ELECTROMAGNETIC ANALYSIS OF PLANAR LAYERED STRUCTURES Fatma Caliskan 169 pages Directed by Dr. Andrew F. Peterson The electrical design of microelectronic devices and their packaging is complicated because of non-ideal attributes of the actual circuit realization. Electromagnetic modeling offers the possibility of accurately predicting the electrical performance of devices and reducing the cost associated with the design process. The proposed research concerns extensions of electromagnetic modeling techniques and their application to microelectronic package design. The method of moments (MoM) is utilized as a technique in modeling and analyzing these designs. Recently, an alternate approach called the locally corrected Nystrm method (LCN) has been applied to solve integral equations in electromagnetics. Recent research suggests that the LCN is well-suited for higher-order implementations and does not require cell-to-cell current continuity in the underlying representation. Thus it may offer advantages over the MoM, especially for problems involving complex 3-D structures. If cell-to-cell continuity is not required, nonconforming meshes may offer simpler geometrical modeling. In this proposal, we consider applying the above techniques to problems in package designs, which often involve multilayer structures, solid or perforated ground planes, embedded passive devices such as capacitors and spiral inductors, and interconnects in horizontal or vertical directions. Several examples will be used to illustrate the modeling.

Based on the combined surface and volume RWG (Rao-Wilton-Glisson) basis functions, a simulator of a patch antenna on a finite dielectric substrate using the Method of Moments (MoM) has been implemented in Matlab. The metal surface is divided into planar triangular elements whereas the (inhomogeneous) dielectric volume is divided into tetrahedral elements. The structure under study is comprised of a typical patch antenna consisting of a single patch above a finite ground plane, and a probe feed. The performance of the solver is studied for different mesh configurations. The results obtained are tested by comparison with the commercial ANSOFT HFSS v8.5 and WIPL-D simulators. The former uses a large number of finite elements (up to 30,000) and adaptive mesh refinement, thus providing the reliable data for comparison. Behavior of the most sensitive characteristic ¡V antenna input impedance ¡V is tested, close to the first resonant frequency. The error in the resonant frequency is estimated at different values of the relative dielectric constant ƒÕr, which ranges from 1 to 20. The reported results show reasonable agreement. However, the solver needs to be further improved.

La cohabitaion, dans une même enceinte, des systèmes électroniques de puissance et decommande est un risque majeur pour le bon fonctionnement des systèmes électroniquesembarqués (Automobile, Aéronautique et Spatial). La maitrise de la CEM de ces systèmes estun défit surtout avec la difficulté de rendre compte des phénomènes EM.Ce mémoire décrit la modélisation des émissions conduites et des émissions EM rayonnéespar le câblage des systèmes électroniques embarqués. On explique d’abord le choix del’utilisation de la méthode de modélisation PEEC puis on décrit sa formulationmathématique et on présente les différents phénomènes EM qu’elle prend en compte dansle contexte de câblage.Dans le deuxième chapitre, un calcul tridimensionnel des émissions EM rayonnées a étéproposé. Il prend en compte de la géometrie des cellules de discrétisation et permet unemeilleure considération de la montée en fréquence. Il permet ainsi de se rapprocher au plusprès de la structure. Une validation de ce calcul a été effectuée en comparant les résultatsde simulation à ceux obtenus par d’autres calculs analytiques, par modélisation « élémentsfinis » et aussi par des mesures en champ proche.Le troisième chapitre est consacré aux applications aux systèmes électroniques embarqués.Dans une première partie, une étude d’un système de câblage, qui reflète les différentsphénomènes rencontrés en automobile, a été faite. Ensuite, un câblage est associé à unconvertisseur DC/DC. Ce convertisseur est utilisé comme un générateur de perturbations EMen Hautes fréquences. Dans cette partie du travail, on présente une modélisation de cesperturbations qui se propagent vers le câblage et qui se transforment en émissions EMrayonnées. / The calculation of radiated electromagnetic (EM) fields from wiring systems such as cables in automotive/aeronautic applications is an important issue for the prediction and the prevention of EM Interference (EMI) and EM Compatibility (EMC) problems. Many numericalmethods of electromagnetic modeling are available and can be used for EM analysis. Amongthese methods, the Partial Element Equivalent Circuit method (PEEC) is particularly wellsuited. Usually the PEEC approach is an efficient way to deal with circuit simulation ofdistributed structures. In automotive/aeronautic applications, the electronic systemsoperate at increasingly higher frequencies. So, the radiation of cables cannot be neglected. A3D model based on the PEEC method was proposed in order to take into account the size ofautomotive wiring systems. It includes resistive, inductive and capacitive effects. This 3Dapproach was shown to accurately predict the conducted disturbances by cables above largeground planes.In this work a new 3D calculation approach based on Maclaurin expansion is developed toevaluate the radiated fields with the PEEC method. The originality of this work is consideringboth radiated and conducted disturbances in the framework of a PEEC method for large sizestrcutures. In a first step the current carried by the conductors are determined from thecircuit model deduced with the 3D PEEC method. Then in a second step the field radiated bythe wiring systems is evaluated using an analytical calculation deduced from the distributionof currents.The third part of this work is to identify the disturbances generated

