Rapid expansion of the information technology (IT) sector, market size and consumer interest for off-line power supply continue to rise, particularly for computers, flat-panel TVs, servers, telecom, and datacenter applications. Normal components of an off-line power supply include an electromagnetic interference (EMI) filter, a power factor correction (PFC) circuit, and an isolated DC-DC converter. For off-line power supply, an isolated DC-DC converter offers isolation and output voltage adjustment. For an off-line power supply, it takes up significantly more room than the rest; thus, an isolated DC-DC converter is essential for enhancing the overall performance and lowering the total cost of an off-line power supply. In contrast, data center server power supplies are the most performance-driven, energy-efficient, and cost-aware of any industrial application power supply. The full extent of data centers' energy consumption is coming into focus. By 2030, it is anticipated that data centers will require around 30,000 TWh, or 7.6% of world power usage. In addition, with the rise of cloud computing and big data, the energy consumption of data centers is anticipated to continue rising rapidly in the near future. In data centers, isolated DC-DC converters are expected to supply even higher power levels without expanding their size and with much greater efficiency than the present standard, which makes their design even more challenging. LLC resonant converters are frequently utilized as DC-DC converters in off-line power supply and data centers because of their high efficiency and hold-up capabilities. LLC converters may reduce electromagnetic interference because the primary switches and secondary synchronous rectifiers (SRs) both feature zero-voltage-switching (ZVS) and zero-current-switching (ZCS) for the SRs. Almost every state-of-the-art off-line power supply uses LLC converters in their DC-DC transformations. However, LLC converters face three important challenges. First, the excessive core loss caused by the uneven flux distribution in planar magnetics, owing to the huge size and high-frequency operation of the core. These factors led to the observation of dimensional resonance within the core and an excessive amount of eddy current circulating within the core, which resulted in the generation of high eddy loss within the ferrite material. This was normally assumed to be negligible for small core sizes and lower frequencies. This dissertation proposes methods to help redistribute the flux in the core, particularly in the plates where the majority of core losses are concentrated, and to provide more paths for the flux to flow so that the plates' thickness can effectively be reduced by half and core losses, particularly eddy loss, are reduced significantly. Second, the majority of power supplies in the IT sector are needed to deliver high-current output, but the transformer is cumbersome and difficult to build because of its high conduction losses. In addition, establishing a modular solution that can be scaled up to greater power levels while attaining a superior performance relative to best practices is quite difficult. By increasing the switching frequency to several hundred kilohertz using wide-band-gap (WBG) transistors, printed circuit board (PCB) windings may include magnetics. This dissertation offers a modular and scalable matrix transformer structure and its design technique, allowing any number of elemental transformers to be integrated into a single magnetic core with significantly reduced winding loss and core loss. It has been shown that the ideal power limitations per transformer for PCB-based magnetics beat the typical litz wire design in all design areas, in addition to the unique advantages of PCB-magnetics, such as their low profile, high density, simplicity, and automated construction. Alternatively, shielding layers may be automatically put into the PCB windings between the main and secondary windings during the production process to reduce CM noise. A method of shielding is presented to reduce CM noise. The suggested transformer design and shielding method are used in the construction of a 3 kW 400V/48 V LLC converter, with a maximum efficiency of 99.06% and power density of 530W/in3. Thirdly, LLC converters with a matrix transformer encounter a hurdle for extending greater power, including the number of transformers needed and the magnetic size. In addition to the necessity of resonant inductors, which increase the complexity and size of the magnetic structure, there is a need for a resonant inductor. By interconnecting the three-phases in a certain manner, three-phase interleaved LLC converters may lower the circulating energy, but they have large and numerous magnetic components. In this dissertation, a new topology for three-phase LLC resonant converters is proposed. Three-phase systems have the advantage of flux cancellation, which may be used to further simplify the magnetic structure and decrease core loss. In addition, a study of the various three-phase topologies is offered, and a criterion for selecting the best suitable topology is shown. Compared to the single-phase LLC, the suggested topology has less winding loss and core loss. In addition, three-phase transformers have a lower volt-second rating, and smaller core sizes may be used to mitigate the impact of eddy loss in the ferrite material. In contrast, three-phase systems offer superior EMI performance, which is shown in the loss and size of the EMI filter, and much less output voltage ripple, which is reflected in the size of the output filter. Finally, several methods of integrating resonant inductors into transformer magnetics are presented in order to accomplish a simple, compact, and cost-effective magnetic architecture. By