Dissertations / Theses on the topic 'Silicon photonics'

Les cristaux photoniques à gradient (CPG) permettent une ingénierie de leur indice effectif, ce qui offre de nouveaux degrés de liberté pour la conception de dispositifs photoniques. Ils s’appréhendent par l’optique à gradient d’indice (GRIN optics), qui décrit des milieux inhomogènes dans lesquels la lumière ne se propage pas rectilignement. Il est ainsi possible d’envisager tout profil d’indice. Les CPG sont donc particulièrement attractifs pour la miniaturisation des composants optiques, notamment en photonique sur Silicium. Ils sont fondés sur la variation d’un paramètre de la maille élémentaire du cristal photonique (CP); ici, c’est le facteur de remplissage qui varie afin que l’indice effectif du CPG réalise le profil d’indice souhaité. Le but de cette thèse est d’explorer le potentiel des CPG en concevant des dispositifs à gradient d’indice sur la "plateforme" Silicium sur isolant (SOI) aux longueurs d’onde pour les télécommunications. C’est la chaine complète qui va de la conception à la caractérisation du dispositif, en passant par la simulation et la fabrication, qui est mise en œuvre. Nous nous sommes principalement concentrés sur deux instruments typiques de l’optique à gradient d’indice : la lentille de Mikaelian et le Half Maxwell Fish Eye (HMFE). Dans cette thèse, nous proposons une nouvelle méthode d’approximation de l’indice effectif adaptée à la "plateforme" SOI, que nous avons validée en concevant une lentille de Mikaelian (à profil d’indice sécante hyperbolique). Pour de tels dispositifs, il faut en effet tenir compte de deux indices effectifs : celui du mode guidé dans la couche de Silicium et celui du CP. Dans cette méthode, l’indice effectif du CP est d’abord calculé pour remplacer l’indice de la couche du mode guidé ; puis l’indice effectif de cette couche est calculé. Les résultats de simulation obtenus au moyen d’un logiciel commercial (méthode FDTD) montrent que la lentille ainsi conçue satisfait les prévisions analytiques, contrairement à ce que donnent les méthodes couramment utilisées. Nous l’avons alors appliquée au HMFE. Les dispositifs ont ensuite été fabriqués en salle blanche par lithographie par faisceau d’électrons (EBL) et par gravure plasma (ICP). Les différents CPG fabriqués consistent en des trous d’air répartis périodiquement dans la couche de Silicium, dont le diamètre minimal est d’environ 40 nm. Puis, ils ont été caractérisés en deux temps, notamment par microscopie en champ proche (SNOM). L’épaisseur de ces dispositifs est de quelques longueurs d’onde (3 ou 5 λ_0 environ), tandis la largeur de leur tâche focale est proche de la limite de diffraction (0.5 λ_0 environ). Ils fonctionnent sur une plage de longueurs d’onde de 150 nm environ. Les résultats de la lentille de Mikaelian ont été utilisés pour développer un convertisseur de taille de mode (taper) effectif sur quelques longueurs d’onde. Il est dix fois plus court qu’un convertisseur classique. Dans cette thèse, nous montrons aussi comment il est possible d’interpréter la propagation de l’onde EM dans ces composants à gradient d’indice sur "plateforme" SOI au moyen du principe de l’interféromètre multimode. En se propageant, les différents modes accumulent une différence de phase, qui se traduit par un battement qui modifie la distribution du champ EM, conduisant à la focalisation. La longueur caractéristique de ce battement est égale à la distance focale. Tous ces dispositifs sont étudiés pour s’intégrer dans des circuits de photonique intégrée
Gradient photonic crystals (GPhCs) enable the engineering of their effective index, opening up new degrees of freedom in photonic device design. They can be understood through gradient index optics (GRIN optics), which describe inhomogeneous media in which light does not propagate along straight paths. This makes it possible to consider any index profile. This makes GPhCs particularly attractive for the miniaturization of optical components, especially in silicon photonics. They are based on the variation of a parameter of the photonic crystal elemental cell (PhC); here, the filling factor is varied so that the effective index of the GPhC achieves the desired index profile. The aim of this thesis is to explore the potential of GPhCs by designing graded-index devices on the Silicon-On-Insulator (SOI) "platform" at telecom wavelengths. The complete chain from design to device characterization, including simulation and manufacturing, is implemented. We focused on two typical gradient index optics instruments: the Mikaelian lens and the Half Maxwell Fish Eye (HMFE). In this thesis, we propose a new effective index approximation method for the SOI "platform", which we have validated by designing a Mikaelian lens (with a hyperbolic secant index profile). For such devices, two effective indices need to be taken into account: that of the guided mode in the Silicon layer and that of the PhC. In this method, the effective index of the PhC is first calculated to replace the index of the guided mode layer; then the effective index of this layer is calculated. Simulation results obtained using commercial software (FDTD method) show that the lens designed in this way satisfies the analytical predictions, contrary to the results obtained with commonly used methods. We then applied it to HMFE.The devices were then fabricated in the cleanroom by electron beam lithography (EBL) and plasma etching (ICP). The individual GPhCs consisted of periodically distributed air holes in the Silicon layer, with a minimum diameter of around 40 nm. They were then characterized in two stages, notably by near-field microscopy (SNOM). These devices are only a few wavelengths thick (approx. 3 or 5 λ_0), while their focal spot width is close to the diffraction limit (approx. 0.5 λ_0). They operate over a wavelength range of around 150 nm. The Mikaelian lens results have been used to develop a mode size converter (taper), which is effective over a few wavelengths. It is ten times shorter than a conventional converter. In this thesis, we also show how it is possible to interpret EM wave propagation in these graded-index components on the SOI platforms using the multimode interferometer principle. As they propagate, the different modes accumulate a phase difference, resulting in a mode beat that modifies the EM field distribution, leading to focusing. The characteristic length of this mode beat is equal to the focal length. All these devices are studied for integration into integrated photonics circuits

