Ultra precision single point diamond turning (SPDT) is a machining process used to produce optical grade surfaces in a wide range of materials. Aluminum is of primary interest as a workpiece material because it is easily diamond turnable, highly reflective and corrosion resistant. The cutting tool used is made from a single crystal diamond honed to a very sharp cutting edge. The machines used in this process are extremely precise and stiff. The nature of the cutting parameters used in SPDT changes the process physics substantially over conventional machining. The underlying reason relates to the relative size of the uncut chip thickness and the cutting edge radius of the tool in comparison to the grain size of the workpiece. When performing SPDT, there is a functional limit to the achievable surface finish. This is predominately due to material side flow and the opening up of material defects. Thus the machined surfaces have to undergo post processing operations like lapping or polishing, which increase cost and production time. Thus, the objective of this study was to improve the surface finish of the SPDT process to minimize the amount of post processing. The approach involved addressing the ratio between the tool cutting edge radius and the microstructure. Realizing the limitations associated with sharpening a diamond tool further, efforts have been made to mechanically or thermo-mechanically induce dislocations into the workpiece to refine the microstructure and in so doing enhance machinability. As dislocations act as a point of defect, it is observed that higher dislocation density offers less side flow and leads to better surface roughness. A special tool with a flat secondary edge was then developed to address the remaining side flow issue for planar surfaces. The combination of thermo-mechanically produced ultra fine grained material with the special tool provided a substantial reduction in surface roughness from values typically reported at 3nm [Roblee, 2007] Ra to 0.75nm R0 • In addition to this the use of the custom designed tool can improve the productivity associated with machining a flat face by a factor of one hundred times by allowing the feed rate to be increased while still achieving the desired surface finish. / Thesis / Doctor of Philosophy (PhD)

The pursuit of dense monolithic integration and higher operating speed continues to push the integrated circuit (IC) fabrication technologies to their limits. The increasing process variation, associated with aggressive technology scaling, is having a negative impact on circuit yield in current IC technologies, and the problem is likely to become worse in the future. Circuit solutions that are more tolerant of the process variations are needed to fully utilize the benefits of technology scaling. The primary goal of this research is to develop high-frequency circuits that can deliver consistent performance even under the threat of increasing process variation. These circuits can be used to build ``self-healing" systems, which can detect process imperfections and compensate accordingly to optimize performance. In addition to improving yield, such adaptive circuits and systems can provide more robust and efficient solutions for a wide range of applications under varying operational and environmental conditions.Silicon-germanium (SiGe) BiCMOS technology is an ideal platform for highly integrated systems requiring both high-performance analog and radio-frequency (RF) circuits as well as large-scale digital functionality. This research is focused on designing circuit components for a high-frequency wideband self-healing receiver in SiGe BiCMOS technology. An adaptive image-reject mixer, low insertion-loss switches, a wideband low-noise amplifier (LNA), and a SiGe complementary LC oscillator were designed. Healing algorithms were developed, and automated self-healing of multiple parameters of the mixer was demonstrated in measurement. A monte-carlo simulation based methodology was developed to verify the effectiveness of the healing procedure. In summary, this research developed circuits, algorithms, simulation tools, and methods that are useful for building "self-healing" systems.

