The exploitation of the superprism phenomenon for optical demultiplexing using a slab photonic crystal on the silicon on insulator platform is the main subject of this thesis. The S-vector and k-vector superprisms are considered. Design equations for the S-vector superprism demultiplexer which fully take into account the nonlinear spectral dependence of beam propagation and dispersion are introduced. This allows wide-band coarse wavelength division multiplexing (CWDM) demultiplexers to be designed. Selecting minimum prism area as a metric, the best photonic crystal lattice, design parameters and prism geometry is sought. A full 3-D modeling approach using the plane wave expansion method is employed to ensure the practicality of the design. We show that the slab 1-D photonic crystal can provide the smallest superprism. Based on our result, an area of 1367 mum2 is sufficient to resolve 4 standard CWDM channels (20nm channel spacing). We extend this approach by proposing a stratified photonic crystal which has 5 times less area for an 8 channel CWDM design. / We then propose the first fully integrated k-vector superprism layout. Design rules and equations are presented and we use these to obtain the design parameters that result in a minimum prism area. We show that an optimized 1-D photonic crystal k-vector superprism with the area of less than 0.1 mm2 is sufficient to resolve 32 standard dense wavelength division multiplexing (DWDM) channels (100GHz channel spacing). The resulting chip size is approximately 4.5 times less than an equivalent etched grating demultiplexer. / We also demonstrate that fast lenses can be made using slab 1-D photonic crystal with an periodicity. / We introduce an analytical approximation technique for slab 1-D photonic crystals based on the weighted index method. The variational nature of the method leads to acceptable results for moderate refractive index contrast materials. The method can also be extended to 2-D cases and to nonlinear systems. / The plane wave expansion (PWE) method and field matching have been combined to obtain a new method which is capable of obtaining all types of modes including the leaky modes of slab 1-D photonic crystals. The method requires fewer plane waves than the conventional PWE method but provides a better approximation. We compare our results with an accurate finite element method as a benchmark. / A report of our first attempt for the fabrication, post-possessing and optical characterization of the proposed k-vector superprism demultiplexer is also presented. We recommend the development of a cladding, and more accurate fabrication procedures for future investigations.

This dissertation presents both detailed experimental and extensive theoretical studies of optical wavemixing in nonlinear absorptive Kerr media. Non-degenerate two wave mixing (NDTWM) is a simple and powerful technique widely used to study the nonlinear refractive index and the grating decay time. It is unique due to its ability to separate the phase and the absorptive grating contributions. The work included in this thesis has evolved from trying to explain the unexplained results on the symmetric component of the NDTWM gain reported as "anomalous" behavior in ruby. We note that all previous theories for NDTWM have taken the approximation that the strong pump beam intensity was constant. In this approximation, the origin of unsymmetry in energy exchange between the two interacting beams was due only to the absorption grating of the weak probe beam. / We have shown both experimentally and theoretically that the contribution of the absorption grating can be neglected but not the nonlinear bias absorption--just opposite to what had been common practice. Our approach also accounts for a range of inconsistencies related to intensity dependence of both the NDTWM gain and ratio of the imaginary to the real part of the nonlinear refractive index. Weakening the probe beam, earlier believed to improve accuracy of the approximation of the constant pump beam, actually destroys the symmetry of the energy exchange between the beams, enhancing the weak probe beam at the expense of the pump beam due to nonlinear absorption. / This same two-beam coupling analysis has also been applied to interpret non-degenerate four wave mixing. Amplification of the phase conjugate signal is thus obtained. The spectral response in the frequency domain results in a tunable notch filter which can be controlled by the incident intensities of the pump beams.

