Dissertations / Theses on the topic 'Gnn'

La Dynamique des Fluides Numérique (CFD) joue un rôle essentiel dans la prédiction de divers phénomènes physiques, tels que le climat, l'aérodynamique ou la circulation sanguine. Au coeur de la CFD se trouvent les équations de Navier-Stokes régissant le mouvement des fluides. Cependant, résoudre ces équations à grande échelle reste fastidieux, en particulier lorsqu'il s'agit des équations de Navier-Stokes incompressibles, qui nécessitent la résolution intensive d'un problème de Poisson de Pression, garantissant la contrainte d'incompressibilité. De nos jours, les méthodes d'apprentissage profond ont ouvert de nouvelles perspectives pour améliorer les simulations numériques. Parmi ces approches, les Graph Neural Networks (GNNs), conçus pour traiter des données de type graphe tels que les maillages, se sont révélés prometteurs. Cette thèse vise à explorer l'utilisation des GNNs pour améliorer la résolution du problème de Poisson de Pression. Une contribution clé implique l'introduction d'une nouvelle architecture GNN qui respecte intrinsèquement les conditions aux limites tout en exploitant la théorie des couches implicites pour ajuster automatiquement le nombre de couches GNN nécessaires à la convergence : ce nouveau modèle présente des capacités de généralisation améliorées, gérant efficacement des problèmes de Poisson de différentes tailles et formes. Néanmoins, ses limitations actuelles le restreignent aux problèmes à petite échelle, insuffisants pour les applications industrielles qui nécessitent souvent plusieurs milliers de noeuds. Pour mettre à l'échelle ces modèles, cette thèse explore la combinaison des GNNs avec les méthodes de Décomposition de Domaines, tirant parti des calculs en parallèle sur GPU pour produire des solutions d'ingénierie plus efficaces
Computational Fluid Dynamics (CFD) plays an essential role in predicting various physical phenomena, such as climate, aerodynamics, or blood flow. At the core of CFD lie the Navier-Stokes equations governing the motion of fluids. However, solving these equations at scale remains daunting, especially when dealing with Incompressible Navier-Stokes equations. Indeed, the well-known splitting schemes require the costly resolution of a Pressure Poisson problem that guarantees the incompressibility constraint. Nowadays, Deep Learning methods have opened new perspectives for enhancing numerical simulations. Among existing approaches, Graph Neural Networks (GNNs), designed to handle graph data like meshes, have proven to be promising. This thesis is dedicated to exploring the use of GNNs to enhance the resolution of the Pressure Poisson problem. One significant contribution involves introducing a novel physics-informed GNN-based model that inherently respects boundary conditions while leveraging the Implicit Layer theory to automatically adjust the number of GNN layers required for convergence. This results in a model with enhanced generalization capabilities, effectively handling Poisson problems of various sizes and shapes. Nevertheless, its current limitations restrict it to small-scale problems, insufficient for industrial applications that often require thousands of nodes. To scale up these models, this thesis further explores combining GNNs with Domain Decomposition methods, taking advantage of batch parallel computing on GPU to produce more efficient engineering solutions

Many people - especially during their elderly - consume multiple drugs for the treatment of complex or co-existing diseases. Identifying side effects caused by polypharmacy is crucial for reducing mortality and morbidity of the patients which will lead to improvement in their quality of life. Since there is immense space for possible drug combinations, it is infeasible to examine them entirely in the lab. In silico models can offer a convenient solution, however, due to the lack of a sufficient amount of homogenous data it is difficult to develop both reliable and scalable models in its ability to accurately predict Polypharmacy Side Effect. Recent advancement in the field of representational learning has utilized the power of graph networks to harmonize information from the heterogeneous biological databases and interactomes. This thesis takes advantage of those techniques and incorporates them with the state-of-the-art Graph Neural Network algorithms to implement a Deep learning pipeline capable of predicting the Adverse Drug Reaction of any given paired drug combinations.

Le reti neurali artificiali hanno visto, negli ultimi anni, una crescita vertiginosa nelle loro applicazioni e nelle architetture dei modelli impiegati. In questa tesi introduciamo le reti neurali su domini euclidei, in particolare mostrando l’importanza dell’equivarianza di traslazione nelle reti convoluzionali, e introduciamo, per analogia, un’estensione della convoluzione a dati strutturati come grafi. Inoltre presentiamo le architetture dei principali Graph Neural Network ed esponiamo, per ognuna delle tre architetture proposte (Spectral graph Convolutional Network, Graph Convolutional Network, Graph Attention neTwork) un’applicazione che ne mostri sia il funzionamento che l’importanza. Discutiamo, ulteriormente, l’implementazione di un algoritmo di classificazione basato su due varianti dell’architettura Graph Convolutional Network, addestrato e testato sul dataset PROTEINS, capace di classificare le proteine del dataset in due categorie: enzimi e non enzimi.

While there are existing methods of gamma ray-track reconstruction in specialized detectors such as AGATA, including backtracking and clustering, it is naturally of interest to diversify the portfolio of available tools to provide us viable alternatives. In this study some possibilities found in the field of machine learning were investigated, more specifically within the field of graph neural networks. In this project there was attempt to reconstruct gamma tracks in a germanium solid using data simulated in Geant4. The data consists of photon energies below the pair production limit and so we are limited to the processes of photoelectric absorption and Compton scattering. The author turned to the field of graph networks to utilize its edge and node structure for data of such variable input size as found in this task. A graph neural network (GNN) was implemented and trained on a variety of gamma multiplicities and energies and was subsequently tested in terms of various accuracy parameters and generated energy spectra. In the end the best result involved an edge classifier trained on a large dataset containing a 10^6 tracks bundled together into separate events to be resolved. The network was capable of recalling up to 95 percent of the connective edges for the selected threshold in the infinite resolution case with a peak-to-total ratio of 68 percent for a set of packed data with a model trained on simulated data including realistic uncertainties in both position and energy.
Trots att det existerar en mängd metoder för rekonstruktion av spår i specialiserade detektorer som AGATA är det av naturligt intresse att diversifiera och undersöka nya verktyg för uppgiften. I denna studie undersöktes några möjligheter inom maskininlärning, närmare bestämt inom området neurala grafnätverk.  Under projektets gång simulerades data av fotoner i en ihålig, sfärisk geometri av germanium i Geant4. Den simulerade datan är begränsad till energier under parproduktion så datan består av reaktioner genom den fotoelektriska effekten och comptonspridning. Den variabla storleken på denna data och dess spridning i detektorns geometri lämpar sig för ett grafformat med nod och länkstruktur. Ett neuralt grafnätverk (GNN) implementerades och tränades på data med gamma av variabel multiplicitet och energi och evaluerades på ett urval prestandaparametrar och dess förmåga att generera ett användbart spektra. Slutresultatet involverade en länkklassificerings modell tränat på data med 10^6 spår sammanslagna till händelser. Nätverket återkallade 95 procent av positiva länkar för ett val av tröskelvärde i fallet med oändlig upplösning med ett peak-to-total på 68 procent för packad data behandlad med osäkerhet i energi och position motsvarande fallet med begränsad upplösning.

While there are existing methods of gamma ray-track reconstruction in specialized detectors such as AGATA, including backtracking and clustering, it is naturally of interest to diversify the portfolio of available tools to provide us viable alternatives. In this study some possibilities found in the field of machine learning were investigated, more specifically within the field of graph neural networks. In this project there was attempt to reconstruct gamma tracks in a germanium solid using data simulated in Geant4. The data consists of photon energies below the pair production limit and so we are limited to the processes of photoelectric absorption and Compton scattering. The author turned to the field of graph networks to utilize its edge and node structure for data of such variable input size as found in this task. A graph neural network (GNN) was implemented and trained on a variety of gamma multiplicities and energies and was subsequently tested in terms of various accuracy parameters and generated energy spectra. In the end the best result involved an edge classifier trained on a large dataset containing a 10^6 tracks bundled together into separate events to be resolved. The network was capable of recalling up to 95 percent of the connective edges for the selected threshold in the infinite resolution case with a peak-to-total ratio of 68 percent for a set of packed data with a model trained on simulated data including realistic uncertainties in both position and energy.
Trots att det existerar en mängd metoder för rekonstruktion av spår i specialiserade detektorer som AGATA är det av naturligt intresse att diversifiera och undersöka nya verktyg för uppgiften. I denna studie undersöktes några möjligheter inom maskininlärning, närmare bestämt inom området neurala grafnätverk.  Under projektets gång simulerades data av fotoner i en ihålig, sfärisk geometri av germanium i Geant4. Den simulerade datan är begränsad till energier under parproduktion så datan består av reaktioner genom den fotoelektriska effekten och comptonspridning. Den variabla storleken på denna data och dess spridning i detektorns geometri lämpar sig för ett grafformat med nod och länkstruktur. Ett neuralt grafnätverk (GNN) implementerades och tränades på data med gamma av variabel multiplicitet och energi och evaluerades på ett urval prestandaparametrar och dess förmåga att generera ett användbart spektra. Slutresultatet involverade en länkklassificerings modell tränat på data med 10^6 spår sammanslagna till händelser. Nätverket återkallade 95 procent av positiva länkar för ett val av tröskelvärde i fallet med oändlig upplösning med ett peak-to-total på 68 procent för packad data behandlad med osäkerhet i energi och position motsvarande fallet med begränsad upplösning.