The mid-infrared wavelength region (2-20 µm) is of great utility for a number of applications, including chemical bond spectroscopy, trace gas sensing, and medical diagnostics. Despite this wealth of applications, the on-chip mid-IR photonics platform needed to access them is relatively undeveloped. Silicon is an attractive material of choice for the mid-IR, as it exhibits low loss through much of the mid-IR. Using silicon allows us to take advantage of well-developed fabrication techniques and CMOS compatibility, making the realization of on-chip integrated mid-IR devices more realistic. The mid-IR wavelengths also afford the opportunity to exploit Si's high third-order optical nonlinearity for nonlinear frequency generation applications. In this work, we present a Si-based platform for mid-IR photonics, with a special focus on micro-resonators for strong on-chip light confinement in the 4-5 μm range. Additionally, we develop experimental optical characterization techniques to overcome the inherent difficulties of working in this wavelength regime. First, we demonstrate the design, fabrication, and characterization of photonic crystal cavities in a silicon membrane platform, operational at 4.4 μm (Chapter 2). By transferring the technique known as resonant scattering to the mid-IR, we measure quality (Q) factors of up to 13,600 in these photonic crystal cavities. We also develop a technique known as scanning resonant scattering microscopy to image our cavity modes and optimize alignment to our devices. Next, we demonstrate the electro-optic tuning of these mid-IR Si photonic crystal cavities using gated graphene (Chapter 3). We demonstrate a tuning of about 4 nm, and demonstrate the principle of on-chip mid-IR modulation using these devices. We then investigate the phenomenon of optical bistability seen in our photonic crystal cavities (Chapter 4). We discover that our bistability is thermal in origin and use post-processing techniques to mitigate bistability and increase Q-factors. We then demonstrate the design, fabrication, and characterization grating-coupled ring resonators in a silicon-on-sapphire (SOS) platform at 4.4 μm, achieving intrinsic Q-factors as high as 278,000 in these devices (Chapter 5). Finally, we provide a quantitative analysis of the potential of our SOS devices for nonlinear frequency generation and describe ongoing experiments in this regard (Chapter 6).
Engineering and Applied Sciences

Thanks to its compatibility with the current CMOS technology and its potential of seamless integration with electronics, silicon photonics has been attracting an ever-increasing interest in recent years from both the academia and industry. By applying silicon photonic technology in microwave photonics, on-chip integration of microwave photonic systems could be implemented with improved performance including a much smaller size, better stability and lower power consumption. This thesis focuses on developing silicon-based photonic integrated circuits for microwave photonic applications. Two types of silicon-based on-chip devices, waveguide Bragg gratings and optical micro-cavity resonators, are designed, developed, and characterized, and the use of the developed devices in microwave photonic applications is studied. After an introduction to silicon photonics and microwave photonics in Chapter 1 and an overview of microwave photonic signal generation and processing in Chpater2, in Chapter 3 a silicon-based on-chip phase-shifted waveguide Bragg grating (PS-WBG) is designed, fabricated and characterized, and its use for the implementation of a photonic temporal differentiator is experimentally demonstrated. To have a waveguide grating that is wavelength tunable, in Chapter 4 a tunable waveguide grating is proposed by incorporating a PN junction across the waveguide grating, to use the free-carrier plasma dispersion effect in silicon to achieve wavelength tuning. The use of a pair of wavelength-tunable waveguide gratings to form a wavelength-tunable Fabry-Perot resonator for microwave photonic signal processing is studied. Thanks to its electrical tunability, a high-speed electro-optic modulator, a tunable fractional-order photonic temporal differentiator and a tunable optical delay line are experimentally demonstrated. To increase the bandwidth of a waveguide grating, in Chapter 5 a linearly chirped waveguide Bragg grating (LC-WBG) is designed, fabricated and evaluated. By incorporating two LC-WBGs in two arms of a Mach-Zehnder interferometer (MZI) structure, an on-chip optical spectral shaper is produced, which is used in a photonic microwave waveform generation system based on spectral-shaping and wavelength-to-time (SS-WTT) mapping for linearly chirped microwave waveform (LCMW) generation. To enable the LC-WBG to be electrically tuned, in Chapter 6 a lateral PN junction is introduced in the grating and thus an electrically tunable LC-WBG is realized. By incorporating two tunable LC-WBGs in a Michelson interferometer structure, an electrically tunable optical spectral shaper is made. By applying the fabricated spectral shaper in an SS-WTT mapping system, a continuously tunable LCMW is experimentally generated. Compared with a waveguide Bragg grating device, an on-chip optical micro-cavity resonator usually has a much smaller dimension, which is of help to increase the integration density and reduce the power consumption. Different on-chip optical micro-cavity resonators are studied in this thesis. In Chapter 7, an on-chip symmetric MZI incorporating multiple cascaded microring resonators is proposed. By controlling the radii of the rings, the MZI could be designed to have a spectral response with a linearly-varying free spectral range (FSR), which could be used in photonic generation of an LCMW, and to have a multi-channel spectral response with identical channel spacing, which could be used in the implementation of an independently tunable multi-channel fractional-order temporal differentiator. To further reduce the footprint of an optical micro-cavity resonator, in Chapter 8 an ultra-compact microdisk resonator (MDR) with a single-mode operation and an ultra-high Q-factor is proposed, fabricated and evaluated, and its use for the implementation of a microwave photonic filter and an optical delay line is experimentally demonstrated. To enable the MDR to be electrically tunable, in Chapter 9 an electrically tunable MDR is realized by incorporating a lateral PN junction in the disk. The use of the fabricated MDR in microwave photonic applications such as a high-speed electro-optic modulator, a tunable photonic temporal differentiator and a tunable optical delay line is experimentally demonstrated.

Chip scale photonic integrated circuits can provide important new functions in communications, signal processing and sensing. Recent research on microwave photonics (MWPs) and integrated optical sensors using the silicon photonic devices has opened up new opportunities for signal processing and sensing applications. MWPs brings together the world of microwave engineering and optoelectronics, which provides solutions for processing high frequency microwave signals. It has attracted significant interest in many different areas including communications, sensors, radar systems and defence applications. The use of photonic integrated circuit enhances functionalities and flexibilities as well as enabling a reduction of size and weight for MWP applications. The high integratablity of the photonic circuit not only boosts the filtering, time delay and phase shifting functionalities, but also enables the sensing applications in the nano-scale range. Integrated sensors are under high demand in many environmental chemical and biomedical applications. The mass fabricated integrated sensor provides opportunities for multi-functional sensor array with minimized volume. The research work presented in this thesis aims to investigate silicon photonics applications in MWP signal processing and different sensing circumstances. Firstly, the MWP filter based on the SOI microring resonator with phase compensation method is demonstrated. In addition, instantaneous frequency measurement based on frequency to time mapping is presented. Then, a novel integrated optical sensor system based on SOI add drop microring resonator structure is presented. The MWP techniques for high performance sensing application is explored. Lastly, to address the multi-functionality of silicon photonics based sensor, an application of integrated ultrasound optical sensor is demonstrated. It is expected the work provided in this thesis can assist in the emergence of real-world silicon photonic applications. (1992 out of 2000 characters)

