The ever-increasing demand for high date rate is beyond what is provided by the present wireless and wired access networks. Radio-over-fiber (RoF) technology which can provide broadband wireless access has been considered the most practical and efficient solution. In recent years, RoF with coherent detection has been shown to have better performance than that with direct detection in terms of receiver sensitivity and spectral efficiency. However, RoF with coherent detection suffers from phase noise introduced from both the transmitter and local oscillator (LO) laser sources, which degrades the performance significantly. This study is focused on coherent RoF links for broadband wireless access networks. The thesis consists of four parts. In the first part, a new approach to cancel the phase noise and the unstable frequency difference introduced from the transmitter and LO laser sources based on digital signal processing (DSP) in an RoF link with coherent detection is presented. The proposed schemes rival the RoF link with direct detection in complexity while maintaining a high receiver sensitivity. In addition, a high spectral efficiency coherent RoF link with phase noise cancellation, which can detect both intensity- and phase- modulated signals carried by the same optical carrier, is studied and demonstrated. In the second part, to achieve full-duplex transmission and increase the capacity of an RoF link, radio over wavelength division multiplexing passive optical network (WDM-PON) is studied. To eliminate the requirements of light sources and wavelength management at the optical network units (ONUs), which reduces the cost and eases the installation for a radio over WDM-PON system, a new approach to reuse the downstream wavelength at the ONU with coherent detection and DSP at the optical line terminal (OLT) is investigated. The performance in terms of error vector magnitude (EVM) and bit rate error (BER) is evaluated for both downlink and uplink. In the scheme, the coherent detection improves the receiver sensitivity for the uplink and compensate for the degraded data transmission performance due to the utilization of a wavelength-reused downstream optical signal. Furthermore, since the future internet traffic will become highly symmetric, a symmetrical radio over a colorless WDM passive optical network (PON) with wavelength reuse based on polarization multiplexing and coherent detection is proposed and studied. In the third part, a coherent RoF link based on optical single sideband with no optical carrier (OSSB) modulation with low-cost free-running laser sources for ultra-dense wavelength division multiplexing passive optical networks (UDWDM-PONs) is studied. In a UDWDM-RoF-PON, the channel spacing is very small, thus a WDM filter may not be able to de-multiplex the ultra-dense channels. However, through coherent detection, the channel separation can be realized by using electrical filters at the output of the coherent receiver. In addition, to utilize the spectrum in each channel more efficiently, OSSB modulation is employed. In the proposed scheme, an RoF signal based OSSB modulation with coherent detection is experimentally demonstrated. The channel spacing can be as narrow as 3 GHz. Finally, for 5th generation wireless systems (5G), multi-input and multi-output (MIMO) is a key technology which can multiple the capacity. To seamlessly integrate MIMO into RoF links, it is required that an RoF link can transmit multiple wireless signals over a single wavelength. To enable 4 × 4 MIMO, in the fourth part, an RoF link to transmit four wireless signals with an identical microwave center frequency without using frequency-division multiplexing (FDM) over a single optical wavelength based on optical independent sideband (OISB) modulation and optical orthogonal modulation incorporating optical coherent detection and digital signal processing (DSP) is studied. To increase the spectral efficiency further, a novel high spectral efficiency (20.62 bit/s/Hz) RoF link based on coherent detection and DSP with the spectral efficiency improved by employing both intensity and phase modulation and polarization multiplexing to transmit four microwave signals over a single optical carrier is investigated.

The field of X-ray imaging and X-ray diffraction have been combined in a new technique called Tomographic Energy Dispersive Diffraction Imaging (TEDDI). This diffraction imaging technique allows 3D sample images to be obtained, non-destructively, where each imaged point contains the atomic structural information associated with its diffraction pattern. The main drawback of the TEDDI technique is the long collection times needed to produce the images. In order to overcome this obstacle the rapid TEDDI (rTEDDI) system has been developed at the University of Manchester's Material Science Centre. The research and development of rTEDDI has been the focus of this PhD thesis. A proof of concept for the rTEDDI imaging technique was obtained using thin samples on station 7.6 SRS Daresbury. In this case a first generation array collimator was used in conjunction with an energy resolving Si pixelated detector. Structural information such as lattice parameters, crystal system and phase identiffcation were obtained for metal, polymer and deer antler bone samples. The use of high Z semiconductor detector material was investigated in order to increase the potential of TEDDI for larger and more dense samples. To enable penetration of larger samples high energy X-rays needed to be utilized. In order to detect these higher energies with a good efficiency the detector media was changed from Si to CdZnTe (CZT).The second generation rTEDDI, using CZT as the detection media, was intended to be used under high flux/high energy synchrotron radiation conditions. Testing of the system under these conditions on station 16.3 SRS Daresbury showed an inability to produce diffraction imaging. An in depth investigation into detector and collimator array performance showed a two fold cause. The ERD2004 detector was unable to handle the high countrates experienced during high flux/high energy synchrotron radiation conditions. The MK1.2 collimator array was found to become partially transparent to X-ray energies around the absorption edge of W resulting in the swamping of the diffraction signal under high flux/high energy synchrotron radiation conditions. A new detector Application Specific Integrated Circuit (ASIC) design, developed by the detector division and the Rutherford Appleton Laboratory, and Data Aquisition (DAQ) system, developed by Aspect Systems, as well as a number of new collimator array designs were developed and tested. Testing of the new collimator array structures have shown positive results and the new HEXITECdetector which was designed to be able to handle high countrates, have shown an unprecedented inter pixel uniformity and energy resolution which have been attributed to the ASIC performance and the use of better quality CZT material.

