GaInN has the potential to revolutionize the solar cell industry, enabling higher efficiency solar cells with its wide bandgap range spanning the entire solar spectrum. However, material quality issues stemming from the large lattice mismatch between its binary endpoints and questionable range of p-type doping has thus far prevented realization of high efficiency solar cells. Nonetheless, amorphous and multi-crystalline forms of GaInN have been theorized to exhibit a defect-free bandgap, enabling GaInN alloys at any indium composition to be realized. But the range of possible p-type doping has not yet been determined and no device quality material has been demonstrated thus far. Nonetheless, a Schottky barrier design (to bypass the p-type doping issue) on single-crystal GaInN can be used to provide some insight into the future of amorphous and micro-crystalline GaInN Schottky barrier solar cells. Through demonstration of a functional single crystalline GaInN Schottky barrier solar cell and comparison of the results to the best published reports for more conventional p-i-n GaInN solar cells, this work aims to establish the feasibility of amorphous and multi-crystalline GaInN solar cells. / Ph. D.

In electric power systems, the existence of constant power loads such as output-regulated power converters may bring instability problem to AC or DC distributed systems. Impedance based stability criteria has been proven a good tool for small-signal stability analysis. This works focuses on the developing of a comprehensive software tool which can extract DC or three phase AC impedances, and apply stability analysis. An algorithm is designed to select FFT window and adjust perturbation frequencies. This feature enables the software to accurately measure impedances even in existence of system line harmonics. Furthermore, multi-tone approach is developed to improve simulation time. The complete software tool is tested with simulation models and experiment results, to show the effectiveness. When multi-tone approach is applied on nonlinear loads, it gives incorrect results. The reason is that perturbation frequency will have overlapping with side-band harmonics. An algorithm is designed to avoid this problem. The algorithm is tested with 12-pulse diode rectifier simulation model, and 6-pulse diode rectifier simulation model and experimental test bed. Both simulation and experiment results verifies the concept. / Master of Science

The detection and measurement of gas properties has become essential to meet rigorous criteria of environmental unfriendly emissions and to increase the energy production efficiency. Although low cost devices such as pellistors, semiconductor gas sensors or electrochemical gas sensors can be used for these applications, they offer a very limited lifetime and suffer from cross-response and drift. On the contrary, gas sensors based on optical absorption offer fast response, zero drift, and high sensitivity with zero cross response to other gases. Hence, over the last forty years, diode laser spectroscopy (DLS) has become an established method for non-intrusive measurement of gas properties in scientific as well as industrial applications. Wavelength modulation spectroscopy (WMS) is derivative form of DLS that has been increasingly applied for making self-calibrated measurements in harsh environments due to its improved sensitivity and noise rejection capability compared to direct absorption detection. But, the complexity in signal processing and higher scope of error (when certain restrictions on operating conditions are not met), have inhibited the widespread use of the technique. This dissertation presents a simple and novel strategy for practical implementation of WMS with commercial diode lasers. It eliminates the need for pre-characterization of laser intensity parameters or making any design changes to the conventional WMS system. Consequently, sensitivity and signal strength remain the same as that obtained from traditional WMS setup at low modulation amplitude. Like previously proposed calibration-free approaches, this new method also yields absolute gas absorption line shape or absorbance function. Residual Amplitude Modulation (RAM) contributions present in the first and second harmonic signals of WMS are recovered by exploiting their even or odd symmetric nature. These isolated RAM signals are then used to estimate the absolute line shape function and thus removing the impact of optical intensity fluctuations on measurement. Uncertainties and noises associated with the estimated absolute line shape function, and the applicability of this new method for detecting several important gases in the near infrared region are also discussed. Absorbance measurements from 1% and 8% methane-air mixtures in 60 to 100 kPa pressure range are used to demonstrate simultaneous recovery of gas concentration and pressure. The system is also proved to be self-calibrated by measuring the gas absorbance for 1% methane-air mixture while optical transmission loss changes by 12 dB. In addition to this, a novel method for diode laser absorption spectroscopy has been proposed to accomplish spatially distributed monitoring of gases. Emission frequency chirp exhibited by semiconductor diode lasers operating in pulsed current mode, is exploited to capture full absorption response spectrum from a target gas. This new technique is referred to as frequency chirped diode laser spectroscopy (FC-DLS). By applying an injection current pulse of nanosecond duration to the diode laser, both spectroscopic properties of the gas and spatial location of sensing probe can be recovered following traditional Optical Time Domain Reflectometry (OTDR) approach. Based on FC-DLS principle, calibration-free measurement of gas absorbance is experimentally demonstrated for two separate sets of gas mixtures of approximately 5% to 20% methane-air and 0.5% to 20% acetylene-air. Finally, distributed gas monitoring is shown by measuring acetylene absorbance from two sensor probes connected in series along a single mode fiber. Optical pulse width being 10 nanosecond or smaller in the sensing optical fiber, a spatial resolution better than 1 meter has been realized by this technique. These demonstrations prove that accurate, non-intrusive, single point, and spatially distributed measurements can be made in harsh environments using the diode laser spectroscopy technology. Consequently, it opens the door to practical implementation of optical gas sensors in a variety of new environments that were previously too difficult. / Ph. D.