The development of monolithic microwave integrated circuits (MMICs) has enabled the expansion of multiple circuit elements on a single piece of semiconductor, enclosed in a package with connecting leads. Attributable to the widespread use of wireless circuits and sub-systems, MMICs meet stringent demands for smaller chip area, low loss and low cost. These require highly integrated MMICs with compact features. This thesis provides valuable insight into the design of compact multifunctional MMICs using three-dimensional (3-D) multilayer technology. The proposed technology offers compact, hence low-cost solutions, where all active and passive components are fabricated vertically on the same substrate and no expensive via hole or backside processing is required. The substrate used in this work contains pre-fabricated 0.5 µm pseudomorphic High Electron Mobility Transistor (pHEMT) GaAs active devices. The performances of the uncommitted and committed pHEMTs are compared in terms of their DC, small-signal and large-signal RF measurements and modelling results. Committed pHEMT refers to the pHEMT that is connected to multilayer circuit, whereas uncommitted pHEMT is not. The effect of integrating committed pHEMTs with multilayer passive components is studied and the suitability of the multilayer fabrication processing is assessed. Using this technology, two pHEMT Schottky diodes with 120 µm and 200 µm gate widths are designed, fabricated and extensively characterised by I-V, C-V and S-parameter measurements. The information gained from the measurements is then used to extract all unknown equivalent circuit model parameters using high-frequency on-wafer microwave probing. The measured results showed good agreement with the modelled ones over the frequency range up to 40 GHz. Preliminary demonstrations of the use of these pHEMT Schottky diodes in microwave limiter and detector circuit applications are also discussed, showing promising results. Finally, the implementation of 3-D multilayer technology is shown for the first time in single-pole single-throw (SPST) and single-pole double-throw (SPDT) switches design by utilising the pre-fabricated pHEMTs. The design and analysis of the switches are demonstrated first through simulation using TriQuint's Own Model - Level 3 (TOM3). Three optimised SPST and SPDT pHEMT switching circuits which can address applications ranging from L to X bands are successfully fabricated and tested. The performance of the pHEMT switches is comparable to those of the current state-of-the-art, while simultaneously offering compact circuits with the advantages of integration with other MMIC components. All works reported in this thesis should facilitate foundry design engineers towards further development of 3-D multilayer technology.

Radio-frequency (RF) switches are widely used in electrical systems, telecommunications, and wireless applications. In RF systems, it is often desirable to change the signal path effectively, by us-ing couplers, duplexers, and RF switches for signal division and combining. Typically, in modern RF systems, the RF switch is mostly capitalized in order to reduce the RF footprint but with efficient switch characteristics. A simple method to reduce transceiver space requirement is to integrate RF switches with the frontend module on a single chip. Recent advances in Gallium Nitride (GaN) technology allows RF designers to design faster, smaller, and efficient components using this technology. With high data rates in demand for wireless communication systems, wideband characteristics are needed in modern systems [1]. Therefore, it is desirable to design wideband circuits; such as, mixers, amplifiers, and switches. In this work, a comprehensive study of NRC GaN150 HEMT is conducted to design broadband RF switches. Single pole and double pole switch topologies operating at 1-12 GHz are designed to evaluate GaN 0.15μm RF switches. The main objectives were to design compact sized switches, while having high power handling, low insertion loss, high isolation and high return loss. Additionally, a transmit-receive switch is designed for integration into a frontend module and further fabricated to operate at 10 GHz. There are many applications of RF switches in an RF transceiver, one of which is an impedance tuner. Impedance tuner are attractive for many applications where mobile devices are used for wireless communications. As mobile technology continues to evolve, they are designed to be com-pact, leaving minimal space for the antenna. Consequently, the radiating element is often electrically small and sensitive to near-field coupling requiring tuning. Matching networks aim to tune matching conditions; for example, loading effects due to human hand [2]. For such situations, specialized matching networks can be designed to account for specific loading environmental effects. However, for mobile systems, the environment is unknown; thereby, yielding unpredictable antenna loading, especially for electrically small antennas that have rapidly changing real and imaginary impedance. As a result, it is necessary to design a reconfigurable impedance-matching network to account for possible load impedances. In this work, a 16-bit reconfigurable impedance tuner design comprising of passive microwave components and NRC GaN 0.15μm FET operating at X-band is presented to evaluate its performance for integration with the frontend module on a single chip to reduce cost and increase efficiency of the system.