It is widely accepted that the continued increase of processor performance requires at least partial replacement of electronic interconnects with their photonic counterparts. The implementation of optical interconnects requires the realization of a silicon-based light source, which is challenging task due to the low emission efficiency of silicon. One of the main approaches to address this challenge is the use of doping of silicon based matrices with optical centers, including erbium ions. Erbium ions incorporated in various hosts assume the trivalent state (Er[super]3+) and demonstrate a transition at 1.54 micrometer], coinciding with optical transmission windows in both silicon and silica. Due to the low absorption cross-section and discrete energy levels of the Er[super]3+ ion, indirect excitation is necessary. In late 90s it was demonstrated that the incorporation of excess silicon in erbium-doped silica results in strong erbium sensitization, leading to an increase of the effective absorption cross-section by orders of magnitude. The sensitization was considered to occur via silicon nanocrystals that formed at high annealing temperatures. While a large increase of the absorption cross-section was demonstrated, the incorporation of Si nanocrystals was found to result in a low concentration of excited erbium, as well as silicon related free-carrier absorption. The focus of this dissertation is the investigation of the nature of the sensitization mechanism of erbium in silicon-rich silica. The results presented in the dissertation demonstrate that erbium in silicon-rich silica is predominantly excited by silicon-excess-related luminescence centers, as opposed to the commonly considered silicon nanocrystals. This is a remarkable conclusion that changes the view on the exact origin of erbium sensitization, and that resolves several technical challenges that exist for nanocrystal-based sensitization.; The work shows that in order to sensitize erbium ions in silicon-rich silica there is no need for the presence of silicon nanocrystals, and consequently lower fabrication temperatures can be used. More importantly, the results strongly suggest that higher gain values can be acquired in samples annealed at lower temperature (without silicon nanocrystals) as compared to samples annealed at high temperatures (with silicon nanocrystals). In addition, the maximum gain is predicted to be relatively independent of excitation wavelength, significantly relaxing the requirements on the pump source. Based on the experimental results it is predicted that relatively stable performance of erbium-doped silicon-rich silica is possible up to typical processor operating temperatures of ~ 80 - 90[degrees]C making it a viable material for on-chip devices. The results suggest that low temperature annealed erbium-doped silicon-rich silica is a preferable material for on-chip photonic devices as compared with its high temperature annealed counterpart.; The work shows that the density of indirectly excited erbium ions is significantly larger in samples without silicon nanocrystals (annealed at T[less than]1000[degrees]C) as opposed to samples with silicon nanocrystals (annealed at T[greater than]1000[degrees]C). The density of indirectly excited erbium ions, defining the maximum achievable gain, was demonstrated to be approximately excitation wavelength independent, while the effective erbium absorption cross-section was shown to significantly depend on the excitation wavelength. The excitation mechanism of erbium by luminescence centers was shown to be fast (less than] 30 ns) and capable of erbium sensitization to different energy levels. This multilevel nature of erbium excitation was demonstrated to result in two different mechanisms of the excitation of the first excited state of erbium: fast (less than]30 ns) direct excitation by the luminescence centers, and slow (greater than]2.3 microseconds]) excitation due to the relaxation of erbium ions excited into higher energy levels to the first excited state. Based on photoluminescence studies conducted in the temperature range 15-300K it was shown that the relaxation efficiency of erbium from the second excited state to the first excited state (responsible for the slow excitation mechanism) is temperature independent and approaches unity. The relative stability of the optical properties demonstrated in the temperature range 20-200[degrees]C, implies that relatively stable optical gain can be achieved under realistic on-chip operating conditions. The optimum Si excess concentration corresponding to the highest density of sensitized Er[super]3+ ions is shown to be relatively insensitive to the presence of Si nanocrystals and is ~ 14.5 at.% and ~ 11.5 at.% for samples without and with Si nanocrystals respectively. The presented results and conclusions have significant implications for silicon photonics and the industrial application of Er-doped SiO[sub]2. The work shows that in order to sensitize erbium ions in silicon-rich silica there is no need for the presence of silicon nanocrystals, and consequently lower fabrication temperatures can be used. More importantly, the results strongly suggest that higher gain values can be acquired in samples annealed at lower temperature (without silicon nanocrystals) as compared to samples annealed at high temperatures (with silicon nanocrystals). In addition, the maximum gain is predicted to be relatively independent of excitation wavelength, significantly relaxing the requirements on the pump source. Based on the experimental results it is predicted that relatively stable performance of erbium-doped silicon rich silica is possible up to typical processor operating temperatures of ~ 80 - 90[degrees]C making it a viable material for on-chip devices. The results suggest that low temperature annealed erbium doped silicon-rich silica is a preferable material for on-chip photonic devices as compared with its high temperature annealed counterpart.
ID: 029094291; System requirements: World Wide Web browser and PDF reader.; Mode of access: World Wide Web.; Thesis (Ph.D.)--University of Central Florida, 2010.; Includes bibliographical references.
Ph.D.
Doctorate
Optics and Photonics

Silicon photonics technologies have the potential to overcome the bandwidth limitations inherent in electrical interconnect technology. Modulation technology which is efficient both in terms of size and energy is required if silicon photonics are to replace electronics for interconnect communications. Silicon germanium technologies have the potential to not only improve the performance of current semiconductor devices but to also extend the reach of semiconductor technology into new areas such as development of a room temperature THz laser. A novel process that allows easy fabrication of Ohmic contacts to moderately doped n-type Germanium has been developed. This process has the potential to allow the realization of new devices which have been previously hampered by non-Ohmic contacts or dopant segregation problems. This work reported in this thesis also includes the design and fabrication of Ge/SiGe QCSE devices. Thin barrier QCSE designs have been put forward as a potential way to produce a more energy efficient modulator. Simulations of the devices show that a design with 16 nm Ge QWs and 8 nm SiGe barriers can provide effective modulation covering the entire optical communications C band with less than 3 V DC offset and achieve a contrast ratio across the band of over 3 dB. It was also shown that despite the thin barriers the wavefunctions remain well confined to the QWs suggesting that even thinner barriers are possible. MQW structures with thin barriers were grown and photodiodes fabricated from them. While the wafers did not have barriers as thin as designed they were thinner than devices previously demonstrated. From photocurrent measurements it was shown that these MQW structures were able to effectively modulate light near the 1550 nm wavelength with better performance than devices found in the literature.