Orientador: Marcus Aloizio Martinez de Aguiar / Tese (doutorado) - Universidade Estadual de Campinas, Instituto de Física Gleb Wataghin / Made available in DSpace on 2018-08-22T04:47:13Z (GMT). No. of bitstreams: 1 Viscondi_ThiagodeFreitas_D.pdf: 5908171 bytes, checksum: 62e83e5e2d7f988db884e3964fd40971 (MD5) Previous issue date: 2013 / Resumo: A propagação semiclássica consiste na elaboração e aplicação de métodos para a resolução aproximada da equação de Schrödinger dependente do tempo, assumindo como hipótese que a ação clássica do sistema possui valor bastante superior à constante de Planck. Dentro deste contexto, o propagador quântico representa um elemento de interesse central, uma vez que esta grandeza corresponde à amplitude de probabilidade para a transição entre dois estados do sistema físico. Em um estágio preliminar de nosso trabalho, empregamos o conceito generalizado de estados coerentes para reformular o propagador quântico em termos de uma integral de caminho. Em seguida, com a utilização do método do ponto de sela, realizamos uma dedução detalhada para a aproximação semiclássica do propagador correspondente a uma ampla classe de grupos dinâmicos. A aplicação deste resultado formal está subordinada à resolução de equações clássicas de movimento sob condições de contorno, considerando um espaço de fase com dimensão duplicada. De maneira geral, a busca por trajetórias clássicas sujeitas a valores de contorno demonstra elevado custo computacional e complexidade técnica. Por esta razão, desenvolvemos três diferentes aproximações semiclássicas determinadas exclusivamente por condições iniciais. Em uma primeira situação, elaboramos um método de propagação constituído por uma integral sobre soluções clássicas no espaço de fase duplicado. No segundo caso, com a formulação do operador semiclássico de evolução temporal, eliminamos a necessidade pela duplicação dos graus de liberdade clássicos. A terceira abordagem, designada por propagador clássico corrigido, está definida pela contribuição de uma única trajetória. Com o propósito de avaliar a precisão e eficiência das expressões semiclássicas adquiridas, exemplificamos a aplicação destas ferramentas teóricas para os estados coerentes de SU(2) e SU(3). Por fim, apresentamos uma extensa discussão sobre as vantagens introduzidas pelo espaço de fase duplicado na implementação de uma aproximação semiclássica. Deste modo, verificamos que soluções clássicas tunelantes possuem uma importante participação na descrição precisa da penetração parcial de um pacote de onda em uma barreira de potencial finita. Além disto, mostramos que o fenômeno quântico de reflexão por um potencial atrativo está diretamente associado à ocorrência de trajetórias com comportamento não-clássico. / Abstract: The semiclassical propagation comprises the development and application of methods for obtaining approximate solutions to the time-dependent Schrödinger equation, assuming the hypothesis that the classical action of the system is much greater than the Planck constant. In this context, the quantum propagator represents an element of central interest, since this quantity corresponds to the probability amplitude for the transition between two states of thephysical system. In a preliminary stage of our work, we employ the generalized concept of coherent states to reformulate the quantum propagator in terms of a path integral. Then, with use of the saddlepoint method, we conduct a detailed derivation of the semiclassical approximation for the propagator corresponding to a wide class of dynamical groups. The application of this formal result depends on the resolution of classical equations of motion under boundary conditions, considering a phase space with doubled dimension. Generally, the search for classical trajectories subject to boundary values demonstrates high computational cost and technical complexity. For this reason, we have developed three distinct semiclassical approximations exclusively determined by initial conditions. In a first situation, we elaborate a propagation method composed of an integral over classical solutions in the doubled phase space. In the second case, with the formulation of the semiclassical time-evolution operator, we eliminate the need for the duplication of the classical degrees of freedom. The third approach, designated as corrected classical propagator, is defined by the contribution of a single trajectory. In order to evaluate the accuracy and efficiency of the obtained semiclassical expressions, we exemplify the application of these theoretical tools for the coherent states of SU(2) and SU(3). At last, we present an extensive discussion on the advantages introduced by the doubled phase space in implementing a semiclassical approximation. In this way, we find that tunneling classical solutions have an important participation in the accurate description of the partial penetration of a wave packet in a finite potential barrier. Furthermore, we show that the quantum phenomenon of reflection by an attractive potential is directly associated to the occurrence of trajectories with non-classical behavior. / Doutorado / Física / Doutor em Ciências

La turbulence axisymétrique est un écoulement bidimensionnel trois-composantes. L’étude de ce type de turbulence est motivée par le fait que celle-ci représente la limite asymptotique des écoulements anisotropes, et qu’elle a été le sujet des investigations théoriques dans le passé. Dans ce manuscrit, la turbulence axisymétrique a étudié en géométrie fermée en utilisant des simulations numériques spectrales et pseudo-spectrales.