Decision trees have long been a prevalent area within machine learning. With streaming data environments as well as large datasets becoming increasingly common, researchers have developed decision tree algorithms adapted to streaming data. One such algorithm is SPDT, which approaches the streaming data problem by making use of workers on a network combined with a dynamic histogram approximation of the data. There exist several implementations for decision trees on GPU, but those are uncommon in a streaming data setting. In this research, conducted at RISE SICS, the possibilities of accelerating the SPDT algorithm on GPU is investigated. An implementation is successfully created using the CUDA platform. The implementation uses a set number of data samples per layer to better fit the GPU platform. Experiments were conducted to investigate the impact on both accuracy and speed. It is found that the GPU implementation performs as well as the CPU implementation in terms of accuracy, suggesting that using small subsets of the data in each layer is sufficient for making accurate split decisions. The GPU implementation is found to be up to 113 times faster than the reference Scala CPU implementation for one of the tested datasets, and 13 times faster on average over all the tested datasets. Weak parts of the implementation are identified, and further improvements are suggested to increase both accuracy and runtime performance. / Beslutsträd har länge varit ett betydande område inom maskininlärning. Strömmandedata och stora dataset har blivit allt vanligare, vilket har lett till att forskare utvecklat algoritmer för beslutsträd anpassade till dessa miljöer. En av dessa algoritmer är SPDT. Denna algoritm använder sig av flera arbetare i ett nätverk kombinerat med en dynamisk histogram-representation av data. Det existerar flera implementationer av beslutsträd på grafikkort, men inte många för strömmande data. I detta forskningsarbete, utfört på RISE SICS, undersöks möjligheten att snabba upp SPDT genom att accelerera beräkningar med hjälp av grafikkort. En lyckad implementation skriven i CUDA beskrivs. Implementationen anpassar sig till grafikkortsplattformen genom att använda sig utav ett bestämt antal datapunkter per lager. Experiment som undersöker effekten på noggrannhet och hastighet har genomförts. Resultaten visar att GPU-implementationen presterar lika väl som CPU-implementationen vad gäller noggrannhet, vilket påvisar att användandet av en mindre del av data i varje lager är tillräckligt för goda resultat. GPU-implementationen är upp till 113 gånger snabbare jämfört med en existerande CPU-implementation skriven i Scala, och är i medel 13 gånger snabbare. Svagheter i implementationen identifieras, och vidare förbättringar till implementationen föreslås för att förbättra både noggrannhet och hastighetsprestanda.

[ES] La Teoría de los Modos Característicos, formulada originalmente en 1968, permite obtener un conjunto de valores propios reales y modos de corriente reales que se pueden utilizar para expandir la corriente total en un cuerpo conductor o dieléctrico. La información proporcionada por las corrientes características (modos) y los valores propios se puede utilizar para realizar el diseño de la antena de una manera sistemática. La gran ventaja de la Teoría de los Modos Característicos sobre otros métodos de diseño tradicionales es la clara visión física que se proporciona sobre los fenómenos que contribuyen a la radiación de la antena. Esta visión física permite comprender mejor el funcionamiento de la antena, de modo que el diseño y la optimización de la misma se pueden llevar a cabo de forma rápida y coherente. El objetivo de esta Tesis es extender el uso de la Teoría de los Modos Característicos al diseño de antenas reconfigurables de manera sistemática. El uso de antenas reconfigurables se ha vuelto cada vez más importante en los últimos años, como una forma de cumplir con los requisitos impuestos y el rendimiento deseado en los nuevos sistemas de comunicaciones inalámbricas. El enfoque de hacer una antena reconfigurable implica la capacidad de una antena para ajustar sus características, como la frecuencia de operación, el diagrama de radiación y la polarización. Diversos componentes activos como diodos varactor, diodos PIN, MEMS, interruptores de RF, elementos fotoconductores o materiales inteligentes pueden usarse para cambiar estas características básicas de la antena. La reconfiguración dinámica de los parámetros operativos de una antena conduce a un sistema de comunicación flexible. Las antenas reconfigurables por frecuencia permiten el mantenimiento de múltiples estándares mientras se mantienen las mismas características físicas de las antenas. La antena reconfigurable por radiación y la diversidad de polarización permiten aumentar la eficiencia de las