L’enregistrement des données de neurones a connu d’énormes progrès au cours des dernières années ; il aide à diagnostiquer les maladies à l’intérieur du cerveau comme la maladie de Parkinson et la dépression clinique. Un grand nombre de patients atteints de Parkinson utilisent un implant neuronal pour réduire les tumeurs et le mouvement rigide. Afin de contrôler le mouvement, une petite électrode est placée sur le cerveau pour réduire et même éliminer les symptômes de Parkinson au moment où une simulation électrique arrive. Le système d’enregistrement de données de neurones exige un lien complet. En utilisant des microélectrodes, on prend les données provenant des neurones dans le cerveau, on les convertis en données numériques et ensuite on transmet ces données numérisées en utilisant une liaison sans fil. Dans ce travail, nous nous concentrons sur l’envoi de données de neurones à partir d’un dispositif implanté à travers la peau en utilisant la lumière. Il y’a différentes façons de transmettre les données sans fil, soit avec antenne, soit avec un émetteur optique ; nous discutons à propos de ces méthodes dans le chapitre de la revue de la littérature. Nous avons choisi de travailler avec Émettant VCSEL ou Vertical Cavity Surface lasers ; une diode laser spécialisée avec une meilleure efficacité et une vitesse élevée par rapport à d’autres dispositifs optiques. La première partie de la recherche était d’étudier la meilleure façon de transmettre des données à travers la peau humaine, le mode de transmission et les propriétés du milieu à travers lequel la lumière se propage. Après avoir choisi le mode de transmission, nous avons conçu un lien intégré en utilisant la technologie de 0,18 um CMOS. Ce lien intégré est constitué de deux parties, du côté de l’émetteur, qui est un moteur apte à entraîner le VCSEL avec un dB bande passante à 3 de 1,3 GHz et une faible consommation de puissance de 12 mW, et un côté récepteur qui se compose d’une photodiode reliée à un VCSEL CMOS amplificateur d’adaptation d’impédance à gain élevé (90 dB) et haute vitesse de (250 Mbps). La deuxième partie était de construire une liaison optique discrète avec des composants à faible coût commercial, donc nous avons conçu deux PCB (Printed Circuit Board) pour le côté émetteur ainsi que le côté récepteur, et nous avons conçu un système mécanique pour aligner l’émetteur et la photodiode. Nous avons ensuite testé notre liaison optique, ce qui a démontré la capacité de transmettre des données par le biais de 3 mm de tissu de porc à un débit binaire de 20 Mbps avec une faible consommation d’énergie de 3MWen utilisant OOK (On Off Keying) la transmission de données, et enfin nous avons fait une comparaison entre nos résultats et d’autres oeuvres. / Neural data recording has seen huge progress during the past few years; it helps for diagnosing diseases inside the brain like Parkinson disease and clinical depression. A big number of Parkinson’s patients use a neural implant to lessen tumors and rigid movement. A small electrode will be placed on the brain. It helps to control motion and when an electrical simulation happens, it helps reduce and even eliminate Parkinson symptoms. The neural data recording system requires a complete link starting by recording neural data using electrodes, convert this data onto digital data and transmit the digitized data using a wireless link. In this work we are focusing on sending neural data from an implanted device through the skin using light. There are different ways to transmit data wirelessly with either antenna or with an optical transmitter; we discuss about those methods in the literature review chapter. We choose to work with VCSEL or Vertical Cavity Surface Emitting Lasers; a specialized laser diode with improved efficiency and high speed compared to other optical devices. The first part of the research was to study the best way to transmit data through the human skin, the method of transmission and the properties of the medium through which the light will propagate. After choosing the method of transmission, we designed an integrated link using 0.18 um CMOS technology. This integrated link consists of two parts, the transmitter side which is a VCSEL driver able to drive the VCSEL with a 3 dB bandwidth of 1.3 GHz and low power-consumption of 12 mW, and a receiver side that consists of a photodiode connected to a CMOS transimpedance amplifier with high gain (90 dB) and high speed of (250 Mbps). The second part was to build a discrete optical link with commercial low cost components, so we designed two PCBs (Printed Circuit Board) for the transmitter and receiver side, and we designed a mechanical system to align the transmitter and the photodiode. We then tested our optical link, and it demonstrated the capability to transmit data through 3 mm of pork tissue at a bit-rate of 20 Mbps with low power consumption of 3 mW using OOK (On Off Keying) data transmission, and finally we did a comparison between our results and other works.