Études antérieures concernant la génération des structures cohérentes, obtenues dans les écoulements en déclin libre, sont considérées ici dans le contexte des écoulements statistiquement stationnaires, où l’énergie est injectée soit par un forçage spectralement localisé ou par une rotation des disques en haut et en bas du cylindre. On montre que les structures observées sont conformes aux prédictions théoriques.Lorsqu’un protocole de forçage anisotrope est utilisé, une bifurcation est observée entre un état non-tourbillonnant (bidimensionnel deux-composantes, 2D2C) et un écoulement tourbillonnant turbulent (bidimensionnel trois-composante, 2D3C). Cette transition est modélisée à travers un système de deux équations différentielles ordinaires (ODE), et on montre que ce modèle retient la physique essentielle de cette transition. La transition de l’écoulement axisymétrique à un écoulement tridimensionnel (3D3C) est ensuite étudiée à l’aide d’une dimension non-entière, en introduisant de façon continue la variation azimutale dans le système. On montre que la limite 2D2C est singulière et qu’une petite variation azimutale permet une redistribution d’énergie sur les différentes composante énergétiques. Le modèle ODE est adapté pour ce système et on montre que pour l’écoulement considéré la corrélation pression-déformation est responsable d'un niveau approximativement proportionnel à la dimension non-entière. Des Simulations des Grandes Echelles sont réalisées pour évaluer la robustesse des observations à grands nombres de Reynolds. / Axisymmetric turbulence is a two-dimensional three-component flow. The investigation of this type of turbulence is motivated by the fact that it represents the asymptotic limit of anisotropic flows and since it has been the subject of theoretical investigations in the past. In the present manuscript such a flow is investigated in wall-bounded cylindrical geometry using spectral and pseudo-spectral numerical simulations.Previous results on the generation of coherent structures, obtained for freely decaying flow, are here assessed in the context of statistically steady flow, where the energy is supplied by either a spectrally localized forcing, or by moving top and bottom plates of the cylinder. It is shown that the observed structures are consistent with theoretical predictions.When an anisotropic forcing protocol is used, a bifurcation is observed from a non-swirling (two-dimensional two-component, 2D2C) flow to a swirling (two-dimensional three-component 2D3C) turbulent flow. This transition is modelled by a system of two ordinary differential equations (ODE), and it is shown that this model retains the essential physics of the transition.The transition of the axisymmetric flow to three-dimensional (3D3C) flow is then studied using non-integer dimension, by smoothly introducing azimuthal variation into the system. It is shown that the 2D2C limit is singular and that small azimuthal variation allows a redistribution of energy over the different energy components. The ODE model is adapted for this system and it is shown that for the considered flow the pressure-strain correlation is responsible for a swirl-level approximately proportional to the non-integer dimension. Large-Eddy Simulations are carried out to assess the robustness of the observations at higher Reynolds number.

River confluences - locations where rivers join one another - are fundamental components of natural drainage networks. Differences in topography, geology, soils, land use, and human activities within watersheds upstream of confluences can produce differences in thermal or chemical properties of river flows and in the materials transported by these flows. Mixing is initiated along the mixing interface (MI) that develops between the two incoming streams with different properties. Therefore, the understanding of fluvial processes at confluences is important for determining river mixing both at and downstream of individual confluences and at the scale of drainage networks. The primary goal of this thesis is to describe the main mechanisms that control mixing and transport at river confluences and the role played by the complex flow structures in the flow and how they change with planform geometry and other flow and geometrical parameters. The study is carried out using Computational Fluid Dynamics modeling based on the state of the art Detached Eddy Simulation approach and High Performance Computing. By starting with a mixing layer between parallel streams with simple geometry, the model is validated based on laboratory experiment data. Moreover, some hypotheses regarding the growth of the mixing layer are amended with the extensive data provided by the model, which is a valuable supplement to the experiment. By performing a detailed parametric study in very long and wide domains for simplified cases one can focus on the spatial development of the MI and the large scale coherent structures forming within and in the vicinity of the MI without the complications of other factors. More specifically, the influence of velocity and density difference of the two streams, flow depth, inflow conditions and angles between the two streams on the spatial development of the MI is analyzed. The data resulting from these simulations conducted in simple geometries is a unique set of data which can be used to test and improve theoretical models used to predict global parameters describing flow and mixing at natural river confluences. In particular, this research uses for the first time well resolved Large Eddy Simulation based techniques to investigate how density differences between the incoming streams affect the spatial development of the mixing interface and mixing downstream of the confluence apex. In order to investigate flow dynamics, mixing processes and effects of temperature stratifications at natural river confluences with discordant bed, a series of simulations is performed for the confluence of the Ebro and Segre Rivers in Spain, which is one of the most studied confluences in Europe. With the detailed survey data of the confluence bed and flow conditions data provided, the goal is to understand the main mechanisms responsible for mixing at a confluence with a large bed discordance and how the velocity ratio between the two incoming streams affects mixing. Besides, more insights are provided that if temperature stratification effects affect significantly flow structure and mixing based on real conditions recorded at a natural confluence. The study provides a comprehensive set of flow data in the confluence including velocity, temperature distribution etc. It serves as important supplement to the field measurements, which are generally more difficult to obtain. It also allows estimating scale effects between field conditions and conditions at which laboratory experiments of confluence flow and mixing are conducted.

