Current AlGaN/GaN Heterojunction Field Effect Transistors (HFETs) make use of multiple sources, drains, and gates in parallel to maximize transconductance and effective gain while minimizing the current density through each channel. To connect the sources to a common ground, current practice prescribes the fabrication of air bridges above the gates and drains. This practice has the advantage of a low dielectric constant and low parasitic capacitance, but it is at the expense of manufacturability and robust device operation. In the study described below, the air bridges in AlGaN/GaN HFETs were replaced by a polymer supported metallization bridge with the intention of improving ease of fabrication and reliability. The DC, high frequency, and power performance for several polymer step heights were investigated. The resultant structures were functional and robust; however, their electrical performance was degraded due to high source resistance. The cause of the high source resistance was found to be thinning of the metallization at the polymer step. The effect was more pronounced for higher step heights. / Master of Science

L'évolution des composants d'électronique de puissance se heurte aujourd'hui aux limites physiques du silicium. L'utilisation des semi-conducteurs à large bande interdite permettraient de dépasser ces limites. Parmi ces nouveaux matériaux, le diamant possède les propriétés les plus intéressantes pour l'électronique de puissance : champ de rupture et conductivité thermique les plus élevés parmi les solides, grandes mobilités des porteurs électriques, possibilité de fonctionnement à haute température. Les substrats de diamant synthétisés actuellement par des méthodes de dépôt en phase vapeur ont des caractéristiques cristallographiques compatibles avec l'exploitation de ces propriétés en électronique de puissance. L'utilisation technologique du diamant reste toutefois difficile ; ses propriétés de dureté et d'inertie chimique rendent son utilisation délicate. L'objet de ces travaux est dans un premier temps d'évaluer les bénéfices que pourrait apporter le diamant en électronique de puissance. Ensuite, différentes étapes technologiques nécessaires à la fabrication de composants sur diamant sont étudiées : dépôts de contacts électriques, dopage et gravure ionique. Enfin, une étude sur la fabrication de diodes Schottky est présentée. Les résultats obtenus permettent d'établir les perspectives à ces travaux et les challenges scientifiques et technologiques qu'il reste à relever. / The evolution of power electronic devices is getting more and more limited by the silicon intrinsic properties. This limitation could be overcome by using wide bandgap semiconductors. Among these materials, diamond properties are the more fitted for power electronics: the highest critical electric field and thermal conductivity amongst the solids, high carriers mobility, high temperature operation possibility. At this time, diamond samples grown by chemical vapour deposition methods exhibit crystallographic properties that are suitable for a use in power electronics. Though, the realization of diamond power devices remains difficult due to its hardness and chemical inertness, among others. First, this work aims at determining the profit that could represent diamond for power electronics. Second, different technologic steps that are necessary to the realisation of electronic devices are studied: ohmic contacts deposition, doping and ion etching. Finally, the first devices we realised, Schottky diodes, are presented. Their characterisation allows establishing new objectives for the future developments of our studies.

付記する学位プログラム名: 京都大学卓越大学院プログラム「先端光・電子デバイス創成学」 / 京都大学 / 新制・課程博士 / 博士(工学) / 甲第23909号 / 工博第4996号 / 新制||工||1780(附属図書館) / 京都大学大学院工学研究科電子工学専攻 / (主査)教授 木本 恒暢, 教授 川上 養一, 准教授 安藤 裕一郎 / 学位規則第4条第1項該当 / Doctor of Philosophy (Engineering) / Kyoto University / DFAM

京都大学 / 0048 / 新制・課程博士 / 博士(工学) / 甲第19997号 / 工博第4241号 / 新制||工||1656(附属図書館) / 33093 / 京都大学大学院工学研究科電子工学専攻 / (主査)教授 木本 恒暢, 教授 藤田 静雄, 准教授 船戸 充 / 学位規則第4条第1項該当 / Doctor of Philosophy (Engineering) / Kyoto University / DGAM