Les applications exigeant des très nombreuses données (médias sociaux, diffusion vidéo en continu, mégadonnées, etc.) se développent à un rythme rapide, ce qui nécessite de plus en plus de liaisons optiques ultra-rapides. Ceci implique le développment des transmetteurs optiques intégrés et à bas coût et plus particulirement en photonique sur silicium en raison de ses avantages par rapport aux autres technologies (LiNbO3 et InP), tel que la compatibilité avec le procédé de fabrication CMOS. Les modulateurs optoélectronique sont un élément essentiel dans la communication op-tique. Beaucoup de travaux de recherche sont consacrées au développement de dispositifs optiques haut débit efficaces. Cependant, la conception de modulateurs en photonique sur sili-cium (SiP) haut débit est diffcile, principalement en raison de l'absence d'effet électro-optique intrinsèque dans le silicium. De nouvelles approches et de architectures plus performances doivent être développées afin de satisfaire aux critères réliés au système d'une liaison optique aux paramètres de conception au niveau du dispositif integré. En outre, la co-conception de circuits integrés photoniques sur silicium et CMOS est cruciale pour atteindre tout le potentiel de la technologie de photonique sur silicium. Ainsi cette thèse aborde les défits susmentionnés. Dans notre première contribution, nous préesentons pour la première fois un émetteur phononique sur silicium PAM-4 sans utiliser un convertisseur numérique analog (DAC)qui comprend un modulateur Mach Zehnder à électrodes segmentées SiP (LES-MZM) implémenté dans un procédé photonique sur silicium générique avec jonction PN latérale et son conducteur CMOS intégré. Des débits allant jusqu'à 38 Gb/s/chnnel sont obtenus sans utili-ser un convertisseur numérique-analogique externe. Nous présentons également une nouvelle procédure de génération de délai dans le excitateur de MOS complémentaire. Un effet, un délai robuste aussi petit que 7 ps est généré entre les canaux de conduite. Dans notre deuxième contribution, nous présentons pour la première fois un nouveau fac-teur de mérite (FDM) pour les modulateurs SiP qui inclut non seulement la perte optique et l'efficacité (comme les FDMs précédents), mais aussi la bande passante électro-optique du modulateur SiP (BWEO). Ce nouveau FDM peut faire correspondre les paramètres de conception physique du modulateur SiP à ses critères de performance au niveau du système, facilitant à la fois la conception du dispositif optique et l'optimisation du système. Pour la première fois nous définissons et utilisons la pénalité de puissance du modulateur (MPP) induite par le modulateur SiP pour étudier la dégradation des performances au niveau du système induite par le modulateur SiP dans une communication à base de modulation d'amplitude d'impulsion optique. Nous avons développé l'équation pour MPP qui inclut les facteurs de limitation du modulateur (perte optique, taux d'extinction limité et limitation de la bande passante électro-optique). Enfin, dans notre troisième contribution, une nouvelle méthodologie de conception pour les modulateurs en SiP intégré à haute débit est présentée. La nouvelle approche est basée sur la minimisation de la MPP SiP en optimisant l'architecture du modulateur et le point de fonctionnement. Pour ce processus, une conception en longueur unitaire du modulateur Mach Zehnder (MZM) peut être optimisée en suivant les spécifications du procédé de fabrication et les règles de conception. Cependant, la longueur et la tension de biais du d'éphaseur doivent être optimisées ensemble (par exemple selon vitesse de transmission et format de modulation). Pour vérifier l'approche d'optimisation proposée expérimentale mont, a conçu un modulateur photonique sur silicium en phase / quadrature de phase (IQ) ciblant le format de modulation 16-QAM à 60 Gigabaud. Les résultats expérimentaux prouvent la fiabilité de la méthodologie proposée. D'ailleurs, nous avons augmenté la vitesse de transmission jusqu'à 70 Gigabaud pour tester la limite de débit au système. Une transmission de données dos à dos avec des débits binaires de plus de 233 Gigabit/s/channel est observée. Cette méthodologie de conception ouvre ainsi la voie à la conception de la prochaine génération d'émetteurs intégrés à double polarisation 400+ Gigabit/s/channel. / Data-hungry applications (social media, video streaming, big data, etc.) are expanding at a fast pace, growing demand for ultra-fast optical links. This driving force reveals need for low-cost, integrated optical transmitters and pushes research in silicon photonics because of its advantages over other platforms (i.e. LiNbO3 and InP), such as compatibility with CMOS fabrication processes, the ability of on-chip polarization manipulation, and cost effciency. Electro-optic modulators are an essential component of optical communication links and immense research is dedicated to developing effcient high-bitrate devices. However, the design of high-capacity Silicon Photonics (SiP) transmitters is challenging, mainly due to lack of inherent electro-optic effect in silicon. New design methodologies and performance merits have to be developed in order to map the system-level criteria of an optical link to the design parameters in device-level. In addition, co-design of silicon photonics and CMOS integrated circuits is crucial to reveal the full potential of silicon photonics. This thesis addresses the aforementioned challenges. In our frst contribution, for the frst time we present a DAC-less PAM-4 silicon photonic transmitter that includes a SiP lumped-element segmented-electrode Mach Zehnder modula-tor (LES-MZM) implemented in a generic silicon photonic process with lateral p-n junction and its co-designed CMOS driver. Using post processing, bitrates up to 38 Gb/s/channel are achieved without using an external digital to analog converter. We also presents a novel delay generation procedure in the CMOS driver. A robust delay as small as 7 ps is generated between the driving channels. In our second contribution, for the frst time we present a new figure of merit (FOM) for SiP modulators that includes not only the optical loss and effciency (like the prior FOMs), but also the SiP modulator electro-optic bandwidth ( BWEO). This new FOM can map SiP modulator physical design parameters to its system-level performance criteria, facilitating both device design and system optimization. For the frst time we define and employ the modulator power penalty (MPP) induced by the SiP modulator to study the system level performance degradation induced by SiP modulator in an optical pulse amplitude modulation link. We develope a closed-form equation for MPP that includes the SiP modulator limiting factors (optical loss, limited extinction ratio and electro-optic bandwidth limitation). Finally in our third contribution, we present a novel design methodology for integrated high capacity SiP modulators. The new approach is based on minimizing the power penalty of a SiP modulator (MPP) by optimizing modulator design and bias point. For the given process, a unit-length design of Mach Zehnder modulator (MZM) can be optimized following the process specifications and design rules. However, the length and the bias voltage of the phase shifter must be optimized together in a system context (e.g., baud rate and modulation format). Moreover, to verify the proposed optimization approach in experiment, we design an in-phase/quadrature-phase (IQ) silicon photonic modulator targeting 16-QAM modulation format at 60 Gbaud. Experimental results proves the reliability of our proposed methodology. We further push the baud rate up to 70 Gbaud to examine the capacity boundary of the device. Back to back data transmission with bitrates more than 233 Gb/s/channel are captured. This design methodology paves the way for designing the next generation of integrated dual- polarization 400+ Gb/s/channel transmitters.

