Les travaux de cette thèse ont été menés dans le cadre d’une collaboration entre les laboratoires SATIE et LTN IFSTTAR. Ils portent principalement, sur l’étude de la robustesse des composants JFET SiC de puissance pour des applications de découpage à haute fréquence, forte puissance surfacique et à haute température lorsqu’ils sont soumis à des régimes extrêmes de fonctionnement. Les travaux présentés traitent également de façon plus générale l’étude de la durée de vie de packaging dédiés à ce type de composants et adaptés à la haute température pour des applications aéronautiques. La robustesse de différents lots des VJFETs SiC d’un fabricant particulier (SemiSouth) a été étudiée en régimes d’avalanche et de court circuit afin de déterminer les énergies que peuvent supporter ces composants dans ces modes de fonctionnement particuliers en cherchant notamment à quantifier la température du cristal et à mettre en évidence les mécanismes physiques à l’origine des défaillances. Nous avons ainsi également développé un modèle éléments finis thermique afin d’estimer la température de jonction du JFET SiC lors des régimes extrêmes pour chercher à relier l’apparition de la défaillance à la température. Finalement, nous décrivons des mécanismes physiques à l’origine des dégradations lors de la répétition de tels régimes extrêmes de fonctionnement expliquant à terme la destruction par vieillissement des transistors. Un substrat céramique à base de Si3N4 a été le support des études menées dans le cadre de cette thèse sur le packaging. Nous avons caractérisé les dégradations de ces substrats par des analyses acoustiques après vieillissement par cyclage thermique de forte amplitude. Un modèle thermomécanique a été développé afin d’estimer les contraintes mécaniques dans l’assemblage et valider les résultats expérimentaux obtenus. Enfin, nous avons également initiés des travaux de diagnostic thermique sur des puces JFET SiC, par des mesures d’impédance thermique pouvant être utilisées pour la détection de défauts de délaminage dans un assemblage de puissance. / The work presented in this thesis was conducted between SATIE and LTN IFSTTAR laboratories. It focuses on the study of the robustness of SiC power components subjected to hard working conditions for high switching frequency, high power density and high temperature applications. The work also presents a study on the robustness of a dedicated package adapted to high temperature applications. The robustness of several SiC VJFETs from a particular manufacturer (SemiSouth) was studied in avalanche and short circuit modes in order to estimate the energies that can withstand these components in these operating modes. The experimental protocol also includes thermal models to quantify the crystal temperature and to highlight the ageing physical mechanisms causing failure. Therefore, we had developed a finite element model to estimate the thermal junction temperature of the SiC JFET in extreme working conditions to try to relate the failure to the maximum temperature reached after each cycle. Finally, we described the physical mechanisms behind the degradations that explain ultimately the destruction of ageing transistors under repetitive avalanche mode. A ceramic substrate made of Si3N4 has been the support of studies conducted in this thesis on the packaging reliability. We characterized the degradation of these substrates by acoustic analysis after ageing by thermal cycling of high amplitude. A thermo-mechanical model was developed to estimate the mechanical stresses in the assembly and validate the experimental results. Finally, we have initiated thermal diagnostic studies on SiC JFET chips. We have shown that thermal impedance measurements can be used for the detection of delamination defects in a power assembly.

