Table of contents

We present a novel approach to the calculation of magnetic anisotropies in crystals and in thin films. Our technique is based on self-consistent real-space recursion calculations on a tight-binding linear-muffin-tin-orbital (TB LMTO)-Hubbard Hamiltonian including spin-orbit coupling and allowing for arbitrary orientations of the local spin-quantization axes. It allows scanning of the magnetic energy continuously as a function of the orientation of the magnetic moment, and thus avoids the computational problems that plague other techniques. Applications are presented for bulk body-centred-cubic iron and for face-centred-cubic Fe monolayers on Cu(100) and Cu(111) substrates. We predict a perpendicular direction of the magnetic moment for Fe/Cu(100) and in-plane orientation of the spins for Fe/Cu(111), with anisotropy energies of the order of 1 to 2 meV atom^-1. The technique is accurate enough for studying the weak in-plane anisotropies in Fe/Cu(111) that are of the order of 1 mu eV atom^-1.

Results of an investigation of the trends and dynamics of excited state production in collisions of inert gas projectiles on Na, Mg, Al and Si surfaces are presented. Measurements of charge fractions and electron spectra are reported. The efficiency of excited auto-ionizing-state production as a function of energy is determined. These results are compared with known gas-phase collision results. The similar general trends suggest that both in gas-phase collisions and in ion-surface scattering, similar primary excitation mechanisms are operative. However, surface-specific electron capture, loss de-excitation and core rearrangement processes lead to a strong modification in final-state distributions and dominant selective production of the lower-lying excited states occurs.

This work develops the semiclassical dielectric theory of energy loss, including retardation effects, for electrons travelling normal to a stratified slab. Starting from a transfer matrix formulation and using computer algebra, we obtain closed formulae for the dispersion relation, the Hertz vector and the energy-loss spectrum, valid for any finite number of layers. Our results are applied to multilayers of Al/Al₂O₃ and are used to investigate the validity of modelling a single interface by a sharp discontinuity in the dielectric function.

This paper develops the dielectric theory of energy loss, including retardation effects, for electrons travelling parallel to the interfaces of a stratified slab. A transfer matrix recurrence relation is introduced to implement the boundary conditions at surfaces and interfaces. This recurrence relation is solved exactly to provide closed formulae for the Hertz vector, the dispersion relation and the energy-loss probability, valid for any position of the beam and any number of layers. The non-retarded limits of these expressions are also examined and compared with the results of a simpler calculation based directly on Poisson's equation. Our final results are illustrated by numerical calculations applied to Al/Al₂O₃ multilayers and a InSb/GaP bilayer.

This work extends the non-retarded semiclassical dielectric theory of electron energy loss to multilayers composed of anisotropic media. A transfer matrix recurrence relation is used to obtain expressions for the potential, dispersion relation and energy-loss spectrum, valid for any finite number of anisotropic layers and any orientation of the anisotropy axes. Both normal and parallel incidence are considered in the non-retarded limit. These theoretical results are illustrated by application to graphite surfaces and graphite-diamond interfaces.

We present an experiment of densification of a 2D horizontal assembly made of regular pentagons. The annealing leads to a dense crystalline arrangement with quasisixfold symmetry in spite of the geometrical characteristics of the grains and of the a priori short-range correlation length in the packing. This configuration yields the maximum average number of side to side contacts between the grains and probably the maximum packing fraction.

We investigate the magnetoplasmon polariton modes and the propagation of electromagnetic waves in a random-thickness superlattice in a magnetic field. The layer thicknesses are distributed in accordance with a special probability function. The magnetic field is parallel to the interfaces, and the propagation of the waves is perpendicular to the field. By use of the transfer-matrix method we show that there exists a resonant optical transmission for waves with particular frequencies. We calculate the dispersion relation of the bulk modes, the resonant frequencies and the transmission coefficients. The transmission coefficients at the resonant frequencies are sensitive to variations in the magnetic field. Possible applications of such a structure in new devices are discussed.

The influence of capping layers on surface phonon polaritons in two-layer superlattices is investigated here theoretically. General analytical expressions are given. A few illustrative applications are presented afterwards, for p-polarized non-retarded polaritons within GaAs-InAs superlattices with or without an InP capping layer. A few dispersion curves for such polaritons are drawn. The variation of the frequencies of these polaritons with the thickness of the capping layer is given.

The spin wave stiffness constant D (q) of unsupported FCC (100), (110), and (111) nickel monolayers are calculated for different directions, q, of spin wave propagation. It is shown that D is isotropic for the (100) and (111) orientations, and highly anisotropic for the (110) orientation. Results indicate that the strength of the exchange coupling between nearest neighbour magnetic moments is quite sensitive to the effective space dimensionality as well as the crystalline structure of the system.