La construction d’animaux transgéniques est une technique clef qui a permis l’étude de nombreux aspects de la biologie du nématode Caenorhabditis elegans. Les animaux transgéniques peuvent être construits soit en injectant l’ADN exogène dans les gonades syncitiales de l’hermaphrodite adulte, soit en bombardant une population de vers avec des microbilles enrobées d’ADN. Dans les deux cas, l’utilisation de marqueurs génétiques est indispensable pour l’identification des individus transgéniques et la maintenance des lignées. Nous avons développé un vecteur d’expression pour les nématodes contenant le gène de résistance à la néomycine (neo), qui fonctionne comme marqueur génétique. Le gène neo confère la résistance au G-418, un antibiotique qui inhibe la synthèse de protéines chez les eucaryotes et qui est létal pour les nématodes sauvages. Nous avons montré que le marqueur neo est un marqueur génétique très puissant qui permet l’identification rapide des animaux transgéniques et qui permet l’enrichissement des populations transgéniques en présence de l’antibiotique, facilitant ainsi la maintenance des lignées. Ce système ne nécessite aucun contexte génétique particulier pour fonctionner et est donc compatible avec des lignées receveuses mutantes, ainsi que des lignées transgéniques ayant été transformées avec d’autres marqueurs génétiques. De plus, le gène neo est sous le contrôle du promoteur du gène de C. elegans rps-27, codant pour une protéine ribosomale dont la séquence est hautement conservée entre les nématodes. Nous avons utilisé ce gène comme marqueur génétique pour la transgénèse de l’espèce Caenorhabditis briggsae, ce qui suggère que le système neo pourrait aussi être utilisé pour d’autres espèces de la famille Caenorhabditis. Finalement, nous avons aussi montré que le système neo peut être utilisé dans le contexte des techniques d’ingénierie génétique basées sur le transposon Mos1. En conclusion, la sélection en présence de G-418 offre des nouvelles possibilités d’expériences pour la transgénèse de C. elegans et d’autres espèces proches. Les avantages du système neo devraient ainsi contribuer à développer des techniques de transgénèse du ver plus flexibles et efficaces. / The generation of transgenic animals has been instrumental to study many biological aspects of Caenorhabditis elegans biology. Transgenic animals can be obtained by either microinjection of the exogenous DNA into the syncitial gonad of the hermaphrodite or by bombardment of a population of worms with DNA coated microparticles. Both techniques rely on the use of genetic markers to facilitate the recovery of transformed animals and the maintenance of transgenic lines. We developed a nematode expression vector carrying the neomycin resistance gene (neo) as a selection marker. This gene confers resistance to G-418, an antibiotic that normally inhibits protein synthesis in eukaryotes and is lethal for wild-type nematodes. We showed that the neo marker is a potent tool that allows a clear-cut selection of transgenic animals and hands-off maintenance of non-integrated populations on G-418 plates. This system does not imply any prerequisite on the original genotype of the recipient strain and can therefore be used on mutants lines as well as transgenic strains obtained with common markers. Moreover, we placed the neo gene under the control of the C. elegans rps-27 promoter, a highly conserved ribosomal protein throughout the nematode phylogeny. We were able to provide resistance to Caenorhabditis briggsae using this vector; this likely indicates that neo can be used in any species from the Caenorhabditis family. Finally, we demonstrated that this powerful selection system can be used in the context of Mos1 transposon excision-repair methods. Therefore, the neo system offers a wide range of new possibilities for transgenesis both in C. elegans and in other related species. We therefore believe that the benefits of the neo system should contribute to the development of more flexible and efficient techniques for nematode transgenesis.