Ce mémoire est dédié à l’imagerie par diffraction cohérente. Dans un premier temps, nous présentons la conception et à la mise en oeuvre expérimentale d’un système d’imagerie compact fonctionnant sur ce principe. Il est composé d’unediode UV (λ = 400 nm), d’une caméra CCD, et d’une plate-forme pour placer l’échantillon à observer. Le faisceau cohérent issu de la diode éclaire l’échantillon, et la figure de diffraction est enregistrée par la caméra. La rétro-propagation du champ détecté permet, en principe, de déterminer le profil de l’échantillon. Néanmoins, la phase du champ, perdue lors de la détection, nous contraint à employer desméthodes de « reconstruction de la phase », cette quantité étant nécessaire à l’opération d’inversion. Plusieurs techniques ont été utilisées. L’holographie par Transformée de Fourier, par exemple, est une méthode déterministe qui consiste à utiliser une référence circulaire (ou rectangulaire) gravée à côté de l’échantillon. La phase est encodée dans la figure de diffraction, sous la forme de franges d’interférences issues de l’objet et de la référence. Une simple Transformée de Fourier du signal permet alors de retrouver le profil de l’échantillon. Uneméthode itérative a également été mise en oeuvre, basée sur un jeu de contraintes dans les espaces réel et réciproque. En particulier, l’objet éclairé doit être « isolé », i.e. plus petit que le faisceau incident. Bien que cette méthode soit non-déterministe, nous verrons toutefois qu’elle est plus robuste et permet d’obtenir de meilleures résolutions spatiales qu’en holographie. Cette étude est un point de départ à l’observation d’objets tridimensionnels. Nous présentons une première méthode déterministe, basée sur l’holographie par Transformée de Fourier. Pour ce faire, une « pupille holographique » est utilisée et sert de support à une première reconstruction 2D du champ. Celui-ci est ensuite rétro-propagé vers l’échantillon placé à proximité, afin de réaliser une mise au point entièrement numérique de ce dernier. La contrainte « d’isolation » de l’objet est alors levée par l’utilisation de cette pupille. Avec cette méthode, le champ latéral est toutefoislimité par le diamètre de la pupille. Pour l’observation d’échantillons plus larges, la technique d’holographie « en ligne » a également été exploitée. Elle consiste à éclairer l’objet avec une onde sphérique et à enregistrer les franges d’interférences (ou « hologramme »). Une rétro-propagation est ensuite effectuée pour faire la mise au point sur l’échantillon. Le caractère divergent du faisceau permet de disposer d’un champ latéral de plusieurs millimètres. Le problème de « l’image jumelle », inhérent à cette configuration, est résolu via unalgorithme itératif couplé à la rétro-propagation. Des reconstructions tridimensionnelles ont été effectuées sur divers échantillons, avec cesdeux méthodes — reconstruction pupillaire et holographie en ligne. Pour chacune d’entre elles, des interfaces de reconstruction ont été mises au point et fonctionnent pendant la détection, afin d’observer l’objet en temps réel. Nous disposons alors d’un prototype d’imagerie sans lentille compact et complet. Enfin, nous présentons l’application d’une technique de reconstruction de la phase, appelée LIFT (pour LInearized Focal plane Technique), appliquée à un analyseur de front d’onde Shack-Hartmann. Usuellement, de tels capteurs ont une résolution spatiale limitée par le pas des micro-lentilles : seules les pentes locales (tip/tilt) sont déterminées. Le LIFT consiste à déterminer la phase à l’échelle de chaque micro-lentille, en exploitant le profil du spot correspondant. Des matrices d’interaction sont calculées afin de linéariser la relation entre l’espace réel (micro-lentilles) et l’espace réciproque (matrice CCD), et une boucle itérative permet d’étendre cedomaine de linéarité. Un gain de résolution spatiale de l’ordre de 3, au niveau de chaque micro-lentille, est attendu avec cette technique. / This dissertation is dedicated to coherent diffractive imaging. Firstly, we present the conception and experimental implementation of a compact imaging system, working on this principle. It is made of an UV laser diode (λ = 400 nm), a CCD camera,and a platform to place the sample. The coherent beam coming from the diode illuminates the sample, and the diffraction pattern is recorded by the camera. Back-propagating the detected field should allow, in principle, to derive the sample’s profile. Nevertheless, the field’s phase, lost during the detection, forces us to use “phase retrieval” methods, this quantity being necessary to the inversion process. Several techniques have been used for that purpose. Fourier Transform Holography (FTH), for example, is a deterministic method thatconsists in using a circular reference, closely drilled nearby the sample. The phase is encoded