The magnetic behaviour of the inhomogeneous alloyed film, Cu_0.94Co_0.06, produced by electrodeposition has been investigated in the temperature range 2 to 300 K and in magnetic fields of up to 5 T using SQUID magnetometry. The films exhibit either hysteretic or reversible behaviour depending upon the temperature. This is interpreted as being due to the presence of small, FCC, Co-rich particles of FCC Co in the non-magnetic matrix, giving rise to superparamagnetism at higher temperatures. With the help of low-field susceptibility measurements made on samples cooled in zero field we have been able to determine the distribution of blocking temperatures and thus make an estimate of the size distribution of the particles in the Cu matrix; these have been found to range in diameter from 6 to 12 nm. These results have been compared with values obtained from a measurement of the temperature dependence of the remanence. We have found that, even in such a relatively dilute magnetic system as Cu_0.94Co_0.06, interactions between magnetic clusters are not negligible. The interaction energy is estimated to be of the order of 1.7*10^-3 eV.

The giant magnetoresistance effect in magnetic superlattices for the current perpendicular to and in the layer planes is studied within a unified semiclassical approach that is based on the Boltzmann equation with exact boundary conditions for the spin-dependent distribution functions of electrons. Interface processes responsible for the magnetoresistance are found to be different in these geometries, and this can result in an essential difference in general behaviour between the in-plane magnetoresistance and the perpendicular-plane one. A correlation between the giant magnetoresistance and the multilayer magnetization is also discussed.

The total backscattering coefficient eta ₊ has been measured for 1 keV positrons incident normally on a number of elemental targets with atomic numbers Z in the range 6-82. The measurements require the minimization of surface contamination on the samples and as a result data analysis has to allow for the formation of positronium at the sample surface. Pronounced structure in eta ₊(Z) is observed, and it is suggested that this may be correlated with the effective atomic electron density.

We study the electronic states at the Gamma point of a heterostructure, formed by a semiinfinite substrate (AlAs), a well material (GaAs) and a semi-infinite superlattice (AlAs/GaAs), grown along the (001) direction. We employ an empirical tight binding (ETB) Hamiltonian together with the surface Green function matching (SGFM) method. We study the effect of the well layer width and the relative thicknesses of the layers forming the superlattice on the energy values and spatial localization of the different states (near the main gap at the Gamma point) of this heterostructure.

The crystalline structure of poly(3-octylthiophene) as a function of pressure up to 3 GPa (30 kbar) has been studied by energy dispersive x-ray diffraction. The distance between neighbouring thiophene chains in the direction normal to the thiophene rings, b_{perpendicular to} , has been found to decrease with increasing pressure. This is believed to increase the planarity of the polymer main chains and explains why the thermochromic transition is inhibited under pressure. The crystallographic repetition distance a in the direction along the octyl side chains is found to be constant during increment of pressure. This shows how the alkyl side chains in cooperation act as rigid interchain spacers even if, independently, they are flexible and rather disordered. A lack of change in the interchain charge complex transfer integral can be understood from the pressure behaviour of a. The measurements indicate that the crystalline fraction of this paracrystalline polymer increases with pressure, and that the size of the crystalline domains increases slightly.

The relation between the singularities of multifractal energy spectral measures and the behaviours of the Green function is studied in the framework of a tight-binding Hamiltonian. If the measure mu (E) has a seating behaviour at energy E of the form Delta mu (E)= mu (E+ delta )- mu (E- delta ) varies as delta ^{alpha (E)}, it is proved that the imaginary part of the Green function P(E, epsilon ) scales as P(E, epsilon ) varies as epsilon ^{beta (E)} with beta (E)= alpha (E)+1, the reverse being also true. This is exemplified in the case of the density of states and the local density of states of the one-dimensional Fibonacci quasicrystalline chain.