comunicaciones interiores al reducir significativamente el nivel de interferencia. En esta tesis, la investigación gira en torno al análisis y diseño de estructuras de antenas con la Teoría de los Modos Característicos de manera que los parámetros de las antenas sean sintonizables, por lo que el diseño final de la estructura es una antena reconfigurable que modifica dinámicamente su rendimiento de una forma razonada y coherente. Se analizarán los modos característicos y sus propiedades para diseñar antenas reconfigurables, que combinarán diferentes modos y producirán diferentes características de radiación. En este trabajo se implementará un método sistemático para obtener las características requeridas de la antena. Además, se analizarán y evaluarán diferentes técnicas de conmutación para evaluar su efecto real en la antena y en los modos característicos. El trabajo se centrará en el diseño de pequeñas antenas reconfigurables para aplicaciones IoT y sistemas 5G. / [CA] La Teoria dels Modes Característics, formulada originalment el 1968, permet obtenir un conjunt de valors propis reals i modes de corrent reals que es poden utilitzar per expandir el corrent total en un cos conductor o dielèctric. La informació proporcionada pels corrents característics (modes) i els valors propis es pot utilitzar per fer el disseny de l'antena d'una manera sistemàtica. El gran avantatge de la Teoria dels Modes Característiques sobre altres mètodes de disseny tradicionals és la clara visió física que es proporciona sobre els fenòmens que contribueixen a la radiació de l'antena. Aquesta visió física permet comprendre millor el funcionament de l'antena, de manera que el disseny i l'optimització de la mateixa es poden dur a terme de forma ràpida i coherent. L'objectiu d'aquesta Tesi és estendre l'ús de la Teoria dels Modes Característics al disseny d'antenes reconfigurables de manera sistemàtica. L'ús d'antenes reconfigurables s'ha tornat cada cop més important en els darrers anys, com una forma de complir amb els requisits imposats i el rendiment desitjat en els nous sistemes.temes de comunicacions sense fil. L'enfocament de fer una antenna reconfigurable implica la capacitat duna antena per ajustar les seves característiques, com la frequ¿ència d'operació, el diagrama de radiació i la polarització. Diversos components actius com va donar dos varactor, díodes PIN, MEMS, interruptors de RF, elements fotoconductors o materials intel·ligents poden utilitzar-se per canviar aquestes característiques bàsiques de l'antena. La reconfiguració dinàmica dels paràmetres operatius duna antena condueix a un sistema de comunicació flexible. Les antenes reconfigurables per frequ¿ència permeten el manteniment de múltiples estàndards mentre es mantenen les mateixes característiques físiques de les antenes. L'antena reconfigurable per radiació i diversitat de polarització permeten augmentar l'eficiència de les comunicacions interiors en reduir significativament el nivell dinterferència. En aquesta tesi, la investigació gira al voltant de l'anàlisi i el disseny de estructures d'antenes amb la Teoria dels Modes Característics de manera que els paràmetres de les antenes siguin sintonitzables, per això que el disseny final de lestructura és una antena reconfigurable que modifica dinàmicament el seu rendiment d'una forma raonada i coherent. S'analitzaran els modes característics i les propietats per dissenyar antenes reconfigurables, que combinaran diferents maneres i produiran diferents característiques de radiació. En aquest tractament baix s'implementarà un mètode sistemàtic per obtenir les característiques requerides de l'antena. A més, s'analitzaran i avaluaran diferents tècniques de commutació per avaluar el seu efecte real a la antena i en les maneres característiques. El treball se centrarà en el disseny de petites antenes reconfigurables per a aplicacions IoT i sistemes 5G. / [EN] The Theory of Characteristic Modes, originally formulated in 1968, allows to obtain a set of real eigenvalues and real current modes that can be used to expand the total current on a conducting or dielectric body. The information provided by characteristic eigencurrents (modes) and eigenvalues can be used to perform antenna design in a systematic way. The great advantage of the Theory of Characteristic Modes over other traditional design methods is the clear physical vision provided about the phenomena that contribute to the radiation of the antenna. This physical vision allows to better understand the operation of the antenna, so that the design and the optimization of it can be carried out quickly and coherently. The objective of this Thesis is to extend the use of the Theory of Characteristic Modes to the design of reconfigurable antennas in a systematic way. The use of reconfigurable antennas has become more and more important during