increasing the switching frequency to 500 kHz, all six transformers and six inductors may be achieved using four-layer PCB winding. To decrease CM noise, additional 2-layer shielding may be implemented. A 500 kHz, 6-8 kW, 400V/48V, three-phase LLC converter with the suggested magnetic structure achieves 99.1% maximum efficiency and a power density of 1000 W/in3. This dissertation addresses the issues of analysis, magnetic design, expansion to higher power levels, and electromagnetic interference (EMI) in high-frequency DC/DC converters used in off-line power supply and data centers. WBG devices may be effectively used to enable high-frequency DC/DC converters with a hundred kilohertz switching frequency to achieve high efficiency, high power density, simple yet high-performance, and automated manufacture. Costs will be minimized, and performance will be considerably enhanced. / Doctor of Philosophy / The IT industry, market size, and customer interest in off-line power supply continue to grow quickly, especially for computers, flat-panel TVs, servers, telecom, and datacenter applications. Off-line power supplies usually have a DC-DC converter, an EMI filter, and a PFC circuit. A DC-DC converter is needed for an off-line power supply. An isolated DC-DC converter makes an off-line power supply work better and cost less, even though it takes up more space than the rest. But power supplies for data center servers are the most performance-driven, energy-efficient, and cost-conscious industrial applications. It's becoming clear how much energy data centers use. By 2030, data centers will use 7.6% of the world's power, or 30,000 TWh. With the rise of cloud computing and big data, energy use in data centers is likely to go up by a lot. In data centers, isolated DC-DC converters are expected to have much more power without getting bigger and to be much more efficient than the current standard. This makes their design even harder. LLC resonant converters are often used as DC-DC converters in data centers and off-line power supplies because they are very efficient and easy to control. LLC converters may have less electromagnetic interference because both the primary switches and the secondary synchronous rectifiers (SRs) have zero-voltage-switching (ZVS) and zero-current-switching (ZCS). Almost every modern off-line power supply uses LLC converters for DC-DC stage. LLC converters have to deal with three big problems. Due to the large size of the core and the high frequency of operation, the uneven distribution of flux in planar magnetics causes too much core loss. This dissertation suggests ways to redistribute flux in the core, especially in the plates where most core losses are concentrated and provide more flux paths to reduce plate thickness by half and core losses, especially eddy loss. Second, most IT power supplies need to put out a lot of current, but transformers are bulky and hard to build because they lose a lot of current. It is hard to make a modular solution that can scale up to higher levels of power and perform better than best practices. With wide-band-gap (WBG) transistors, the switching frequency can be raised to several hundred kilohertz so that magnetics can be added to PCB windings. This dissertation describes a modular and scalable matrix transformer structure and design method that lets any number of elemental transformers be put into a single magnetic core with much less winding loss and core loss. PCB-based magnetics have a low profile, a high density, are easy to build, and can be built automatically. Their ideal power limits per transformer beat the typical litz wire design in every way. Shielding layers can be added automatically between the main and secondary PCB windings to cut down on CM noise. CM noise is lessened by shielding. The suggested transformer design and shielding method are used to build a 3 kW 400V/48 V LLC converter with a maximum efficiency of 99.06% and a power density of 530W/in3. Third, LLC converters with matrix transformers can't get more power without more transformers and a bigger magnetic size. Resonant inductors, which add to the size and complexity of a magnetic structure, are also needed. By connecting the three phases, three-phase interleaved LLC converters use less energy, but they have a lot of magnetic parts. In this paper, a three-phase LLC resonant converter topology is proposed. In three-phase systems, flux cancellation makes magnetic structures easier to understand and reduces core loss. There is also a study of three-phase topologies and a set of criteria for choosing one. Compared to the single-phase LLC, the topology cuts down on winding and core loss. Three-phase transformers have a lower volt-second rating, and ferrite material eddy loss can be reduced by making the core smaller. The size and loss of the EMI filter show that three-phase systems have less output voltage ripple and better EMI performance. Finally, several ways of putting resonant inductors into the magnetics of a transformer are shown to make a magnetic architecture that is simple, small, and cheap. At 500 kHz, all six transformers and all six inductors can be wound on a four-layer PCB. CM noise can be cut down with 2-layer shielding. With the suggested magnetic structure, a 500 kHz, 6-8 kW, 400V/48V, three-phase LLC converter can reach 99.1% maximum efficiency and 1000 W/in3. This dissertation presents analysis, magnetic design, expanding to higher power levels, and electromagnetic interference (EMI) in high-frequency DC/DC converters used in off-line power supplies and data centers. WBG devices can be used to make high-frequency DC/DC converters with a switching frequency of a few hundred kilohertz that are powerful, easy to use, and can be automated. Both cost and performance will get better.