Measurements of dislocation resistivity and the deviations from Matthiessen's rule (DMRs) are reported for polycrystalline Cu samples with various dislocation densities, impurities and phonons. Since this DMR can reach the same order of magnitude as the dislocation resistivity itself, its quantification is essential in using the electrical resistivity for dislocation density measurement. In the framework of the classical two-group model, a formula for the DMR is derived for the three independent scatterers. The fits by this formula are satisfactory provided that isotropic phonon scattering is chosen for samples with high dislocation densities. While samples with a high purity (residual resistivity ratio (RRR)>or=300) show dislocation scattering anisotropies A_dis approximately=0.1 irrespective of dislocation density, the more impure material (purity 99.99%) (RRR=130) exhibits a significant decrease in A_dis at low dislocation densities, indicating some contribution of enhanced strain fields in the dislocation structure. This leads to an enhanced DMR contribution in this material and makes the determination of the dislocation density N by the electrical resistivity method less reliable. Nevertheless, in the whole purity range considered, the electrical resistivity yields at least the same measuring resolution as TEM of about Delta N approximately=1*10⁹ cm^-2, but a better performance for high dislocation densities up to the order of 10¹⁴ cm^-2.

Lattice constants and bond lengths in crystalline Si_1-xGe_x alloys are calculated by the bond orbital model and radial force model. The results are compared with theoretical results obtained by the valence force-field model calculation and experimental results obtained by X-ray diffraction and extended X-ray absorption fine-structure measurements, and are found to be in good agreement with experiments. The trends in elastic constants and electronic band structure as functions of composition x are obtained in the simple virtual-crystal approximation.

The intrasubband and intersubband plasmons in the long-wavelength limit are calculated numerically for GaAs/AlAs parallel quantum-well wires (PQWW) produced by the deposition of GaAs and AlAs fractional layers on (001) vicinal GaAs substrates. One intrasubband plasmon mode is found, the energy of which approaches zero as the in-plane wave vector q to 0. The dispersion relations of the intrasubband plasmons are anisotropic with h(cross) omega _p=(( sigma ²q_x²+ alpha ₀²q_y²)/q)¹2/. The constant alpha ₀ increases monotonically with the electron density n_e of the PQWW, while a increases with n_e when the Fermi energy E_F is in the electron conduction subbands; it decreases with n_e when E_F is in the gaps of the subbands. The intersubband plasmons are found to be multi-branch structured with non-zero energies as q to 0. The calculations show that, for the GaAs/AlAs PQWW we consider where the lateral periods L_x of the PQWW are much less than 200 AA, the far-infrared optical absorptions are dominated by single-electron direct transitions between different electron conduction subbands. For PQWW with L_x larger than 300 AA, the absorptions are due to intersubband plasmon absorptions. Between these two regions, both single-electron direct transitions and intersubband plasmon absorptions can make important contributions to the absorptions, depending on the structures of the PQWW.

In this paper, weak localization of a new kind is studied. Within the framework of the new developing theory, electron exchange is shown to alter the angular distribution of scattered electrons. The range of energies of incident electrons for which exchange effects are important is estimated.

Mutual screening of charged impurities in bulk compensated semiconductors is described. Several models for analytical treatment are discussed. Monte Carlo simulations of impurity relaxation in disordered systems at non-zero temperatures are performed. Pair distribution functions are obtained. A structure of the screened acceptor-donor-electron complex is studied. Friedel-like oscillations in a screening charge distribution are observed. The matrix element for electron scattering and electron mobility is calculated. Remarkable behaviour at small wavevectors is observed. Essential corrections against standard models at the limit of zero temperature are found.

Quasi-two-dimensional hydrogenic impurity levels in quantum wells under the influence of magnetic fields are investigated. Experimentally, 1s to np (n=2, 3, 4) transition energies in quantum wells of various widths are measured by the photoconductivity technique for magnetic fields up to 9 T. Theoretically, these energies are calculated by a new variational method that is valid for the whole range of magnetic field strengths, from the zero-field limit to the high-field limit. Good agreement is achieved for every transition for well widths up to approximately 300 AA, for which the confinement effect is appreciable. Consistent discrepancies found for 450 AA are expected to be accounted for by including subband mixing because of the wide width and small subband separation. Our study demonstrates that the simple variational method can produce good results for highly excited levels in magnetic fields provided that the confinement due to the quantum well is sufficiently strong.

Under suitable approximations, it is shown that mobility of electrons in two-dimensional systems is always zero for a finite value of disorder, irrespective of the type of scattering potential.

The transmission through a two-dimensional (2D) channel, which connects semi-infinite 2D electron gas reservoirs and can be used as a model of both the few-atom-size (microscopic) and mesoscopic contact is investigated under the assumption of the zero-field ballistic motion of electrons in the system. An analytical theory is advanced to a nearly exact analytical expression for the transmission probability verified by numerical calculations for a vast variety of channel parameters. On these bases a detailed analytical and numerical analysis of size effects in the quantum point contact (QPC) transmission is performed with emphases put on basic distinctions between atomic-size and mesoscopic-size QPC. The QPC properties are visualized in calculations of transmission against Fermi energy dependences for some representative channel parameters. The results obtained are directly applicable to the description of coherent excitation energy transfer in relevant molecular 2D geometries. The suggested formalism can be also used for an analytical analysis of some other problems of current dispute, in particular, of non-linear field effects and 3D QPC properties.