Interest in wide bandgap semiconductors such as silicon carbide (SiC), gallium nitride (GaN), gallium oxide (Ga 2 O 3 ) and diamond has increased due to their ability to deliver high power, high switching frequency and low loss electronic devices for power conversion applications. To meet these requirements, semiconductor material defects, introduced during growth and fabrication, must be minimized. Otherwise, theoretical limits of operation cannot be achieved. In this dissertation, the non-ideal current- voltage (IV) behavior of GaN-based Schottky diodes is discussed first. Here, a new model is developed to explain better the temperature dependent performance typically associated with a multi-Gaussian distribution of barrier heights at the metal-semiconductor interface [Section 3.1]. Application of this model gives researches a means of understanding not only the effective barrier distribution at the MS interface but also its voltage dependence. With this information, the consequence that material growth and device fabrication methods have on the electrical characteristics can be better understood. To show its applicability, the new model is applied to Ru/GaN Schottky diodes annealed at increasing temperature under normal laboratory air, revealing that the origin of excess reverse leakage current is attributed to the low-side inhomogeneous barrier distribution tail [Section 3.2]. Secondly, challenges encountered during MOCVD growth of low-doped GaN drift layers for high-voltage operation are discussed with focus given to ongoing research characterizing deep-level defect incorporation by deep level transient spectroscopy (DLTS) and deep level optical spectroscopy (DLOS) [Section 3.3 and 3.4]. It is shown that simply increasing TMGa so that high growth rates (>4 µm/hr) can be achieved will cause the free carrier concentration and the electron mobilities in grown drift layers to decrease. Upon examination of the deep-level defect concentrations, it is found that this is likely caused by an increase in 4 deep level defects states located at E C - 2.30, 2.70, 2.90 and 3.20 eV. Finally, samples where the ammonia molar flow rate is increased while ensuring growth rate is kept at 2 µm/hr, the concentrations of the deep levels located at 0.62, 2.60, and 2.82 eV below the conduction band can be effectively lowered. This accomplishment marks an exciting new means by which the intrinsic impurity concentration in MOCVD-grown GaN films can be reduced so that >20 kV capable devices could be achieved. / Doctor of Philosophy / We constantly rely on electronics to help assist us in our everyday lives. However, to ensure functionality we require that they minimize the amount of energy lost through heat during operation. One contribution to this inefficiency is incurred during electrical power conversion. Examples of power conversion include converting from the 120 V wall outlet to the 5 V charging voltage used by cellphones or converting the fluctuating voltage from a solar panel (due to varying sun exposure) to the 120 V AC power found in a typical household. Electrical circuits can be simply designed to accomplish these conversions; however, consideration to every component must be given to ensure high efficiency. A popular example of an electrical power conversion circuit is one that switches the input voltage on and off at high rates and smooths the output with an inductor/capacitor network. A good analogy of this process is trying to create a small stream of water from a fire hydrant which can either be off or on at full power. Here we can use a small cup but turning the fire hydrant on and trying to fill the cup will destroy it. However, if the fire hydrant is pulsed on and off at very short intervals (1 µs), its possible to fill the cup without damaging it or having it overflow. Now, under ideal circumstances if a small hole is poked in the bottom of the cup and the interval of the fire hydrant is timed correctly, a small low power stream of water is created without overflowing the cup and wasting water. In this analogy, a devices capable of switching the stream of water on and off very fast would need to be implemented. In electrical power conversion circuits this device is typically a transistor and diode network created from a semiconducting material. Here, similar to the fire hydrant analogy, a switch would need to be capable of holding off the immense power when in the off position and not impeding the powerful flow when in the on position. The theoretical limit of these two characteristics is dependent on the material properties of the switch where typically used semiconductors include silicon (Si), silicon carbide (SiC), or gallium nitride (GaN). Currently, GaN is considered to be a superior option over Si or SiC to make the power semiconductor switching device, however research is still required to remove non-ideal behavior that ultimately effects power conversion efficiency. In this work, we first examine the spurious behavior in GaN-based Schottky diodes and effectively create a new model to describe the observed behavior. Next, we fabricated Ru/GaN Schottky diodes annealed at different temperatures and applied the model to explain the room-temperature electrical characteristics. Finally, we grew GaN under different conditions (varying TMGa and ammonia) so that quantum characteristics, which have been shown to affect the overall ability of the device, could be measured.

