Thermal conductivity is a powerful tool to probe the phonon and electron exitations in a solid, especially in superconductors were one can basically tune the respective electronic and phononic contributions by applying a magnetic field below Tc. / After a short review on the concepts of superconductivity, thermal conductivity and amorphous matter, we present a study of the thermal conductivity of an exotic material, the amorphous metallic superconductor Fe0.5Ni 0.5Zr2. The results indicate an unexpected dominant electonic contribution to the thermal conductivity across the superconducting transition, in accordance with an inhomogeneous sample composed of a bulk normal phase with inhomogeneous superconducting phases.

The RMn6T6-xXx family of compounds (R = rare earth; T = Ge, Sn; X = Ga, In) has seen a lasting and intensive series of studies over the past several years. In these systems, a spin reorientation process, which is a pure rotation of the magnetic structure relative to the crystal axes, can be used to determine the anisotropic contributions to the transferred hyperfine fields at the Sn sites. The anisotropic contribution has been shown to be substantial in the MnSn2 and FeSn2 compounds, and is an important fraction of the overall transferred hyperfine field. A spin reorientation transition can be either temperature-induced, or field-induced (spin-flop). The temperature-induced spin reorientation generally results from a competition between the magnetocrystalline anisotropies of the rare earth and Mn sublattices. The substitution of Sn with Ga strongly affects the anisotropy, shown here to decrease the spin reorientation temperature with increasing x by 255 +/- 18 K/Ga in TbMn6Sn 6-xGax. However, the sublattice anisotropies seem unaffected by In substitution, and the spin reorientation temperature is nearly constant throughout a large range of In concentration. A field-induced spin-flop can be achieved by applying a large enough field perpendicular to the direction of the moments. / In this study, we show how both the temperature-induced SR and field-induced SF allow for the anisotropic field to be isolated from the isotropic contribution. The consistency between the two measurements of the anisotropic field indicates that the magnitude of the anisotropic contribution is independent of the driving force of the reorientation. We show that a complete 90° spin reorientation occurs in the ErMn6Sn5.89Ga0.11 and TbMn 6Sn6-xGax compounds (0.2 ≤ x ≤ 0.8), as well as in TbMn 6Sn5.46In0.54 at room temperature for an applied field of 0.57(3) T. The site preference for Ga substitution is investigated and compared with former results. Finally, the anisotropic contribution at one of the Sn sites is shown to exceed 40% in all of the compounds investigated, and this site assignment is confirmed.

The structural and magnetic properties of Diamagnetic/Ferromagnetic: Cu/Fe multilayers, prepared by DC-magnetron sputtering, were studied as a function of Fe layer thickness tFe. Structural characterization reveals the successful growth of high-quality layered structures along the film growth direction. However, the increasing contribution of the interface roughness was visible with decreasing tFe . X-ray diffraction data indicate the dissolution of Fe atoms in fcc Cu medium due to interfacial mixing. / The magnetic transition from ferromagnetism to paramagnetism with decreasing tFe was confirmed by conversion-electron Mossbauer spectroscopy (CEMS), vibrating sample magnetometry, and magnetotransport measurements. The monotonic decrease in the average hyperfine field with t Fe indicates that the transition is a gradual process as t Fe decreases from 34 A down to 7 A. The isomer shift of the singlets in the CEM spectra suggests the presence of fcc Fe, while the observed doublet is assigned to the Cu-Fe alloy phase at the interfaces. / The variation of magnetoresistance (MR) with t Cu indicates that for a nominal t Fe of 20 A the multilayer has a well-defined superlattice structure, whereas multilayers with nominal tFe, of 5 A have a granular-alloy-like structure. AC susceptometry provided direct evidence for island formation for nominal tFe = 5 A by exhibiting the blocking characteristics of superparamagnets. The temperature-dependence of the magnetization suggests that for nominal tFe = 5 A, 75% of the Fe atoms are in the superparamagnetic bcc phase, leaving the remaining 25% in the Cu-Fe alloy and fcc Fe phases. / The observed magnetic transition is likely due to superparamagnetic relaxation rather than a structural transition from bcc Fe to fcc Fe as t Fe decreases below a critical thickness tC at which a multilayer structure becomes an island structure.