Wide-bandgap (WBG) semiconductor materials such as silicon carbide (SiC) and gallium-nitride (GaN) allow higher voltage ratings, lower on-state voltage drops, higher switching frequencies, and higher maximum temperatures. All these advantages make them an attractive choice when high-power density and high-efficiency converters are targeted. Two different gate-driver designs for SiC power devices are presented. First, a dual-function gate-driver for a power module populated with SiC junction field-effect transistors that finds a trade-off between fast switching speeds and a low oscillative performance has been presented and experimentally verified. Second, a gate-driver for SiC metal-oxide semiconductor field-effect transistors with a short-circuit protection scheme that is able to protect the converter against short-circuit conditions without compromising the switching performance during normal operation is presented and experimentally validated. The benefits and issues of using parallel-connection as the design strategy for high-efficiency and high-power converters have been presented. In order to evaluate parallel connection, a 312 kVA three-phase SiC inverter with an efficiency of 99.3 % has been designed, built, and experimentally verified. If parallel connection is chosen as design direction, an undesired trade-off between reliability and efficiency is introduced. A reliability analysis has been performed, which has shown that the gate-source voltage stress determines the reliability of the entire system. Decreasing the positive gate-source voltage could increase the reliability without significantly affecting the efficiency. If high-temperature applications are considered, relatively little attention has been paid to passive components for harsh environments. This thesis also addresses high-temperature operation. The high-temperature performance of two different designs of inductors have been tested up to 600_C. Finally, a GaN power field-effect transistor was characterized down to cryogenic temperatures. An 85 % reduction of the on-state resistance was measured at −195_C. Finally, an experimental evaluation of a 1 kW singlephase inverter at low temperatures was performed. A 33 % reduction in losses compared to room temperature was achieved at rated power. / <p>QC 20160922</p>

L’interaction de molécules organiques avec les surfaces semiconductrices permet de contrôler les propriétés physiques de ces dernières et ce, soit à travers une modification locale en utilisant des molécules individuelles, soit par la passivation de la surface par une mono-couche complète. Aussi, le contrôle de l’interaction moléculaire nous permet de modifier les propriétés intrinsèques des molécules à travers un découplage électronique partiel ou complet entre les orbitales moléculaires et la surface. Pour atteindre ces objectifs, cette thèse présente l’étude expérimentale de l’adsorption de molécules sur la surface semiconductrice à large gap de 6H-SiC(0001)-3x3. Les expériences ont été réalisées à l’aide d’un microscope à effet tunnel opérant dans les conditions d’Ultra-Haut Vide et de température ambiante (UHV RT-STM). Les résultats ont été comparés à des études théoriques employant des calculs selon la théorie de la fonctionnelle de la densité (DFT). Trois molécules on été étudié durant ce travail de thèse : C60, Caltrope et Trima. Les études STM et DFT montre que les molécules individuelles de C60 sont chimisorbé à la surface de carbure de silicium SiC(0001)-3x3 à travers la formation d’une seule liaison Si-C avec un seul adatome de silicium, contrairement aux autres surfaces semiconductrices où la molécule se chimisorbe en formant plusieurs liaisons. Trois sites d’adsorption par rapport à l’adatome de Si de la maille de surface ont été observés. Pour expliquer les observations STM, les forces de Van der Waals entre la molécule de C60 et les atomes de la surface voisins ont du être pris en compte dans les calculs DFT. Il a été observé aussi que les molécules de C60 forment de petits clusters même à de faibles taux de couverture ce qui indique la présence d’un état précurseur de la molécule et des interactions intermoléculaires non négligeable. La molécule de Caltrope, nouvellement synthétisée, a été étudié aussi bien sur la surface de Silicium que celle de SiC. Le dépôt de cette molécule complexe ne peut être réalisé selon la méthode d’évaporation classique sans induire sa dissociation et a donc nécessité l'emploi de techniques d’évaporation spécifiques. Nos résultats expérimentaux montrent un comportement remarquable: le dépôt de molécule individuelle