in the diffraction pattern, in the form of interference fringes coming from the object and the reference. Then, a simple inverse Fourier Transform of the signal leads the profile of the sample. An iterative method has also been implemented, based on a set of constraints in the real and reciprocal spaces. In particular, the illuminated object must be “isolated”, i.e. smaller than the incident beam. Although this method is non-deterministic, we will see thatit is more robust and gives better resolutions than the holographic cases. This study is the starting point of three-dimensional imaging. We present a first deterministic method, based on FTH. For this purpose, a “holographic pupil” is used and serves as a support for a first 2D reconstruction of the field. The latter is then back-propagated towards the sample closely placed, in order to realize an entirely numerical focusing on it. The “isolation constraint” is then removed by the use of this pupil. However, with this method, the field of view is limitedby the pupil’s diameter. In order to observe larger samples, the “in-line holography” technique has been exploited as well. It consists in illuminating the object with a spherical wave and recording the interference fringes (or “hologram”). A back-propagation is made after the fact in order to do the focusing on the sample. The divergent nature of the beam allows for reaching several millimeters for the lateral field of view. The “twin image problem”, inherent to this configuration, is solved via an iterative algorithm coupled to the back-propagation process. Three-dimensional reconstructions have been made on varied samples, with these two methods — pupil reconstruction and in-line holography. In both cases, reconstruction interfaces have been implemented and work during the detection, in order to observe the object in real time. We then have a compact and complete lens-less imaging prototype. Finally, we present the application of a phase retrievaltechnique, named LIFT (LInearized Focal plane Technique), applied to a Shack-Hartmann wavefront sensor. Usually, such sensors have a spatial resolution that is limited by the micro-lenses size : only the local slopes, i.e. tip and tilt, are retrieved. The LIFT consists in determining the phase at the scale of each micro-lens, by exploiting the corresponding spot profile. Interaction matrices are calculated in order to linearize the relation between the real space (micro-lenses) and the reciprocal space (CCD chip), and an iterative loop allows for increasing this linearity domain. With this technique, a gain in spatial resolution by a factor 3 is expected.

Dans cette thèse, nous étudions les effets cohérents associés à la propagation d’ondes dans les milieux diffusants, en particulier les ondes électromagnétiques.En milieux faiblement désordonnés, l'intensité lumineuse fluctue spatialement sur de grandes distances. Ce phénomène est le résultat d'effets cohérents mésoscopiques complexes, qui se produisent à une échelle microscopique. Nous montrons que ces fluctuations mésoscopiques cohérentes de la lumière induisent des forces de rayonnement d'un nouveau genre. L'amplitude de ces forces fluctuantes est déterminée par un paramètre unique et facilement réglable, la conductance adimensionnée, qui dépend à la fois de la géométrie et des propriétés de diffusion du milieu. Notre découverte devrait donc avoir des applications intéressantes, telles que de nouveaux capteurs pour la matière molle ou la biophysique.Du point de vue méthodologique, nous utilisons une approche à la Langevin pour décrire les fluctuations lumineuses cohérentes, où un bruit précisément calculé rend compte des effets cohérents mésoscopiques. Nous montrons comment inclure systématiquement les corrections cohérentes dans le terme de bruit, afin de reproduire les fluctuations d'intensité. Cette description permet de comprendre les fluctuations cohérentes comme résultant d’un flux lumineux hors équilibre, caractérisé par deux paramètres seulement, le coefficient de diffusion et la mobilité, qui sont par ailleurs liés par une relation d’Einstein. Un avantage évident de cette méthode est sa dépendance à deux paramètres seulement, ce qui fournit une description à la fois compacte et précise des riches effets cohérents sous-jacents. De plus, la correspondance que nous présentons entre la lumière cohérente et l'hydrodynamique hors d'équilibre est facilement généralisable à une large classe de problèmes d'ondes quantiques ou classiques.Pour les perspectives futures, cette connexion entre les effets cohérents mésoscopiques et les processus stochastiques hors équilibre devraient intéresser les communautés de la mésoscopie et de la mécanique statistique. Pour les premiers, le lien avec l'hydrodynamique hors équilibre fournit un nouvel éclairage sur la physique mésoscopique, ainsi