Diamond is a semiconductor with many superior material properties such as high breakdown field, high saturation velocity, high carrier mobilities and the highest thermal conductivity of all materials. These extreme properties, as compared to other (wide bandgap) semiconductors, make it desirable to develop single-crystalline epitaxial diamond films for electronic device and detector applications. Future diamond devices, such as power diodes, photoconductive switches and high-frequency field effect transistors, could in principle deliver outstanding performance due to diamond's excellent intrinsic properties. However, such electronic applications put severe demands on the crystalline quality of the material. Many fundamental electronic properties of diamond are still poorly understood, which severely holds back diamond-based electronic device and detector development. This problem is largely due to incomplete knowledge of the defects in the material and due to a lack of understanding of how these defects influence transport properties. Since diamond lacks a shallow dopant that is fully thermally activated at room temperature, the conventional silicon semiconductor technology cannot be transferred to diamond devices; instead, new concepts have to be developed. Some of the more promising device concepts contain thin delta-doped layers with a very high dopant concentration, which are fully activated in conjunction with undoped (intrinsic) layers where charges are transported. Thus, it is crucial to better understand transport in high-quality undoped layers with high carrier mobilities. The focus of this doctoral thesis is therefore the study of charge transport and related electronic properties of single-crystalline plasma-deposited (SC-CVD) diamond samples, in order to improve knowledge on charge creation and transport mechanisms. Fundamental characteristics such as drift mobilities, compensation ratios and average pair-creation energy were measured. Comparing them with theoretical predictions from simulations allows for verification of these models and improvement of the diamond deposition process.

<p>Recent developments in semiconductor device technology have seen the advent of wide-bandgap (WBG) based devices that enable operation at high switching frequencies. These devices, such as silicon carbide (SiC) metal-oxide-semiconductor field-effect transistors (MOSFETs), are becoming a favored choice in inverters for electric drive systems because of their lower switching losses and higher allowable operating temperature. However, the fast switching of such devices implies increased voltage edge rates (high <i>dv/dt</i>) that give rise to various undesirable effects including large common-mode currents, electromagnetic interference, transient overvoltages, insulation failure due to the overvoltages, and bearing failures due to</p> <p>microarcs. With increased use of these devices in transportation and industrial applications, it is imperative that accurate models and efficient simulation tools, which can predict these high-frequency effects and accompanying system losses, be established. This research initially focuses on establishing an accurate wideband model of a surface-mount permanent-magnet</p> <p>ac machine supplied by a WBG-based inverter. A new multirate simulation framework for predicting the transient behavior and estimating the power losses is then set forth. In this approach,</p> <p>the wideband model is separated into high- and low-frequency models implemented using two different computer programs that are best suited for the respective time scales. Repetitive execution of the high-frequency model yields look-up tables for the switching losses in the semiconductors, electric machine, and interconnecting cable. These look-up tables are then incorporated into the low-frequency model that establishes the conduction</p> <p>losses. This method is applied to a WBG-based electric drive comprised of a SiC inverter and permanent-magnet ac machine. Comparisons of measured and simulated transients are provided.</p>