We present a variational estimate for the binding energy of a Frenkel exciton in the insulating cuprate superconductors. Starting from the three-band Hubbard model we perform a canonical transformation to O(t²), where t is the bare nearest-neighbour copper-oxygen hopping integral. An effective Hamiltonian is then derived to describe the hopping of the exciton through the copper oxide plane. The critical parameter in the model is the nearest-neighbour copper-oxygen Coulomb repulsion, V. It is found that a critical value of V is needed to observe bound Frenkel excitons, and that these excitons have the same symmetry as the parent copper orbital, d(x²-y²). We determine the critical value of V using a variational approach, and attempt to fit the parameters of the model to known experimental results.

We have re-examined the infrared absorption bands due to the O-H bond-stretching vibration and the polarization characteristics in undoped and MgO-doped LiNbO₃ using well-characterized crystals. It has been found that the O-H bond stretching vibrational frequency nu (OH) has a strong correlation with Nb concentration in the crystals. We have also determined the position where hydrogen enters using Novak's empirical relationship between the values of nu (OH) and the length of the hydrogen bond and the structure analysis data for the undoped crystals. On the basis of those results and the polarization characteristics, we have examined the intrinsic and the extrinsic defect structure models in LiNbO₃. It has been clarified that the behaviour of nu (OH) reflects the defect structures. The behaviour of nu (OH) supports the Li-site vacancy model as the intrinsic defect structure model, and the corresponding extrinsic defect model. A brief discussion is also given of the behaviour of nu (OH) in crystals simultaneously doped with two kinds of impurity.

Four emission bands, all belonging to the main thallium centres, have been detected in CsI:Tl, and their spectral, polarization and kinetic characteristics have been studied at 0.35-320 K. A new model is proposed for the excited states responsible for the luminescence of thallium-doped caesium iodide. In this model, the main thallium centre is considered as a cluster consisting of a Tl⁺ ion and at least 12 I^- and 11 Cs⁺ ions. Two weak ultraviolet emission bands of CsI:Tl (3.31 and 3.09 eV) are ascribed to electronic transitions from trigonal and tetragonal Jahn-Teller minima of the triplet relaxed excited state, whose structure is similar to that of Tl⁺ centres in FCC alkali halides. Two intense visible bands (2.55 and 2.25 eV) are assumed to arise from two different ("weak" and "strong") off-centre configurations of a selftrapped exciton perturbed by the TI⁺ ion. The minima responsible for all the emission bands are located on the same adiabatic potential energy surface. The excitonic-like nature of visible emission could explain the high scintillation efficiency of CsI:Tl.

Using the linear-response dielectric theory and local electron density approximation, we calculate the electronic energy loss of low- and intermediate-energy protons in collision with a solid atom as a function of the impact parameter. In these studies, the correlation and exchange interaction of the electron gas in solids is taken into consideration by using a local field correction dielectric function. Based on the formalism electronic stopping cross-sections of protons in solids are calculated, and the results are compared with experimental data and empirical results, as well as with other previous theoretical results.

alpha "-Fe₁₆N₂ single-crystal films were prepared with a facing targets sputtering (FTS) system onto single-crystal NaCl substrates, at a deposition rate as high as 2.0 AA s^-1. In the process of deposition, the partial pressures of argon and nitrogen gases were held at 0.3 Pa and 0.05 Pa, respectively. The effect of the substrate temperature on the formation of Fe-N phases was investigated in detail. The crystal structure of the Fe-N films was examined with an x-ray diffractometer and a transmission electron microscope. At T_s=150 degrees C, single-crystal alpha "-Fe₁₆N₂ grows epitaxially on alpha -Fe with an orientation relationship of alpha "-Fe₁₆N₂ (110)// alpha Fe (110), Fe₁₆N₂ (111)// alpha -Fe(111). When T_s=100 degrees C and 200 degrees C, the Fe-N films contain more alpha '-martensite and gamma '-Fe₄N, respectively. The integrated electron diffraction pattern of the alpha "-Fe₁₆N₂ single crystal in the (111) direction was clearly observed. The film revealing a single-crystal electron diffraction pattern exhibits a high saturation magnetization M_s up to 2050 emu cc^-1, the high M_s of the alpha "-Fe₁₆N₂ phase.