We have performed a theoretical study into silicon-on-insulator photonic waveguide arrays. Such waveguides are capable of high levels of light confinement which reinforces the already strong nonlinear response of silicon, making systems involving the waveguidcs ideal for the study of non-linear effects. This study is focussed on two nonlinear processes in relation to the waveguide arrays: optical soli tans and modulational instability, which are often related effects themselves. Optical solitons are pulses localised ill Due or more spatial and/or temporal dimensions which propagate through media -in a. robust, self-reinforcing manner. They require a balance between nonlinearity, diffraction and dispersion. Modulational instability is related to wave-mixing processes whereby photons of a certain frequency arc converted to photons of different frequencies, depending on phase matching and conservation laws. The instability causes the growth of spectral sidebands about a pump pulse, and is often found to occur during soliton propagation. In this thesis a study of the propagation of light within arrays of waveguides is presented, wherein conditions are tuned to promote soliton formation and an emphasis is placed on investigating discrete spatiotemporal solitons. Advantages and disadvantages of employing silicon waveguides for soliton formation are noted with suggestions given to enable minimising of the latter. It is shown that silicon-on-insulator waveguides can provide an excellent medium for supporting discrete spatiotemporal solitons, and where applicable theoretical results have been related to experimental ones performed in tandem . Similar arrays to used to study modulational instability. It is shown that, through exploitation of the supermodes supported by a waveguide array, different degrees of instability, quantified by an amount of 'gain', are possible within the same array. Depending on the initial excitation conditions it is possible for a pulse to experience either large or insignificant amounts of the gain.

El silicio es la plataforma más prometedora para la integración fotónica, asegurando la compatibilidad con los procesos de fabricación CMOS y la producción en masa de dispositivos a bajo coste. Durante las últimas décadas, la tecnología fotónica basada en la plataforma de silicio ha mostrado un gran crecimiento, desarrollando diferentes tipos de dispositivos ópticos de alto rendimiento. Una de las posibilidades para continuar mejorando las prestaciones de los dispositivos fotónicos es mediante la combinación con otras tecnologías como la plasmónica o con nuevos materiales con propiedades excepcionales y compatibilidad CMOS. Las tecnologías híbridas pueden superar las limitaciones de la tecnología de silicio, dando lugar a nuevos dispositivos capaces de superar las prestaciones de sus homólogos electrónicos. La tecnología híbrida dióxido de vanadio/ silicio permite el desarrollo de dispositivos de altas prestaciones, con gran ancho de banda, mayor velocidad de operación y mayor eficiencia energética con dimensiones de la escala de la longitud de onda. El objetivo principal de esta tesis ha sido la propuesta y desarrollo de dispositivos fotónicos de altas prestaciones para aplicaciones de conmutación. En este contexto, diferentes estructuras basadas en silicio, tecnología plasmónica y las propiedades sintonizables del dióxido de vanadio han sido investigadas para controlar la polarización de la luz y para desarrollar otras funcionalidades electro-ópticas como la modulación.
Silicon is the most promising platform for photonic integration, ensuring CMOS fabrication compatibility and mass production of cost-effective devices. During the last decades, photonic technology based on the Silicon on Insulator (SOI) platform has shown a great evolution, developing different sorts of high performance optical devices. One way to continue improving the performance of photonic optical devices is the combination of the silicon platform with another technologies like plasmonics or CMOS compatible materials with unique properties. Hybrid technologies can overcome the current limits of the silicon technology and develop new devices exceeding the performance metrics of its counterparts electronic devices. The vanadium dioxide/silicon hybrid technology allows the development of new high-performance devices with broadband performance, faster operating speed and energy efficient optical response with wavelength-scale device dimensions. The main goal of this thesis has been the proposal and development of high performance photonic devices for switching applications. In this context, different structures, based on silicon, plasmonics and the tunable properties of vanadium dioxide, have been investigated to control the polarization of light and for enabling other electro-optical functionalities, like optical modulation.
El silici és la plataforma més prometedora per a la integració fotònica, assegurant la compatibilitat amb els processos de fabricació CMOS i la producció en massa de dispositius a baix cost. Durant les últimes dècades, la tecnologia fotònica basada en la plataforma de silici ha mostrat un gran creixement, desenvolupant diferents tipus de dispositius òptics d'alt rendiment. Una de les possibilitats per a continuar millorant el rendiment dels dispositius fotònics és per mitjà de la combinació amb altres tecnologies com la plasmònica o amb nous materials amb propietats excepcionals i compatibilitat CMOS. Les tecnologies híbrides poden superar les limitacions de la tecnologia de silici, donant lloc a nous dispositius capaços de superar el rendiment dels seus homòlegs electrònics. La tecnologia híbrida diòxid de vanadi/silici permet el desenvolupament de dispositius d'alt rendiment, amb gran ample de banda, major velocitat d'operació i major eficiència energètica en l'escala de la longitud d'ona. L'objectiu principal d'esta tesi ha sigut la proposta i desenvolupament de dispositius fotònics d'alt rendiment per a aplicacions de commutació. En este context, diferents estructures basades en silici, tecnologia plasmònica i les propietats sintonitzables del diòxid de vanadi han sigut investigades per a controlar la polarització de la llum i per a desenvolupar altres funcionalitats electró-òptiques com la modulació.
Sánchez Diana, LD. (2016). High performance photonic devices for switching applications in silicon photonics [Tesis doctoral no publicada]. Universitat Politècnica de València. https://doi.org/10.4995/Thesis/10251/77150
TESIS