the last years, as a way to meet the imposed requirements and desired performance of novel wireless communication systems. The approach of making an antenna reconfigurable implies the capability of an antenna system to adjust its characteristic such as operating frequency, radiation pattern and polarization. Active components such as varactor diodes, PIN diodes, MEMS, RF switches, photoconductive elements or smart materials can be used to change these base characteristics. The dynamic reconfigurability of the operating parameters of an antenna leads to a flexible communication system. Frequency reconfigurable antennas enable multi-standard performances to be maintained while preserving the same physical characteristics of the antennas. Radiation reconfigurable and polarisation diversity antennas enable to increase the efficiency of indoor communications by significantly decreasing the level of interference. In this thesis, the research is turning around the analysis and design of antenna structures with the Theory of Characteristic Modes in a way that the antenna parameters will be tuneable, so the final design of the structure is a reconfigurable antenna that modifies dynamically its performance in a reasoned and coherent way. Characteristic modes and their properties will be analysed in order to design reconfigurable antennas, which combine different modes and produce different radiation characteristics. A systematic method to obtain the required characteristics of the antenna will be implemented during this work. Moreover, different switching techniques will be analysed and evaluated in order to asses their real effect on the antenna and on the characteristic modes. In this work, small reconfigurable antenna prototypes will be designed for IoT applications and 5G systems. / This study was supported by the Spanish Ministry of Science and Innovation (Ministerio Ciencia e Innovación) under project no. PID2019-107885GB-C32 and Generalitat Valenciana under project no. AICO/2019/018. This work has also been supported by the Spanish Ministry of Science, Innovation and Universities (Ministerio Ciencia, Innovación y Universidades) under the project TEC2016- 78028-C3-3-P and by “Centre National pour la Recherche Scientifique et Technique (CNRST, Maroc)” as part of the research excellence scholarships program / Mahlaoui Boudallaa, Z. (2023). Analysis and Design of Reconfigurable Antennas Using the Theory of Characteristic Modes [Tesis doctoral]. Universitat Politècnica de València. https://doi.org/10.4995/Thesis/10251/201555

L'introduction des technologies MEMS ("Micro Electro Mechanical Systems") dans les modules hyperfréquences répond au besoin croissant en systèmes de communications intégrables, reconfigurables et présentant d'excellentes performances électriques jusqu'aux fréquences millimétriques. Le développement de nouvelles architectures intelligentes jusque là inaccessibles aux technologies traditionnelles est également envisageable grâce à ces composants. Cela dit, la conception multi-physique de ces circuits alliant des aspects électrostatiques, mécaniques et électromagnétiques reste difficile à mettre en Suvre et complique leur optimisation. De plus, peu de recherches se sont focalisées sur la tenue en puissance de ces composants, pourtant primordiale pour envisager leur intégration dans des chaînes d'émission radio fréquences. Nos travaux de thèse ont porté sur la conception et la caractérisation de micro-interrupteurs MEMS de puissance et de circuits hyperfréquences les intégrant et opérant en bande X (10GHz). Le premier chapitre présente une méthodologie multi-physique de conception de commutateurs MEMS RF électrostatiques à contact capacitif réalisés au Laboratoire d'Analyse et d'Architecture des Systèmes. Cette méthodologie, associée à une topologie optimale de micro-commutateurs, a permis un prototypage efficace de commutateurs MEMS et a été validée expérimentalement grâce à des structures montrant d'excellentes performances hyperfréquences. Le second chapitre s'intéresse à l'optimisation en puissance de commutateurs MEMS RF. Nous avons développé une méthodologie de prédiction de la puissance maximale de fonctionnement pour ces composants. Cette méthodologie a été ensuite utilisée pour loptimisation en puissance du commutateur développé dans le chapitre 1. Un dimensionnement ainsi que lensemble des résultats de simulations sont présentés et validés expérimentalement. Enfin le dernier chapitre présente la mise en application des méthodologies décrites aux deux premi ers chapitres pour la conception d'un circuit de commutation de puissance large bande 6-18 GHz basé sur des topologies série et parallèle d'interrupteurs MEMS. Les structures ont ainsi été optimisées, fabriquées et mesurées. Elles présentent dexcellentes performances RF sur une large gamme de fréquence.