There is a pressing need to develop alternative communications links due to a number of physical phenomena, limiting the bandwidth and energy efficiency of wire-based systems or economic factors such as cost, material-supply reliability and environmental costs. Networks have moved to optical connections to reduce costs, energy use and to supply high data rates. A primary concern is that current optical-detection devices require high optical power to achieve fast data rates with high signal quality. The energy required therefore, quickly becomes a problem. In this thesis, advances in single-photon avalanche diodes (SPADs) are utilised to reduce the amount of light needed and to reduce the overall energy budget. Current high performance receivers often use exotic materials, many of which have severe environmental impact and have cost, supply and political restrictions. These present a problem when it comes to integration; hence silicon technology is used, allowing small, mass-producible, low power receivers. A reconfigurable SPAD-based integrating receiver in standard 130nm imaging CMOS is presented for links with a readout bandwidth of 100MHz. A maximum count rate of 58G photon/s is observed, with a dynamic range of ≈ 79dB, a sensitivity of ≈ −31.7dBm at 100MHz and a BER of ≈ 1x10−9. We investigate the properties of the receiver for optical communications in the visible spectrum, using its added functionality and reconfigurability to experimentally explore non-ideal influences. The all-digital 32x32 SPAD array, achieves a minimum dead time of 5.9ns, and a median dark count rate (DCR) of 2.5kHz per SPAD. High noise devices can be weighted or removed to optimise the SNR. The power requirements, transient response and received data are explored and limiting factors similar to those of photodiode receivers are observed. The thesis concludes that data can be captured well with such a device but more electrical energy is needed at the receiver due to its fundamental operation. Overall, optical power can be reduced, allowing significant savings in either transmitter power or the transmission length, along with the advantages of an integrated digital chip.

Time resolved imaging is concerned with the measurement of photon arrival time. It has a wealth of emerging applications including biomedical uses such as fluorescence lifetime microscopy and positron emission tomography, as well as laser ranging and imaging in three dimensions. The impact of time resolved imaging on human life is significant: it can be used to identify cancerous cells in-vivo, how well new drugs may perform, or to guide a robot around a factory or hospital. Two essential building blocks of a time resolved imaging system are a photon detector capable of sensing single photons, and fast time resolvers that can measure the time of flight of light to picosecond resolution. In order to address these emerging applications, miniaturised, single-chip, integrated arrays of photon detectors and time resolvers must be developed with state of the art performance and low cost. The goal of this research is therefore the design, layout and verification of arrays of low noise Single Photon Avalanche Diodes (SPADs) together with high resolution Time-Digital Converters (TDCs) using an advanced silicon fabrication process. The research reported in this Thesis was carried out as part of the E.U. funded Megaframe FP6 Project. A 32x32 pixel, one million frames per second, time correlated imaging device has been designed, simulated and fabricated using a 130nm CMOS Imaging process from ST Microelectronics. The imager array has been implemented together with required support cells in order to transmit data off chip at high speed as well as providing a means of device control, test and calibration. The fabricated imaging device successfully demonstrates the research objectives. The Thesis presents details of design, simulation and characterisation results of the elements of the Megaframe device which were the author’s own work. Highlights of the results include the smallest and lowest noise SPAD devices yet published for this class of fabrication process and an imaging array capable of recording single photon arrivals every microsecond, with a minimum time resolution of fifty picoseconds and single bit linearity.