Silicon based integrated circuits has been dominating the electronics technology industry in the last few decades. As the telecommunications and the computing industry slowly converges together, the need for a material to build photonics integrated circuits (PIC) that can be cost-effective and be produced in mass market has become very important. This thesis describes and outlines the characteristics of high index contrast waveguides as a building blocks that can be designed, fabricated and employed on devices in silicon photonics. Initially in this work, a fully vectorial H-field based finite element method has been used to obtain the modal characteristics of high index contrast bent waveguide to get a better understanding of the curved section. Through the beam propagation method, the propagation losses and the spot-size along the propagation distance are obtained when a mode from the straight guide is launched into a bent guide. It is also learnt that mode beating exists at the junction of a straight-to-bent waveguide, in which higher order modes will also be generated. It will be shown in this work that power do exchange between the two polarization states, therefore the polarization conversion, the power losses and the bending losses will be investigated. It will also shown in here that by applying lateral offsets with coupled waveguides of unequal widths, the insertion loss can be reduced. Secondly, for a high index contrast waveguide such as the silicon strip waveguide with a nanoscale cross-section, modes in such waveguide are not purely TE or TM but hybrid in nature, with all the six components of their E and H-fields being present. Therefore a detail analysis of the modal field profiles along with the Poynting vector profile will be shown. The effects of waveguide's width and height on the effective indices, the hybridness, the modal effective area and the power confinement in the core or cladding has been studied. Furthermore the modal birefringence of such strip waveguide will be shown. It will be presented that for a strip waveguide with height of 260 nm, single mode exists in the region of the width being 200 nm to 400 nm and that the modal effective is at its minimum when width is around 320 nm for both polarization states. Thirdly, a compact polarization rotator with an asymmetric waveguide structure design, suitable for fabrication that does not require a slanted side wall or curved waveguide is considered in this work. It will be shown in here that due to the hybrid nature of the asymmetric waveguide design, maximum polarization rotation (from TE to TM) will be achieve by enhancing the non-dominant field profile of both polarized fundamental mode. As the modal hybridness and the propagation constants of both polarized modes will be obtained, the half-beat length, polarization conversion and polarization cross-talk will be calculated by using the FEM and the least squares residual boundary method (LSBR). It is learnt that a compact single stage polarization rotator with a device length of 48 μm with more than 99% of polarization conversion is achieved in this work. Finally, a study of vertical and horizontal slot waveguide will be shown. Based on silicon strip waveguide, a detail modal characteristics of E and H-fields along with the Poynting vectors are presented. It will be shown that for slot waveguide, high power confinement and power density will be achieved in the slot area. It will be presented that by optimising the waveguide and slot dimension, the performance of the power confinement and power density in the slot region can be improved.

Thesis: Ph. D., Massachusetts Institute of Technology, Department of Electrical Engineering and Computer Science, 2015.
This electronic version was submitted by the student author. The certified thesis is available in the Institute Archives and Special Collections.
Cataloged from student-submitted PDF version of thesis.
Includes bibliographical references (pages 173-183).
As raw compute power of a single chip continues to scale into the multi-teraflop regime, the processor I/O communication fabric must scale proportionally in order to prevent a performance bottleneck. As electrical wires suffer from high channel losses, pin-count constraints, and crosstalk, they are projected to fall short of the demands required by future memory systems. Silicon-photonic optical links overcome the fundamental tradeoffs of electrical wires; dense wavelength division multiplexing (DWDM) - where multiple data channels share a single waveguide or fiber to greatly extend bandwidth density - and the potential to combine at chip-scale with a very large scale integrated (VLSI) CMOS electrical chip make them a promising alternative for next-generation processor I/O. The key device for VLSI photonics is the optical microring resonator, a compact micrometer-scale device enabling energy-efficient modulation, DWDM channel selection, and sometimes even photo-detection. While these advantages have generated considerable interest in silicon-photonics, present-day integration efforts have been limited in scale owing to the difficulty of integration with advanced electronics and the sensitivity of microring resonators to both process and thermal variations. This thesis develops and demonstrates the pieces of a photonically-interconnected processor-to-memory system. We demonstrate a complete optical transceiver platform in a commercial 45 nm SOI process, showing that optical devices can be integrated into an advanced, commercial CMOS SOI process even without any changes to the manufacturing steps of the native process. To show that photonic interconnects are viable even for commoditized and cost-sensitive memory, we develop the first monolithic electronic-photonic links in bulk CMOS. As the stabilization of ring resonators is critical for use in VLSI systems, we contribute to the understanding of process and thermal variations on microring resonators, leading to the demonstration of a complete auto-locking microring tuning system that is agnostic to the transmitted data sequence and suitable for unencoded low-latency processor-to-memory traffic. Finally, the technology and methods developed in this work culminate in the demonstration of the world's first processor chip with integrated photonic interconnects, which uses monolithically integrated photonic devices to optically communicate to main memory.
by Chen Sun.
Ph. D.

Thesis (S.M.)--Massachusetts Institute of Technology, Dept. of Electrical Engineering and Computer Science, 2013.
Cataloged from PDF version of thesis.
Includes bibliographical references (pages 137-142).
Silicon photonics is moving fast toward industrialization. It satisfies the increasing demand for higher speed, larger bandwidth communication. Thus it has a wide range of applications including high-performance computing, data center, telecom etc.. However, the on-chip (waveguide) and off-chip (fiber) components for silicon photonics have quite different characteristics for the polarizations of light. The polarization dependence of on-chip silicon photonics components still remains a bottleneck for the real application of it. Efficient devices for manipulating polarizations are highly demanded. Herein, we present the designs of adiabatic polarization rotator (PR) and polarization splitter and rotator (PSR) to deal with this issue. With their adiabatic nature, larger bandwidth (>100 nm) and better fabrication tolerance have been achieved. Besides, the effort toward the realization of a full-functional two-input two-output PBS, which is an exact correspondent of the traditional cube PBS in free space is presented. The structure was fabricated in a commercial state-of-art CMOS foundry and has a bandwidth of over 150 nm and less than -10dB crosstalk level. Though its application in traditional communication can be replaced by PS, PR or PSR, its application in more accurate systems such as polarization-entangled states generation and manipulation in quantum optics or on-chip heterodyne interferometers. Moreover, original compact-ring resonator based even-dropping optical bus system is proposed and analyzed in detail. Large free-spectral range offered by small radius micro-ring gives more communication channels to fully utilize the power of wavelength division multiplexing (WDM). Furthermore, a multi-channel WDM broadcasting system is proposed using the optical bus design. We have demonstrated a two-channel broadcasting system, which can be further increased to more than 16 channels.
by Zhan Su.
S.M.

