Table of contents

The AC conductance G(E, omega ) (at Fermi energy E and frequency omega ) of a quasi-one-dimensional system of finite length L governed by a tight-binding model Hamiltonian is expressed in terms of the pertinent transfer matrices, Consequently, one can study dynamic response functions of disordered systems within the powerful framework of random (transfer) matrices, which proved to be extremely useful in the analysis of DC conductance. We employ this formalism to investigate the low-frequency behaviour of In G(E, omega ). As expected, the linear term vanishes, whereas the quadratic term can be expressed in terms of the eigenvalues of the DC transfer matrix. In a strictly one-dimensional case it can be written in terms of the DC conductance and its energy derivatives. The thermodynamic limit L to infinity cannot be taken term by term. It is conjectured that, in general, (lnG(E, omega )) is not self-averaging, in agreement with previous numerical results.

We have measured polarized Raman spectra in a 2.0 mu m GaN epitaxial layer of high quality, grown on a sapphire substrate. All symmetry-allowed optical phonons in GaN have been assigned as follows: A₁(LO), 735 cm^-1; A₁(TO), 533 cm^-1; E₁(LO), 743 cm^-1; E₁(TO), 561 cm^-1; E₂, 144 and 569 cm^-1. Using the Lyddane-Sachs-Teller relation, the static dielectric constants of GaN for the ordinary and extraordinary directions have been estimated as epsilon _{perpendicular to 0}=9.28 and E/sub //0/=10.1. We have also observed quasi-LO phonons in GaN. A brief discussion on these will be given.

The nanoscopic structure of annealed Ge₂₀Te₈₀ glasses has been observed using atomic-force microscopy (AFM). The AFM images reveal to some extent the structure of amorphous semiconductors. AFM images as large as 20*20-50*50 nm² have shown many particles with diameters 2.5-3-5 nm-of the same order of magnitude as the interparticle distance. Since the peaks for each alignment lie at roughly the same height, it has been possible to image a couple of linear alignments 15-20 nm in length by AFM. The two-dimensional alignments generated a pseudo-surface (or infant surface) around 15*10 nm² in area. With the aid of our previous work on the quasi-atomic-scale images observed by scanning tunnelling microscopy and on the diffraction data from neutron scattering, the nanoscopic structure of annealed Ge₂₀Te₈₀ glasses can be interpreted in terms of short-range order (SRO), medium-range order (MRO) and supramedium-range order (SMRO) (the regions are approximately=0.5 nm, approximately=2.5-3.5 nm and approximately=15-20 nm, respectively). The SMRO should be established by pseudo-ordered alignment or rings of the particles characterized by MRO of diameter 2.5-3.5 nm; the MRO should be established by pseudo-ordered arrangements among the structural units characterized by SRO.

We consider a model of a crystal monolayer, which consists of a monatomic step on an anisotropic substrate of a periodic substrate interaction potential. The formation of the surface mono-atomic step is described by a surface misfit phenomenon involving large-amplitude displacements of atoms on the crystal surface. These large-amplitude displacements are misfit dislocations whose dynamics are governed by a two-dimensional double-sine-Gordon equation. Some solutions of this equation, namely the bi-kink soliton solutions, are found with emphasis on effects of the surface boundaries. It is shown that the soliton density, which measures the bi-kink fraction in a multi-soliton configuration, can be conserved in two dimensions.

We report the results of ab initio Hartree-Fock calculations of the exciton relaxation in the second layer of the (100)NaCl surface which demonstrate a possible mechanism of the alkali-atom emission in the primary process of decay of an electronic excitation in alkali halides.

We have investigated a solution of tetrabutylammonium iodide (TBAI) in the polar solvent formamide (FA) using angle-resolved photoelectron spectroscopy at the Berlin Electron Storage Ring for Synchrotron Radiation (BESSY). The concentration was set to 0.5 molality. We have evaluated the signals from C 1s pertaining to TBAI and FA separately, varying the photon energy between 310 and 540 eV as well as the e-emission angle with respect to the surface normal. The combination of these data with earlier results from ARXPS obtained by Siegbahn and co-workers allowed us to establish a concentration-depth profile of the surface-active salt TBAI. For this purpose we have employed a new theoretical formulation and developed a new elaborate fitting program based on a genetic algorithm. The concentration of the salt could be followed down to a depth of about 45 AA below the surface. Cross sections for the inelastic energy loss of electrons in the liquid could be established as a function of electron energy. Comparison with conventional surface tension measurements allowed us to derive absolute values for the cross sections.

