistivity behaviour can be varied quite significantly by varying the quenching rate.

The synthesis and physical characterization of Nd_1-xLa_xNiO_{3- delta} (from x=0 to 1) samples are reported. Nd_1-xLa_xNiO_{3- delta} samples are orthorhombic for x<or=0.4 and rhombohedral for x>0.4. Resistivity and magnetic susceptibility measurements from 4.2 K to 300 K are also reported. The increase of La content up to x=0.4 induces the decrease of metal-insulator (MI) phase transition temperature and compounds with x>or=0.5 are metallic. The distortion degree of the ideal perovskite structure, directly related to the Ni-O-Ni angle, is correlated with the MI phase transition temperature. The metallic phase is characterized by a linear dependence of rho (T) with a residual resistivity that decreases with x, being near zero for LaNiO₃. The conductivity in the semiconducting phases at low temperatures follows a variable range hopping mechanism. The magnetic susceptibility shows a Curie-Weiss contribution from the Nd sublattice and a Pauli susceptibility of similar magnitude for all samples. The analysis of the experimental data shows large electronic correlations in these compounds.

The use of energy functionals based on charge density as the basic variable is advocated for ab initio molecular dynamics. It is demonstrated that the constraint of positivity of density can be incorporated easily by using the square root of the density for minimization of the energy functional. An ad hoc prescription for including non-local pseudopotentials for plane wave based calculations is proposed and is shown to yield improved results. Applications are reported for equilibrium geometries of small finite systems, viz. dimers and trimers of simple metal atoms like Na and Mg, which represent a rather stringent test for approximate kinetic energy functionals involved in such calculations.

Using an early theory by Houghton and Maki based on the BCS model, which assumes a temperature-independent relaxation time tau , we compute the flux-flow magnetoresistivity rho _S(T,H) of layered superconductors near H(c₂) in the clean limit. Then we use our results to analyse the magnetoresistivity curves of single-crystal YBaCuO. Extending the validity of the calculation to a temperature-dependent tau , we find that for constant tau , or for relaxation times obtained from experiments, the theory reproduces qualitatively the saturation-like behaviour of the observed magnetoresistance, and also the broadening of the resistive transition, as well as the changes in the slope d rho _S(T)/dT with field. However, quantitative agreement with experiments is not achieved. This allows an estimate to be made of the contribution of the thermal vortex line fluctuations to the total dissipation; we find the former to be about 10% of the latter.

The influence of plastic deformation on magnetic properties has been studied in Ni_75+xAl_25-x (0<or=x<or=2) and Ni_77-xAl₂₃Pd_x (0<or=x<or=3) intermetallic compounds. The temperature dependence of the spontaneous magnetization, M_S(T), has been analysed according to the self-consistent renormalization (SCR) theory of spin fluctuation. The qualitative agreement between the experimental result and the theoretical one is satisfactory in the plastically deformed samples. The value of M_S(0) decreases slightly on compression; the reduction of M_S(0) attains to 10% in the samples of 20% strain. However, the Curie temperature is constant within experimental error. The SCR theory indicates that plastic deformation gives an influence especially to the derivative of the density of states at the Fermi level. The dislocation distribution is observed under the electron microscope and its density, rho , is obtained for several samples. The relationship between M_S(0) and rho suggests that the non-magnetic region spreads around the APB and CSF ribbons at least as far as 100 atomic distances, which are introduced between superpartial dislocations by plastic deformation.

By making use of the Landau-Ginzburg energy expanded up to the sixth power of magnetization density, the spontaneous volume magnetostriction and dependence on pressure P of the Curie temperature T_c of an itinerant-electron ferromagnet with negative mode-mode coupling among spin fluctuations are discussed. The spontaneous moment is found to show first-order and second-order transitions at T_c, according to the values of the Landau coefficients. It is shown that the difference between the spontaneous volume magnetostrictions at T=0 and T_c is large and the P-dependence of T_c becomes anomalously large when a certain condition is satisfied by the Landau coefficients. The results obtained in the present theory could explain the anomalous T-dependence of the thermal expansion and the strong P-dependence of T_c observed in Invar alloys, some Laves phase Co compounds and others.

The influence of local order around the titanium site on Ti 2p X-ray absorption spectra is studied for rutile, anatase and brookite. This is done by a configuration interaction based calculation of Ti L_2,3 X-ray absorption edge shapes, explicitly taking into account the exact first-neighbour surroundings of the cation acting through crystal field splitting and hopping terms. The evolution of the peak splittings and intensities from rutile to anatase is well reproduced for the L₂ edge and can be attributed to the small changes in the cation site symmetry. In contrast the splitting of the second peak of the L₃ edge cannot be attributed to differences in the first coordination shell of the cation between the titanium oxides.

The lattice dynamics of ice Ih (H₂O, D₂O and HDO) have been investigated using inelastic incoherent neutron scattering in the energy transfer range 2-500 meV (i.e. 16-4025 cm^-1) using TFXA and HET on ISIS. The spectra of D₂O containing different concentrations of H₂O from 0.5% to 30% by weight were measured. Three main peaks at approximately 105, approximately 185 and approximately 408 meV associated with three localized modes, mu ₁, mu ₂ and mu ₃, of the H defect centre have been observed in the neutron spectra. Two Gaussian components can be resolved in the mu ₁ mode, at 103+or-0.5 meV and 107+or-0.5 meV, indicating the existence of two different force constants for the H bonds. A lattice dynamical calculation for the H defect in a D₂O ice Ih host lattice, based on our two H bond force constants model, shows the three modes in very similar positions to those observed in the measured data and, in particular, predicts the splitting of the mu ₁ peak into modes at 103 and 107 meV.

The structural phase sequence of (C₃H₇NH₃)₂CuCl₄ (abbreviated to C₃CuCl) has been determined between 20 and 300 K by optical and x-ray methods. The Jahn-Teller compound undergoes two types of reconstructive phase transition between antiferrodistortive (AF) and ferrodistortive (F) structures: the first may be characterized by a spontaneous 90 degrees switching of the long Jahn-Teller axes of the octahedra within every second layer ('octahedron-axes switching' (OAS)), and the second by a spontaneous reorientation of the hydrogen-bonding pattern ('hydrogen-bond switching' (HBS)) in the whole crystal. With these mechanisms, the highly unusual sequence of the phases delta (Pbca, Z=4), epsilon (P2₁/n11, Z=2), zeta (IC(P2₁/n11)), and eta (P2₁/c11, Z=4) is delta -180 K(OAS) to zeta -132 K(HBS) to eta , delta from 192 K(OAS)- epsilon from 180 K- zeta from 141 K(HBS)- eta . On cooling, the OAS transition at 180 K leads directly to the incommensurate (IC) zeta phase with a modulation wavevector q_zeta =0.58t₂*+0.08t₃*. On heating, the modulation disappears at the same temperature and, as a new phase, the symmetrical epsilon phase occurs; at 192 K, the compound switches back to the room-temperature delta phase. The structure of the low-temperature eta phase, generated by the HBS mechanism, may be considered a distorted version of the twin structure of the delta phase. No symmetry relation exists between the IC zeta phase and the delta and eta phases. We find that the complex transition sequence of C₃CuCl is in fact the result of two interfering phase sequences, namely those of the AF and F states, respectively. X-ray data on the high-temperature IC gamma phase are also presented.