This report is done in collaboration with WirelessCar for the master of science thesis at Halmstad University. Many different parameters influence fuel consumption. The objective of the report is to evaluate if Graph neural networks are a practical model to perform fuel consumption prediction on areas. The model uses a partitioning of geographical locations of trip observations to capture their spatial information. The project also proposes a method to capture the non-stationary behavior of vehicles by defining a vehicle node as a separate entity. The model then captures their different features in a dense layer neural network and utilizes message passing to capture context about neighboring nodes. The model is compared to a baseline neural network with a similar network architecture as the graph neural network. The data is partitioned to define an area with Kmeans and static gridnet partition with and without terrain details. This partition is used to structure a homogeneous graph that is underperforming. The practical drawbacks of the initial homogeneous graph are inspected and addressed to develop a heterogeneous graph that can outperform the neural network baseline.

This diploma thesis describes and implements the design of a graph neural network usedfor 2D segmentation of neural structure. The first chapter of the thesis briefly introduces the problem of segmentation. In this chapter, segmentation techniques are divided according to the principles of the methods they use. Each type of technique contains the essence of this category as well as a description of one representative. The second chapter of the diploma thesis explains graph neural networks (GNN for short). Here, the thesis divides graph neural networks in general and describes recurrent graph neural networks(RGNN for short) and graph autoencoders, that can be used for image segmentation, in more detail. The specific image segmentation solution is based on the message passing method in RGNN, which can replace convolution masks in convolutional neural networks.RGNN also provides a simpler multilayer perceptron topology. The second type of graph neural networks characterised in the thesis are graph autoencoders, which use various methods for better encoding of graph vertices into Euclidean space. The last part ofthe diploma thesis deals with the analysis of the problem, the proposal of its specific solution and the evaluation of results. The purpose of the practical part of the work was the implementation of GNN for image data segmentation. The advantage of using neural networks is the ability to solve different types of segmentation by changing training data. RGNN with messaging passing and node2vec were used as implementation GNNf or segmentation problem. RGNN training was performed on graphics cards provided bythe school and Google Colaboratory. Learning RGNN using node2vec was very memory intensive and therefore it was necessary to train on a processor with an operating memory larger than 12GB. As part of the RGNN optimization, learning was tested using various loss functions, changing topology and learning parameters. A tree structure method was developed to use node2vec to improve segmentation, but the results did not confirman improvement for a small number of iterations. The best outcomes of the practical implementation were evaluated by comparing the tested data with the convolutional neural network U-Net. It is possible to state comparable results to the U-Net network, but further testing is needed to compare these neural networks. The result of the thesisis the use of RGNN as a modern solution to the problem of image segmentation and providing a foundation for further research.

L'industrie 4.0 est un changement important et un défi de taille pour chaque segment industriel. La recherche utilise d'abord l'analyse de la littérature pour trier la littérature et énumérer la direction du développement et l'état d'application de différents domaines, ce qui se consacre à montrer un rôle de premier plan pour la théorie et la pratique de l'industrie 4.0. La recherche explore ensuite la tendance principale de l'offre à plusieurs niveaux dans l'industrie 4.0 en combinant l'apprentissage automatique et les méthodes traditionnelles. Ensuite, la recherche examine la relation entre l'investissement et l'emploi dans l'industrie 4.0 pour examiner la dépendance interrégionale de l'industrie 4.0 afin de présenter un regroupement raisonnable basé sur différents critères et de faire des suggestions et une analyse de la chaîne d'approvisionnement mondiale pour les entreprises et les organisations.De plus, notre système d'analyse jette un coup d'oeil sur la macroéconomie. La combinaison du traitement du langage naturel dans l'apprentissage automatique pour classer les sujets de recherche et de la revue de la littérature traditionnelle pour enquêter sur la chaîne d'approvisionnement à plusieurs niveaux améliore considérablement l'objectivité de l'étude et jette une base solide pour des recherches ultérieures. L'utilisation de réseaux et d'économétrie complexes pour analyser la chaîne d'approvisionnement mondiale et les problèmes macroéconomiques enrichit la méthodologie de recherche au niveau macro et politique. Cette recherche fournit des analyses et des références aux chercheurs, aux décideurs et aux entreprises pour leur prise de décision stratégique
Industry 4.0 is a significant shift and a tremendous challenge for every industrial segment, especially for the manufacturing industry that gave birth to the new industrial revolution. The research first uses literature analysis to sort out the literature, and focuses on the use of “core literature extension method” to enumerate the development direction and application status of different fields, which devotes to showing a leading role for theory and practice of industry 4.0. The research then explores the main trend of multi-tier supply in Industry 4.0 by combining machine learning and traditional methods. Next, the research investigates the relationship of industry 4.0 investment and employment to look into the inter-regional dependence of industry 4.0 so as to present a reasonable clustering based on different criteria and make suggestions and analysis of the global supply chain for enterprises and organizations. Furthermore, our analysis system takes a glance at the macroeconomy. The combination of natural language processing in machine learning to classify research topics and traditional literature review to investigate the multi-tier supply chain significantly improves the study's objectivity and lays a solid foundation for further research. Using complex networks and econometrics to analyze the global supply chain and macroeconomic issues enriches the research methodology at the macro and policy level. This research provides analysis and references to researchers, decision-makers, and companies for their strategic decision-making

This thesis proposes an introduction to the fundamental concepts of supervised deep learning. Starting from Rosemblatt's Perceptron we will discuss the architectures that, in recent years, have revolutioned the world of deep learning: graph neural networks, which led to the formulation of geometric deep learning. We will then give a simple example of graph neural network, discussing the code that composes it and then test our architecture on the MNISTSuperpixels dataset, which is a variation of the benchmark dataset MNIST.

Thesis (DTech (Electrical Engineering))--Cape Peninsula University of Technology, 2017.
Gunn diodes play a prominent role in the development of low-cost and reliable solid-state oscillators for diverse applications, such as in the military, security, automotive and consumer electronics industries. The primary focus of the research presented here is the optimisation of GaAs and GaN Gunn diodes for mm-wave operations, through rigorous Monte Carlo particle simulations. A novel, empirical technique to determine the upper operational frequency limit of devices based on the transferred electron mechanism is presented. This method exploits the hysteresis of the dynamic velocity-field curves of semiconductors to establish the upper frequency limit of the transferred electron mechanism in bulk material that supports this mechanism. The method can be applied to any bulk material exhibiting negative differential resistance. The simulations show that the upper frequency limits of the fundamental mode of operation for GaAs Gunn diodes are between 80 GHz and 100 GHz, and for GaN Gunn diodes between 250 GHz and 300 GHz, depending on the operating conditions. These results, based on the simulated bulk material characteristics, are confirmed by the simulated mm-wave performance of the GaAs and GaN Gunn devices. GaAs diodes are shown to exhibit a fundamental frequency limit of 90 GHz, but with harmonic power available up to 186_GHz. Simulated GaN diodes are capable of generating appreciable output power at operational frequencies up to 250 GHz in the fundamental mode, with harmonic output power available up to 525 GHz. The research furthermore establishes optimised doping profiles for two-domain GaAs Gunn diodes and single- and two-domain GaN Gunn diodes. The relevant design parameters that have been optimised, are the dimensions and doping profile of the transit regions, the width of the doping notches and buffer region (for two-domain devices), and the bias voltage. In the case of GaAs diodes, hot electron injection has also been implemented to improve the efficiency and output power of the devices. Multi-domain operation has been explored for both GaAs and GaN devices and found to be an effective way of increasing the output power. However, it is the opinion of the author that a maximum number of two domains is feasible for both GaAs and GaN diodes due to the significant increase in thermal heating associated with an increase in the number of transit regions. It has also been found that increasing the doping concentration of the transit region exponentially over the last 25% towards the anode by a factor of 1.5 above the nominal doping level enhances the output power of the diodes.