With data throughput increasing exponentially in wireless communication networks, the limited radio frequency (RF) spectrum is unable to meet the future data rate demand. As a promising complementary approach, optical wireless communication (OWC) has gained significant attention since its licence-free light spectrum provides a considerable amount of communication bandwidth. In conventional OWC systems, the information-carried signal has to be real-valued and non-negative due to the incoherent light output of the conventional optical transmitter, light emitting diode (LED). Therefore, an intensity modulation and direct detection (IM/DD) system is used for establishing the OWC link. Some modified orthogonal frequency division multiplexing (OFDM) schemes have been proposed to achieve suitable optical signals. In previous research, three OFDM-based schemes have been presented, including DC-biased optical orthogonal frequency division multiplexing (DCO-OFDM), asymmetrically clipped optical orthogonal frequency division multiplexing (ACO-OFDM) and unipolar orthogonal frequency division multiplexing (U-OFDM). Basic concepts of SPAD receivers are studied and a novel application in OWC is proposed for a permanent downhole monitoring (PDM) system in the gas and oil industry. In this thesis, a complete model of the SPAD-based OWC system is presented, including some related SPAD metrics, the photon counting process in SPAD and a specific nonlinear distortion caused by passive quenching (PQ) and active quenching (AQ) recharged circuits. Moreover, a practical SPAD-based visible light communication (VLC) system and its theoretical analysis are presented in a long-distance gas pipe with a battery-powered LED and a basic on-off keying (OOK) modulation scheme. In this thesis, two novel optical orthogonal frequency division multiplexing (O-OFDM) technologies are proposed: non-DC-biased orthogonal frequency division multiplexing (NDCOFDM) and OFDM with single-photon avalanche diode (SPAD). The former is designed for optical multiple-input multiple-output (O-MIMO) systems based on the optical spatial modulation (OSM) technique. In NDC-OFDM, signs of modulated O-OFDM symbols and absolute values of the symbols are separately transmitted by different information carrying units. This scheme can eliminate clipping distortion in DCO-OFDM and achieve high power efficiency. Furthermore, as the indices of transmitters carry extra information bits, NDC-OFDM gives a significant improvement in spectral efficiency over ACO-OFDM and U-OFDM. In this thesis, SPAD-based OFDM systems with DCO-OFDM and ACO-OFDM are presented and analysed by considering the nonlinear distortion effect of PQ SPAD and AQ SPAD. A comprehensive digital signal processing of SPAD-based OFDM is shown and theoretical functions of the photon counting distribution in PQ SPAD and AQ SPAD are given. Moreover, based on Bussgang theorem, a conventional method for analysing memoryless distortion, close-formed bit-error rate (BER) expressions of SPAD-based OFDM are derived. Furthermore, SPAD-based OFDM is compared with conventional photo-diode (PD) based OFDM systems, and a gain of 40 dB in power efficiency is observed.

This thesis explores Single-Photon Avalanche Diodes (SPADs), which are solid-state devices for photon timing and counting, and concentrates on SPADs integrated in nano-scale CMOS. The thesis focuses on: the search for new theory regarding Geiger-mode operation; proving the utility of calibrated Technology Computer- Aided Design (TCAD) tools for accurately simulating SPADs for the first time; the investigation of how manufacture influences device operation; and the integration of high performance SPADs into CMOS which rival discrete devices. The accepted theories of SPAD operation are revisited and it is discovered that previously neglected minority carriers have many significant roles such as determining: after-pulsing, Dark Count Rate (DCR), bipolar “SPAD latch-up,” nonequilibrium DCR, and “quenching”. The “quenching” process is revisited and it is concluded that it is the “probability time” of ≈100-200ps, and not the previously thought latching current that is important. SPADs are also found to have transient negative differential resistance. The new theories of SPADs are also supported by steady-state 1D, 2D and 3D TCAD simulations as well as novel transient simulations and videos. It is demonstrated as possible to simulate DCR, Photon Detection Efficiency (PDE), guard ring performance, breakdown voltage, breakdown voltage variation, “quenching,” and transient operation of SPADs with great accuracy. The manufacture of SPADs is studied focusing on the operation and optimisation of guard rings and it is found that ion implantation induced asymmetry from the tilt and rotation/twist is critical. Where symmetric, guard rings fail first along the <100> directions due to enhanced mobility. Process integration rules are outlined for obtaining high performance SPADs in CMOS while maintaining compatibility with transistors. The minimisation of tunnelling with lightly-doped junctions and the reduction of ion implantation induced defects by additional annealing are found essential for achieving low DCR. The thesis demonstrates that it is possible to realise high performance SPADs in CMOS through the innovation of a “Deep SPAD” which achieves record PDE of ≈72% at 560nm with >40% PDE from 410-760nm, combined with 18Hz DCR, <60ps FWHM timing resolution, and <4% after-pulsing which is demonstrated to have potential for significant further improvement. The findings suggest that CMOS SPAD-based micro-systems could outperform existing photon timing and counting solutions in the future.