Integrated photonics is a promising technology which has applications in many areas such as biomedical, aero-space, radar, distributed computing, sensors and high speed signal processing. The aim is to reduce the cost, size and power consumption of optical devices and components by integrating them on a nano-scale. Currently silicon (Si) is the main material of choice in integrated photonics, however it has drawbacks which limit its uses in particular for high efficiency optical sources and detectors. The focus of this thesis will therefore be on finding an optimized integrated photonics platform which can support the integration of an optical source, photodetector, optical modulator and waveguides with matched confinement factors, whilst maintaining compatibility with existing CMOS fabrication processes and techniques used in the microelectronics industry. Some potential materials include gallium arsenide (GaAs), silicon carbide (SiC), indium phosphide (InP), gallium nitride (GaN) and silicon germanium (SiGe).

The American Institute for Manufacturing Integrated Photonics (AIM Photonics) is focused on developing an end- to- end integrated photonics ecosystem in the U.S., including domestic foundry access, integrated design tools, automated packaging, assembly and test, and workforce development. This paper describes how the institute has been structured to achieve these goals, with an emphasis on advancing the integrated photonics ecosystem. Additionally, it briefly highlights several of the technological development targets that have been identified to provide enabling advances in the manufacture and application of integrated photonics.

This thesis presents experimental work developing silicon nanocrystal based light emitting devices for silicon photonics. The chapters are organized as follows: In chapter 2, fabrication and characterization of silicon nanocrystal based devices are presented. In collaboration with Intel Corporation and Bruno Kessler Foundation and thanks to the support of European Commission through the project No. ICT-FP7-224312 HELIOS and through the project No. ICT-FP7-248909 LIMA, it is shown that layers and devices containing silicon nanocrystals can be formed in a production silicon-fab on 4 and 8 inch silicon substrates via PECVD and subsequent thermal annealing. Devices produced by single layer and multilayer deposition are studied and compared in terms of structural properties, conduction mechanisms and electroluminescence properties. Power efficiency is evaluated and studied in order to understand the relation between exciton recombination and electrical conduction. A band gap engineering method is proposed in order to better control carrier injection and light emission in order to enhance the electroluminescence power efficiency. In chapter 3, the power efficiency of silicon nanocrystal light-emitting devices is studied in alternating current regime. An experimental method based on impedance spectroscopy is proposed and an electrical model based on the constant phase element (CPE) is derived. It is, then, given a physical interpretation of the electrical model proposed by considering the disordered composition of the active material. The electrical model is further generalized for many kinds of waveforms applied and it is generalized for the direct current regime. At the end, time-resolved electroluminescence and carrier injection in alternate current regime are presented. In chapter 4, erbium implanted silicon rich oxide based devices are presented. The investigation of opto-electrical properties of LED in direct current and alternate current regime are studied in order to understand the injection mechanism and estimate the energy transfer between silicon nanocrystals and erbium. At the end a device layout and process flow for an erbium doped silicon nanocrystal based laser structure are shown. In chapter 5, some other applications of silicon nanocrystal are presented. An example of all-silicon solar cell is shown. The photovoltaic properties and carrier transport of silicon nanocrystal based solar are studied. At the end, the combination of emitting and absorbing properties of silicon nanocrystal based LED are used to develop an all-silicon based optical transceiver.

This thesis presents experimental work developing silicon nanocrystal based light emitting devices for silicon photonics. The chapters are organized as follows: In chapter 2, fabrication and characterization of silicon nanocrystal based devices are presented. In collaboration with Intel Corporation and Bruno Kessler Foundation and thanks to the support of European Commission through the project No. ICT-FP7-224312 HELIOS and through the project No. ICT-FP7-248909 LIMA, it is shown that layers and devices containing silicon nanocrystals can be formed in a production silicon-fab on 4 and 8 inch silicon substrates via PECVD and subsequent thermal annealing. Devices produced by single layer and multilayer deposition are studied and compared in terms of structural properties, conduction mechanisms and electroluminescence properties. Power efficiency is evaluated and studied in order to understand the relation between exciton recombination and electrical conduction. A band gap engineering method is proposed in order to better control carrier injection and light emission in order to enhance the electroluminescence power efficiency. In chapter 3, the power efficiency of silicon nanocrystal light-emitting devices is studied in alternating current regime. An experimental method based on impedance spectroscopy is proposed and an electrical model based on the constant phase element (CPE) is derived. It is, then, given a physical interpretation of the electrical model proposed by considering the disordered composition of the active material. The electrical model is further generalized for many kinds of waveforms applied and it is generalized for the direct current regime. At the end, time-resolved electroluminescence and carrier injection in alternate current regime are presented. In chapter 4, erbium implanted silicon rich oxide based devices are presented. The investigation of opto-electrical properties of LED in direct current and alternate current regime are studied in order to understand the injection mechanism and estimate the energy transfer between silicon nanocrystals and erbium. At the end a device layout and process flow for an erbium doped silicon nanocrystal based laser structure are shown. In chapter 5, some other applications of silicon nanocrystal are presented. An example of all-silicon solar cell is shown. The photovoltaic properties and carrier transport of silicon nanocrystal based solar are studied. At the end, the combination of emitting and absorbing properties of silicon nanocrystal based LED are used to develop an all-silicon based optical transceiver.

Thesis (S.M.)--Massachusetts Institute of Technology, Dept. of Electrical Engineering and Computer Science, 2013.
Cataloged from PDF version of thesis.
Includes bibliographical references.
In 2020, 1Tb/s on-/off-chip communication bandwidth and ~100fJ/bit total energy in a point to point link is predicted by Moore's law for high performance computing applications. These requirements are pushing the limits of on-chip silicon CMOS transistors and off-chip VCSELs technology. The major limitation of the current systems is the lack of ability to enable more than a single channel on a single wire/fiber. Silicon photonics, offering a solution on the same platform with CMOS technology, can enable Wavelength Division Multiplexed (WDM) systems. However, Silicon photonics has to overcome the wafer level, fabrication variations and dynamic temperature fluctuations, induced by processor cores with low-energy high-speed resonators. In this work, we offer a solution, called as Automated Wavelength Recovery (AWR), to these limitations. In order to demonstrate AWR, we design and demonstrate high performance active silicon resonators. A microdisk modulator achieved open eye-diagrams at a data rate of 25Gb/s and error-free operation up to 20Gb/s. A thermo-optically tunable microdisk modulator with Low power modulation (1 If/bit) at a data rate of 13-Gb/s, a 5.8-dB extinction ratio, a 1.22-dB insertion loss and a record-low thermal tuning (4.9-[mu].W/GHz) of a high-speed modulator is achieved. We demonstrated a new L-shaped resonant microring (LRM) modulator that achieves 30 Gb/s error-free operation in a compact (< 20 [mu]m²) structure while maintaining single-mode operation, enabling direct WDM across an uncorrupted 5.3 THz FSR. We have introduced heater elements inside a new single mode filter, a LRM filter, successfully. The LRM filter achieved high-efficiency (3.3[mu]W/GHz) and high-speed ([tau]f ~1.6 [mu]s) thermal tuning and maintained signal integrity with record low thru to drop power penalty (<1.1 dB) over the 4 THz FSR and <0.5dB insertion loss. We have integrated a heater driver and adiabatic resonant microring (ARM) filter in a commercial bulk CMOS deep-trench process for the first time. The proposed AWR algorithm is implemented with an ARM multiplexer. An advanced method for AWR is also introduced and demonstrated with passive resonators.
by Erman Timurdogan.
S.M.