Cette étude, syntaxique et textuelle, se penche sur le problème de la définition de l’apposition. Les grammaires de l’allemand et les études linguistiques consacrées à l’apposition font apparaître de grandes divergences quant à l’extension de la notion qui reste encore empreinte du poids de la tradition grammaticale. Les critères relevés dans les différentes définitions (accord casuel, accolage, identité référentielle, non restrictivité, suppressibilité, parenté avec une structure propositionnelle ou relation prédicative) sont soumis à un examen critique qui motive une redéfinition en intension et en extension de l’objet d’étude : la construction détachée à support nominal est une prédication supplémentaire averbale dans laquelle l’apport GN2 peut morphosyntaxiquement toujours se positionner directement à droite du support GN1. Cette définition permet de distinguer la construction détachée de phénomènes considérés comme appositifs (le GN en als et wie, la dislocation, la construction absolue).La mise à distance de GN2 par rapport à GN1 s’explique par la répartition et la hiérarchisation de l’information dans l’énoncé et au-delà, dans le texte. La construction détachée apporte les éléments pertinents pour la compréhension du texte. Cette fonction explicative instaure et entretient l’ethos coopératif du locuteur-scripteur dont celui-ci peut jouer à des fins argumentatives.Le corpus d’étude (volume 2) est composé de textes de presse et de textes littéraires contemporains
This syntactic and textual study examines the various ways apposition can be defined. German grammars and linguistic studies devoted to apposition vary widely in the range of that notion, which is still very much under the influence of grammatical tradition. The various criteria which appear in definitions (case agreement, adjacency, referential identity, non-restrictiveness, deletability, similarity to a grammatical clause, subject-predicate relation) are critically examined, which leads to a redefinition of the object of our study, both intensionally and extentionally: noun-based detached constructions can be seen as additional averbal predications in which the appositive NP2 can always be morphosyntactically positioned immediately to the right of the NP1 base. Our definition makes it possible to distinguish detached constructions from phenomena which are usually considered as appositive (als- or wie-introduced NPs, dislocations, absolute constructions).NP2 displacement away from NP1 results from the distribution and prioritization of information within utterances, and further at the textual level. Detached constructions provide relevant elements for textual comprehension. This explanatory function creates and maintains a high level of mutual comprehension between the speaker/writer and the reader, which the speaker can use for argumentative purposes.The corpus of our study (vol. 2) is made up of newspaper and magazine articles and contemporary literary texts

The growth of thin-film metal nitride materials from elemental metal targets by plasma-assisted pulsed laser deposition (PLD) has been explored and analysed. A new UHV PLD growth system has been installed and assembled and its system elements were calibrated. A series of GaN thin films have been grown to calibrate the system. In-situ RHEED indicated that the films were single crystal and that growth proceeded in a three-dimensional fashion. SEM images showed heavy particulation of film surfaces that was not in evidence for later refractory metal nitride films. This may be connected to the fact that Ga targets were liquid while refractory metals were solid. Most GaN films were not continuous due to insufficient laser fluence. Continuous films did not exhibit photoluminescence. HfN films have been grown by PLD for the first time. Films grown have been shown to have high reflectivity in the visible region and low resistivity. These factors, along with their crystal structure, make them suitable candidates to be used as back-contacts in GaN LEDs and could also serve as buffer layers to enable the integration of GaN and Si technologies. Growth factors affecting the films’ final properties have been investigated. Nitrogen pressure, within the operating range of the plasma source, has been shown to have little effect on HfN films. Substrate temperature has been demonstrated to have more influence on the films’ properties, with 500 °C being established as optimum. ZrN films have also been grown by PLD. Early results indicated that they exhibit reflectivities 50 % ± 5 % lower than those of HfN. However, further growth and characterisation would be required in order to establish this as a fundamental property of ZrN as nitride targets were mostly used in ZrN production. Single-crystal epitaxial GdN and SmN films have been produced by PLD. This represents an improvement in the existing quality of GdN films reported in the literature, which are mostly polycrystalline. In the case of SmN, these are the first epitaxial films of this material to be grown. Film quality has been monitored in-situ by RHEED which has allowed growth to be tailored to produce ever-higher crystal quality. Post-growth analyses by collaborators was also of assistance in improving film growth. Substrate temperatures and nitrogen plasma parameters have been adjusted to find optimum values for each. In addition, laser fluence has been altered to minimise the presence of metal particulates in the films, which interfere with magnetic measurements carried out in analyses. Capping layers of Cr, YSZ or AlN have been deposited on the GdN and SmN prior to removal from vacuum to prevent their degradation upon exposure to atmospheric water vapour. The caps have been steadily improved over the course of this work, extending the lifetime of the nitride films in ambient. However, they remain volatile and this may persist since water vapour can enter the film at the edge regardless of capping quality. Optical transmission has shown an onset of absorption at 1.3 eV for GdN and 1.0 eV for SmN.

The panelists will be sharing information about gun violence in the U.S, proposed firearm legislation in Tennessee, research examining gun violence patterns in other states, and education to help parents and others prevent unintentional shootings by children.

O objetivo do trabalho é fazer a avaliação global de alguns dos modos energéticos de transporte do gás natural (gasodutos, GNL e GNC), inclusive como energia secundária (eletricidade e GTL). Para a escolha do melhor modo de transporte, além dos custos de cada projeto, serão considerados outros três fatores, que são: impactos ambientais, impactos sociais e riscos políticos. Para isto foi escolhida uma ferramenta de análise conhecida como Avaliação de Custos Completos (ACC), sendo que esta considera os quatro fatores citados anteriormente como de mesma importância na avaliação para um planejamento que vise a sustentabilidade. Para melhor visualização e entendimento do trabalho, foi feito um estudo de caso para o transporte de energia (gás natural e/ou eletricidade) para atender mercados consumidores na região de fronteira entre o Brasil e a Bolívia. Fundamentalmente o estudo do transporte foi feito em solo boliviano, correspondente entre os campos produtores de gás, localizados em Margarita, até a região de fronteira próxima a Corumbá no Mato Grosso do Sul (MS), e também próxima a uma região com grandes jazidas de minério de ferro conhecida como El Mutún. Metodologicamente dentro das quatro dimensões consideradas, foram usados alguns subcritérios mais relevantes para esta análise. Para todos os sub-critérios foram atribuídas notas específicas, de acordo com a região e com o tipo de transporte de energia, e após isto os dados foram passados para um software de tomada de decisão chamado Decision Lens, obtendo assim um rank das melhores opções transporte para a região. Como resultado, o mais coerente para atender a região, são gasodutos de transporte para as demandas elétricas e de gás natural (consumo do MS mais Mutún) e trens carregados com derivados líquidos do gás natural (GTL) para atender a demanda de diesel do MS.
The objective of this work is to make a global evaluation of some energy ways of transport of natural gas (gas-pipelines, LNG and CNG), as well as secondary energy (electricity and GTL). For the selection of a better way of transport, beyond the costs of each project, others three factors were considered: environmental impacts, social impacts and political risks. To perform this, a tool of analysis known as Evaluation of Complete Costs (ACC in portuguese) was chosen. This tool considers the four factors above mentioned as having the same importance in the evaluation process for planning that aims towards sustainability. For better visualization and understanding of the work, a case study for energy transport was made (natural gas and/or electricity) to supply the consumer markets in the region bordering Brazil and Bolivia. Basically the study of the transport was made in Bolivian territory; at the producing gas fields located in Margarita as far as the region bordering Corumbá in the Mato Grosso do Sul area, and also next to a region with great known iron ore deposits known as El Mutún. Methodologically, within the four dimensions above, some more relevant sub-criteria for the analysis were used. For all the sub-criteria considered specific ranks were taken, in accordance with the region and type of transported energy, and after this the data was fed to a software called Decision Lens, thus getting a rank of the best options to carry gas to the region. The results, most coherent way to supply gas to the region, happens to be gas-pipelines of transport for the electric and natural gas demands (consumption in MS plus Mutún) and trains with liquid derivatives of natural gas (GTL) to supply care of the demand of diesel of MS.

Gallium Nitride based devices due to their inherent material properties are considered as one of the most promising devices to realize high power, high frequency transistors with lower power consumption in next-generation applications. Although the technology has been studied since early 1970s, there is still a vast room and expectations in its yet unachieved findings. In present work the GaN technology is explored and state-of-the-art studies of GaN based HEMTs and their application in MMICs are presented. Different designs are presented and evaluated and the results are reported. In particular the HEMT performance is studied in terms of DC in addition to large signal conditions, where the device’s performance becomes function of power levels it is driven with. The peculiarities and challenges of building an automated Load-Pull setup are outlined and analysis for further improvements is presented.