que des outils utiles pour étudier les quantités jusqu'ici difficiles d'accès, telles que les fonctions de corrélation d'intensité d'ordres supérieurs. Pour les seconds, ces travaux devraient motiver une étude plus approfondie des processus indépendants du temps inspirés de la mésoscopie. / In this work, we study coherent effects associated to wave propagation in scattering media, in particular electromagnetic waves.In weakly disordered media, light intensity fluctuates spatially over large distances. This phenomenon is the result of complex mesoscopic coherent effects, which occur at a microscopic scale. We show that these mesoscopic coherent fluctuations of light induce radiation forces of a new kind. The strength of these fluctuating forces is determined by a single and easily tunable parameter, the dimensionless conductance, which depends on both the geometry and the scattering properties of the medium. Our findings should therefore have interesting applications such as new sensors in soft condensed matter or biophysics.On the methodological viewpoint, we use a hydrodynamic Langevin approach to describe the coherent light fluctuations, where a properly tailored noise accounts for mesoscopic coherent effects. We show how to systematically include the coherent corrections in the noise term, in order to reproduce the intensity fluctuations. This description allows to understand coherent light fluctuations as resulting from a non equilibrium light flow, characterized by two parameters only, the diffusion coefficient and the mobility, otherwise related by an Einstein relation. A clear asset of this method is its dependence upon two parameters only, which provides a compact yet accurate description of the rich underlying coherent effects. Moreover, the mapping we present between coherent light and out of equilibrium hydrodynamics is easily generalizable to a large class of quantum or classical wave problems.For future perspectives, this connection between coherent effects in mesoscopics and non equilibrium stochastic processes should be of interest in both the mesoscopics and statistical mechanics communities. For the former, the mapping to non equilibrium hydrodynamics provides a new insight to mesoscopic physics as well as useful tools to study quantities so far difficult to access, such as higher orders intensity correlation functions. For the latter, this work should motivate further study of time independent processes inspired from mesoscopics.

Real-time coherent Grazing-Incidence Small-Angle X-ray Scattering was used to investigate the average kinetics and the fluctuation dynamics during self-organized ion beam nano-patterning of two semiconductor surfaces: silicon at room temperature and germanium heated above its recrystallation temperature. For silicon nano-patterning, initially flat samples at room temperature were bombarded by a broad collimated beam of 1keV Ar+ and Kr+ ions at 65° polar angle, leading to the amorphization of the ion-irradiated surfaces and the spontaneous formation of nanoscale ripples. The temporal evolution of the average X-ray scattering intensity shows the evolution of average kinetics, while the fluctuation dynamics can be investigated by correlation of X-ray speckles. The surface behavior at early times can be explained within a linear theory framework. The transition away from the linear theory behavior is observed in the dynamics since the intensity correlation function quickly evolves into a compressed exponential decay on length scales corresponding to the peak wavelength and a stretched exponential decay on shorter length scales. The correlation times for silicon nano-patterning are maximum at the ripple wavelengths while they are smaller at other wavelengths. This has notable similarities and differences with the phenomenon of de Gennes narrowing. Overall, this dynamics behavior is found to be consistent with the simulations of a nonlinear growth model by Harrison et al. Following the formation of self-organized nano-ripples, they move across the surface. Homodyne X-ray alone cannot detect the motion, but because of the gradient of ion flux across the sample, we were able to measure in-situ the corresponding ripple velocity gradient by cross-correlating speckles and tracking their movements. For germanium nano-patterning at an elevated temperature, flat germanium samples kept at 300°C were bombarded by 1keV Ar+ ions at normal incidence. Unlike the case when surfaces are amorphizated during room temperature bombardment, the crystalline nano-pattern formation occurs mainly due to a surface instability caused by the Ehrlich-Schwoebel barrier. By using a linear theory analysis on the X-ray scattering intensities in the early times, we measured the contribution of the Ehrlich-Schwoebel barrier to the crystalline nano-patterning kinetics.