Thesis: Ph. D., Massachusetts Institute of Technology, Department of Electrical Engineering and Computer Science, 2017.
Cataloged from PDF version of thesis.
Includes bibliographical references (pages 109-114).
We present results on the development of integrated erbium-doped aluminum oxide lasers on a silicon photonics platform. A key achievement in this work is a scalable laser design for high output power and ultra-narrow linewidth performance. Using a novel wavelength-insensitive design, a CMOS compatible waveguide structure is proposed to achieve high confinement factor and intensity overlap for both the pump (980 nm) and signal (1550 nm) wavelengths. Laser operation in the C- and L- bands of the erbium gain spectrum is obtained with both a distributed Bragg reflector and a distributed feedback structure. We demonstrate power scaling with output power greater than 75 mW and obtain an ultra-narrow linewidth of 5.3 t 0.3 kHz. We investigate the influence of gain film thickness uniformity in distributed feedback laser performance and show a compensation scheme based on a curved cavity design. We then consider the application in optical communications by demonstrating a multiwavelength cascaded laser to generate wavelength division multiplexing (WDM) light sources. Finally, we propose an integration scheme of laser in full silicon photonics platform by using an erbium trench. The approach is alignment free and allows the erbium-doped film deposition to be the last backend process, providing a pathway to a scalable CMOS compatible laser device.
by Purnawirman.
Ph. D.

Thesis: Ph. D., Massachusetts Institute of Technology, Department of Materials Science and Engineering, 2018.
This electronic version was submitted by the student author. The certified thesis is available in the Institute Archives and Special Collections.
Cataloged from student-submitted PDF version of thesis.
Includes bibliographical references.
A complete photonic chip must include the following components: a light source, usually lasers, an isolator, a waveguide, a modulator and a photodetector. Limited by material intrinsic properties, silicon alone cannot realize all the above mentioned functions. The development of silicon photonics has found its way through exploiting novel materials as hybrid platforms to manufacture various devices and systems. In this thesis, we focus on the development of novel material in emerging needs of broadband coherent light source and optical isolators. Chalcogenide glass stands out among the candidates for light generation and sensing due to its large non-linear figure of merit and wide transparency window in the IR spectral range, while magnetic garnet still presents the best device performance among magneto-optical isolators owing to the ease of phase formation and relatively low material absorption. We first investigated the fabrication technology of chalcogenide glass and developed a process flow to produce low loss planar chalcogenide glass waveguides. Using electron beam lithography to minimize sidewall roughness and reactive ion etch to achieve vertical sidewalls. We managed to demonstrate a record low loss of 0.5 dB/cm in single mode core chalcogenide waveguides. Based on this low loss platform, we integrated a supercontinuum light source onto a sensor chip. Our work presented a step forward towards miniaturization photonic sensor chips. Next, we focused on a hybrid platform of chalcogenide glass and magnetic garnet. By carefully designing device architecture, a monolithically integrated TM polarized magneto-optical (MO) isolator with 3 dB insertion loss and 40 dB isolation ratio was demonstrated. Both parameter sets record among current monolithically integrated on-chip MO isolators. Meanwhile, we also demonstrated a monolithically integrated MO isolator with TE polarization featuring 11.5 dB insertion loss and 20 dB isolation ratio. Lastly, we leveraged cavity enhanced spectroscopy platform to study radiation induced effect on SiNx, a-Si and SiC materials. We found a refractive index modulation to the order of 10⁻³ after receiving 10 Mrad Gamma radiation dose.
by Qingyang Du.
Ph. D.

Silicon photonics is an expanding domain of research since its booming a decade ago. Leveraging decades of research and development from the electronics industry, silicon photonics is the ideal candidate to overcome the bottleneck that microelectronics is facing with regard to the interconnect bandwidth limitations. In addition to being a platform compatible for both photonics and electronics, silicon is transparent in the mid-infrared regime, has a high refractive index for tight light confinement (i.e., small footprints), and presents a high nonlinear refractive index that is of high interest for optical signal processing applications. However, the integration of a silicon photonic layer is still challenging due to either the deposition flexibility or the material quality. In this thesis, a technique is presented to integrate a silicon layer with the deposition flexibility of amorphous silicon and the material quality of crystalline silicon, whilst being low-cost and having a thermal budget of < 450 °C making it compatible with the CMOS technology. Using a laser treatment, the as-deposited amorphous silicon is locally crystallised into polycrystalline silicon, a composite material of made of crystalline silicon crystallites surrounded by an amorphous silicon matrix. Both film and wire structures are processed and the material quality is assessed through optical microscopy, Raman spectroscopy, and X-ray diffraction. The optical quality of the polycrystalline wire waveguides is also investigated in the linear and nonlinear regime. In parallel to the planar silicon photonics geometry, silicon core fibres are also investigated in this work. These novel fibres offer the capability to integrate the functionality of silicon photonics platforms directly inside fibre architectures. Amorphous core fibres can be drawn with the lowest losses but as for the planar geometry, the material lacks electronics capabilities. On the other hand, polycrystalline silicon core fibres, which are suitable for both optical and electrical applications, can be drawn but their propagation losses remain high. In this work, two silicon core fibre material improvement methods, based on laser recrystallisation and tapering, are presented. To assess the material improvement, fibres are analysed through optical microscopy, Raman spectroscopy and X-ray diffraction. Finally, the optical losses of the improved fibres are measured on an optical transmission setup.