Innovative nano-scale devices have been developed to enhance future circuit design to overcome physical barriers hindering complementary metal-oxide semiconductor (CMOS) technology. Among the emerging nanodevices, resonant tunneling diodes (RTDs) have demonstrated promising electronic features due to their high speed switching capability and functional versatility. Great circuit functionality can be achieved through integrating heterostructure field-effect transistors (HFETs) in conjunction with RTDs to modulate effective negative differential resistance (NDR). However, RTDs are intrinsically suitable for implementing threshold logic rather than Boolean logic which has dominated CMOS technology in the past. To fully take advantage of such emerging nanotechnology, efficient design methodologies and design automation tools for threshold logic therefore become essential. In this thesis, we first propose novel programmable logic elements (PLEs) implemented in threshold gates (TGs) and multi-threshold threshold gates (MTTGs) by exploring RTD/ HFET monostable-bistable transition logic element (MOBILE) principles. Our three-input PLE can be configured through five control bits to realize all the three-variable logic functions, which is, to the best of our knowledge, the first single RTD-based structure that provides complete logic implementation. It is also a more efficient reconfigurable circuit element than a general look-up table which requires eight configuration bits for three-variable functions. We further extend the design concept to construct a more versatile four-input PLE. A comprehensive comparison of three- and four-input PLEs provides an insightful view of design tradeoffs between performance and area. We present the mathematical proof of PLE's logic completeness based on Shannon Expansion, as well as the HSPICE simulation results of the programmable and primitive RTD/HFET gates that we have designed. An efficient control bit generating algorithm is developed by using a special encoding scheme to implement any given logic function. In addition, we propose novel techniques of formulating a given threshold logic in conjunctive normal form (CNF) that facilitates efficient SAT-based equivalence checking for threshold logic networks. Three different strategies of CNF generation from threshold logic representations are implemented. Experimental results based on MCNC benchmarks are presented as a complete comparison. Our hybrid algorithm, which takes into account input symmetry as well as input weight order of threshold gates, can efficiently generate CNF formulas in terms of both SAT solving time and CNF generating time. / Master of Science

Lean, premixed combustion has enjoyed increased application due to the need to reduce pollutant emissions. Unfortunately, operating the flame at lean conditions increases susceptibility to thermoacoustic (TA) instability. Self-excited TA instabilities are a result of the coupling of the unsteady heat release rate of the flame with the acoustics of the combustion chamber. The result is large pressure oscillations that degrade performance and durability of combustion systems. Industry currently has no reliable tool to predict instabilities a priori. CFD simulations of full-scale, turbulent, reacting flows remain unrealizable. The work in this paper is part of a study that focuses on developing compact models of TA instabilities, i.e. acoustics and flame dynamics. Flame dynamics are defined as the response in heat release to acoustic perturbations. Models of flame dynamics can be coupled with models of combustor enclosure acoustics to predict TA instabilities. In addition, algorithms to actively control instabilities can be based on these compact models of flame dynamics and acoustics. The work outlined in this thesis aims at determining the flame dynamics model experimentally. Velocity perturbations are imparted on laminar and turbulent flames via a loudspeaker upstream of the flame. The response of the flame is observed through two measurements. Hydroxyl radical (OH*) chemiluminescence indicates the response in chemical reaction rate. Tunable Diode Laser Absorption Spectroscopy (TDLAS), centered over two water absorption features, allows a dynamic measurement of the product gas temperature. The response in product gas temperature directly relates to the enthalpy fluctuations that couple to the acoustics. Experimental frequency response functions of a laminar, flat-flame burner and a turbulent, swirl-stabilized combustor will be presented as well as empirical low-order models of flame dynamics. / Master of Science