The advances in nanoscience and nanotechnology in recent decades have renewed the interests in the optical properties of metals. Today, the field known as Plasmonics studies the control and manipulation of the electromagnetic near-fields of metallic nanostructures in order to realize novel subwavelength optical applications. In particular, this thesis explores the phenomenon of plasmon resonance for molecular sensing. Surface plasmon resonance (SPR) on flat metal surfaces is used for index-of-refraction sensing and localized surface plasmon resonance (LSPR) on metal nanospheres can produce surface-enhanced Raman scattering (SERS). The operation principles and the experimental evaluation of two SPR sensing devices are presented. An integrated sensor and a 2D wavelength-angle modulated version were estimated to have an angular sensitivity of 126°/RIU and 91°/RIU, respectively. Furthermore, through an implementation of a full vector multiple-multipole light scattering method, useful for the calculations of the field focusing efficiency between assemblies of metal nanospheres, we showed that optical frequency electric fields can be enhanced in gold nanoparticle assemblies by an order of 450 in nanometer volumes. Keywords: surface plasmons, plasmonics, optical biosensors, surface-enhanced Raman scattering, optics of metals, nanophotonics, nanomaterials, classical electrodynamics / L'avancée des nanosciences et de la nanotechnologie au cours des dernières décennies a suscité un renouvellement de l'intérêt pour les propriétés optiques des métaux. Aujourd'hui, la Plasmonique cherche à contrôler les champs proches électromagnétiques des nanostructures métalliques afin de bénéficier des nouvelles applications reposant sur l'optique de sous-longueur d'onde. En particulier, ce mémoire explore l'utilisation de la résonance à plasmons pour la capture des molécules. L'étude se divise en la résonance à plasmons de surface (SPR) pour des surfaces métalliques planes et pour des nanosphères métalliques. Ces deux méthodes permettent de créer des capteurs sensibles aux variations d'indice de réfraction et d'autres qui reposent sur des effets non-linéaires tels que la diffusion Raman exaltée de surface (SERS). Suite à l'introduction des bases, l'opération ainsi que des résultats expérimentaux de deux capteurs à plasmons de surface sont présentés. Un capteur intégré et une version 2D avec modulation de longueur d'onde et de l'angle possèdent une sensibilité angulaire d'environ 126°/RIU et 91°/RIU, respectivement. Par la suite, la réalisation de la méthode du multiple-multipole, utile pour évaluer l'efficacité de la concentration des champs entre des nanosphères métalliques, est discutée. Une amélioration de la concentration de champ optique de l'ordre de 450 par des nanospheres d'or est présentée. Mots-clés: plasmon de surface, plasmonique, biocapteurs optiques, diffusion Raman exaltée de surface, optique des métaux, nanophotonique, nanomatériel, électro-dynamique classique

Presented in this thesis is a novel systematic approach to determine the general characteristics and stability of an EDFL, and a thorough experimental investigation of a mode-locked dual-wavelength EDFL. / The novel method to characterize an EDFL consists of performing an extensive set of measurements, and from these measurements a set of parameters is calculated. These parameters quantifiably describe the laser in terms of its temporal and spectral characteristics, signal power and wavelength variations, and number of occurrences for which the laser ceases its multi-wavelength operation. / An experimental investigation of a free-running mode-locked dual-wavelength EDFL with different signal wavelength spacing and pulse chirps is performed with the characterization method proposed in this thesis. / The original conclusions that are drawn from the experimental investigation of an EDFL are that the stability of the EDFL exponentially decreases as a function of the signal wavelength spacing, and the stability of an EDFL is improved with linearly chirped FBGs.