Le nitrure de gallium (GaN) est un matériau en plein essor depuis le début des années 1990 pour des applications dans le domaine de l'optoélectronique telles que les diodes électroluminescentes (DELs) bleues ou blanches, les diodes lasers (DLs) bleues ou les détecteurs ultra-violets. L'activité épitaxie de GaN par la technique de croissance HVPE (Epitaxie en phase vapeur par la méthode aux hydrures), a vu le jour au LASMEA en 1998. Les premières études expérimentales et de modélisation réalisées sur des échantillons de faibles dimensions (surface d'environ 1 à 3 cm2) ont conduit à la mise en évidence des mécanismes de croissance et à la maîtrise du procédé. Le développement de ce matériau à l'échelle industrielle a nécessité de travailler sur des surfaces de dimension plus grandes de l'ordre de deux pouces. Un nouveau dispositif expérimantal HVPE a été conçu dans ce sens, mis en place au sein du laboratoire et le procédé a été validé. De nouvelles investigations ont été menées sur l'étude de l'épitaxie sélective de GaN pour la réalisation de structures périodiques de faible dimensionnalité à morphologies contrôlées. Des structures de morphologies poutres et pyramidales de GaN de 1 à 2 µm de large ont ainsi pu être épitaxiées par la technique HVPE. Une analyse systématique de la variation des conditions de croissance est effectuée, visant à maîtriser l'ensemble des paramètres qui influent sur les morphologies et les dimensions des structures. Cette étude est couplée à la compréhension des mécanismes de croissance mis en jeu au cours de l'épitaxie sélective de GaN

Very few studies have explored the relationship between the availability of certain types of firearms and gun density on both gun aggravated assaults and gun homicides. Nonetheless, research by Koper (2001) discovered that the availability of more lethal types of firearms, not gun density, was directly related to an increase in gun homicide rates for Dallas. However, this study did not take into account certain social and economic variables that may strengthen or weaken the determined relationship. The current study uses data previously analyzed by Koper (2001) and includes social and economic variables that have been linked to lethal violence while using gun aggravated assaults and gun homicides as the dependent variables. The results will help ascertain to what extent the impact of firearm availability on gun crimes is contingent on contextual factors.
M.A.
Department of Sociology
Sciences
Applied Sociology MA

Thesis (M.A.)--University of Missouri-Columbia, 2007.
The entire dissertation/thesis text is included in the research.pdf file; the official abstract appears in the short.pdf file (which also appears in the research.pdf); a non-technical general description, or public abstract, appears in the public.pdf file as well as 10 media.jpg files. Title from title screen of research.pdf file (viewed on November 12, 2008) Includes bibliographical references.

Monte Carlo modelling is a common technique in numerous fields, and is widely used in semiconductor device simulation. This thesis describes the application of Monte Carlo modelling to the simulation of Gunn diode devices, focusing on devices composed of Gallium Arsenide (GaAs) and Gallium Nitride (GaN). Gunn diodes are simple structures that take advantage of negative differential resistance to act as a source of high frequency radiation, from 10 GHz to over 100 GHz in GaAs devices. It has been theorised that GaN should exhibit negative differential resistance and a GaN Gunn diode could produce radiation of even higher frequency, within the terahertz band. Gunn diodes have the advantage of being cheap and portable, and so are worth exploring as such a source. Unfortunately, GaN devices have a high electron density and so they tend to generate heat quickly. It therefore becomes important to include modelling of heat generation and flow in simulations of these devices. This is uncommon in Monte Carlo models of Gunn diodes, as in less highly doped devices thermal effects can usually be assumed to result in the device reaching an equilibrium temperature of about 100 K above the ambient. This thesis describes the creation of a model to track the generation and distribution of heat during operation of a GaN device. Simulations found that thermal effects within the device were significant. Heat generation occurred to the extent that the device could only be operated in pulsed mode, with on pulses of 2 ns requiring 50 ns of cooling for sustainable operation. The increased temperature within the device also lead to deleterious changes in the Gunn diode's operating frequency. In the simulated device, a 150 K change in temperature lead to a decrease in operating frequency of 40 GHz, from an initial frequency of 280 GHz. At the end of 2 ns of operation, the mean temperature within the device had increased by 120 K. The high accidental doping level in GaN also means the use of a doping notch to act as a nucleation point for dipoles within a Gunn diode, a common technique in other materials, becomes less feasible. As an alternative to a notch, a device was simulated incorporating a doping spike to nucleate the dipole. The use of a doping spike is not novel, however its use in GaN has not been previously explored. Simulations found that a fully-depleted p-type doping notch of length 2.1 nm, doped at 1x1024 m-3 would act as a nucleation point for dipole operation. The device was compared to a simulated device incorporating a doping notch of width 0.25 µm doped at 0.5x1023 m-3 and found to operate at a similar frequency and RF efficiency, making it a viable substitute. One limitation of Gunn diodes is that when operated in transit-time mode, the operating frequency is determined by the length of the diode's transit region and so is well-defined and fixed. This means that traditional Gunn diodes are not as useful a source of radiation for spectroscopic applications as might be desirable. Recent experimental results for planar devices have shown a broadening in operation frequency and even multiple frequencies. This thesis explores the hypothesis that such a broadening might be achieved in a vertical structure via the incorporation of an additional notch into the Gunn diode's transit region, effectively incorporating two transit regions into the device. Results showed that this novel device structure did show multiple modes of operation. Under a DC applied voltage, the device showed spontaneous switching behaviour, oscillating between dipole and accumulation layer operation from the second notch. Changes in the frequency of an applied RF voltage would shift the device from operating from the first or second notch, in dipole and accumulation layer mode respectively.

At the Department of Technology and Built environment at the University of Gävle there was an interest to study GNU Radio, which is an "open source radio project. The project is based on that most of the radio signal processing is made in an ordinary PC. The idea behind this degree project was that in a laptop there are several radio transmitters/receivers that takes space, generates heat and transmit in varied frequency band etcetera.

 

All these radio transmitters/receivers could be replaced with a Software Defined Radio system. It means that one common, general radio hardware is used to different communications such as: WLAN, Bluetooth, GPRS, 3G etcetera. The waveform is generated in the software, which makes the system very flexible. To transmit and receive radio signals a USB-based hardware is required, for example from Ettus Research LLC.

 

During this degree project two PC:s was used for the signal processing and the signal transferring. The operating system that was used on the computers, were the Linux based Ubuntu 8.04. To generate the signals, to modulate/demodulate the signals and to get the communication on the sound cards in/out-port working, the different packages in the GNU Radio software was used and for programming the high level language, Python, was used.

 

In this degree project a lot of experiments where made, for example a sine wave was generated in computer 1 and the signal was amplitude modulated and transferred to computer 2, through the sound card. In computer 2 the signal was demodulated and filtrated, before it was saved to the hard drive. When the signal was saved on computer 2, it could be sent out on the sound card and be studied on an oscilloscope. This transfer between the computers was made with a stereo cable, but also with a radio link equipment on the University of Gävle.

 

The result of this degree project was satisfying, because the signal was possible to modulate, transfer, demodulate and save. In the wire transfer a lot of noise was generated on to the signal, mostly because of the sound cards. When the wireless transfer was made it appeared more noise, because of the quality of the receiver, the transmitter and the antennas.

 

This work can be developed to more advanced systems.

Vid Högskolan i Gävle på institutionen för Teknik och Byggd miljö (ITB) fanns ett intresse att undersöka GNU Radio, som är ett open source radio-projekt. Projektet bygger på att den största delen av radiosignalbehandlingen sker i en vanlig PC. Idén som låg till grund för detta examensarbete var att det i en laptop finns ett stort antal radiosändar- och mottagarkretsar som tar plats, genererar värme och sänder på olika frekvensband med mera.

 

Alla dessa radiosändar- och mottagarkretsar skulle kunna ersättas med ett Software Defined Radio-system. Vilket innebär att en gemensam, generell radiohårdvara används för olika kommunikationer som: WLAN, Bluetooth, GPRS, 3G med flera. Vågformerna genereras i mjukvaran, vilket gör systemet mycket flexibelt. För att kunna ta emot och sända radiosignaler behövs en hårdvara. Denna hårdvara har bland annat Ettus Research LLC tagit fram, med USB-anslutning.

 

Under examensarbetet har två stycken PC använts för behandling av signaler, samt överföring mellan dessa. Operativsystemet som användes på datorerna var det Linuxbaserade Ubuntu 8.04. För att generera signaler, modulation/demodulation av dessa signaler samt för att få kommunikation med ljudkortets in-/utgång att fungera, användes de olika paketen i mjukvaran GNU Radio och för programmering användes högnivåspråket Python.

 

I detta examensarbete utfördes ett flertal experiment, bland annat genererades en sinussignal i dator 1 och signalen amplitudmodulerades och överfördes till dator 2 via ljudkortet. På dator 2 demodulerades denna signal och filtrerades, innan den sparades på hårddisken. Signalen kunde sedan skickas ut på ljudkortet och studeras med ett oscilloskop. Överföringen mellan datorerna gjordes med en stereokabel, men också med en radiolänkutrustning som fanns på Högskolan i Gävle.

 

Resultatet var tillfredställande då signalen kunde moduleras, överföras samt demoduleras och sparas. I den trådbundna överföringen uppstod mycket brus i signalen, till största delen berodde detta på ljudkorten. När den trådlösa överföringen gjordes uppstod mera brus, vilket berodde på kvalitén hos mottagare, sändare och antennerna.