FLIM is branch of microscopy mainly used in biology which is quickly improving thanks to a rapid enhancement of instrumentation and techniques enabled by new sensors. In FLIM, the most precise method of measuring fluorescent decays is called TCSPC. High voltage PMT detection devices together with costly and bulky optical setups which scan the sample are usually required in TCSPC instrumentation. SPADs have enabled a big improvement in TCSPC measurement setup, providing a CMOS compatible device which can be designed in wide arrays format. However, sensors providing in-pixel TCSPC do not scale in size and in large array like the time-gated SPAD pixel sensors do. Time-gated pixels offer a less precise lifetime estimation, discarding any photon information outside a given time window, but this loss in photon-efficiency is offset by gains in pixel size. This work is aimed at the development of a wide field TCSPC sensor with a pixel size and fill factor able to reduce the cost of such devices and to obtain a high resolution time-resolved fluorescence image in the shortest time possible. The study focuses on SPAD and pixel design required to maximise the fill factor in sub 10 μm pixel pitch. Multiple pixel designs are proposed in order to reduce pixel area and so enable affordable wide array TCSPC systems. The first proposed pixel performs the CMM lifetime estimation in order to reduce the frame rate needed to stream the data out of the SPAD array. This pixel is designed in a 10 μm pitch and attains with the most aggressive design a fill factor of 10:17 %. A second design proposes an analogue TCSPC which consists in a S/H TAC circuitry. This simpler pixel can achieve a higher fill factor of 19:63% as well as smaller pitch of 8 μm thanks to the adoption of SPAD n-well and electronics area sharing. This last design is implemented in a 320 x 256 SPAD array in which is included part of a novel ADC aimed at reduction of the processing time required to build a TCSPC histogram. A more conventional analogue readout is used to evaluate the pixel performance as well as a more fine TCSPC histogram. The device was used to measure the fluorescence lifetime of green micro-spheres while the 2b flash ADC is used to demonstrate rapid resolution and separation of two different fluorescence decays.

This thesis works through the design of a radar-based system for imaging snowpacks remotely and over large areas to assist in avalanche prediction. The key to such a system is the ability to image volumes of snow at shallow, spatially-varying angles of incidence. To achieve this prerequisite, the design calls for a ground-based Synthetic Aperture Radar (SAR) capable of generating three-dimensional, ultra-high-resolution images of a snowpack. To arrive at design parameters for this SAR, the thesis works through relevant principles in avalanche mechanics, alpine-snowpack geophysics, and electromagnetic scattering theory. The thesis also works through principles of radar, SAR, antenna, and image processing theory to this end. A preliminary system is implemented to test the feasibility of the overall design. The preliminary system demonstrates ultra-high-resolution, three-dimensional imaging capabilities and the ability to image the volume of multiple alpine snowpacks. Images of these snowpacks display the structural patterns indicative of different layers in the snowpacks. Possible attributions of the patterns to physical properties in the snowpack are explored, but conclusions are not arrived at. Finally, lessons from the implementation of this preliminary system are discussed in terms of opportunities to be capitalized upon and problems to be overcome in future systems that more faithfully realize the complete design set forth in the thesis.

Ma thèse développe essentiellement la phénoménologie d'une forme particulière de relaxation pour le non équilibre ; d'abord dans le cas du plasma magnétisé, où il est identifié en un sens précis un point critique, autour duquel la relaxation est intermittente, puis dans un cadre plus général. Est reconsidéré la résonance entre ondes de Langmuir et particules : "l'amortissement Landau" n'est pas de nature cinétique, pour la simple raison que le champ est moyen. La preuve est apportée par le calcul des susceptibilités. Etablir cette preuve nécessite de recourir à l'hypothèse de linéarité thermodynamique, or l'instabilité est du type Rayleigh, qui développe une turbulence loin de l'équilibre. Il s'avère que dans une description de type Van der Waals, la définition de l'énergie libre est compromise par des coefficients fluctuants : des cycles s'effectuent entre les régions sur-critique et métastable. Par extrapolation de la théorie de Landau, ou avec l'argument d'une orbite de renormalisation chaotique,à cause de la présence d'un champ extérieur, intrinsèque et non nul, on détermine les exposants critiques. Ils sont notoirement différentes de ceux que livrent la théorie statique ou même la théorie dynamique critique. En particulier, z=1. Dans le régime stationnaire, l'hypothèse d'invariance d'échelle permet d'associer une généralisation de la relation de fluctuation-dissipation. Il y a l'amorce de la détermination, assez universelle, des distributions non maxwelliennes.