This thesis consists of a series of theoretical contributions to algorithmic methods for phase retrieval from diffraction intensity measurements. Our work is presented in three related, but somewhat independent parts. / The first part consists of a refinement to the phase propagation methods developed by Bates et al. in 1982. We introduce a correction term for the calculation of the phase difference between actual samples in reciprocal space. We show, numerically, how our method leads to improved image reconstructions in 1D. / In the second part we develop an algorithm for phase retrieval based on the Fourier series expansion of a sharp, square object support. We obtain a series of equations that describe the dependence between different points in reciprocal space, and show that this dependence becomes simpler when only a few terms of the expansion are taken into account. Our algorithm consists of two stages: In the first one, a few of the coupled equations are solved in order to obtain the phase within a localized region. This is followed by a propagation stage in which the rest of the unknown phase values are obtained by means of simple propagation method. We present a numerical example in which we use a downhill minimization method to solve the equations that arise on the first stage. / Finally, we propose a strategy for phase retrieval from x-ray diffraction measurements of a system undergoing the kinetics of a first-order transition following a temperature quench. We use both, a simple theoretical model and numerical simulations to obtain an expression for the average phase-decorrelation time in ordering dynamics. We present an example to show how this result can be used to solve the phase problem faster and with higher convergence rates. / Cette thèse consiste en une série de contributions théoriques à des méthodes algorithmiques pour la récupération de phase à partir de mesures d'intensités de diffraction. Notre travail est présenté en trois parties liées mais indépendantes. / La première partie consiste en un raffinement des méthodes de propagation de phase développée par Bates et al. en 1982. Nous introduisons un terme de correction pour le calcul de la différence de phase entre des échantillons réels dans l'espace réciproque. Nous montrons, numériquement, la façon dont notre méthode mène à des reconstructions d'image améliorées en 1D. / Dans la seconde partie, nous développons un algorithme pour la récupération de phase basé sur le dévelopement en série de Fourier d'un support carré. Nous obtenons une série d'équations qui décrivent la dépendance entre les différents points dans l'espace réciproque, et nous démontrons que cette dépendance devient plus simple lorsque seuls quelques termes de l'expansion sont pris en compte. Notre algorithme est composé de deux étapes: dans la première, quelques-unes des équations couplées sont résolues en vue d'obtenir la phase dans une région localisée. Elle est suivie par une étape de propagation dans laquelle le reste des valeurs de phase inconnues sont obtenues au moyen de la méthode de propagation simple. Nous présentons un exemple numérique dans lequel nous utilisons une méthode de minimisation de descente pour résoudre les équations qui se posent sur la première étape. / Enfin, nous proposons une stratégie pour la récupération de la phase des mesures de diffusion de rayons x d'un système subissant la cinétique d'une transition de premier ordre, après une trempe. Nous utilisons un modèle théorique simple et des simulations numériques afin d'obtenir une expression du temps moyen de décorrélation de phase dans la dynamique d'ordre. Nous présentons un exemple pour montrer comment ce résultat peut être utilisé pour résoudre le problème de la phase plus vite et avec des taux plus élevés de convergence.

This document presents the design, fabrication, characterization and packaging of a stand-alone all-photonic 1X4 switch. All of the designs pertaining to the implementation of the device including those of required custom components are given. The core of the switch is comprised of a "2-bit" electro-optic deflector capable of sub-microsecond deflection times. The packaged switch operates at 1100V and boasts a worst-case fiber-to-fiber insertion loss and crosstalk reading of 5.3dB and -23dB, respectively. Furthermore, a worst-case deflection time of 86ns has been achieved.

No description available.

No description available.

Optical interconnects represent an attractive alternative technology for the implementation of dense, high-speed interconnects, as they do not suffer from many of the problems plaguing electrical interconnects such as frequency-dependent crosstalk and attenuation. / However, optics has still not been accepted commercially as an interconnect technology. There is concern regarding the cost and complexity of the optomechanics needed to achieve the very fine alignments necessary to guarantee that the light emitted from the source actually falls on the receiver. The demonstration of a simple-to-assemble, dense and robust optical interconnect would constitute an important proof of the practicality of this technology. The photonic backplane demonstrator system presented in this thesis addresses these issues through a novel approach; the system uses slow Gaussian beams (f/16) and a clustered design to maximize misalignment tolerances. This in turn relaxes the positioning and packaging requirements for the components, thus simplifying assembly. / This thesis pursues two sets of complementary goals; the first set is concerned with the demonstration of some desirable optomechanical characteristics for optical interconnects such as passive alignment, repeatability and stability while the second set of goals is concerned with a verification of hypotheses often used in the design and implementation of optical interconnects. Such hypotheses are often used in practice to design optical interconnects despite the fact that little data exists in the literature to warrant their use. It therefore makes good sense to spend some time verifying the accuracy of these models. This will provide a solid engineering foundation for the design of future systems.

No description available.