 

Detta arbete kan utvecklas till mer avancerade system.

Ce travail de thèse porte sur l'étude optique de nanofils de GaN et de microcavités d'AlN contenant des boîtes quantiques de GaN. La largeur de raie de l'exciton lié au donneur neutre dans le spectre de photoluminescence des nanofils de GaN crûs par épitaxie par jets moléculaires met en evidence l'homogénéité des contraintes dans le matériau. S'ils ne présentent aucun confinement excitonique, la géométrie filaire permet une relaxation efficace des contraintes et permet d'étudier précisément le bord de bande du GaN relaxé en phase cubique. Par ailleurs, nous infirmons l'attribution de la transition à 3.45 eV observée dans le spectre des nanofils de GaN wurtzite à un satellite à deux électrons. En effet, les règles de sélection de son dipôle, ainsi que son évolution sous champ magnétique intense, montrent que cette transition n'a pas les propriétés d'un satellite à deux électrons. Nous avons également étudié la spectroscopie de microdisques d'AlN contenant des boîtes quantiques de GaN. Des facteurs de qualité record pour les cavités en AlN ont été mesurés autour de 3 eV. Des nanocavités d'AlN contenues dans des guides d'onde unidimensionnels ont également été étudiées. L'attribution de chaque mode au guide d'onde ou à la cavité, prédite par des calculs préliminaires, est confirmée expérimentalement par une localisation différente. Ces structures donnent lieu à d'excellents facteurs de qualité, de 2300 à 3.45 eV, jusqu'à 4400 à 3.14 eV. Si le facteur de Purcell attendu est très élevé (autour de 100), nous n'avons pas réussi à observer l'effet Purcell. Ceci s'explique soit par l'instabilité des modes de cavité et de l'émission des boîtes quantiques sous exposition prolongée, soit par l'importance des recombinaisons non radiatives. Enfin, il apparaît que le frein principal à l'obtention de l'effet laser dans ces structures est l'important champ électrique interne, qui ralentit l'émission spontanée des boîtes quantiques.

Ce travail de thèse porte sur l'étude optique de nanofils de GaN et de microcavités d'AlN contenant des boîtes quantiques de GaN. La largeur de raie de l'exciton lié au donneur neutre dans le spectre de photoluminescence des nanofils de GaN crûs par épitaxie par jets moléculaires met en evidence l'homogénéité des contraintes dans le matériau. S'ils ne présentent aucun confinement excitonique, la géométrie filaire permet une relaxation efficace des contraintes et permet d'étudier précisément le bord de bande du GaN relaxé en phase cubique. Par ailleurs, nous infirmons l'attribution de la transition à 3.45 eV observée dans le spectre des nanofils de GaN wurtzite à un satellite à deux électrons. En effet, les règles de sélection de son dipôle, ainsi que son évolution sous champ magnétique intense, montrent que cette transition n'a pas les propriétés d'un satellite à deux électrons. Nous avons également étudié la spectroscopie de microdisques d'AlN contenant des boîtes quantiques de GaN. Des facteurs de qualité record pour les cavités en AlN ont été mesurés autour de 3 eV. Des nanocavités d'AlN contenues dans des guides d'onde unidimensionnels ont également été étudiées. L'attribution de chaque mode au guide d'onde ou à la cavité, prédite par des calculs préliminaires, est confirmée expérimentalement par une localisation différente. Ces structures donnent lieu à d'excellents facteurs de qualité, de 2300 à 3.45 eV, jusqu'à 4400 à 3.14 eV. Si le facteur de Purcell attendu est très élevé (autour de 100), nous n'avons pas réussi à observer l'effet Purcell. Ceci s'explique soit par l'instabilité des modes de cavité et de l'émission des boîtes quantiques sous exposition prolongée, soit par l'importance des recombinaisons non radiatives. Enfin, il apparaît que le frein principal à l'obtention de l'effet laser dans ces structures est l'important champ électrique interne, qui ralentit l'émission spontanée des boîtes quantiques
This work focuses on the optical study of GaN nanowires and AlN microcavities containing GaN quantum dots. The 1-meV linewidth of the neutral donor-bound exciton line in the photoluminescence spectrum of MBE-grown GaN nanowires evidences that the strain is homogeneous in the material. These nanowires do not exhibit any excitonic confinement, but the efficient strain relaxation allows to grow strain-free zinc-blende GaN nanowires and then to conduct fine spectroscopy on cubic GaN near band edge. Beside, we show that the tentative attribution of the recombination line at 3.45 eV in the spectrum of wurtzite GaN nanowires to a surface-enhanced two-electron satellite does not hold. Indeed, its dipole polarization selection rules and its evolution with intense applied magnetic field do not match that of a two-electron satellite. We also performed the spectroscopy of GaN/AlN quantum dot microdisks. Record quality factors for AlN cavities were measured around 3 eV. GaN/AlN quantum dot nanocavities embedded in photonic crystal waveguides were also investigated. The attribution of each mode either to the waveguide or to the cavity, predicted by calculations, is experimentally confirmed by a different light localization. These structures allow excellent quality factors to be reached, from 2300 at 3.45 eV, up to 4400 at 3.14 eV. Although the expected Purcell factor is very high (around 100), we did not manage to observe the Purcell effect. This originates either from an enhancement of non-radiative recombination channels or from an instability of both the cavity modes and the quantum dot emission under intense exposure. Finally, it appears that the main limiting factor to achieve lasing in these structures is the strong built-in electric field, which slows up the spontaneous emission rate of the quantum dots

Ce travail de recherche porte sur les cristaux de gan massifs et sur les heterostructures gan/algan. Dans le premier chapitre, nous exposons les principales techniques experimentales qui nous serviront a caracteriser les cristaux massifs de gan et les couches homoepitaxiees, a savoir la spectroscopie infrarouge et -raman, la photoluminescence et la double diffraction de rayons x. Dans le second chapitre, nous etudions deux types de cristaux massifs de gan : pour les cristaux non intentionnellement dopes, nous etablissons que la face polarisee (000 $$1)n possede toujours plus de porteurs libres que la face (0001)ga et mettons en evidence la presence d'un gradient de concentration en electrons libres le long de l'axe polaire c (0001). En outre, nous revelons que les defauts natifs de lacunes de gallium sont plus nombreux sur la face (0001)ga. Les cristaux dopes magnesium, a caractere semi-isolant, permettent l'etude des phonons au centre de la zone de brillouin ainsi que les processus a deux phonons. Dans le troisieme chapitre, nous etudions les couches de gan deposees par epvom sur les deux types de cristaux de gan. Nous montrons que la meilleure face des substrats de gan pour l'homoepitaxie est la face (0001)ga. En effet, les couches deposees sur cette face incorporent beaucoup moins d'impuretes donatrices (o et si) et possedent donc moins de porteurs libres que les couches deposees sur la face (000 $$1)n. Dans le dernier chapitre, nous etudions les proprietes du gaz d'electrons 2d dans les heterostructures gan/algan deposees sur gan massif par ejm. Nous obtenons la meilleure mobilite electronique a basse temperature jamais atteinte dans un systeme a base de gan. En outre, nous determinons le facteur de lande en etudiant la dependance angulaire de l'amplitude des oscillations de shubnikov de haas. Enfin, nous concluons par la presentation du premier transistor a effet de champ (hfet) developpe sur cristaux massifs de gan.