L'utilisation récente d'une voie originale de croissance cristalline basée sur les mécanismes vapeur-liquide-solide (VLS) à partir d'un bain Ge-Si a permis des améliorations importantes de la qualité cristalline des couches minces hétéroépitaxiales de SiC-3C sur substrats sur substrat α-SiC(0001). Ce travail a pour but d'approfondir les connaissances sur cette technique de croissance, d'améliorer le procédé et de déterminer les propriétés du matériau élaboré. La première partie est dédiée à la compréhension et la maîtrise des différents mécanismes impliqués dans la croissance de SiC-3C par VLS. Cela a notamment permis la détermination des paramètres limitant la taille des échantillons et la démonstration des avantages à utiliser des alliages fondus contant 50 at% de Ge au lieu de 75 at%. Une étude de la croissance latérale sur substrats patternés a donné des indications intéressantes pouvant être intégrées dans le modèle d'élimination des macles. L'incorporation intentionnelle et non intentionnelle de dopants de type n et p pendant la croissance VLS a été suivie. Pour le dopage n, nous avons démontré l'existence d'un lien clair entre l'impureté N et la stabilisation du polytype SiC-3C. En outre, nous avons réussi à abaisser le dopage résiduel n des couches en dessous de 1x1017 cm-3. Pour le dopage p, le meilleur élément n'est pas le Ga mais l'Al, même s'il doit être ajouté à un alliage de type Si-Ge pour éviter l'homoépitaxie. Enfin, ces couches ont été caractérisées optiquement et électriquement par différentes techniques. Les mesures C-V et G-V ont permis d'estimer une concentration très faible (7×109 cm-2) de charges fixes dans l'oxyde SiO2 ainsi qu'une densité d'états d'interface aussi basse que 1.2×1010 cm-2eV-1 à 0.63 eV sous la bande de conduction. Ces valeurs record sont une très bonne base pour le développement d'un composant de type MOSFET en SiC-3C / Recently, the use of an original growth approach based on vapour-liquid-solid (VLS) mechanism with Ge-Si melts has led to significant improvement of the crystalline quality of the 3C-SiC thin layers heteroepitaxially grown on α-SiC(0001) substrate. This work tries to deepen the knowledge of such specific growth method, to improve the process and to determine the properties of the grown material. The first part was dedicated to the understanding and mastering of the various mechanisms involved in 3C-SiC growth by VLS mechanism. This led to the determination of the parameters limiting sample size and the demonstration of the benefits of using 50 at% Ge instead of 75 at% Ge melts. A study of lateral enlargement on patterned substrates gave some interesting hints which can be integrated in the model of twin defect elimination. The incorporation of non intentional and intentional n- and p-type dopants during VLS growth was studied. For n-type doping, a clear link between N impurity and 3C polytype stability was demonstrated. Besides, high purity layers with residual n-type doping below 1x1017 cm-3 were achieved. For p-type doping, the best element was shown to be Al and not Ga, even if it has to be alloyed with Ge-Si melts to avoid homoepitaxial growth. Finally, these layers were characterised by several optical and electrical means like Raman spectroscopy, low temperature photoluminescence, deep leveltransient spectroscopy and MOS capacitors measurements. Very low concentrationsof fixed oxide charges estimated about 7×109 cm-2 and interface states densities Dit equal to 1.2×1010 cm-2eV-1at 0.63 eV below the conduction band have been achieved. These record values are a very good base toward 3C-SiC MOSFET

As pastagens são o principal uso da terra no mundo, ocupando dois terços de sua área agricultável e três quartos da área agricultável do Brasil. Quando bem manejadas, as pastagens possuem alta capacidade de estocar C no solo, porém podem perder até 50% do C do solo quando em avançado estágio de degradação. Os objetivos desse trabalho foram quantificar o potencial de sequestro de C e avaliar a dinâmica da matéria orgânica no solo em locais de conversão de pastagens degradadas para bem manejadas no Brasil. Para tanto, foram avaliadas as alterações dos estoques de C no solo nas pastagens do Brasil através de uma meta-análise, bem como os fatores de manejo e as taxas de variações do estoque de C no solo em diferentes status de pastagens. Em sete cronossequências situadas em Vila Bela da Santíssima Trindade (MT), Nova Xavantina (MT), Conquista D\'Oeste (MT), Dueré (TO), Carmolândia (TO) e Paraíso (TO), foram determinadas as alterações na quantidade e na qualidade da Matéria Orgânica do Solo (MOS). Foram determinados os estoques de C e N e as taxas de variação locais. A qualidade da MOS foi avaliada por meio de análise isotópica, fracionamento físico, Índice de Manejo do C (IMC), grau de humificação (HFIL) e teor de C na Biomassa Microbiana (C-BM). Através da meta-análise, estimou-se que no Brasil os solos sob pastagens degradadas apresentam redução dos estoques de C de 0,13 Mg ha-1 ano-1. As pastagens nominais foram capazes de aumentar o estoque de C no solo, enquanto as pastagens melhoradas nem sempre mantiveram os estoques semelhantes aos da vegetação nativa. A recuperação das pastagens promove acúmulo de C no solo na faixa de 0,40 Mg ha-1 ano-1. O cumprimento da meta de recuperação dos 30 milhões de hectares de pastagens resultará no acumúlo de 12 Tg ha-1 ano-1 de C. A degradação das pastagens nos estados do Mato Grosso e Tocantins reduziu os estoques de C no solo numa faixa de -0,06 Mg ha-1 ano-1, enquanto a recuperação das mesmas proporcionou aumento de 0,12 Mg ha-1 ano-1. Nas pastagens avaliadas, cerca de 54% do C é originado das gramíneas cultivadas. A fração orgânica foi a mais sensível às Mudanças de Uso da Terra (MUT), promovendo diminuição dos estoques de C nessa fração. A implantação de pastagens tem impacto negativo na qualidade da MOS, com diminuição do IMC de até 70%. Contudo, se bem manejadas, o IMC das pastagens pode ser superior ao da vegetação nativa, como acontece quando a pastagem é consorciada com Pueraria spp. O C-BM e o grau de humificação não foram alterados após a implantação ou a recuperação das pastagens, não sendo bons indicadores de qualidade da MOS. As alterações dos estoques de C nas frações da MOS e o IMC foram os indicadores mais eficientes de alterações