est induit sur la surface de manière efficace par la pointe du STM démontrant ainsi l’idée d’imprimerie moléculaire. Suite à son adsorption sur la surface de silicium à travers une seule liaison, la molécule de Caltrope se comporte comme un moteur moléculaire activé thermiquement. La troisième molécule a être étudié est la molécule de Trima. Elle a été sélectionnée à cause de sa taille comparable à la distance des ad-atomes de silicium de la surface de SiC. La structure chimique de la molécule qui se termine par un groupement cétone rend possible la fonctionnalisation de la surface. Ceci est révélé par les calculs DFT de la densité de charge. La distribution de charge montre qu’il n’y a pas de partage entre les atomes d’oxygènes de la molécule et les ad-atomes de la surface et donc nous avons un évidence claire pour la formation d’une liaison dative. / The interaction of organic molecules with a semiconductor surface enables the physical properties of the surface to be controlled, from a local modification using individual isolated molecules to passivation using a complete monolayer. Controlling the molecular interaction also allows us to modify the intrinsic properties of the molecules by partial or complete electronic decoupling between the molecular orbitals and the surface. To this end, this thesis presents experimental studies of the adsorption of molecules on the wide band gap 6H-SiC(0001)-3×3 substrate. The experiments were performed using Ultra-High Vacuum Room Temperature Scanning Tunneling Microscopy (UHV RT STM) and the results were compared with comprehensive theoretical Density Functional Theory (DFT) calculations. Three different molecules were studied in this thesis: C60, Caltrop and Trima. The STM and DFT studies show that individual C60 fullerene molecules are chemisorbed on the silicon carbide SiC(0001)-3×3 surface through the formation of a single Si-C bond to one silicon adatom, in contrast to multiple bond formation on other semiconducting surfaces. We observed three stable adsorption sites with respect to the Si adatoms of the surface unit cell. To explain the STM observations, Van der Waals forces between the C60 molecule and the neighboring surface atoms had to be included in the DFT calculations. The C60 molecules are also observed to form small clusters even at low coverage indicating the presence of a mobile molecular precursor state and non negligible intermolecular interactions. The second newly designed Caltrop molecule was studied on both the Si and SiC surfaces. Intact adsorption of this complex organic molecule cannot be realized using classical adsorption methods and requires the use of specific evaporation techniques. Our experimental results show remarkable behavior: The STM tip efficiently deposits single molecules one at a time, demonstrating the concept of single molecule printing. After adsorption on the Si surface through one bond, the Caltrop operates as a thermally activated molecular rotor. The third molecule to be studied is the Trima molecule. This molecule was chosen because it is commensurable in size with the surface Si adatom distance. The chemical termination of the molecule with a ketone group enables the successful functionalization of the SiC surface. The Trima molecule provides a rare and clear-cut example of the formation of two dative bonds between the oxygen atoms of the carbonyl groups and the Si adatoms of the SiC surface. This is revealed by the DFT calculations of the charge density. The charge distribution shows that there is no sharing of electrons between the oxygen atoms of the molecule and the surface which is clear evidence for the formation of a dative bond.

A large variety of material systems for light emission and detection were studied: from very small band gap semiconductors for infra-red (IR) detectors to wide band gap semiconductors for ultra violet (UV) emission as well as CdSe/ZnSe QDs for single photon emitters and rare earth doped oxides for laser fabrication. The growth and characterization aspects were tackled. This work will focus on the relations between the growth procedures and the optical properties. The information that can be gained from optical studies as well as the limitations of those ones will be explained in each case. Following that, a number of projects will be presented. The main one will be based on how to circumvent the problems linked with p-type doping of wide bandgap semiconductors. This project, based on field effect hole injection in wide band-gap semiconductors addresses the major challenge of fabricating efficient deep UV emitters.