Made available in DSpace on 2014-06-11T19:35:45Z (GMT). No. of bitstreams: 0 Previous issue date: 2011-09-22Bitstream added on 2014-06-13T20:07:27Z : No. of bitstreams: 1 gomes_mc_dr_bauru.pdf: 4074042 bytes, checksum: d6522228622688cc96df0482f8a71eca (MD5)
Fundação de Amparo à Pesquisa do Estado de Mato Grosso (FAPEMAT)
Este trabalho é resultado de um estudo teórico sobre o GaN e o Semicondutor Magnético Ga1-xMnxN, tanto para bulk (sólido) como para as superfícies nanoestruturadas mais estáveis, dada sua importância para o desenvolvimento como material spintrônico. Analisamos deste material suas propriedades estruturais, energéticas e eletrônicas, a partir de cálculos periódicos baseados na teoria do Funcional da Densidade (DFT), como o funcional híbrido B3LYP, e também apresentamos resultados preliminares do estudo das propriedades magnéticas deste material. Realizamos um estudo pormenorizado das estruturas de bandas e da densidade de estados, do Ga1-xMnxN bulk (x~0,02 a 0,18) quanto em superfícies (x~0,0 a 0,17) para os modelos de supercélula de 32 e 96 átomos bulk e modelos de superfícies com 12 e 24 camadas. Os resultados obtidos nos mostram que para concentrações acima de 6% ocorre um acréscimo na distância de ligação Mn-N na direção c, pelo fato do Mn apresentar um raio atômico superior ao Ga e propriedades físicas e químicas distintas. Para os cálculos com superfícies, foi realizada a substituição do Ga por Mn em diferentes posições relativas na superfície, sub-superfície e core, ocorrendo o aumento da energia total conforme os átomos de Mn se movem para os sítios mais internos da superfície e ao considerar a forma de equilíbrio baseada na estabilidade termodinâmica, os valores das energias superficiais das superfícies (1010) e (1120) do GaN wurtzita são as mais estáveis para a concentração de ~8%. Com o aumento da concentração, ocorre nas estruturas de bandas uma diminuição do gap, tanto para o bulk quanto para as superfícies, porém ele se mantém direto no bulk, com exceção para concentração de 18% e, na superfície (1010), enquanto que na superfície (1120) o gap é indireto...
This work is the result os a theoretical study concerning GaN and the Magnetic Semiconductor Ga1-xMnxN, in both the bulk and the most stable nanoestructured surfaces, due to its importance in the development as spintronic material. We analyze the structural, energetic and electronic properties of this material, by means of periodic calculations based on the Functional Density Theory (DFT), at the hybrid functional B3LYP level, and also present the preliminary results of the study of the magnetic properties of this material. We carried out a detailed study of the band structures and the density of states, for both the Ga1-xMnxN bulk (x~002 a 0,18) and its surfaces (x~0,0 a 0,17) using supercell models constitued by 32 and 96 atoms for the bulk and 12 and 24 atomic layers for the surface slab model. Our results show that for Mn concentrations above 6% there is an increase in the Mn-N bond distance in the c direction, due to the fact that the Mn has an atomic radium greater than that of the Ga and different physical and chemical properties. For the surface calculations, we substituted the Ga for the Mn in different positions relative to the external surface, sub-surface and corre, it was observed that the total energy increased as the Mn atoms moved from the surface layer to the interior sites and when we consider the equilibrium shape based in the thermodynamic stability, the most stable surface energies for the (1010) and (1020) planos of wurtzite GaN are found for the ~8% Mn concentration. When the Mn concentration increases, the band gap decreases for the bulk as well as for the surfaces, the gap being direct for the bulk, except for the 18% concentration, and for the (1010) surface, whereas the gap is found indirect for the (1120) surface for the concentrations 6 and 17%. The analysis of the density... (Complete abstract click electronic access below)

Peut-on se servir d'une langue africaine comme le gὲngbe pour enseigner les sciences et la technologie modernes? A cette question, certains répondent par la négative prétextant que les langues africaines sont trop pauvres en vocabulaire de spécialité et inaptes à l'enseignement de ces disciplines. Dans ce travail, l'auteur s'inscrit en faux contre ce préjugé. Après avoir effectué une étude sociolinguistique de la langue, il conclut que pour servir de véhicule à l'enseignement des sciences et de la technologie modernes, le gὲngbe doit se doter à la fois d'une orthographe stable et normée puis d'un lexique enrichi. La thèse comprend donc deux grandes parties. Une première consacrée à la modernisation de l'orthographe et une deuxième à celle du lexique où l'auteur démontre que le gὲngbe dispose suffisamment de ressources susceptibles de l'aider à dire toutes les innovation techniques et scientifiques
Can an African language such as Gὲngbe be used for teaching science and modern technology? Some say no, arguing that African languages are too poor in specialist vocabulary and therefore unfit for teaching these subjects. In this work, the author challenges such a prejudice. After a sociolinguistic survey of the language, he concludes that in order to serve as a vehicle for teaching science and modern technology, Gὲngbe must develop a standardised orthography and enrich its lexicon. The thesis is therefore divided into two parts. The first part deals with orthography and the second with the lexicon where the author shows that Gὲngbe does indeed have sufficient resources at its disposal to articulate any and all technological and scientific terminology

Zur Zukunft der Katalogisierung werden zur Zeit zwei große Themenkomplexe diskutiert. Einmal der Aufbau einer Gemeinsamen Normdatei (GND) und die Einführung eines neuen Regelwerks, „Resource Description and Access (RDA)“. Das Regelwerk soll die bislang geltenden RAK ablösen und nicht nur für Bibliotheken, sondern auch für Archive und Museen zur Verfügung stehen. Da die verschiedenen Normdateien auf unterschiedlichen Datenmodellen und Formaten basieren, wird an einer Zusammenführung der Normdatenbanken in eine einheitliche Normdatei gearbeitet.