[ES] Los sistemas de transmisión óptica no-coherente se emplean actualmente en las redes ópticas de corto alcance (< 80 km), como son las redes de ámbito metropolitano. La implementación más común en el estado del arte se basa en sistemas que emplean multiplexación por división en longitud de onda (WDM, wavelength division multiplexing) de cuatro longitudes de onda (¿) proporcionando un régimen binario de 100 Gbps (4¿×25 Gbps). En los últimos años, los sistemas de transmisión ópticos no-coherentes están evolucionando desde 100 Gbps a 400 Gbps (4¿×100 Gbps). Dado que este mercado comprende un gran número de sistemas, el coste es un parámetro importante que debe ser lo más bajo posible. El objetivo de esta tesis es investigar distintos aspectos del procesado de señal en general y, específicamente, investigar nuevas técnicas de procesado digital de señal (DSP, digital signal processing) que puedan ser utilizadas en sistemas de transmisión óptica no-coherentes empleando la modulación por amplitud de pulsos (PAM, pulse-amplitude modulation). Para que una técnica DSP sea interesante en el contexto de una red óptica WDM no-coherente, esta debe mitigar de manera efectiva al menos una de las tres limitaciones principales que afectan a estos sistemas: limitaciones de ancho de banda, limitaciones por dispersión cromática (CD), y el ruido. En esta tesis se proponen y examinan una serie de algoritmos cuyo su rendimiento es analizado mediante simulación y experimentalmente en laboratorio: - Feed-forward equalizer (FFE): este es el esquema de ecualización más común que se emplea principalmente en las transmisiones ópticas no-coherentes de alto régimen binario. Puede compensar grandes limitaciones en el ancho de banda. - Estimación de la secuencia de máxima verosimilitud (MLSE): el MLSE es un detector óptimo y, por lo tanto, proporciona las mejores prestaciones en detección cuando se abordan las limitaciones por CD y de ancho de banda. - Conformación geométrica de la constelación: en los esquemas de modulación de intensidad óptica multinivel, la distancia entre los niveles de amplitud puede ajustarse adecuadamente (de manera que no son equidistantes) a fin de aumentar la tolerancia de la señal frente al ruido. - Conformación probabilística: técnica diseñada específicamente para esquemas de modulación multinivel. Esta técnica ajusta la probabilidad de cada nivel de amplitud de modo que se incrementa la tolerancia al ruido óptico. - Señalización de respuesta parcial (PRS, partial signaling response): este es un enfoque basado en DSP donde una interferencia entre símbolos (ISI, inter-symbol interference) controlada es introducida intencionalmente de tal manera que la señal resultante requiere menos ancho de banda. La técnica PRS puede adaptarse para combatir también el efecto de CD. - Pre-énfasis digital (DPE, digital pre-emphasis): esta técnica consiste en aplicar el inverso de la función de transferencia del sistema a la señal en el transmisor, lo que reduce el impacto de las limitaciones de ancho de banda en el receptor. - Modulación con codificación Trellis (TCM, Trellis-coded modulation): esquema de modulación que combina elementos de corrección de errores (FEC, forward error correction) con técnicas de partición en conjuntos y modulación multidimensional para generar una señal más resistente al ruido. - Modulación multidimensional por partición en conjuntos: muy similar a TCM, pero sin ningún elemento FEC. Tiene menos ganancias que TCM en términos de tolerancia al ruido, pero no es tan sensible al ISI. Utilizando estas técnicas, esta tesis demuestra que es posible lograr una transmisión óptica con régimen binario de 100 Gbps/¿ empleando componentes de bajo coste. En esta tesis también demuestra regímenes binarios de más de 200 Gbps, lo que indica que la transmisión óptica no-coherente con modulación PAM puede ser una solución viable y eficiente en coste / [CAT] Actualment, s'utilitzen sistemes òptics no coherents en xarxes òptiques de curt abast ( < 80 km), com són les xarxes d'àmbit metropolità. La implementació més comuna que podem trobar en l'estat de l'art es correspon amb sistemes emplenant multiplexació per divisió en longitud d'ona (WDM, wavelength division multiplexing) de quatre longituds d'ona (¿) proporcionant un règim binari de 100 Gbps (4¿×25 Gbps). En els últims anys, els sistemes de transmissió òptica no-coherents han evolucionat des de 100 Gbps cap a 400 Gbps (100 Gbps/¿). Atès que el mercat de sistemes de curt abast compren un gran volum de dispositius òptics instal·lats, el cost unitari és molt important i ha de ser el més baix possible. L'objectiu d'aquesta tesi és analitzar aspectes del processament de senyal en general i, específicament, investigar noves tècniques de processament digital de senyal (DSP, digital signal processing) que puguen ser utilitzades en sistemes de transmissió òptica no-coherent que utilitzen la modulació per amplitud d'impulsos (PAM, pulse-amplitude modulation). Per tal que una tècnica DSP es considere interessant per a una xarxa òptica WDM no-coherent, aquesta ha de mitigar efectivament almenys una de les tres principals limitacions que afecten aquests sistemes: limitacions d'ample de banda, limitacions per dispersió cromàtica (CD), i el soroll. En aquesta tesi s'examinen una sèrie d'algoritmes, el seu rendiment s'analitza per simulació i experimentalment en laboratori: - Feed-forward equalizer (FFE): aquest és l'esquema d'equalització més comú i s'utilitza bàsicament en les transmissions òptiques no coherents d'alt règim binari. Pot compensar grans quantitats de limitacions d'ample de banda. - Estimació de la seqüència de probabilitat màxima (MLSE): el MLSE és un detector òptim i, per tant, proporciona el millor rendiment quan es tracta de limitacions d'ample de banda i de CD. - Conformació geomètrica de la constel·lació: en esquemes de modulació òptica d'intensitat multinivell es pot ajustar la distància entre els nivells d'amplitud (de manera que ja no són equidistants) per augmentar la tolerància del senyal al soroll. - Conformació probabilística: una