Transport Properties of Wide Band Gap Semiconductors Louis Tirino III 155 pages Directed by Dr. Kevin F. Brennan The objective of this research has been the study of the transport properties and breakdown characteristics of wide band gap semiconductor materials and their implications on device performance. Though the wide band gap semiconductors have great potential for a host of device applications, many gaps remain in the collective understanding about their properties, frustrating the evaluation of devices made from these materials. The model chosen for this study is based on semiclassical transport theory as described by the Boltzmann Transport Equation. The calculations are performed using an ensemble Monte Carlo simulation method. The simulator includes realistic, numerical energy band structures derived from an empirical pseudo-potential method. The carrier-phonon scattering rates and impact ionization transition rates are numerically evaluated from the electronic band structure. Several materials systems are discussed and compared. The temperature-dependent, high-field transport properties of electrons in gallium arsenide, zincblende gallium nitride, and cubic-phase silicon carbide are compared. Since hole transport is important in certain devices, the simulator is designed to simulate electrons and holes simultaneously. The bipolar simulator is demonstrated in the study of the multiplication region of gallium nitride avalanche photodiodes.

The recent application of semiconductor materials, such as GaN, to power electronics has led to the development of a new generation of devices, which promise lower losses, higher operating frequencies and reductions in equipment size. The aim of this research is to study the capabilities of emerging GaN power devices, to understand their advantages, drawbacks, the challenges of their implementation and their potential impact on the performance of power converters. The thesis starts by presenting the development of a simple model for the switching transients of a GaN cascode device under inductive load conditions. The model enables accurate predictions to be made of the switching losses and provides an understanding of the switching process and associated energy flows within the device. The model predictions are validated through experimental measurements. The model reveals the suitability of the cascode device to soft-switching converter topologies. Two GaN cascode transistors are characterised through experimental measurement of their switching parameters (switching speed and switching loss). The study confirms the limited effect of the driver voltage and gate resistance on the turn-off switching process of a cascode device. The performance of the GaN cascode devices is compared against state-of-the-art super junction Si transistors. The results confirm the feasibility of applying the GaN cascode devices in half and full-bridge circuits. Finally, GaN cascode transistors are used to implement a 270V - 28V, 1.5kW, 1 MHz phase-shifted full-bridge isolated converter demonstrating the use of the devices in soft-switching converters. Compared with a 100 kHz silicon counterpart, the magnetic component weight is reduced by 69% whilst achieving a similar efficiency of 91%.

Les convertisseurs à base de Si atteignent leurs limites. Face à ces besoins, le GaN, avec sa vitesse de saturation des électrons et le champ électrique de claquage élevés est candidat idéal pour réaliser des redresseurs, surtout s’il est épitaxié sur substrat à bas cout. Ce travail est dédié au développement des diodes Schottky sur AlGaN/GaN. Une couche de SiNx en faible traction a été obtenue. Un contact ohmique de Ti/Al avec une gravure partiel a donné une Rc de 2.8 Ω.mm avec une résistance Rsh de 480 Ω/□. Des diodes Schottky avec les étapes issues de ces études ont été fabriqué. La diode recuite à 400 °C avec 30 nm de profondeur de gravure a montré une hauteur de barrière de 0,82 eV et un facteur d'idéalité de 1,49. La diode a présenté une très faible densité de courant de fuite de 8.45x10-8 A.mm-1 à -400 V avec une tension de claquage entre 480 V et 750 V. / Si-based devices for power conversion applications are reaching their limits. Wide band gap GaN is particularly interesting due to the high electron saturation velocity and high breakdown electric field, especially when epitaxied on low cost substrates such as Si. This work was dedicated to the development and fabrication of the Schottky diode on AlGaN/GaN on Si. SiNx passivation in very low tensile strain is used. Ti (70 nm)/Al (180 nm) partially recessed ohmic contacts annealed at 800 ºC exhibited a 2.8 Ω.mm Rc with a sheet resistance of 480 Ω/sq. Schottky diodes with the previously cited passivation and ohmic contact were fabricated with a fully recessed Schottky contact annealed at 400 ºC. A Schottky barrier height of 0.82 eV and an ideality factor of 1.49 were obtained. These diodes also exhibited a very low leakage current density (up to -400 V) of 8.45x10-8 A.mm-1. The breakdown voltage varied between 480 V and 750 V.