Enhancing performance of power electronics converters has always been an interesting topic in the power electronics community. Over the years, researchers and engineers are developing new high performance component, novel converter topologies, smart control methods and optimal design procedures to improve the efficiency, power density, reliability and reducing the cost. Besides pursuing high performance, researchers and engineers are striving to modularize the power electronics converters, which provides redundancy, flexibility and standardization to the end users. The trend of modularization has been seen in photovoltaic inverters, telecommunication power supplies, and recently, HVDC applications. A systematic optimal design approach for modular power converters is developed in this dissertation. The converters are developed for aerospace applications where there are stringent requirement on converter form factor, loss dissipation, thermal management and electromagnetic interference (EMI) performance. This work proposed an optimal design approach to maximize the nominal power of the power converters considering all the constraints, which fully reveals the power processing potential. Specifically, this work studied three-phase active front-end converter, three-phase isolated ac/dc converter and inverter. The key models (with special attention paid to semiconductor switching loss model), detailed design procedures and key design considerations are elaborated. With the proposed design framework, influence of key design variables, e.g. converter topology, switching frequency, etc. is thoroughly studied. Besides optimal design procedure, control issues in paralleling modular converters are discussed. A master-slave control architecture is used. The slave controllers not only follow the command broadcasted by the master controller, but also synchronize the high frequency clock to the master controller. The control architecture eliminates the communication between the slave controllers but keeps paralleled modules well synchronized, enabling a fully modularized design. Furthermore, the implementation issues of modularity are discussed. Although modularizing converters under form factor constraints adds flexibility to the system, it limits the design space by forbidding oversized components. This work studies the influence of the form factor by exploring the maximal nominal power of a double-sized converter module and comparing it with that of two paralleled modules. The tradeoff between modularity and performance is revealed by this study. Another implementation issue is related to EMI. Scaling up system capacity by paralleling converter modules induces EMI issues in both signal level and system level. This work investigates the mechanisms and provides solutions to the EMI problems. / Ph. D. / As penetration of power electronics technologies in electric power delivery keeps increasing, performance of power electronics converters becomes a key factor in energy delivery efficacy and sustainability. Enhancing performance of power electronics converters reduces footprint, energy waste and delivery cost, and ultimately, promoting a sustainable energy use. Over the years, researchers and engineers are developing new technologies, including high performance component, novel converter topologies, smart control methods and optimal design procedures to improve the efficiency, power density, reliability and reducing the cost of power electronics converters. Besides pursuing high performance, researchers and engineers are striving to modularize the power electronics converters, enabling power electronics converters to be used in a “plug-and-play” fashion. Modularization provides redundancy, flexibility and standardization to the end users. The trend of modularization has been seen in applications that process electric power from several Watts to Megawatts. This dissertation discusses the design framework for incorporating modularization into existing converter design procedure, synergically achieving performance optimization and modularity. A systematic optimal design approach for modular power converters is developed in this dissertation. The converters are developed for aerospace applications where there is stringent v requirement on converter dimensions, loss dissipation, and thermal management. Besides, to ensure stable operation of the onboard power system, filters comprising of inductors and capacitors are necessary to reduce the electromagnetic interference (EMI). Owning to the considerable weight and size of the inductors and capacitors, filter design is one of the key component in converter design. This work proposed an optimal design approach that synergically optimizes performance and promotes modularity while complying with the entire aerospace requirement. Specifically, this work studied three-phase active front-end converter, three-phase isolated ac/dc converter and three-phase inverter. The key models, detailed design procedures and key design considerations are elaborated. Experimental results validate the design framework and key models, and demonstrates cutting-edge converter performance. To enable a fully modularized design, control of modular converters, with focus on synchronizing the modular converters, is discussed. This work proposed a communication structure that minimizes communication resources and achieves seamless synchronization among multiple modular converters that operate in parallel. The communication scheme is demonstrated by experiments. Besides, the implementation issues of modularity are discussed. Although modularizing converters under form factor constraints adds flexibility to the system, it limits the design space by forbidding oversized components. This work studies the impact of modularity by comparing performance of a double-sized converter module with two paralleled modules. The tradeoff between modularity and performance is revealed by this study.