This thesis studies the diffuse x-ray scattering from rough interfaces. We review the scaling hypothesis of the height difference correlation function of rough interfaces, and the relationship between the roughness and the specular reflectivity and diffuse scattering cross-section in the distorted wave Born approximation (DWBA). We study the properties of the position sensitive detector (PSD), particularly its dark counts and noise level. The conventional setup and off-plane scan setup are compared for their advantages and disadvantages. We use a polished silicon surface to exemplify the data processing. We find that the parameters which fit the detector scan data can fit all data from the specular reflectivity, the rocking scan and the offset scan very well. The polished silicon surface is well described by an exponential form of the height-height correlation function which satisfies the scaling hypothesis.

Monte Carlo simulations used for representing dynamical physical phenomena are studied in terms of a Markov chain operator acting on the probability distrubution of the states of a given system. The most general transition rule satisfying detailed balance and leading to a canonical ensemble probability distribution is derived using this formalism. The explicit Markov chain representing the two most commonly used canonical algorithms, the Metropolis and the Glauber transition rules, is then constructed and numerically applied to the states of an Ising model. The dynamical properties of the system are studied for each algorithm. Various measures, such as time-time correlation functions, are estimated for different system sizes and finite-size sealing is applied. In particular, the effects of the transition rule on the dynamic critical exponent is investigated. / We at first examine one- and two-dimensional systems using periodic boundary conditions. Systems with free boundary conditions were also studied, and their results were equivalent with respect to the dynamical critical properties of the system. The effects of conservation laws were also investigated and both conserved and non-conserved systems were studied. Both local and non-local spin-exchange dynamics were investigated for conserved systems. Finally, our approach was used to simulate quenches on small systems. / This method is them used to analyze phenomenological transformations done by dynamical renormalization-group (RG) methods. It is found that, when the RG transformation is linear in probability space, there exists a corresponding Markov chain generating the time sequence of the renormalized systems. An example is given for the one-dimensional Ising model.

We study the dynamics of two reaction-diffusion phenomena driven by chemical activation and thermal dissipation and evolving, respectively, on a randomly distributed or continuous medium. The first system describes the process of slow combustion of a randomly distributed reactant. It is studied by a phase-field model built up from first principles and describes the evolution of thermal and reactant concentration fields. Our combustion model incorporates thermal diffusion, activation and dissipation. We examine it in a manner which makes a connection between the propagation of combustion fronts, their kinetic roughening and the percolation transition. In so doing, we examine slow combustion in the context phase transitions. The second system describes propagation of reaction fronts arising in transformations obeying the Arrhenius law of chemical reactions. It too is modelled by a set of phase-field equations describing the dynamics of both thermal and concentration fields. A typical example of this transformation is the crystallization of an amorphous material. In addition to the features of our combustion model, this model also incorporated a realistic treatment of mass diffusion. Front propagation of our model is shown to undergo period doubling bifurcations as one varies the background temperature at which the system is maintained. The signature of these bifurcations is the same as those of the logistics map. We study how the bifurcation structure changes as a function mass diffusion, focusing on changes of the background temperature for which period doubling first emerges. This temperature is the most easily obtained experimentally.

In this thesis, we performed a detailed study of the thermal conductivity in a wide range of quasicrystals. Three systems were studied: AlPdMn, AlCuFe and AlPdRe, and the samples were in both single and polycrystalline form. A further variable was added by introducing a controlled level of defects. / At low temperatures (below 20K), the thermal conductivity is defect limited, being controlled by boundary scattering, two level systems, stacking faults and dislocations. At high temperatures (above 20K), we find the thermal conductivity is limited by intrinsic properties of the quasicrystalline structure and phonon-phonon scattering. / From fitting the thermal conductivity to a detailed model we are able to predict the maximum thermal conductivity of a perfect quasicrystal.