Phonon spectra of CdTe and HgTe surfaces and interfaces are calculated within a Green function formalism. The dependence of phonon density of states spectra on the surface orientation is studied. Of the two compounds considered, HgTe shows the most drastic modification of its phonon structure as a surface is created. (001) and (111) polar interfaces show very different vibrational properties: while the former exhibits a two-mode or persistent behaviour with a vanishing phonon gap for any polarization, the latter exhibits a one-mode behaviour depending on the vibration polarization. For the (011) interface a two-mode behaviour is obtained, where bulk optical modes of one compound penetrate some distance into the other compound. The obtained phonon spectra show interface induced effects in a frequency range where Raman scattering data in CdTe/HgTe superlattices were previously reported but not explained.

Near-edge X-ray absorption fine structure (NEXAFS) has been applied to the elucidation of the geometrical structures for condensed layers of Si(CH₃)_nCl_4-n (0<or=n<or=4). For compounds containing both methyl and chlorine groups (n=1,2,3), it was observed that the NEXAFS at the Si K edge split into two sigma * resonances which correspond to the excitations from Si 1s to the unoccupied molecular orbitals of the S-Cl and Si-C bonds, respectively. The polarization dependence of the intensity of the two resonances was observed only for Si(CH₃)₃Cl. On the basis of the incidence-angle dependences of NEXAFS peak intensities, it was concluded that the Si-Cl bond in Si(CH₃)₃Cl is parallel to the surface and the average angle between surface normal and three Si-C bonds is about 43 degrees .

The asymptotic form of the image interaction is derived for a classical external point charge at a distance z₁ outside a periodic metallic surface, generalizing to real metals the analytical result of Lang and Kohn (1973) for jellium. The centre of gravity z_c of the induced charge coincides with the position of the image plane z₀ in the limit of linear response. However, whereas z_c is shown to depend through non-linear response on the magnitude of the external charge q, z₀ is independent of q. We show that surface periodicity does not modulate z₀, but adds to the image interaction a periodic component decaying exponentially with z. In addition, the long-ranged effect of non-linearity on the interaction energy is a term attractive to a positive external charge and proportional to q³/(z₁-z₀)⁴. We report first-principles pseudopotential calculations of the interaction energy with three aluminium surfaces: (111), (100) and (110). A supercell geometry is used. A discrete classical model without adjustable parameters reproduces the effect of the surface periodicity on the image interaction at each of the three surfaces.

It is demonstrated using three-dimensional computer simulations that some simple non-interacting electron models that include electron scattering by grain boundaries exhibit coexistence of large persistent currents and small conductances, similar to that observed experimentally in isolated micrometre-scale gold rings. Models with simple grain boundaries, and models with small numbers of regularly stepped or atomically rough dilute grain boundaries, have been studied and found to yield similar results, which differ markedly, however, from the predictions of models that assume only random impurity scattering. This difference is due to the fact that equilibrium persistent currents and non-equilibrium transport coefficients are physically different things and depend in different ways on the topology of the defect structure in a conducting ring. Experiments on metal and semiconductor rings that should be able to determine whether this is the explanation of the effects observed by Chandrasekhar et al. (1991) are proposed.

The method of surface Green function matching (SGFM) is extended to the case of an arbitrary number N of interfaces so that all the SGFM formulae are obtained in a compact N*N matrix (or supermatrix) form ready for practical calculations. As an example, shear (surface horizontal) waves in a Ni-W-Al-Ni structure are studied by a straightforward application of the general formulae. The range of applications of physical interest is stressed.

To understand the mechanism of defect initiated atomic emissions induced by electronic excitation of semiconductors, a semiempirical approach is used to study the nature of the relaxed excited states for an Si adatom on the Si(111) surface. We found an antibonding excited state embedded on the continuum and intermediate excited states, of which the relaxed states show bond weakening. Our results suggest that a series of excitation-relaxation processes can finally lead to the emission of weakly bonded atoms associated with defects on the semiconductor surface.

The quantum mechanical problem of tunnelling and reflection of holes from barriers and wells in semiconductor heterojunctions has been analysed. By using the Luttinger Hamiltonian to describe the dynamics of holes, the inherent anisotropy, non-parabolicity and heavy- and light-hole mixing at the top of the valence band has been taken into account. The symmetries of the hole scattering matrix are worked out, from which very general relations among the reflectivities and transmissivities are obtained. These analytical results are complemented by numerical calculations for several different situations; within the range of parameters investigated, no resonant unity transmission above the barrier is found. We do find that the transmission is zero (anti-resonances) in the well configuration for particular values of the energy of the incoming heavy hole.