Le nitrure de gallium (GaN) et ses alliages ternaires et quaternaires suscitent de plus en plus d’intérêt dans les communautés scientifiques et industrielles pour leur potentiel d’utilisation dans des dispositifs électroniques haute fréquence, dans les transistors à forte mobilité électroniques, dans la photo-détection UV et les cellules solaires de nouvelles générations. L’aboutissement de ces nouveaux composants reste entravé à l’heure actuelle, entre autre, par la non maîtrise des techniques d’établissement de contacts électriques. C’est dans ce cadre général que s’inscrivent les travaux de cette thèse. Même si l’objectif principal de cette thèse concerne l’étude des défauts électriquement actifs dans les alliages de semiconducteurs à grand gap B(AlGa)N et de leur rôle dans les propriétés de transport, la réalisation des contacts ohmiques et des contacts Schottky constitue une étape essentielle dans la réalisation des dispositifs à étudier. Pour les contacts ohmiques, nous avons déposé des couches de type Ti/Al/Ti/Au (15/200/15/200) par évaporation thermique. Des résistances spécifiques des contacts de l’ordre de 3x10-4Wcm2 ont été déterminées par les méthodes des TLM linéaires et confirmées par les TLM circulaires. Une modélisation théorique a été entreprise dans ce sens pour analyser les mesures expérimentales. Ensuite on a réalisé des diodes Schottky en déposant des contacts métalliques de Platine (Pt) d’épaisseur 150 nm. Des facteurs d’idéalité de 1.3 et une hauteur de barrière de 0.76 eV ont été obtenus et d’une manière reproductible. Une fois ces dispositifs réalisés, une étude des mécanismes de transport a été entreprise et nous a permis de mettre en évidence l’existence des effets tunnel direct et assisté par le champ, en plus de l’effet thermoïonique classique. Ceci a été mis en évidence par des mesures de courant et de capacité en fonction de la température. Pour les photodétecteurs, nous avons réalisés les mêmes mesures de courant et de capacité à l’obscurité et sous illumination à des longueurs d’ondes adaptées. Ces mesures nous ont permis de comprendre les phénomènes de gain qu’on a observés sur ces échantillons et aussi de mettre en évidence des mécanismes thermiquement actifs, dont les énergies d’activation ont été déterminées par la technique de l’Arrhenius. L’étude des défauts électriquement actifs a été menée par la technique transitoire de capacité de niveaux profonds, la (DLTS). Cette technique a été récemment mise en oeuvre au laboratoire et nous a permis d’effectuer des mesures sous différentes conditions incluant diverses polarisations de repos, différentes fréquences, et différentes hauteurs et largeurs d’impulsion de polarisation. Un des résultats importants est la possibilité de caractérisation à la fois des pièges à majoritaires et des pièges à minoritaire en changeant simplement les conditions de polarisation et contrairement aux procédures habituelles où une excitation optique supplémentaire est souvent nécessaire pour augmenter la concentration des porteurs minoritaires. Il a ainsi été mis en évidence, en accord avec la plupart des résultats de la littérature, l’existence de 6 pièges à électrons, tous situés en dessous de 0.9 eV de la bande de conduction, de trois pièges à trous dans l’intervalle 0.6 - 0 .7 eV au dessus de la bande de valence et un piège à trous distribué à l’interface. Une procédure rigoureuse de fit a été mise au point et a permis de confirmer nos résultats obtenus par la procédure classique de l’Arrhenius / Gallium nitride (GaN) and its ternary and quaternary alloys are attracting more and more interest in the scientific and industrial communities for their potential for use in high frequency electronic devices, for transistors with high electronic mobility, for UV photo-detection and new-generation solar cells. The outcome of these new components is still be seen to be limited in many areas, mainly due to the lack of control of electrical contacts implementation techniques. It is in this context that this thesis takes place.Although the main objective of this thesis deals with the study of the electrically active defects in high band gap B(AlGa)N semiconductor alloys and their role in the transport properties, the production of ohmic and Schottky contacts is an essential step in the realization of the devices under study. For the Ohmic contacts, we have deposited Ti/Al/Ti/Au (15/200/15/200) layers by thermal evaporation. Using the Transfer Length Method (TLM), we obtained specific contact resistances in the order of 3x10-4Wcm2. The Circular TLM has also confirmed this result. Besides, a