Les techniques de diagnostic actuellement utilisées en endoscopie digestive ne permettentpas une observation complète et précise de l’intestin grêle. Les capsules endoscopiquestentent de répondre à cette problématique, mais le médecin n’a aucune maîtrise de leurdéplacement une fois qu’elles sont avalées par les patients. Dans ce contexte, les travauxdéveloppés dans cette thèse portent sur l’utilisation des champs magnétiques afin demanipuler un objet (comme une capsule) sans contact sur un large espace de travail. Pourcela, nous étudions les systèmes de manipulation magnétique à électroaimants mobiles.Nous étudions dans un premier temps la manière de modéliser le champ magnétiquecréé par un électroaimant. Nous développons un modèle hybride (analytique recalé surdes mesures) prenant en compte le noyau ferromagnétique des électroaimants. Dans unsecond temps, nous nous intéressons à un système générique de manipulation magnétique,composé de n électroaimants pouvant se déplacer dans tout l’espace. Nous développons unmodèle de ce système, en intégrant les mobilités de chacun des électroaimants. Ce modèleest ensuite linéarisé, ce qui nous permet d’introduire une commande linéarisante. Cettecommande permet une gestion indépendante du déplacement de chacun des électroaimants,ainsi que du courant qui les traverse. Enfin, les développement théoriques sont testés ensimulation et sur un dispositif expérimental. / Diagnostic techniques currently used in gastrointestinal endoscopy do not allow a completeand accurate observation of the small bowel. Endoscopic capsules were designed to solvethis problem, but the doctor cannot control their movement once they are swallowed bypatients. In that context, the work developed in this thesis focuses on the use of magneticfields to manipulate an object (such as a capsule) without contact, on a large workspace.For this, we study magnetic systems with mobile electromagnets. We study first how tomodel the magnetic field created by an electromagnet. We develop an analytical modeltaking into account the ferromagnetic core of electromagnets. Then, we study a genericsystem composed of n electromagnets which can move throughout space. We develop amodel of the system, integrating the mobility of each of the electromagnets. This modelis then linearized, allowing us to introduce a linearizing control. This command allowsindependent management of the movement of each of the electromagnets, and the currentflowing in it. Finally, the model and the command are tested in simulation and on anexperimental device.Keywords: Magnetism,