La genèse des travaux présentés dans ce mémoire se situe dans le domaine de l'astrophysique des très hautes énergies. Il y a un siècle les scientifiques de la planète ont identifié un nouveau type de messagers venant de l'espace : les rayons cosmiques. Ce rayonnement est constitué de particules (photons et autres) de très haute énergie qui bombardent la terre en permanence. Le passage de rayonnements cosmiques dans l'atmosphère terrestre se traduit par la génération de brefs éclairs lumineux (5ns) d'une intensité très faible (1pW) : le flash Tcherenkov, devenant alors visible du sol.<br /><br />Dans l'état de l'art le meilleur détecteur de lumière est aujourd'hui le Photomultiplicateur (PMT), grâce à ses caractéristiques de sensibilité et de vitesse. Mais il présente quelques inconvénients : faible efficacité quantique, coût, poids etc. Nous présentons dans cette thèse une nouvelle technologie alternative : les compteurs de photons sur semi-conducteur, constitués de photodiodes polarisées en mode Geiger.<br />Ce mode de fonctionnement permet d'obtenir un effet de multiplication au moins identique à celui des PMT. Un modèle physique et électrique a été développé pour reproduire le comportement de ce détecteur.<br />Nous présentons ensuite dans ce travail de thèse un procédé technologique original permettant la réalisation de ces dispositifs dans la centrale de technologie du LAAS-CNRS, avec la simulation de chaque opération du processus.<br />Nous avons mis au point une fiche pour la caractérisation électrique des dispositifs, du mode statique au mode dynamique, et vérifié la conformité aux simulations SILVACO, et au modèle initial. Les résultats obtenus sont déjà excellents, compte tenu qu'il s'agit d'une première étape de prototypage, et comparables avec les résultats publiés dans la littérature.<br />Ces composants sur silicium peuvent intervenir dans toutes les applications où il y a un photomultiplicateur, et le remplacer. Les applications sont donc très vastes et la croissance du marché très rapide. Nous présentons une première expérience d'astrophysique installée au Pic du Midi qui a détecté des flashs Tcherenkov de rayons cosmiques avec cette nouvelle technologie à semi-conducteur.

Les avalanches de neige sont un des risques naturels majeurs des régions montagneuses. Les grandeurs physiques de l'avalanche telles que la vitesse, l'épaisseur et la densité sont difficiles à évaluer au sein même de l'écoulement, alors qu'elles sont à la base des modèles d'avalanche. C'est pour ces raisons que la détermination des sollicitations exercées par l'avalanche de neige sur les ouvrages de type galerie paravalanche s'avère problématique. Le niveau de protection de ce type d'ouvrage est donc difficilement quantifiable. Dans une première partie de ce travail, nous avons fait un inventaire des ouvrages de protection et des modèles d'avalanches couramment utilisés. La deuxième partie de notre travail porte sur la connaissance et la détermination des sollicitations avalancheuses et, leurs variations spatio-temporelles sur les structures de type galerie paravalanche. Pour ceci, une expérimentation a été réalisée sur une plaque instrumentée de capteur de force placée dans un couloir d'avalanche sur le site naturel du Col du Lautaret. Pour compléter ces résultats, une expérimentation d'écoulement granulaire dans un canal incliné a été mise en place. Les résultats expérimentaux montrent que les sollicitations d'impact ont des valeurs très élevées proche de la rupture de pente et qu'elles décroissent rapidement avec l'augmentation de la distance par rapport à cette rupture. La dernière partie de notre travail consiste à étudier avec le logiciel EF Fedeaslab les réponses dynamiques de la structure des galeries paravalanches sous les chargements avalancheux obtenus expérimentalement. L'effet dynamique sur la structure a été quantifié en comparant les réponses de la structure sous chargement dynamique puis statique.