da qualidade da MOS em pastagens. As recuperação de pastagens degradadas promove aumento no estoque de C no solo e melhoria da qualidade da MOS. / Pastures are the main land use in the world, occupying two-thirds of the world\'s arable land and three quarters of agricultural areas of Brazil. When well-managed, pastures have a high capacity to store SOC, but they may lose up to 50 % of stock SOC when in an advanced stage of degradation. The objective of this work was to quantify the potential of carbon sequestration and to evaluate soil organic matter dynamics in the conversion of degraded pastures to well-managed in Brazil. We evaluated changes in stocks SOC in pastures in Brazil through a meta-analysis, determining the management factors and the rates of stocks SOC changes in different pasture status. In seven cronossequences located in Vila Bela da Santíssima Trindade (MT), Nova Xavantina (MT), Conquista D\'Oeste (MT), Dueré (TO), Carmolândia (TO) and Paraíso (TO) were determined the changes in the quantity and quality of soil organic matter (SOM). C and N stocks and rates of local variations were determined. The quality of the SOM was evaluated through isotopic analysis, physical fractionation, C management index (CMI), degree of humification (HFIL) and C content in microbial biomass (MB-C). Through the meta-analysis it was estimated that in Brazil degraded pastures present a reduction of stocks SOC of 0.13 Mg ha-1 year -1. Nominal pastures were able to increase stock SOC, while improved pastures did not always maintain stocks similar to native vegetation. The recovery of pastures promotes the accumulation of C in the soil at the rate of 0.40 Mg ha-1 year -1. Meeting the recovery goal of 30 million ha of pasture will result in the accumulation of 12 Tg C ha-1year -1. The degradation of pastures in the states of Mato Grosso and Tocantins reduces C stocks in the soil at a rate of 0.06 Mg C ha-1 year-1. While the recovery of degraded pastures has the potential to increase C stocks in the soil with rate of 0.12 Mg C ha-1 year-1. In the pastures evaluated, about 54% of the C originates from the cultivated grasses. The organic fraction was the most sensitive to LUC, promoting the decrease of C stocks in this fraction. Pasture implantation has a negative impact on SOM quality, with a reduction in the CMI of up to 70%; however, if managed well, the CMI of pastures may be higher than that of native vegetation. The MB-C and degree of humification were not altered after the implantation or the recovery of the pastures. Changes in C stocks in SOM fractions and CMI were the most efficient indicators of changes in SOM quality in pastures. The recovery of degraded pastures promotes an increase in C stock in the soil and an improvement in the quality of the MOS.

Apresentamos neste trabalho um estudo do efeito da depleção do silício policristalino e da corrente de tunelamento em dispositivos com óxidos de porta finos. Utilizamos curvas características da capacitância em função da tensão de porta (C-V), para analisar a degradação causada por estes efeitos.Quanto ao efeito da depleção do silício policristalino a capacitância total na região de inversão apresenta uma redução conforme a concentração de dopantes do silício policristalino diminui. Este efeito foi observado em curvas C-V tanto de alta como de baixa freqüência, sendo esta última mais afetada. A corrente de tunelamento através do óxido de porta apresentou uma influência na largura da região de depleção no silício, que aumentou devido ao tunelamento de portadores do substrato. Como resultado, uma diminuição na capacitância do silício foi observada, fazendo a curva C-V diminuir na região de inversão. Quando considerado o efeito de depleção no silício policristalino junto com o efeito do tunelamento, observou-se que na região da porta houve um excesso de portadores, causando uma diminuição na região de depleção do silício policristalino. Neste caso a curva C-V sofreu uma maior redução, tornando-se difícil separar os dois efeitos. A curva C-V de baixa freqüência foi a mais atingida, pois como os portadores tem tempo de resposta, pode-se observar a influência da corrente de tunelamento nas cargas de inversão. Apresentamos ainda um novo método para a determinação da concentração de dopantes no substrato e no silício policristalino, através de curvas C-V de alta freqüência. Simulações numéricas bidimensionais e medidas experimentais foram utilizadas para validação do método. Os resultados obtidos indicam que o método proposto apresenta um grande potencial, tendo como principal vantagem a simplicidade de aplicação. / In this work we present the study of polysilicon depletion and the gate tunneling current effects in thin-gate oxide devices. Characteristic curves of capacitance as a function of the gate voltage (C-V) were used to analyze the degradation caused for these effects. Regarding the poly depletion effect, a reduction of the total capacitance in the inversion region was verified as the polysilicon doping concentration decreases. This effect was observed in C-V curves in high and low frequency, being the last one more affected. The gate tunneling current presented an influence on the width of the depletion silicon region, which increased due to the carriers tunneling from the substrate. As a result, a reduction in the silicon capacitance was observed, causing the C-V curve reduction in the inversion region. When the polysilicon depletion effect is considered together with the tunneling effect, it was observed that there is a carriers excess in the gate region, causing a reduction of the polysilicon depletion region width. In this case, the C-V curve suffered a larger reduction, making difficult to separate both effects. The most affected characteristic was the C-V curve at low frequency, due to existence of the carrier response time that allows observing the influence of the tunneling current in inversion charges. A new method for the determination of the doping concentration of substrate and polysilicon was also presented, through C-V curves at high frequency. Two-dimensional simulations and experimental measurements were used to validate the method. The obtained results indicate that the propose method present a higher potential, having as principal advantage the simplicity of application.