Mon travail de thèse porte sur la modélisation électromagnétique et l'imagerie de structures périodiques désorganisées. Un certain motif dans une subdivision élémentaire (une "cellule") est répété dans les autres cellules de la structure dans certaines directions de l'espace. Cette répétition est désorganisée par un changement des propriétés des matériaux et/ou géométries des parties constitutives, dans une ou plusieurs cellules. Au premier niveau de modélisation, ces panneaux sont une succession de plaques planes l'une sur l'autre. Chacun se compose d'un agencement linéaire régulier de longs cylindres avec mêmes sections circulaires finies, tous orientés dans la même direction: nous les appelons "fibres", chaque cylindre résultant de l'hypothèse d'un faisceau de fibres de petite taille. Le matériau constitutif des fibres est différent du matériau d'enrobage (matrice) et le renforce. Chaque plaque est constituée de fibres avec différents axes pour assurer la robustesse. Il y a peu ou beaucoup de plaques, avec la répétition d'une petite pile de plaques. Pour les panneaux conducteurs (à base de carbone), l'imagerie est MHz ; pour des panneaux sans pertes ou à faibles pertes (à base de verre), l'imagerie est micro-onde (quelques dizaines de GHz, voire plus, THz). Il pourrait y avoir des cylindres manquants ou déplacés à l'intérieur d'une plaque, avec des changements conséquents dans éventuellement plusieurs cellules, adjacentes ou non. Des dommages locaux peuvent également se produire, entraînant des changements dans la forme ou les propriétés électromagnétiques d'un ou plusieurs cylindres dans une ou plusieurs cellules dans une ou plusieurs plaques. Un caractère aléatoire de la distribution des inclusions pourrait tenir compte des incertitudes de positionnement par rapport aux géométries supposées. Illuminer correctement les structures et la collecte des champs résultant (dans le champ proche espérons-le, peut-être dans le champ lointain) devraient permettre leur imagerie et concourir à leur diagnostic. Ainsi, si une structure périodique sous interrogation est désorganisée, on souhaite imager cette structure tout en prenant soin au mieux de l'information préalable sur la périodicité et la désorganisation, sur les systèmes de détection, et, évidemment, à propos des besoins et des limites de l'essai. / My PhD work is about electromagnetic modeling and imaging of disorganized periodic structures. A certain pattern in an elementary subdivision (a “cell”) is repeated in the other cells of the structure into certain directions of space. This repetition is disorganized by a change of material properties and/or geometries of the constitutive parts, within one or more cells. At first level of modeling, these panels are a succession of planar plates one over the other. Each consists of a regular linear arrangement of long cylinders with same finite circular sections, all orientated into the same direction: we call them “fibers”, each cylinder resulting from the assumption of a bundle of small-size fibers. The constitutive material of the fibers differs from the embedding material (matrix) that they reinforce. Each plate is made of fibers with different axes for sturdiness. There are few or many plates, with repetition of a small stack of plates. For conductive panels (carbon-based), imaging is MHz; for lossless or weakly lossy panels (glass based), imaging is microwave (a few tens GHz, possibly more). There might be missing/displaced cylinders inside a plate, with consequent changes in possibly several cells, adjacent or not. Local damages might occur also, leading to changes in shape or electromagnetic properties of one or more cylinders in one or more cells in one or more plates. Randomness in distribution of the inclusions might account for uncertainties of positioning with respect to assumed geometries. Properly illuminating the structures and collecting the resulting fields (in the near-field hopefully, possibly in the far-field) should allow their imaging and concur to their diagnostics. So, a periodic structure under interrogation is disorganized. One wishes to successfully image the structure while taking care at best of prior information on periodicity and disorganization, on sensing systems, and obviously of needs and limitations of the testing. The PhD benefits from a grant from the Chinese Scholarship Council.

Esta dissertação tem como objetivo o projeto de um mancal magnético para rodas de reação com aplicação na malha de controle de atitude de satélites. Mancais magnéticos são alternativas aos mancais tradicionais tais como os de esferas ou de lubrificação seco pois trabalham sem contato mecânico entre o rotor e o estator, minimizando assim a fricção entre ambas as partes. Além da minimização do atrito, o ganho em confiabilidade e vida útil da roda de reação é considerável por não apresentar desgastes mecânicos. Devido às consequências de qualquer fricção no movimento relativo entre a inércia (parte rotativa da roda de reação) e o satélite, o mancal torna-se um componente crítico da roda de reação. A fricção se traduz não apenas num maior consumo de potência elétrica, como também na introdução de uma zona morta de atuação em torque, bem como na limitação da vida útil da roda de reação devido ao gradual desgaste do mancal. O mancal proposto possui dois graus de liberdade axiais ativamente controlados e faz uso de ímãs para a estabilização passiva dos demais graus de liberdade. Ao longo do desenvolvimento são apresentados modelos não lineares dos campos magnéticos e das forças atuantes no mancal são encontrados. Com esses modelos, uma otimização é realizada a fim de encontrar melhores características. Um modelo não linear da dinâmica do rotor é desenvolvido e um controle PID capaz de estabilizar o rotor em seu ponto de equilíbrio é apresentado com o objetivo de demonstrar a viabilidade da topologia proposta. / The main objective of this work is to project a magnetic bearing for reaction wheels with application in satellite attitude control. Magnetic bearings are alternatives to traditional bearings such as ball or dry lubrication because they work without mechanical contact between the rotor and the stator thereby minimizing friction between both parts. In addition to minimizing friction, the gain in reliability and lifetime of the reaction wheel is considerable as a consequence of the absence of wear. Because of the consequences of any friction in the relative movement between the inertia (of the reaction wheel) and the satellite ( which is rigidly connected to the satellite body), the bearing becomes a critical component of the reaction wheel. The friction gives rise not only to a greater consumption of electric power, as well as the introduction of a torque dead zone operation, in a reduced lifetime of the reaction wheel due to gradual wear of the bearing. The proposed bearing has two axial degrees of freedom actively controlled and makes use of magnets for the passive stabilization of other degrees of freedom. Nonlinear models of magnetic fields and forces acting on the bearing are presented. With these models, an optimization is performed to find the best bearing characteristics. A nonlinear model rotor dynamics is developed and a PID control capable of stabilizing the active degrees of freedom presented.