We present an in situ resistance study of evaporated Co₆₅Fe₃₅/Ag alloys, which present a giant magnetoresistance (GMR) effect. We have observed the influence of the substrate and evaluated the initial perturbed layer in coevaporated alloys, in multilayers with thick layers, and in multilayers with very thin layers. When very thin 'layers' of each metal are deposited, a granular structure is observed. This particular preparation mode allows us to observe, at the beginning of the deposition, a preferential sticking of the Co₆₅Fe₃₅ alloy on the substrate rather than on the already deposited Ag. This result shows the dependence of the granular structure on the sample thickness and corroborates GMR results.

Effective dynamic equations describing weakly non-linear exchange-magnetostatic excitations are derived for the antiferromagnetic film with an 'easy-plane' anisotropy. The existence of 'algebraic' solitons has been predicted. The conditions for the excitation of soliton states are investigated for various values of magnetic field perpendicular to the surface of the film and film thickness. The soliton relaxation is discussed.

We report on the temperature dependence of high-resolution angle-resolved photoemission spectra from d-like (Tamm) surface states on Cu₃Au(100) and Cu(100). The temperature was varied between T=39 and 600 K, the surface states were investigated in the vicinity of the M point of the corresponding surface Brillouin zones. On both samples, the temperature influences the Tamm states in a very similar way. We propose that d-like surface states may be used as sensitive probes to monitor the T-dependent potential changes in the outermost layer.

The irradiation effect of the Al₆₂Cu_25.5Fe_12.5 icosahedral quasicrystal (IQC) induced by 120 keV Ar ions and 1 MeV electrons has been studied in situ using high-voltage electron microscope-ion accelerator dual-irradiation facilities. The effect is dose and temperature dependent and we can define three critical temperatures T_c, T_d and T_p. When irradiated at T<T_c the IQC is transformed into a crystalline microtwin oriented according to icosahedral symmetry. When T>T_c, the IQC remains an IQC but with a higher (when T>T_d) or lower (when T<T_d) degree of perfection. The planar defects originally existing in the Al₆₂Cu_25.5Fe_12.5 IQC foils are stable when T>T_p but disappear when T<T_p. For the Al₆₂Cu_25.5Fe_12.5 IQC, we found that T_c approximately=400 K, T_d approximately=450 K (for 1 MeV electrons) or 550 K (for 120 keV Ar ions) and T_p approximately=600 K.

Thin films of Fe_1-xNi_xS₂ (0<or=x<or=1) solid solutions have been prepared by sulphuration of Fe-Ni layers previously flash evaporated. Their structural properties have been investigated by X-ray diffraction (XRD) as a function of the sulphuration temperature and Ni/Fe atomic ratio. The dependence of the lattice parameter on the Ni/Fe atomic ratio shows that Ni increases the dimensions of the unit cell, which appears to be initially compressed.

The isomer shifts of undoped and transition-metal-doped compounds such as Co, Ni and Mn-doped beta -FeSi₂ are discussed on the basis of molecular orbital calculation. From the LCAO SCF cluster calculation, the electron charge densities calculated p(0) at iron nuclei of Fe I and Fe II sites are found to be consistent with isomer shifts as experimentally determined. On the basis of the Mulliken population analysis, it is disclosed that the 3d screening effect also plays an important role in determining the isomer shifts of both Fe sites in beta -FeSi₂. Concerning Co-doped specimen, increases in the isomer shifts of both Fe sites with increasing Co concentration is observed; this experimental result can be interpreted as the lattice expansion due to the Co dopant, considering that the calculated p(0) of both Fe sites is reduced on lattice expansion. However, the calculated isomer shift of the Fe I site results in a value larger than the experimental value; this implies that lattice expansion is anisotropic rather isotropic. Increases in the isomer shifts of both Fe sites in Ni or Mn-doped beta -FeSi₂ are also observed, which is also interpreted as due to lattice expansion.

Experimental studies on shallow acceptor and donor field ionization dynamics in MBE-grown GaAs epitaxial layers are presented. The experiments have been carried out with a new non-resonant spectroscopic technique, transient tunnelling spectroscopy. It has been found that intense hole tunnelling from shallow acceptors begins at electric fields higher than 6000 V cm^-1, while electron tunnelling from shallow compensating donors to the conduction band occurs at much lower electric fields, of the order of 500 V cm^-1. The dependence of the acceptor to valence band tunnelling time on the electric field strength has been measured for the first time, From the analysis of the experimental results it is concluded that initially the acceptor to valence band tunnelling dynamics is dominated by a light-hole mass.