Le mos maiorum est fondé sur la mémoire et a une nature orale. Le droit Quiritaire était essentiellement fondé sur les mores. Ceux-ci règlent les vies des familiae et des gentes patriciennes. Jusqu’à la rédaction des Douze Tables, le droit était oral et la connaissance et l’interprétation du droit étaient les prérogatives des pontifes. La nature orale de la tradition rend possible sa manipulation politique de la part des patriciens et, dans un deuxième temps, de la noblesse patricio-plébéienne. Ensuite, nous éclairons le passage du droit pontifical à la iurisprudentia laïque qui apparaît vers le IIIe siècle. Nous étudions aussi l’évolution des rapports entre l’auctoritas de la classe dirigeante et la revendication d’équité dans le domaine juridique adressée par le peuple. Nous mettons en lumière la crise de la tradition à l’époque tardo-républicaine. L’appel au mos maiorum par Cicéron et Salluste s’inscrit dans le projet de mettre en œuvre un renouvellement politique de la res publica sur un fondement moral. Le mos maiorum est évoqué dans l’espoir de revenir à la res publica des maiores. Toutefois, à cette époque, l’État s’est effondré et les citoyens ne sont pas à la hauteur de leurs ancêtres. Aussi bien Cicéron que Salluste sont conscients que la tradition ne peut pas constituer un modèle à leur époque. En effet, cette dernière a profondément changé au fil du temps. D’ailleurs, l’évocation du mos maiorum, est mise place aussi bien par les optimates que par les populares dans des buts politiques opposés. La tendance à l’individualisme répandue dans la société tardo-républicaine est la cause de l’effondrement de la res publica. / Mos maiorum is based on memory and it has an oral nature. The Quiritary Law was primarily based on mores. They regulated the the way of living both of familiae and patrician gentes. Prior to the writing of Twelve Tables, the law was oral and the knowledge as well as the interpretation of the law were the prerogatives of the pontiffs. The oral nature of the tradition makes it possible for the patricians and, then, for the patrician-plebeian nobility to provoke its political manipulation. Thereafter, we will enlighten the transition from the Pontifical Law to secular iurisprudentia, which appears approximately at the 3rd century BC. We will also clarify the evolution of relations between the auctoritas of the ruling class and the claim for equity in the legal area appealed by the people. We will shed light on the crisis of tradition in the late republican era. The calling to the mos maiorum by Cicero and Sallust is a part of the project to implement a political renewal of the res publica on a moral basis. Mos maiorum is mentioned in the hope of returning to the res publica of maiores. However, at that time, the state collapses and citizens are not equal to their ancestors. Both Cicero and Sallust are aware that tradition can not be a model in their time. Indeed, the latter has changed profoundly over time. Moreover, the evocation of the mos maiorum is put both by the optimates and by the populares for opposite political purposes. A tendency to widespread individualism in the late-republican society is the cause of the collapse of the res publica.

Apresentamos neste trabalho um estudo do efeito da depleção do silício policristalino e da corrente de tunelamento em dispositivos com óxidos de porta finos. Utilizamos curvas características da capacitância em função da tensão de porta (C-V), para analisar a degradação causada por estes efeitos.Quanto ao efeito da depleção do silício policristalino a capacitância total na região de inversão apresenta uma redução conforme a concentração de dopantes do silício policristalino diminui. Este efeito foi observado em curvas C-V tanto de alta como de baixa freqüência, sendo esta última mais afetada. A corrente de tunelamento através do óxido de porta apresentou uma influência na largura da região de depleção no silício, que aumentou devido ao tunelamento de portadores do substrato. Como resultado, uma diminuição na capacitância do silício foi observada, fazendo a curva C-V diminuir na região de inversão. Quando considerado o efeito de depleção no silício policristalino junto com o efeito do tunelamento, observou-se que na região da porta houve um excesso de portadores, causando uma diminuição na região de depleção do silício policristalino. Neste caso a curva C-V sofreu uma maior redução, tornando-se difícil separar os dois efeitos. A curva C-V de baixa freqüência foi a mais atingida, pois como os portadores tem tempo de resposta, pode-se observar a influência da corrente de tunelamento nas cargas de inversão. Apresentamos ainda um novo método para a determinação da concentração de dopantes no substrato e no silício policristalino, através de curvas C-V de alta freqüência. Simulações numéricas bidimensionais e medidas experimentais foram utilizadas para validação do método. Os resultados obtidos indicam que o método proposto apresenta um grande potencial, tendo como principal vantagem a simplicidade de aplicação. / In this work we present the study of polysilicon depletion and the gate tunneling current effects in thin-gate oxide devices. Characteristic curves of capacitance as a function of the gate voltage (C-V) were used to analyze the degradation caused for these effects. Regarding the poly depletion effect, a reduction of the total capacitance in the inversion region was verified as the polysilicon doping concentration decreases. This effect was observed in C-V curves in high and low frequency, being the last one more affected. The gate tunneling current presented an influence on the width of the depletion silicon region, which increased due to the carriers tunneling from the substrate. As a result, a reduction in the silicon capacitance was observed, causing the C-V curve reduction in the inversion region. When the polysilicon depletion effect is considered together with the tunneling effect, it was observed that there is a carriers excess in the gate region, causing a reduction of the polysilicon depletion region width. In this case, the C-V curve suffered a larger reduction, making difficult to separate both effects. The most affected characteristic was the C-V curve at low frequency, due to existence of the carrier response time that allows observing the influence of the tunneling current in inversion charges. A new method for the determination of the doping concentration of substrate and polysilicon was also presented, through C-V curves at high frequency. Two-dimensional simulations and experimental measurements were used to validate the method. The obtained results indicate that the propose method present a higher potential, having as principal advantage the simplicity of application.