Dans le contexte des transports plus électriques, les parties mécaniques tendent à être remplacées par leurs équivalents électriques plus petits. Ainsi, le composant lui-même doit supporter un environnement plus sévère et de lourdes contraintes (haute tension, haute température). Les composants silicium deviennent alors inappropriés. Depuis la commercialisation des premières diodes Schottky en 2001, le carbure de silicium est le matériau reconnu mondialement pour la fabrication de dispositifs haute tension avec une forte intégration. Sa large bande d'énergie interdite et son fort champ électrique critique permettent la conception de transistors à effet de champ avec jonction (JFET) pour les hautes tensions ainsi que les diodes associées. Les structures étudiées dépendent de nombreux paramètres, et doivent ainsi être optimisées. L'influence d'un paramètre ne pouvant être isolée, des méthodes mathématiques ont été appelées pour trouver la valeur optimale. Ceci a conduit à la mise en place d'un critère d'optimisation. Ainsi, les deux grands types de structures de JFET verticaux ont pu être analysés finement. D'une part, la recherche d'une structure atteignant les tensions les plus élevées possible a conduit à l'élaboration d'un procédé de fabrication complexe. D'autre part, un souci de simplification et de stabilisation des procédés de fabrication a permis le développement d'un composant plus simple, mais avec une limite en tension un peu plus modeste. / In the context of more electrical transports, mechanical devices tend to be replaced by their smaller electrical counterparts. However the device itself must support harsher environment and electrical constraints (high voltage, high temperature) thus making existing silicon devices inappropriate. Since the first Schottky diode commercialization in 2001, Silicon Carbide (SiC) is the favorite candidate for the fabrication of devices able to sustain high voltage with a high integration level. Thanks to its wide band gap energy and its high critical field, 4H-SiC allows the design of high voltage Junction Field Effect Transistor (JFET) with its antiparallel diode. Studied structures depends of many parameters, that need to be optimized. Since the influence of the variation of each parameter could not be isolated, we tried to find mathematical methods to emphase optimal values leading to set an optimization criterion. Thus, two main kinds of JFET structure were finely analyzed. In one hand, the aim of the structure that can sustain a voltage as high as possible leads to a complex fabrication process. In the other hand, the care of a simplification and a stabilization of manufacturing process leads to the design of simpler device, but with a bit less sustain capabilities.

Semiconductor diode lasers are successful tools in atomic spectroscopy. They are routinely used in frequency conversion applications to develop devices that access regions of the spectrum not directly available. This thesis describes the practical application of novel violet diode laser systems and their possible inclusion in spectroscopic systems. The design, assembly and successful operation of a doubly resonant optical parametric oscillator is described. There is discussion of the spectral behaviour of the device and the potential for pumping with a violet diode laser. Methods to adapt the output from the solitary diode devices are demonstrated with the use of microlensed diode lasers and extended cavity configurations. Details of the current tuning, linewidth and smooth tuning characteristics of a number of the lasers used are given. A commercial violet diode laser is used within an extended cavity to measure the hyperfine structure of atomic indium from a hollow cathode galvatron source at room temperature. Stabilisation of the diode laser to a line from the indium spectrum is attempted. The remainder of the thesis is concerned with the development of techniques to deliver clearer and more precise spectral information about trace species. Microlensed red and violet diode lasers are used to generate light at 254nm via sum frequency generation for the direct detection and modulation spectroscopy of mercury vapour, with microlensed lasers with modulation allowing more accurate discrimination between spectral features than direct absorption measurements. In addition Raman tweezers modulation spectroscopy is undertaken to investigate polymer microspheres and biological cell samples where the use of the modulation technique demonstrated improvements in the acquisition time and clarity of spectra through increased signal to noise and rejection of background fluorescence effects. A comparison between the direct and modulation techniques for all the systems indicates the greater sensitivity of the modulation technique.