Ataxia telangiectasia: Identifizierung und Charakterisierung nicht-konservativer Spleißmutationen und deren Auswirkungen im ATM-Gen. Ein hoher Anteil der bisher im ATM-Gen identifizierten Mutationen (>350, www.vmresearch.org/atm.htm) stellt Deletionen oder Insertionen direkt an den Exongrenzen dar; viele dieser Aberrationen wurden allerdings nur auf cDNA-Ebene detektiert. Sollte es sich hierbei in den meisten Fällen um Mutationen an den Spleiß-Konsensussequenzen handeln, läge der Anteil der Spleißmutationen im ATM-Gen beträchtlich höher (~35 Prozent) als in anderen betroffenen Genen (~15 Prozent). Um der Frage nachzugehen, ob im ATM-Primärtranskript aufgrund einer erhöhten Labilität gegenüber Spleißmutationen auch Veränderungen weniger konservierter Positionen innerhalb der Donor- oder Akzeptor-Spleißstellen zu aberrantem Spleißen führen, wurden 20 AT-Zellinien mittels „Protein Truncation Test“ nach Deletionen oder Insertionen an den Exongrenzen durchsucht. Die 7 neu identifizierten Spleißmutationen wurden anschließend unter Verwendung eines „Splice Scoring“- Systems näher charakterisiert, die Penetranz der jeweiligen Mutation durch semiquantitative PCR evaluiert und die Auswirkungen auf Proteinebene durch Western Blotting überprüft. Obwohl nur eine der 7 neu identifizierten Spleißmutationen eine schwächer konservierte Position der Spleißsequenzen betraf, konnten im Rahmen einer Kooperation mit der Medizinischen Hochschule Hannover (Arbeitsgruppe Dr. T. Dörk) weitere Spleißmutationen an den Intronpositionen +3, +5 und -6 identifiziert werden, die ebenfalls in völlig aberrantem Spleißen resultieren. Daten weiterer Arbeitsgruppen lassen vermuten, daß tatsächlich ~ 50 Prozent aller Spleißaberrationen im ATM-Gen auf Mutationen außerhalb der konservierten Dinukleotidbereiche (gt und ag) zurückführen sind. Nijmegen Breakage Syndrom (NBS): Suche nach Genen, die einen NBS-ähnlichen Phänotyp auslösen. Über 90 Prozent aller NBS-Patienten tragen Mutationen im NBS1-Gen, dessen Translationsprodukt im Komplex mit MRE11 und RAD50 eine zentrale Rolle in DNA-DSB-Reparatur und Zellzykluskontrolle spielt. Weitere Mutationen bei Patienten mit NBS-ähnlichem Phänotyp wurden im Gen der DNA-Ligase IV identifiziert, die zusammen mit weiteren Angehörigen des NHEJ-Reparaturweges (XRCC4, DNA-PKcs, Ku70, Ku80) ebenfalls in DNA-DSB-Reparatur involviert ist. Zellen von Patienten mit NBS-ähnlichem Phänotyp wurden daher durch direkte Sequenzierung und/oder Western Blotting auf Defekte in den oben genannten Genen/Proteinen untersucht. In einem parallel durchgeführten unabhängigen Mutationsscreening (Medizinische Hochschule Hannover, Arbeitsgruppe Dr. T. Dörk) wurden in Fibroblasten einer Patientin mit NBS-ähnlichem Phänotyp Mutationen im RAD50-Gen identifiziert. Die Auswirkungen der RAD50-Defizienz auf die zelluläre Lokalisation der beiden Komplexpartner NBS1 und MRE11 sowie deren Fähigkeit zur Focibildung nach DNA-Schädigung wurde im Rahmen dieser Arbeit durch Immunfluoreszenzstudien untersucht: während NBS1 vorwiegend nukleäre Lokalisation aufwies, war MRE11 zu etwa gleichen Anteilen zwischen Nukleus und Zytoplasma verteilt; beide Proteine waren nach Bestrahlung der Zellen nicht mehr zur Focibildung fähig. Da in MRE11-defizienten Zellen keine nukleäre NBS1-Lokalisation beobachtet wird, scheint der Kerntransport des NBS1 von funktionellem MRE11, nicht aber von RAD50 abhängig zu sein. Fanconi Anämie (FA): Untersuchung einer möglichen Verbindung zwischen den FA-Proteinen und der RAD51-Familie. FA-Zellen aller bisher bekannter Komplementationsgruppen zeichnen sich durch Hypersensitivität gegenüber DNA-„interstrand crosslinks“ (ICLs) aus, zu deren Behebung u.a. die homologe Rekombinationsreparatur (HRR) eingesetzt wird, bei der das RAD51(A)-Protein eine zentrale Rolle spielt. Aufgrund schwacher Homologien werden 5 weitere Proteine (RAD51B, C, D, XRCC2 und 3) der RAD51-Familie zugeordnet. Da Knockout-Zellinien aller RAD51-Familienmitglieder ebenfalls hohe Sensitivität gegenüber ICLs aufweisen, wurde eine mögliche Verbindung zwischen den FA-Proteinen FANCA, C, G und der RAD51-Familie getestet. Unter Verwendung des "Yeast Two Hybrid" (Y2H)-Systems konnten zunächst mehrere Interaktionen zwischen FA- und RAD51-Proteinen detektiert werden. Zur Bestätigung einer funktionellen Verbindung wurden die FA- und RAD51-Proteine in humanen 293-Zellen überexprimiert. Aufgrund der focibildenden Eigenschaften des RAD51-Proteins wurden die FA-Proteine und die RAD51-Familie auf Focibildung nach DNA-Schädigung sowie etwaige Kolokalisationen getestet; mögliche physikalische Interaktionen wurden durch Koimmunpräzipitationsstudien überprüft. Die RAD51-Familie zeigten keinerlei Focibildung nach DNA-Schädigung während die FA-Proteine in einigen Experimenten eine Lokalisation in nukleäre Foci zeigten, die sich jedoch in Größe und Homogenität deutlich von denen klassischer DNA-Reparaturfoci unterschieden und nicht mit RAD51 kolokalisierten. Die häufige Beschränkung der FA-Foci auf Bereiche besonders dicht gepackten Chromatins kann möglicherweise als weiterer Hinweis auf die postulierte Rolle der FA-Proteine bei Chromatin Remodelling Mechanismen interpretiert werden. Bei Überexpression in HEK293-Zellen konnte keine der im Y2H-System identifizierten Interaktionen zwischen FA- und RAD51-Proteinen durch Koimmunpräzipitationen detektiert werden. Dennoch erscheinen seit der Identifizierung des FANCD2, das durch den FA-Komplex aktiviert wird und mit dem RAD51-Interaktor BRCA1 in nukleäre Foci kolokalisiert, weitere Untersuchungen einer Verknüpfung der FA-Proteine mit den Angehörigen des HRR-Weges durchaus sinnvoll
Ataxia telangiectasia: Identification and characterization of mutations at non-conserved splice positions within the ATM gene. Since the identification of ATM, more than 350 different mutations have been identified (www.vmresearch.org/atm.htm) with an unusual frequency of exon skipping or deletions/insertions at the exon boundaries. Although in many studies only the observed aberrations in cDNA were reported, they presumably represent splice mutations. These findings suggest that the rate of splicing defects in the ATM gene may be substantially higher than that reported for other human genetic disorders. In order to investigate this phenomenon, we have asked the question whether the high frequency of splice mutations in the ATM gene might be caused by an increased liability of the ATM transcript towards mutations at less conserved positions within the splice site consensus sequences. Mutation screening of 20 AT cell lines using the protein truncation test revealed 10 different splice mutations, seven of which had not been reported previously. A splice scoring system was used to estimate the penetrance of each mutation. Potentially leaky splicing was evaluated by semiquantitative PCR. The implications of each mutation on the protein level were investigated by western blotting. One out of seven new splice mutations was positioned at a less conserved consensus site. Furthermore, in cooperation with the Medizinische Hochschule Hannover (group of Dr. T. Doerk), point mutations at intron positions +3, +5 and –6 were identified, each resulting in completely aberrant splicing. In addition, there is accumulating evidence from other studies that indeed ~50 per cent of splice mutations in the ATM gene are positioned outside the highly conserved donor or acceptor dinucleotides, gt or ag, respectively. Nijmegen breakage syndrome (NBS): Search for genetic defects resulting in an NBS- or NBS-like phenotype. More than 90 per cent of NBS patients carry mutations in the NBS1 gene. The NBS1 protein, in a complex with MRE11 and RAD50, is known to be involved in DNA DSB repair via NHEJ and cell cycle control. In cells from patients with an NBS-like phenotype, defects in another component of the NHEJ pathway, DNA ligase IV, have been identified very recently. Therefore, mutation screening in genes known to be involved in NHEJ (NBS1, MRE11, RAD50, DNA ligase IV, XRCC4, DNA-PKcs, Ku70, Ku80) was carried out in cell lines from NBS-like patients using direct sequencing and/or western blotting. In one of these cell lines, an independently performed screening at the Medizinische Hochschule Hannover (group of Dr. T. Doerk) revealed two mutant alleles of the RAD50 gene. In the present work, the cellular RAD50-deficient phenotype was characterized in response to induced DNA damage. In contrast to normal controls, the RAD50-deficient fibroblasts were not able to form IR-induced foci. Localization of NBS1 was still predominantly nuclear, whereas MRE11 was distributed equally between nucleus and cytoplasm. In contrast, MRE11-deficient cells do not show much nuclear localization for NBS1, suggesting that the nuclear localization of NBS1 depends on functional MRE11, but not on the presence of RAD50. Fanconi anemia (FA): Are the FA proteins linked to DNA DSB repair via the RAD51 family? FA cells of all complementation groups are hypersensitive towards DNA interstrand crosslinks (ICLs). Removal of ICLs includes the generation of a DNA DSB. Their repair is then presumably carried out by the homologous recombination (HRR) pathway, an essential component of which is the RAD51 protein. Due to weak homologies five other proteins have been assigned to the RAD51 family (RAD51B, C, D, XRCC2 and 3). Strikingly, knockout mutants of all RAD51 paralogs also show high sensitivity to ICL-inducing agents. In order to investigate a possible connection between the FA pathway and the RAD51 family, the yeast two-hybrid system was used to test for any interactions between the FA proteins FANCA, C, G and the RAD51 family members. Several interactions were detected. Further characterization was done by study of nuclear foci and immunoprecipitation. FA and RAD51 proteins were overexpressed in human 293 cells. Since RAD51 localizes into nuclear foci after DNA damage, FA proteins and RAD51 paralogs were tested for their ability to (co-) localize into nuclear foci after MMC-treatment. Whereas none of the RAD51 paralogs showed nuclear foci formation, in some experiments FA proteins moved into nuclear speckles following DNA damage. However, size and homogeneity of DNA damage-induced FA speckles was quite different from those formed by „classical“ DNA repair proteins; in addition, no colocalization was seen with foci containing RAD51. Notably, FA speckles were often restricted to sites of very tightly packed chromatin, again suggesting a role of the FA-proteins in chromatin remodelling mechanisms. To identify any physical interaction between FA proteins and the RAD51 family, coimmunoprecipitation studies (Co-IPs) were carried out, after overexpression in HEK293 cells. Whereas the reported interaction between FANCA and FANCG could repeatedly be shown, no Co-IP was detected for any combination of an FA protein with a member of the RAD51 family. However, since the identification of FANCD2, a connection between the FA pathway and RAD51 seems all the more plausible: following DNA damage FANCD2 colocalizes with BRCA1, which itself is closely associated with RAD51. Further studies will be necessary in order to unravel the way by which the FA proteins might interact with conserved DNA repair mechanisms like the HRR pathway