O carbeto de silício (SiC) é um semicondutor com propriedades adequadas para substituir o silício em dispositivos eletrônicos em aplicações que exijam alta potência, alta frequência e/ou alta temperatura. Além disso, é possível crescer termicamente um filme de dióxido de silício (SiO2) sobre o SiC de maneira análoga ao silício. Porém, esses filmes apresentam maior densidade de defeitos eletricamente ativos na região interfacial SiO2/SiC que no caso do SiO2/Si, o que limita a qualidade dos dispositivos formados. Assim, compreender a origem da degradação elétrica e desenvolver métodos para passivar os defeitos na região interfacial SiO2/SiC são importantes passos para o desenvolvimento da tecnologia do SiC. Buscando uma melhor compreensão da natureza dos defeitos presentes na região interfacial SiO2/SiC, a interação de estruturas SiO2/SiC com vapor d’água enriquecido isotopicamente (D2 18O) e a interação com monóxido de carbono (CO), um dos subprodutos da oxidação térmica do SiC, foram investigadas. Observou-se que a interação com CO gera cargas positivas na estrutura e que a incorporação de deutério proveniente da água é fortemente dependente da rota de formação do filme de SiO2. Sabendo que a incorporação de nitrogênio e de fósforo na região interfacial SiO2/SiC são eficientes métodos para reduzir o número de defeitos eletricamente ativos nessa região, investigou-se a incorporação de nitrogênio em estruturas de SiC através de tratamentos térmicos em amônia enriquecida isotopicamente (15NH3) e desenvolveu-se um novo método de incorporação de fósforo, fazendo sua deposição por pulverização catódica (sputtering) Os métodos de incorporação propostos resultaram em maiores quantidades de nitrogênio e de fósforo na região interfacial SiO2/SiC do que os encontrados na literatura, tornando-os promissores candidatos na passivação elétrica do SiC. Além da caracterização físico-química utilizando diferentes técnicas, também foi feita a caracterização elétrica de capacitores Metal-Óxido-Semicondutor (MOS) testando filmes de SiO2 obtidos por sputtering ou por crescimento térmico. Adicionalmente, desenvolveu-se uma rota de síntese de padrões de 18O mais estáveis ao longo do tempo para serem utilizados em análises por reação nuclear. Também foi proposta uma metodologia de quantificação de fósforo via análise por reação nuclear. Dos resultados obtidos neste doutorado, uma melhor compreensão da natureza e da origem dos defeitos presentes na região interfacial SiO2/SiC foi alcançada. Também obteve-se uma melhor compreensão de como os elementos passivadores nitrogênio e fósforo interagem nessa região. / Silicon carbide (SiC) is a semiconductor with adequate properties to substitute silicon in electronic devices in applications that require high power, high frequency, and/or high temperature. Besides, a silicon dioxide (SiO2) film can be thermally grown on SiC in a similar way to that on Si. However, these films present higher density of electrical defects in the SiO2/SiC interfacial region when compared to the SiO2/Si interface, which limits the quality of the fabricated devices. Thus, it is important to understand the origin of the electrical degradation and to develop methods to passivate the defects in the SiO2/SiC interfacial region in order to develop the SiC technology. Aiming at a better understanding of the nature of defects at the SiO2/SiC interfacial region, the interaction of SiO2/SiC structures with water vapor isotopically enriched (D2 18O) and the interaction with carbon monoxide (CO), one of the SiC thermal oxidation by-products, were investigated. It was observed that the interaction with CO generates positive charges in the structure and that the deuterium incorporation from the water vapor is strongly dependent on the formation route of the SiO2 film. Knowing that nitrogen and phosphorous incorporation in the SiO2/SiC interfacial region are efficient methods to reduce the number of electrical defects in this region, the nitrogen incorporation in SiC structures by isotopically enriched ammonia (15NH3) annealings was investigated and a new method to incorporate phosphorous, by sputtering deposition was developed The proposed incorporation methods resulted in higher amounts of nitrogen and phosphorous then those found in literature, making them promising candidates to the electrical passivation of SiC. Besides the physico-chemical characterization using different techniques, the electrical characterization of Metal-Oxide-Semiconductor (MOS) capacitors was also performed, testing SiO2 films obtained by sputtering deposition or thermally grown. Additionally, a route to synthesize 18O standards for nuclear reaction analyses that are more stable over time was developed. Besides, a methodology to quantify phosphorous by nuclear reaction analysis was proposed. From the results obtained in this PhD thesis, a better understanding of the nature and the origin of defects present in the SiO2/SiC interfacial region was obtained, as well as a better understanding on how the passivating elements nitrogen and phosphorous interact in this region.