RF/Microwave switches are widely used in microwave measurement systems, telecommunication and radar applications. The main purposes of RF switches are Tx-Rx switching, band select and switching the signal between different paths. Thus, they are key circuits especially in T/R modules. Wideband operation is an important criterion in EW applications. High power handling is also a key feature especially for radars detecting long range. In this study, different types of high power solid state switches operating at X-Band are designed, fabricated and measured. The main objectives are small size and high power handling while keeping good return loss and low insertion loss. The related studies are investigated and analyzed. Solutions for increasing the power handling are investigated, related calculations are done. Better bias conditions are also analyzed. The measurement results are compared with simulations and analysis. Circuit designs and simulations are performed using AWR&reg / and CST&reg / .

This thesis presents realization of two shunt, capacitive contact RF MEMS switches and two RF MEMS SPDT switches for microwave and millimeter-wave applications, two zero-level package structures for RF MEMS switches and development trials of a BCB based zero level packaging process cycle. Two shunt, capacitive contact RF MEMS switches for 26 GHz and 12 GHz operating frequencies are designed, fabricated and consistencies between fabricated devices and designs are shown through RF measurements. For the switch design at 26 GHz and at the operating frequency, return loss in the upstate is measured to be 27.61 dB, insertion loss and isolation in the downstate is measured to be 0.21 dB and 27.16 dB, respectively. For the switch design at 12 GHz and at the operating frequency, return loss in the upstate is measured to be 38.69 dB, insertion loss and isolation in the downstate is measured to be 0.05 dB and 25.84 dB, respectively. Quite accurate circuit models have been obtained for both of the RF MEMS switches. Two RF MEMS SPDT switches, which utilize the shunt, capacitive contact switches as building blocks are designed through circuit simulations. These two designs are fabricated and their RF measurements have been completed. It is shown from circuit model simulations that, the performances of the fabricated devices and desired responses corresponded to each other. For the SPDT switch design at 26 GHz, return loss at the input port is measured to be 12 dB and insertion loss is measured to be 1.24 dB. For the SPDT switch design at 12 GHz, return loss at the input port is measured to be 5.6 dB and insertion loss is measured to be 0.49 dB. The reason behind the unexpectedly bad performances has been investigated and discovered. The bad performances were due to a common mistake in the layouts of both SPDT switches. These mistakes are corrected in the circuit models and expected performances are obtained. Two different zero-level package structures which use high-resistive Si wafers have been suggested and required design changes have been made on the RF MEMS shunt, capacitive contact switches and SPDT switches in order to minimize the package effects. For this purpose polygonal CPW transitions have been designed and integrated into the designs, followed by the necessary tunings in the switch structures for which EM and circuit simulations are utilized. For the suggested package structures to be produced, two possible process cycles have been studied. One of the process flows was based on KOH anisotropic Si etching and the other one was based on DRIE (Deep Reactive Ion Etching). Great progress has been achieved in the latter process cycle, however this process cycle still needs some more study and it could not be completed in the time required for this thesis study.

The primary reason for the wireless technology evolution is towards building capacity and obtaining higher data rates. Enclosed locations, densely populated campus, indoor offices, and device-to-device communication will require radios that need to operate at data rates up to 10 Gbps. In the next few years, a new generation of communication systems would emerge to better handle the ever-increasing demand for much wider bandwidth requirements. Simultaneously, key factors such as size, cost, and energy consumption play a distinctive role towards shaping the success of future wireless technologies. In the perspective of 3GPP 5G next generation wireless communication systems, the X band was explicitly targeted with a vast range of applications in point to point radio, point to multi point radio, test equipment, sensors and future wireless communication. An X-band RF front-end circuit for next generation wireless network applications is presented in this work. It details the design of a low noise amplifier and a power amplifier for X band operation. The designed amplifiers were integrated with a wideband single-pole-double-throw switch to achieve an overall front-end structure for 10 GHz. The design was carried out and sent for fabrication using a GaN 0.15µm process provided by NRC, a novel design kit. Due to higher breakdown voltage, high power density, high efficiency, high linearity and better noise performance, GaN HEMTs are a suitable choice for future wireless communication. Thus, the assumption is to further explore capabilities of this process in front-end design for future wireless communications.