Today the continuous increase of electric power demand is in our society a global concern. Hence, the reduction of the energy consumption has become the main task of modern power electronics. In this context, wide band semiconductors (WBG), such as gallium nitride (GaN) and related alloys, have outstanding physical properties that can enable to overcome the limitations of Silicon, in terms of operating power, frequency and temperature of the devices. An interesting aspects related to GaN materials is the possibility to grow AlGaN/GaN heterostructures, in which a two dimensional electron gas (2DEG) is formed at the heterojunction. Basing on the presence of the 2DEG, AlGaN/GaN heterostructures are particularly interesting for the fabrication of high electron mobility transistors (HEMTs). One of the most challenging aspects on this field is the development of enhancement mode AlGan/GaN HEMT. This devices would offer a simplified circuitry, in combination with favourable operating conditions for device safety. Hence, this thesis is entitled AlGaN/GaN heterostructures for enhancement mode transistors . The aim of this work was to clarify the mechanisms ruling the electronic transport at some relevant interfaces in AlGaN/GaN devices, after surface modification processes used in normally-off technologies. The thesis is divided in 6 chapters. In the first two chapters, the properties of GaN and related AlGaN alloys are described, explaining the formation of the 2DEG and the working principle of HEMT devices. In chapter 3, a nanoscale characterization of modified AlGaN surfaces is presented in order to deplete the 2DEG. Two different approaches have been studied, i.e., the use of a fluorine plasma treatment and the use of a local oxidation process. Even though a depletion of the 2DEG is possible, several reliability concerns need to be investigated before a practical application to devices can be envisaged. Among the possible approaches for enhancement mode transistors using AlGaN/GaN heterostructures, the use of a p-GaN gate contact seems to be the most interesting one. Hence, chapter 4 reports a detailed investigation on the formation of Ohmic contact to p-GaN. The evolution of a Au/Ni bilayer, annealed at different temperatures and in two different atmospheres (Ar or N2/O2) was considered. The electrical measurements of the contacts annealed under different conditions demonstrated a reduction of the specific contact resistance in oxidizing atmosphere. Structural characterizations of the metal layer associated with nanoscale electrical measurements, allowed to give a possible scenario on the Ohmic contact formation mechanisms. Finally, the temperature dependence of the specific contact resistance allowed the extraction of the metal/p-GaN barrier. The fabrication and characterization of AlGaN/GaN transistors with the use of a p-GaN cap layer under the gate contact is presented in chapter 5. The electrical characterization of p-GaN/AlGaN/GaN transistors demonstrated a significant positive shift of the threshold voltage (Vth) with respect to devices without p-GaN gate. A normally-off behaviour of the devices (Vth= +1.4 V) was obtained upon a reduction of the barrier layer thickness and Al concentration. Finally, a preliminary study on the use of nickel oxide (NiO) as a dielectric below the Schottky gate contact in AlGaN/GaN heterostructures is reported in the last chapter. First, a structural and morphological investigation of the NiO layers grown by MOCVD showed continuous epitaxial film. The electrical measurements on devices allowed to extract a value of the dielectric constant for the grown NiO very close to the theoretical one, and a reduction of the leakage current in HEMT structures integrating such a dielectric. The experimental results of this thesis are summarized in the conclusive section, that also briefly describes the remaining open issues and the possible continuation of this research activity.

Ce document porte sur le développement d’un nouveau banc de mesure pour la caractérisation de l’impédance thermique des HEMTs GaN. Le banc développé repose sur la méthode dite « 3ω » qui consiste à mesurer l’harmonique 3 d’un signal électrique véritable image des variations thermiques du composant. Un balayage en fonction de la fréquence d’excitation conduit à l’extraction de l’impédance thermique. Les résultats de mesures ont été validés par les simulations électriques. Des études complémentaires ont été réalisées pour l’identification des effets de pièges en utilisant différentes méthodes permettant l’extraction de la signature des pièges. La réalisation des modèles non-linéaires est présentée pour les transistors HEMT AlGaN/GaN et InAIN/GaN pour des applications d’amplificateur de puissance dans les bandes de fréquences X et K
This report is devoted to the development of a new measurement bench for thermal impedance characterization of GaN HEMTs. This measurement test set uses the so-called « 3ω » technique, which consists to measure the electrical signal at third harmonic real image of the thermal magnitude variations of the device. A sweep in function of the excitation frequency allows extracting of the thermal impedance. The measurement results have been validated by electrical simulation. Other complementary studies were performed to identify the trapping effects using different methods to extract the traps signature. The realization of nonlinear models is presented for AlGaN HEMT / GaN and InAIN / GaN to the power amplification applications in frequency bands X and K

Cette thèse est consacrée à l'étude expérimentale des excitons dans des puitsquantiques polaires fabriqués à partir de semi-conducteurs à large bande interdite. En raison de la structure de ces matériaux à cristaux wurtzite, les électrons et les trous sont séparés le long de l'axe de croissance du puits quantique, de sorte que les excitons peuvent être considérés comme des excitons indirects (IX) : ils forment une famille de quasi-particules bosoniques à longue durée de vie, dont le moment dipolaire est orienté selon l'axe de croissance du puits. Les IX sont considérés comme un système modéle pour l'étude des états collectifs dans les gaz quantiques bosoniques. Ils sont aussi prometteurs pour le développement de dispositifs excitoniques. Leur longue durée de vie, leur répulsion dipolaire, permettent aux IXs de se déplacer sur de grandes distances avant de se recombiner, ce qui offre la possibilité d'étudier le transport d'exciton par imagerie optique. Dans cette thèse, nous abordons le transport des IXs dans des puits quantiques de GaN/(Al,Ga)N et de ZnO/(Mg,Zn)O. Ce choix de matériau est motivé par l'énergie de liaison élevée des IXs ainsi obtenue. Elle est suffisamment élevée pour, en thèorie, stabiliser les IXs jusqu'à la température ambiante. Mais ce choix poseaussi un certain nombre de défis expérimentaux, car (i) le temps de vie radiatifdépend fortement de la densité d'excitons, ce qui rend la mesure de la densitéexcitonique très complexe ; (ii) la recombinaison non radiative activée thermiquement supprime le signal de photoluminescence excitonique à température ambiante ; (iii) la propagation excitonique coexiste avec une propagation photonique le long du plan du puit quantique, ce qui complique l'analyse ; (iv) il existe un fort champ électrique le long de l'axe de croissance, et aussi desuctuations dans l'épaisseur du puits quantique, ce qui crée un fort élargissement inhomogène de l'émission excitonique. Nous avons abordé toutes ces questions et nous démontrons dans ce travail que les excitons se propagent effectivement dans le plan du puits quantique. Nous arrivons à cette conclusion en combinant des expériences de micro-photoluminescence en régime continu avec des mesures de spectroscopie résolues en temps, et en comparant nos données expérimentales avec divers modèles numériques basés sur les équations dedérive et de diffusion. Dans du matériau de qualité, des puits GaN/(Al,Ga)N obtenus sur substrats GaN, nous avons observé une propagation à temprature ambiante sur plus de 10 µm, et sur plus de 20 µm à 4 K. Nos résultats suggérent que la propagation des excitons sous excitation à onde continue est facilitée par l'écrantage du désordre par les excitons. Néanmoins, la propagation excitonique est encore limitée par la diffusion des excitons sur les défautsiii plutôt que par la diffusion exciton-exciton. Ainsi, l'amélioration de la qualité des interfaces du puits quantique pourrait encore permettre une propagation excitonique sur de plus grandes distances
This thesis is devoted to experimental study of excitons in polar quantum wells(QWs) based on wide band-gap semiconductors. Due to wurtzite crystal structureof these materials, electron and hole are separated in the QW growth axis, sothat excitons can be considered as indirect excitons (IX), a family of long-living bosonic quasi-particles with dipole moment oriented along the QW growth axis. IX are considered as a model system for studies of collective states in quantum gases of bosons, and are also promising for the development of excitonic circuit devices. Long lifetimes and dipole repulsion allow IXs to travel over large distances before recombination providing the opportunity to study exciton transport by optical imaging. In this thesis we address IX transport in a set of GaN/(Al,Ga)N and ZnO/(Mg,Zn)O QWs. This choice of IX is motivated by high binding energy, and potential stability up to room temperature, but present a number of experimental challenges, including (i) dramatic dependence of the exciton radiative lifetime on the exciton density that makes exciton density measurement very complex, (ii) thermally activated nonradiative recombination that quenches exciton PL at room temperature,(iii) coexistence of photon propagation with exciton propagation along the QW plane, and strong inhomogeneous broadening of the exciton emission due to strong built-in electric field and the presence of both monolayeructuations of the QW thickness and the fluctuations of alloy composition in the barriers. We have addressed all these issues and demonstrated exciton propagation by combining continuous wave µ-photoluminescence and time-resolved spectroscopy measurements, supplemented by modelling of the exciton transport within drift-diffusion formalism. In the best quality GaN/(Al,Ga)N QWs grown on free-standing GaN substrates we achieved room-temperature propagation over ~10 µm and up to 20 µm at 4 K. Our results suggest that propagation of excitons under continuous-wave excitation is assisted by effcient screening of the in-plane disorder. Nevertheless, exciton propagation is still limited by the exciton scattering on defects rather than by exciton-exciton scatteringso that improving interface quality can boost exciton transport further

Humboldt-Univ., Diss u.d.T.: Deindl, Philipp: Gen-Gen- und Gen-Umwelt-Interaktionsanalysen bei Kindern der Multizentrischen Allergie-Studie--Berlin, 2005.
Hergestellt on demand. Literaturverz. S. 82 - 96.

From a HF propagation point of view, the ionospheric layers act as partially conducting media (plasma) in which a transmitted radio wave can reflect upon.A way of determining whether a radio wave with a given frequency will reflect from the ionosphere or completely penetrate is to utilize special radar instruments know as ionosondes or chirp sounders. The technique is widely used by amateur enthusiasts and military radio users for monitoring available radio channel links between two remote locations and can often serve as a base for HF radio prognoses.The objective of this Bachelor’s Thesis was to explore, implement and test a single channel receiver for monitoring ionospheric sounders. The implementation is based on Software Defined Radio (SDR) technology and relies on the GNU Chirp Sounder (gcs) open source script program.