In this Ph.D thesis the conductance fluctuations of different physical origins in semi-conductor nanostructures were studied using both diagrammatic analytical methods and large scale numerical techniques. In the "mixed" transport regime where both mesoscopic and ballistic features play a role, for the first time I have analytically calculated the non-universal conductance fluctuations. This mixed regime is reached when impurities are distributed near the walls of a quantum wire, leaving the center region ballistic. I have discovered that the existence of a ballistic region destroys the universal conductance fluctuations. The crossover behavior of the fluctuation amplitude from the usual quasi-1D situation to that of the mixed regime is clearly revealed, and the role of various length scales are identified. My analytical predictions were confirmed by a direct numerical simulation by evaluating the Landauer formula. In another direction, I have made several studies of conductance or resistance oscillations and fluctuations in systems with artificial impurities in the ballistic regime. My calculation gave explanations of all the experimental results concerning the classical focusing peaks of the resistance versus magnetic field, the weak localization peak in a Sinai billiard system, the formation of a chaotic billiard, and predicted certain transport features which were indeed found experimentally. I have further extended the calculation to study the Hall resistance in a four-terminal quantum dot in which there is an antidot array. From my numerical data I analyzed the classical paths of electron motion and its quantum oscillations. The results compare well with recent experimental studies on similar systems. Since these billiard systems could provide quantum chaotic dynamics, I have made a detailed study of the consequence of such dynamics. In particular I have investigated the resonant transmission of electrons in these chaotic systems, and found that the level-s

Quantum corrections to the conductivity have been studied at low temperatures down to 0.15K and fields up to 8.8T in two different disordered systems, namely amorphous Ca-Al alloys doped with Ag and Au and icosahedral Al-Cu-Fe alloys. In the former the influence of spin-orbit scattering on the enhanced electron-electron contribution to the resistivity has been, for the first time, clearly displayed. As the spin-orbit scattering rate increases, this contribution decreases rapidly to finally vanish at extremely high spin-orbit scattering rates. Furthermore the analysis shows that the current weak localization theory gives an accurate description of the experiments irrespective of the level of spin-orbit scattering. / In icosahedral Al-Cu-Fe alloys, detailed study of the low temperature resistivity shows that the magnetoresistance and the temperature dependence of the resistivity data are consistent with the predictions of quantum corrections to the conductivity theories. The success of these theories in this alloy system is attributed to intense electron scattering due to disorder. The spin-orbit scattering and the electron wave-function dephasing rates are extracted from fitting the magnetoresistance. The dephasing rate is found to vary as AT$ sp{p}$ with $p sim1.5$; a characteristic of electron-electron scattering in the strong disorder limit. An antilocalization effect has also been directly observed in the temperature dependence of the resistivity in one of the samples.

Time resolved small angle x-ray scattering (SAXS) using synchrotron radiation was applied to the study of the kinetics of spinodal decomposition (SD) in an AlZn binary alloy at critical composition quenched into the immiscible region. These millisecond time scale measurements, performed at the National Synchrotron Light Source (Brookhaven National Labs., N.Y.), constitute the first direct experimental verification in a binary alloy of the theory proposed by Langer, Bar-on and Miller in 1975 for SD. A scheme based on the composition distribution functional is proposed to account for the decomposition taking place during the quench. The interatomic mobility, a free energy gradient coefficient and two coefficients that suffice to determine a coarse-grained (intensive) free energy have been obtained in the framework of this theory. The mobilities obtained compare well with tracer diffusion measurements reported in literature. A dependence of the coarse-grained free energy coefficients on the coarse-graining length is found and a procedure is proposed to uniquely choose the values of these coefficients based on the predicted integrated intensity from the equilibrium concentrations and on the measured integrated intensities. / Late-stage coarsening regimes were also investigated. In these regimes, growth exponents higher than the value 1/3 predicted by the Lifshitz-Slyozov-Wagner theory are obtained. These higher values, comprised between 0.40 and 0.45 are consistent with predictions that alloys in which elastic effects are important can present a transition regime from a $t sp{1/3}$ growth law to a $t sp{1/2}$ law. The structure factors do not quite scale. They also present a shoulder at high wavevectors, a feature not reported before in metallic alloys.