tècnica dissenyada específicament per als esquemes de modulació multinivell; ajusta la probabilitat de cada nivell d'amplitud de manera que augmenta la tolerància al soroll òptic. - Senyalització de resposta parcial (PRS, partial signaling response): és un enfocament basat en DSP on la interferència entre símbols (ISI, inter-symbol interference) controlada s'introdueix intencionalment de manera que el senyal resultant requereix menys ample de banda. La tècnica PRS es pot adaptar per combatre els efectes del CD. - Pre-èmfasi digital (DPE, digital pre-emphasis): aquesta tècnica consisteix a aplicar la inversió de la funció de transferència del sistema a la senyal en el transmissor de manera que es redueix l'impacte de les limitacions d'ample de banda en la senyal en el receptor. - Modulació amb codificació Trellis (TCM, Trellis-coded modulation): esquema de modulació que combina els elements de correcció d'errors avançats (FEC, forward error correction) amb tècniques de partionament de conjunts i modulació multidimensional per generar un senyal més resistent al soroll. - Modulació multidimensional per partició en conjuntes: molt similar a TCM però sense elements FEC. Té guanys menors que TCM en termes de tolerància al soroll, però no és tan sensible a l'ISI. Mitjançant l'ús d'aquestes tècniques, aquesta tesi demostra que és possible aconseguir una transmissió òptica amb un règim binari de 100 Gbps/¿ utilitzant components de baix cost. Esta tesi també demostra règims binaris de més de 200 Gbps, el que indica que la tecnologia no-coherent amb modulació PAM és una solució viable i eficient en cost per a una nova generació de sistemes transceptors òptics WDM funcionant a 800 Gbps (4¿×200 G / [EN] Non-coherent optical transmission systems are currently employed in short-reach optical networks (reach shorter than 80 km), like metro networks. The most common implementation in the state-of-the-art is the four wavelength (¿) 100 Gbps (4¿×25 Gbps) wavelength division multiplexing (WDM) transceiver. In recent years non-coherent optical transmissions are evolving from 100 Gbps to 400 Gbps (4¿×100 Gbps). Since in the short-reach market the volume of optical devices being deployed is very large, the cost-per-unit of the devices is very important, and it should be as low as possible. The goal of this thesis is to investigate some general signal processing aspects and, specifically, digital signal processing (DSP) techniques required in non-coherent pulse-amplitude modulation (PAM) optical transmission, and also to investigate novel algorithms which could be applied to this application scenario. In order for a DSP technique to be considered an interesting solution for non-coherent WDM optical networks it has to effectively mitigate at least one of the three main impairments affecting such systems: bandwidth limitations, chromatic dispersion (CD) and noise (in optical or electrical domain). A series of algorithms are proposed and examined in this thesis, and their performance is analyzed by simulation and also experimentally in the laboratory: - Feed-forward equalization (FFE): this is the most common equalizer and it is basically employed in every high-speed non-coherent optical transmission. It can compensate high bandwidth limitations. - Maximum likelihood sequence estimation (MLSE): the MLSE is the optimum detector and thus provides the best performance when it comes to dealing with CD and bandwidth limitations. - Geometrical constellation shaping: in multilevel optical intensity modulation schemes the distance between amplitude levels can be adjusted (such that they are no longer equidistant) in order to increase the signal's tolerance to noise. - Probabilistic shaping: another technique designed specifically for multilevel modulation schemes; it adjusts the probability of each amplitude level such that the tolerance to optical noise is increased. - Partial response signaling (PRS): this is a DSP-based approach where a controlled inter-symbol interference (ISI) is intentionally introduced in such a way that the resulting signal requires less bandwidth. PRS can be customized to also mitigate CD impairment, effectively increasing transmission distances up to three times. - Digital pre-emphasis (DPE): this technique consists in applying the inverse of the transfer function of the system to the signal at the transmitter side which reduces the impact of bandwidth limitations on the signal at the receiver side. - Trellis-coded modulation (TCM): a modulation scheme that combines forward error correction (FEC) elements with set-partitioning techniques and multidimensional modulation to generate a signal that is more resistant to noise. - Multidimensional set-partitioned modulation: very similar with TCM but without any FEC elements. It has lower gains than TCM in terms of noise tolerance but is not so sensitive to ISI. By using the techniques enumerated above, this thesis demonstrates that is possible to achieve 100 Gbps/¿ optical transmission bitrate employing cost-effective components. Even more, bitrates higher than 200 Gbps are also demonstrated, indicating that non-coherent PAM is a viable cost-effective solution for next-generation 800 Gbps (4¿×200 Gbps) WDM transceivers. / Prodaniuc, C. (2019). Advanced Signal Processing for Pulse-Amplitude Modulation Optical Transmission Systems [Tesis doctoral no publicada]. Universitat Politècnica de València. https://doi.org/10.4995/Thesis/10251/117315 / TESIS

This diploma thesis deals with the algorithms for signal processing in a field of a low-coherence interferometry. The introductory part is devoted to the basic principle of the low-coherence interferometry and to presenting its field of applications. The second part contains a list of selected algorithms supplemented by a description of their functionality. The procedure of experimental measurement of known surface with Michelson interferometer is described in the third part and the results obtained for particular algorithms are presented. Finally, algorithms are compared on the basis of measurement results.