Decay rates for long-wavelength spin fluctuations in the standard x-y model, and a second model described by the sum of the full Heisenberg exchange interaction and a single-site anisotropy, are calculated on the basis of a coupled mode theory in which the isothermal susceptibilities are consistent with corresponding spherical models (in consequence, quantum fluctuations are absent). Results for the decay rates are provided for two- and three-dimensional lattices and ferromagnetic and antiferromagnetic exchange interactions. Their dependence on the wave vector, q, and the inverse correlation lengths are derived from integral equations that admit homogeneous functions as solutions. The decay rates may thus be expressed as q^d2/R₁(q/ kappa ₀) and q^(d+2/2)R₂(q/ kappa ,q/ kappa ₀) where R₁ and R₂ are, respectively, scaling functions for the x-y and ferromagnetic single-site anisotropy models, d is the spatial dimension, and kappa (KO) is the out of plane (in plane) inverse correlation length. Analytical and numerical values of the scaling functions are presented for the limits of main physical interest.

The magnetic susceptibility, hysteresis curves, ZFC and FC magnetic moments and magnetoresistance measurements indicate the appearance of a spin glass state in UPd₂Ge₂ below 87 K, and the existence of a ferromagnetic phase below 80 K. From these studies and taking into account the neutron diffraction data a mixed spin glass-ferro-antiferromagnetic state is suggested in the low-temperature region.

The Neel vector of a small antiferromagnetic particle can resonate between degenerate directions. The total moment of the particle will generally be non-zero, and this resonance cannot then occur without transverse anisotropy. Previous work is extended to include such anisotropy via a two-spin model representing the two sublattice moments. The WKB exponent in the tunnel splitting is found exactly when the sublattice moments are equal, and by a variational method that holds over the entire parameter space when the moments are unequal.

In the literature on nuclear magnetic resonance in small metal particles, three main lines of approach are found which lead to qualitatively different predictions for the behaviour of Knight shift and, where available, for the linewidth. In this paper it is shown how these differences arise as consequences of the different sets of starting hypotheses used in each type of approach. Some characteristics of the nuclear spin-lattice relaxation are discussed as well, and attention is given to experimental considerations.

The temperature- and frequency-dependent dielectric susceptibility of KD₂AsO₄ was studied in the ferroelectric phase. A dispersion which is related to the dipole reversal was found in the high-frequency range (10⁷ Hz). The dispersion is described by a Gaussian distribution of relaxation times which show Vogel-Fulcher temperature dependences. The high value of the dielectric constant below the phase transition is related to the mobility of the domain walls. This mobility steeply decreases near the freezing temperature T_f and therefore a sudden decrease in the susceptibility takes place. A non-linear superposition of the piezoelectric resonance and the relaxational dispersion was observed.

We have studied the frequency and temperature dependence of the dielectric constant parallel to the ferroelectric c axis in the ferroelectric phase of KH₂AsO₄. The dielectric dispersion found in the high-frequency range (1 MHz-1000 MHz) could be described by a Vogel-Fulcher-dependent Gaussian distribution of relaxation times, similarly as in KH₂PO₄. However, the use of a Vogel-Fulcher law is less convincing than in KH₂PO₄. Below the freezing temperature T_f the relaxation times strongly increase resulting in a decrease of the domain wall mobility. This may be due to the non-linear superposition of the piezoelectric resonance and the relaxational dispersion.

Using the transfer matrix method for calculating the band structure, transmission and reflection coefficients of photonic systems, we present a method for calculating photonic dispersion surfaces. We then show how dispersion surfaces are excellent tools in the search for photonic insulators by applying the method to a simple system. We also explore the potential of metals for incorporation in photonic structures.

Cubic Fe³⁺ centres in NaCl single crystals have been studied with electron nuclear double resonance (ENDOR) at T=17 K. The superhyperfine (SHF) and quadrupole interactions of the paramagnetic ion with the nearest set of four Cl^- anions and the second-nearest set of twelve Na⁺. Cations have been determined. The results confirm the structural model of an Fe³⁺ ion at a tetrahedrally symmetric interstitial position with the nearest four Na⁺ cations missing, as suggested previously from the analysis of the SHF structure of their EPR spectra.

Mossbauer spectroscopy has been used to follow the temperature-dependent magnetization in the temperature range 4.2-823 K of the amorphous as well as the nanocrystalline grains in partially crystallized Fe_73.5Cu₁Nb₃Si_16.5B₆ alloy. In nanocrystallized specimens, the magnetization behaviour of both amorphous and nanocrystalline components is significantly different from that of the bulk amorphous or crystalline Fe₃Si. Initially the nanocrystalline grains exhibit a superparamagnetic behaviour above the Curie temperature of the amorphous matrix. With increasing density of nanocrystalline grains, the superferromagnetic interaction between the grains becomes important.