Les exigences de plus en plus élevés concernant la qualité et propriétés de couches minces ont soutenu le développement de nouveaux procédés de pulvérisation. Ainsi, la décharge magnétron conventionnelle en courant continu, une des sources d’atomes la plus utilisée pour le dépôt de couches minces, a été améliorée par le couplage avec une décharge additionnelle de radio fréquence pour obtenir le nouveau procédé RF-IPVD (Radio Frequency-Ionized Physical Vapour Deposition). Ce procédé permet de générer un degré d’ionisation supérieur à celui dans la décharge magnétron classique, nécessaire pour contrôler les propriétés des couches minces. Un procédé alternatif pour augmenter d’avantage l’ionisation consiste à appliquer des impulsions haute puissance sur la cathode HPPMS (High Power Pulsed Magnetron Sputtering), pour des durés courtes de l’ordre de ųs ou dizaines de ųs. L’étude menée porte sur les phénomènes de pulvérisation et de transport des espèces du métal dans ces trois versions de la décharge magnétron par les moyens de spectroscopie laser à l’aide des diodes laser accordables. Le développement récent de ces diodes nous a permis de sonder les niveaux fondamentaux du Titane et de l’Aluminium, et de caractériser la dépendance spatiale de la densité et température ainsi que la fonction de distribution en vitesse de ces atomes. L’effet des paramètres clés, comme l’intensité du courant et la pression du gaz, est étudie et décrit pour la décharge magnétron conventionnelle. La distribution spatiale et angulaire de la fonction de distribution en vitesses a été mesurée dans la région devant la cible magnétron, afin de caractériser les flux du métal et leur comportement dans le volume de la décharge. L’étude sur les atomes du métal dans le procédé RF-IPVD est concentrée sur l’effet de la décharge additionnelle sur le dépeuplement du niveau fondamental. Une efficacité plus grande des processus d’ionisation est trouvée à plus haute pression et plus haute puissance RF injecté. On a montré aussi que les atomes affectés par les processus d’ionisation sont ceux thermalisées, tandis que la distribution de atomes rapides n’est quasiment pas affectés par la décharge additionnelle.Le diagnostic de la décharge pulsée a nécessité le développement d’une nouvelle procédure expérimentale, capable de suivre l’évolution de la densité et de la température des espèces neutres avec une résolution de l’ordre de la ųs. Cette procédure nous a servi pour décrire l’évolution spatio-temporel des atomes du métal (Ti et Al) et les atomes métastables d’Ar. Ces études offrent une vue globale sur le transport de atomes pulvérisés pendant la post décharge, ainsi qu’une description du fonctionnement de la décharge pulsé via la création des métastables d’Argon. / The higher requirements on the thin films quality have supported the development of new sputtering techniques. Thus, the conventional DC magnetron discharge, one of the most widely used source of atoms for thin film deposition, has been improved by the addition of an auxiliary radio frequency discharge - new technique called RF-IPVD (Radio Frequency -Ionized Physical Vapor Deposition). This technique highly increases the ionization degree compared to conventional magnetron discharge, which is necessary for a better control of the thin films properties. An alternative method to increase the ionization is based on the use of high power pulses on the cathode, HPPMS (High Power Pulsed Magnetron Sputtering), for short periods of time ranging from ųs to tens of ųs.The present study focuses on the sputtering phenomena and the transport of metal sputtered species in these three versions of the magnetron discharge, by means of laser spectroscopy using tunable laser diodes. The recent developments of these diodes have allowed to probe the fundamental levels of titanium and aluminum, and to characterize the spatial dependency of the density and temperature as well as the velocity distribution functions of these atoms. The effect of key discharge parameters, such as current intensity and gas pressure, is studied and described for the conventional magnetron discharge. The spatial and angular velocity distribution functions were measured in front of the magnetron target, in order to characterize the metal fluxes and their behavior in the discharge volume.The study on the metal atoms in the RF-IPVD process is focused on the effect of the additional discharge on the depopulation of the ground state level. Higher ionization efficiency is found at relatively high pressure and it increases with the injected RF power. It was also showed that the thermalized atoms are the ones involved in the ionization process, while the distribution of fast atoms is almost unaffected by the additional discharge.The diagnostics of the HPPMS discharge required the development of a novel experimental procedure, able to monitor the density and temperature of neutral species with a time resolution of ųs. This procedure was used to describe the spatiotemporal evolution of metal atoms (Ti and Al) and Ar metastable atoms. These studies provide an overview on the transport of sputtered atoms during the afterglow, and a description of the pulsed discharge operation, via the creation of metastable argon atoms.