Computational modeling offers better insight into a wide range of bioelectrical and biomechanical problems with improved tools for the design of medical devices and the diagnosis of pathologies. Electromagnetic modeling at low and high frequencies is particularly necessary. Modeling electromagnetic, structural, thermal, and acoustic response of the human body to different internal and external stimuli is limited by the availability of numerically efficient computational human models. This study describes the development to date of a computational full-body human model - Visible Human Project (VHP) - Female Model. Its unique feature is full compatibility both with MATLAB and specialized FEM computational software packages such as ANSYS HFSS/Maxwell 3D. This study also describes progress made to date in using the newly developed tools for segmentation. A visualization tool is implemented within MATLAB and is based on customized version of the constrained 2D Delaunay triangulation method for intersecting objects. This thesis applies a VHP - Female Model to a specific application, transcranial Direct Current Stimulation (tDCS). Transcranial Direct Current Stimulation has been beneficial in the stimulation of cortical activity and treatment of neurological disorders in humans. The placement of electrodes, which is cephalic versus extracephalic montages, is studied for optimal targeting of currents for a given functional area. Given the difficulty of obtaining in vivo measurements of current density, modeling of conventional and alternative electrode montages via the FEM has been utilized to provide insight into the tDCS montage performance. An insight into future work and potential areas of research, such as study of bone quality have been presented too.

Le travail de thèse présenté dans ce manuscrit concerne d’une part l’étude théorique des résonances plasmon de nanoparticules métalliques, et d’autre part une étude expérimentale d’inscription de nanostructures métalliques basée sur l’utilisation d’un Microscope à Force Atomique. La partie théorique présente une nouvelle approche phénoménologique permettant l’analyse des modes de résonance propres de particules uniques ainsi que de leur couplage dans des structures simples. Des algorithmes numériques ont été développés afin d’extraire les différents paramètres phénoménologiques à partir du calcul rigoureux du champ diffusé par les particules. Cette méthodologie a été appliquée à divers cas allant de la particule unique à des réseaux à deux dimensions de particules. La partie expérimentale développe une méthode d’inscription de nanostructures métalliques basée sur une réduction électrolytique d’ions métalliques présents dans une couche de silice méso-poreuse, en appliquant une différence de potentiel entre une pointe AFM conductrice et le substrat conducteur supportant la couche. Des structures sont formées de part et d’autre de la couche de silice, avec la possibilité de commuter leur position par simple inversion du potentiel appliqué. De plus, il apparait que cette commutation est accompagnée de modifications dans la conductivité locale de la couche de silice. Une conséquence du processus d’inscription est la formation de filaments métalliques à l’extrémité des pointes AFM. En particulier, des filaments d’or sont obtenus avec des dimensions allant de quelques dizaines à quelques centaines de nanomètres de long pour une épaisseur de quelques nanomètres / The thesis presented in this manuscript concerns firstly the theoretical study of plasmon resonances of metal nanoparticles, and also an experimental study metallic nanostructures inscription based on the use of an Atomic Force Microscope. The theoretical part presents a new phenomenological approach for analyzing the resonant modes of unique particles and their coupling in simple structures. Numerical algorithms have been developed to extract the phenomenological parameters from the rigorous calculation of the field scattered by the particles. This methodology has been applied to various cases from the single particle to two dimensional particle arrays. The experimental section develops a metallic nanostructures inscription method based on electrolytic reduction of metal ions in meso-porous silica thin film, by applying a voltage between a conductive AFM tip and the conductive substrate supporting the film. Structures are formed on both sides of the silica layer, with the possibility to switch their position by a simple reversal of the applied potential. Moreover, it appears that this switching is accompanied by changes in the local conductivity of the silica layer. A consequence of the inscription process is the formation of metal filaments at the ends of AFM tips. In particular, gold filaments are obtained with sizes ranging from tens to hundreds of nanometers long with a few nanometers thick