As pastagens são o principal uso da terra no mundo, ocupando dois terços de sua área agricultável e três quartos da área agricultável do Brasil. Quando bem manejadas, as pastagens possuem alta capacidade de estocar C no solo, porém podem perder até 50% do C do solo quando em avançado estágio de degradação. Os objetivos desse trabalho foram quantificar o potencial de sequestro de C e avaliar a dinâmica da matéria orgânica no solo em locais de conversão de pastagens degradadas para bem manejadas no Brasil. Para tanto, foram avaliadas as alterações dos estoques de C no solo nas pastagens do Brasil através de uma meta-análise, bem como os fatores de manejo e as taxas de variações do estoque de C no solo em diferentes status de pastagens. Em sete cronossequências situadas em Vila Bela da Santíssima Trindade (MT), Nova Xavantina (MT), Conquista D\'Oeste (MT), Dueré (TO), Carmolândia (TO) e Paraíso (TO), foram determinadas as alterações na quantidade e na qualidade da Matéria Orgânica do Solo (MOS). Foram determinados os estoques de C e N e as taxas de variação locais. A qualidade da MOS foi avaliada por meio de análise isotópica, fracionamento físico, Índice de Manejo do C (IMC), grau de humificação (HFIL) e teor de C na Biomassa Microbiana (C-BM). Através da meta-análise, estimou-se que no Brasil os solos sob pastagens degradadas apresentam redução dos estoques de C de 0,13 Mg ha-1 ano-1. As pastagens nominais foram capazes de aumentar o estoque de C no solo, enquanto as pastagens melhoradas nem sempre mantiveram os estoques semelhantes aos da vegetação nativa. A recuperação das pastagens promove acúmulo de C no solo na faixa de 0,40 Mg ha-1 ano-1. O cumprimento da meta de recuperação dos 30 milhões de hectares de pastagens resultará no acumúlo de 12 Tg ha-1 ano-1 de C. A degradação das pastagens nos estados do Mato Grosso e Tocantins reduziu os estoques de C no solo numa faixa de -0,06 Mg ha-1 ano-1, enquanto a recuperação das mesmas proporcionou aumento de 0,12 Mg ha-1 ano-1. Nas pastagens avaliadas, cerca de 54% do C é originado das gramíneas cultivadas. A fração orgânica foi a mais sensível às Mudanças de Uso da Terra (MUT), promovendo diminuição dos estoques de C nessa fração. A implantação de pastagens tem impacto negativo na qualidade da MOS, com diminuição do IMC de até 70%. Contudo, se bem manejadas, o IMC das pastagens pode ser superior ao da vegetação nativa, como acontece quando a pastagem é consorciada com Pueraria spp. O C-BM e o grau de humificação não foram alterados após a implantação ou a recuperação das pastagens, não sendo bons indicadores de qualidade da MOS. As alterações dos estoques de C nas frações da MOS e o IMC foram os indicadores mais eficientes de alterações da qualidade da MOS em pastagens. As recuperação de pastagens degradadas promove aumento no estoque de C no solo e melhoria da qualidade da MOS. / Pastures are the main land use in the world, occupying two-thirds of the world\'s arable land and three quarters of agricultural areas of Brazil. When well-managed, pastures have a high capacity to store SOC, but they may lose up to 50 % of stock SOC when in an advanced stage of degradation. The objective of this work was to quantify the potential of carbon sequestration and to evaluate soil organic matter dynamics in the conversion of degraded pastures to well-managed in Brazil. We evaluated changes in stocks SOC in pastures in Brazil through a meta-analysis, determining the management factors and the rates of stocks SOC changes in different pasture status. In seven cronossequences located in Vila Bela da Santíssima Trindade (MT), Nova Xavantina (MT), Conquista D\'Oeste (MT), Dueré (TO), Carmolândia (TO) and Paraíso (TO) were determined the changes in the quantity and quality of soil organic matter (SOM). C and N stocks and rates of local variations were determined. The quality of the SOM was evaluated through isotopic analysis, physical fractionation, C management index (CMI), degree of humification (HFIL) and C content in microbial biomass (MB-C). Through the meta-analysis it was estimated that in Brazil degraded pastures present a reduction of stocks SOC of 0.13 Mg ha-1 year -1. Nominal pastures were able to increase stock SOC, while improved pastures did not always maintain stocks similar to native vegetation. The recovery of pastures promotes the accumulation of C in the soil at the rate of 0.40 Mg ha-1 year -1. Meeting the recovery goal of 30 million ha of pasture will result in the accumulation of 12 Tg C ha-1year -1. The degradation of pastures in the states of Mato Grosso and Tocantins reduces C stocks in the soil at a rate of 0.06 Mg C ha-1 year-1. While the recovery of degraded pastures has the potential to increase C stocks in the soil with rate of 0.12 Mg C ha-1 year-1. In the pastures evaluated, about 54% of the C originates from the cultivated grasses. The organic fraction was the most sensitive to LUC, promoting the decrease of C stocks in this fraction. Pasture implantation has a negative impact on SOM quality, with a reduction in the CMI of up to 70%; however, if managed well, the CMI of pastures may be higher than that of native vegetation. The MB-C and degree of humification were not altered after the implantation or the recovery of the pastures. Changes in C stocks in SOM fractions and CMI were the most efficient indicators of changes in SOM quality in pastures. The recovery of degraded pastures promotes an increase in C stock in the soil and an improvement in the quality of the MOS.