Les oligophénylènes constituent une classe de molécules centrale dans la conception de semi-conducteurs organiques pour des applications optoélectroniques. Ces travaux portent sur la synthèse et l’étude approfondie de dérivés linéaires et cycliques du fluorène (un biphényle rigidifié par un pont méthylène), fragment constitutif essentiel dans l’électronique organique. Nous nous intéressons en particulier aux relations structure-propriétés de ces systèmes π-conjugués. Dans une première partie, avec comme cadre le développement de matériaux hôtes pour diodes électroluminescentes (PhOLEDs), nous présentons une étude de la régioisomérie de phényl-fluorènes et de phényl-spirobifluorènes. Ses résultats ont permis la préparation de quatre matériaux hôtes purs hydrocarbures, dimères de spirobifluorène, intégrés dans des PhOLEDs bleues à hautes performances. Dans une seconde partie, nous nous intéressons au domaine récent des nano-anneaux moléculaires, objets cycliques présentant une conjugaison π de nature singulière. Après une revue bibliographique portant sur les cycloparaphénylènes et leurs propriétés, nous présentons nos études concernant plusieurs exemples de leurs analogues pontés : les cycloparafluorènes. / Oligophenylenes constitute a major class of molecules in the design of organic semiconductors for optoelectronics applications. This work involves the synthesis and in-depth study of linear and cyclic derivatives of fluorene (a biphenyl rigidified by a methylene bridge), an essential building block in organic electronics. We focus our attention on the structure-property relationships of these π-conjugated systems. In a first part, within the framework of host materials for phosphorescent organic light-emitting diodes (PhOLEDs), we present a regioisomerism study of phenyl-fluorenes and phenyl-spirobifluorenes. Its results enabled the preparation of four pure hydrocarbon host materials, spirobifluorene dimers, used in high-performance blue PhOLEDs. In a second part, we take interest in the emerging field of molecular nanorings, cyclic objects presenting a singular nature of π-conjugation. After a bibliographical review covering cycloparaphenylenes and their properties, we present our studies regarding several examples of their bridged analogues: cycloparafluorenes.

Résumé : Ce travail avait pour but de contribuer à la détermination précise des paramètres de raies du disulfure de carbone en vue d'applications atmosphériques. Il s'agissait de déterminer théoriquement et expérimentalement les coefficients d'élargissement collisionnel de raies de la bande gamma 3 du disulfure de carbone perturbé par l'air atmosphérique ainsi que leur dépendance en température. Pour cela, nous avons déterminé les coefficients d'élargissement collisionnel du disulfure de carbone perturbé par l'azote, l'oxygène et enfin par l'argon, principaux composants de l'air atmosphérique, aussi bien que leur dépendance en température. En outre, nous avons déterminé théoriquement et expérimentalement les coefficients d'auto élargissement collisionnel de raies de la bande gamma 3 - gamma 1 du disulfure de carbone, à température ambiante. La dépendance vibrationnelle des largeurs collisionnelles étant négligeable, les résultats sont valables quel que soit le niveau supérieur de la transition. Enfin, nous avons déterminé les intensités absolues de quelques raies de la bande gamma 3 - gamma 1 du disulfure de carbone qui nous ont permis de mettre en exergue une des applications atmosphériques des paramètres de raie mesurés en laboratoire. Pour la partie expérimentale, un spectromètre diode-laser haute résolution (5x10-4cm-1) a été utilisé pour enregistrer les différents spectres. De ces derniers, nous avons déterminé les différents paramètres de raie d'absorption en ajustant aux profils expérimentaux des modèles de profils théoriques. Du point de vue théorique, les différents coefficients d'élargissement collisionnel ont été calculés sur base du formalisme semi-classique d'Anderson-Tsao-Curnutte amélioré par J. Bonamy et D. Robert, moyennant quelques modifications pour accorder les valeurs à celles mesurées en laboratoire. Ceci a permis de valider les différents potentiels d'interaction pour les différents systèmes moléculaires considérés. Abstract The purpose of this thesis was to contribute to the accurate determination of line parameters of carbon disulfide for atmospheric applications. We have determined experimentally as well as theoretically the collisional broadening coefficients of lines in the gamma 3 band of carbon disulfide perturbed by the atmospheric air and their temperature dependence. To this end, we determined collisional broadening coefficients, as well as their temperature dependence, of carbon disulfide diluted in nitrogen, oxygen and argon; the main components of the atmospheric air. In addition, we determined the self broadening coefficients of lines in the gamma 3 - gamma 1 band of carbon disulfide at room temperature. As the vibrational dependence of collisional widths is negligible, our results are valid whatever the higher level of transition. Finally, we determined the absolute intensities of a few lines in the gamma 3 - gamma 1 band of carbon disulfide, which have enabled us to highlight one of the atmospheric applications of line parameters measured in the laboratory. For the experimental part, a high resolution diode-laser spectrometer (5x10-4cm-1) was used to record the different spectra. Of these, we determined the parameters of absorption line by adjusting theoretical lineshape models to experimental profiles. From a theoretical point of view, different collisional broadening coefficients were calculated on the basis of semi classical formalism of Anderson-Tsao-Curnutte improved by J. Bonamy and D. Robert, with some modifications to make the calculated values more consistent with those measured. This enabled us to validate the different potentials of interactions for the different molecular systems considered.