The use of full wave electromagnetic modeling and simulation tools allows for accurate performance predictions of unique RF structures that exhibit multi-dimensional scales. Full wave simulation tools need to cover the broad range of frequency including RF and terahertz bands that is focused as RF technology is developed. In this dissertation, three structures with multi-dimensional scales and different operating frequency ranges are modeled and simulated. The first structure involves nanostructured solar cells. The silicon solar cell design is interesting research to cover terahertz frequency range in terms of the economic and environmental aspects. Two unique solar cell surfaces, nanowire and branched nanowire are modeled and simulated. The surface of nanowire is modeled with two full wave simulators and the results are well-matched to the reference results. This dissertation compares and contrasts the simulators and their suitability for extensive simulation studies. Nanostructured Si cells have large and small dimensional scales and the material characteristics of Si change rapidly over the solar spectrum. The second structure is a reconfigurable four element antenna array antenna operating at 60 GHz for wireless communications between computing cores in high performance computing systems. The array is reconfigurable, provides improved transmission gain between cores, and can be used to create a more failure resilient computing system. The on-chip antenna array involves modeling the design of a specially designed ground plane that acts as an artificial magnetic conductor. The work involves modeling antennas in a complex computing environment. The third structure is a unique collar integrated zig-zag antenna that operates at 154.5 MHz for use as a ground link in a GPS based location system for wildlife tracking. In this problem, an intricate antenna is modeled in the proximity of an animal. Besides placing a low frequency antenna in a constricted area (the collar), the antenna performance near the large animal body must also be considered. Each of these applications requires special modeling details to take into account the various dimensional scales of the structures and interaction with complex media. An analysis of the challenges and limits of each specific problem will be presented.

Dans ce rapport, un modèle électromagnétique dédié à l’analyse de la diffraction de cibles métalliques placées en milieux forestiers est développé. Le modèle est basé sur une formulation hybride (surface/volume) de la méthode des moments. Le modèle est d’abord développé pour traiter le cas d’une cible métallique placée en espace libre en se basant sur la formulation surfacique de l’équation intégrale du champ électrique. Ensuite, le modèle est modifié afin de prendre en compte la présence du sol en se basant sur la méthode des images couplée aux coefficients de réflexion de Fresnel. Enfin, le code obtenu est couplé à un modèle précédemment développé au laboratoire permettant d’analyser de larges zones forestières. Ce modèle est basé sur la formulation volumique de l’équation intégrale du champ électrique couplée à la méthode CBFM (Characteristics Basis Functions Method). Le code final obtenu permet de traiter le cas d’une cible métallique placée dans une large zone forestière. Le modèle réalisé est implémenté sous MATLAB et Fortran afin de comparer ses résultats avec ceux donnés par le logiciel commercial FEKO. Les résultats sont comparés pour plusieurs cas de figure: espace libre, avec sol et scène complexe. De plus, le code est utilisé afin de mettre en évidence les différents paramètres qui rentrent en jeu dans le mécanisme de diffraction. Enfin, une validation expérimentale réalisée en collaboration avec CCRM (Centre Commun de Ressources en Micro-ondes de Marseille) est présenté. Une mesure du champ diffractée est réalisée sur des maquettes à échelle réduite (1/25) et les résultats obtenus sont comparés avec ceux donnés par notre modèle. / In this report, a model for the analysis of the scattering of metallic target placed in forested area is presented. The developed model is based on a hybrid (surface/volume) formulation of the method of moments. First, a model based on the surface formulation of the electric field integral equation is developed for the case of a metallic target placed in free space. The model is then modified in order to take into account the presence of a ground by using a complex image method coupled with the Fresnel reflection coefficients. Finally, the obtained code is coupled with a model treating the case of large forest areas that was previously developed in the laboratory. This model is based on a volume formulation of the electric field integral equation coupled with the Characteristics Basis Functions Method. The final hybrid model treats the case of metallic object placed in a large forested area. The model is then implemented on MATLAB and Fortran in order to compare its results with those of the commercial software FEKO. The results were compared different cases: free space, above ground and inside a forested area. The code was also used in order to different parameters affecting the diffraction mechanism. Finally, a numerical validation conducted in collaboration with CCRM (Centre Commun de Ressources en Micro-ondes de Marseille) is presented. A measurement complain of the diffracted field was achieve on scaled models (1/25) and the measurements were compared with the results given by the developed model.