theoretical modelling has been carried out to analyse the experimental measurements. Schottky diodes were then produced by depositing 150 nm platinum (Pt) metal contacts. An ideality factor of 1.3 and a barrier height of 0.76 eV were obtained. On the other hand, a study of transport mechanisms has been performed. It allowed us to demonstrate the existence of the direct tunnelling and the Thermionic Field Emission, in addition to the conventional thermionic effect. This result was underpinned by current and capacity measurements as a function of temperature. For photo detectors, we performed the same measurements of current and capacity in darkness and under illumination at suitable wavelengths. These measurements allowed understanding the internal gain that was observed on the samples. Furthermore, they show the effect of the thermally active mechanisms whose activation energies were determined by the Arrhenius technique. Using the Deep-Level Transient Spectroscopy (DLTS) technique followed up the study of the electrically active defects. This technique has recently been implemented in the laboratory. It allowed us to perform measurements under different conditions including various reverse bias, different frequencies, and different voltage pulse amplitudes and durations. One of the important results is the possibility of characterizing both majority and minority traps by simply changing the polarization conditions, as opposed to the usual procedures where an additional optical excitation is often necessary to increase the concentration of the minority carriers. In accordance with most of the encountered literature results, we found 6 electron traps all located below 0.9 eV of the conduction band, 3 hole traps in the 0.6-0.7 eV range above the valence band and one hole trap distributed at the interface. A rigorous procedure was developed and confirmed our results obtained by the standard Arrhenius technique

<p>Advances in modern technology and recent demand for high power applications have motivated great interest in power electronics. Power semiconductor devices are key components that have enabled significant advances in power electronic systems. Historically, silicon has been the material of choice for power semiconductor devices such as diodes, transistors and thyristors. However, silicon devices are now reaching their fundamental limits, and a transition to wide bandgap semiconductors is critical to make further progress in the field. Among them, SiC (silicon carbide) has attracted increasing attention as a power semiconductor to replace silicon due to its superior properties and technological maturity. In fact, SiC power MOSFETs have been commercially available since 2011, and are actively replacing their silicon counterparts at blocking voltages above 1 kV. At these voltages, the specific on-resistance of SiC MOSFETs is 200-300x lower than that of silicon devices. However, conventional vertical SiC MOSFETs are still far from their theoretical performance at blocking voltages below 2 kV. In this regime, the channel resistance is the dominant limitation due to the relatively low channel mobility at the SiO2/4H-SiC MOS interface.<br></p><p> </p><p>In this thesis, the first successful demonstration of a novel power device in 4H-SiC called the trigate power DMOSFET (double diffused metal oxide semiconductor field effect transistor) is presented. This device reduces the channel resistance by a factor of 3-5× compared with the state-of-art commercial power DMOSFETs, without requiring an increase in the channel mobility. The trigate structure is applied to a power MOSFET for the first time along with a self-aligned short channel process. This new structure utilizes both the conventional horizontal surface as well as the sidewalls of a trench to increase the effective width of the channel without increasing the device area. Conceptual design, optimization, process development and electrical results are presented. The trigate power MOSFET with a trench depth of 1 μm designed for a blocking voltage of 650 V has a specific on-resistance of 1.98 mΩcm²and a channel resistance of 0.67 mΩcm².This corresponds to a ∼2× reduction in the total specific on-resistance, and a 3.3× reduction in the specific channel resistance as compared to a conventional DMOSFET with the same blocking voltage rating. This demonstration is a landmark that could help SiC technology compete successfully in the lower blocking voltage regime below 600 V, and access for the first time a completely new segment in the power electronics application space.</p>

No description available.