La photonique silicium est un domaine prolifique de l’optique intégrée. Elle permet de miniaturiser de nombreuses fonctionnalités optiques, l’émission laser (en considérant les stratégies d’intégration hybride), la modulation électro-optique, le routage, la détection, pour les télécoms, les LIDAR ou la spectroscopie, la métrologie, les capteurs et laboratoires sur puce, toute en produisant à grande échelle avec une grande précision et à bas coût (grâce au technologies CMOS de la microélectronique). L’optique quantique, quant à elle, souffre d’une grande sensibilité aux vibrations et à l’environnement. Les montages optiques nécessitent stabilité, alignement parfait et un grand nombre d’éléments optiques, ce qui limite son développement à grande échelle. Inversement, tous ces aspects sont naturels en photonique intégrée. Le développement de la photonique quantique est ainsi susceptible de permettre l’implémentation à large échelle de systèmes de clés de cryptage pour les télécoms et le calcul quantique. Les prérequis de la photonique quantique sont globalement plus sévères que ceux de la photonique classique. La génération d’états quantiques nécessite notamment un niveau de réjection de la pompe de plus de 100 dB ; le niveau de bruit photonique ambiant sur la puce est également un facteur à soigner particulièrement dans la mesure où les paires de photons générées par les processus quantiques sont par principe de très faible puissance. Dans ce contexte, cette thèse aborde le développement de composants et de circuits pour la photonique quantique silicium. Le but est de générer des états intriqués en énergie-temps et de pouvoir les manipuler sur une puce. Cela va de la conception à l’utilisation des paires de photons, en passant par la fabrication des circuits intégrés optiques. La qualification des propriétés quantiques est aussi explorée afin de cerner les limitations de la plateforme silicium pour le domaine applicatif visé. L’esprit de ce travail est également de proposer des solutions restantes compatibles avec les canaux de télécommunications standard (ITU), de n’utiliser que des composants fibrés standards pour les connexions à réaliser, tout en restant compatibles avec les techniques de fabrication industrielle des grandes fonderies microélectroniques afin de permettre une future production à grand échelle des circuits photoniques quantiques
Silicon photonics is a dynamic research field of integrated optics. It allows to miniaturize numerous optical functionalities such as lasers, electro-optical modulators, routers, detectors, for telecom wavelengths, LIDAR, sensor, metrology or even spectroscopy, all while been able to propose large scale production high precision technologies. On another side, quantum optics suffers from difficulties to scale optical systems, requires extreme stability, perfect alignment, and many bulky optical elements, while solving these issues follows a natural path in integrated photonics. Development of integrated quantum photonics can thus open the door to cheap, powerful, and scalable systems for quantum cryptography, telecoms, and computation. In a significant way, quantum requirements are not the ones of classical circuits with respect to photonic components and circuits. The generation of quantum states indeed requires more than 100dB of pump laser rejection, while being able to manage ultra-low useful optical signals and get rid of on-chip optical noise. In this context, this thesis is dedicated to the study, dimension, realization, and characterization of silicon photonic components and circuits for quantum optics on a chip. The target goal is to generate entangled states in energy-time and manipulate them on chip. The qualification of the quantum properties is also explored to better understand the limitations of the silicon platform in the followed objectives. Another choice of this work is to stay in telecoms wavelength and aligned with the standard channels (ITU grid), to only use off-the-shelf components, all while been CMOS compatible and compliant with standard fabrication process, this to allow the possibility to produce on large scale

This thesis work covers scientific and technological advancements in integrated silicon photonics circuits aimed at developing an All-On-Chip device for quantum photonics experiments. The work has been carried out within the framework of project SiQuro, where the Silicon-On-Insulator platform is chosen to integrate all the components of an optical bench necessary for a quantum experiment into a single chip. The problem of generating photon pairs have been addressed by studying second order polarisation effects in strained silicon with the aim to realize a bright photon pairs source based on Spontaneous Down Conversion. The study revealed that processes other than the Pockels effect are responsible for the non-linearity coefficients previously measured, suggesting to look for other candidate processes for the generation of photon pairs, as third order non-linear processes. To provide with the bright coherent source necessary to enable non-linear processes the integration of a hybrid III-V-silicon mode-locked laser has also been studied. During this study, technological novelties have also been developed by modelling the wedge profile obtained during the wet etching of silicon glass materials to engineer 3D structures. In parallel, the physics of whispering gallery mode resonators, both in silicon and in silicon glass materials, have been addressed. Silicon nitride Ultra High-Quality resonators have been demonstrated by using a strip-loaded configuration, while relative tuning of resonant modes has been demonstrated in an all-optical experiment exploiting the thermo-optic effect. This work represents a step forward in the study of the physics and applications of silicon-based lightwave circuits for integrated photonics.

This thesis work covers scientific and technological advancements in integrated silicon photonics circuits aimed at developing an All-On-Chip device for quantum photonics experiments. The work has been carried out within the framework of project SiQuro, where the Silicon-On-Insulator platform is chosen to integrate all the components of an optical bench necessary for a quantum experiment into a single chip. The problem of generating photon pairs have been addressed by studying second order polarisation effects in strained silicon with the aim to realize a bright photon pairs source based on Spontaneous Down Conversion. The study revealed that processes other than the Pockels effect are responsible for the non-linearity coefficients previously measured, suggesting to look for other candidate processes for the generation of photon pairs, as third order non-linear processes. To provide with the bright coherent source necessary to enable non-linear processes the integration of a hybrid III-V-silicon mode-locked laser has also been studied. During this study, technological novelties have also been developed by modelling the wedge profile obtained during the wet etching of silicon glass materials to engineer 3D structures. In parallel, the physics of whispering gallery mode resonators, both in silicon and in silicon glass materials, have been addressed. Silicon nitride Ultra High-Quality resonators have been demonstrated by using a strip-loaded configuration, while relative tuning of resonant modes has been demonstrated in an all-optical experiment exploiting the thermo-optic effect. This work represents a step forward in the study of the physics and applications of silicon-based lightwave circuits for integrated photonics.

In this thesis, second order optical nonlinearities in silicon waveguides are studied. At the beginning, the strained silicon platform is investigated in detail. In recent years, second order nonlinearities have been demonstrated on this platform. However, the origin of these nonlinearities was not clear. This thesis offers a clear answer to this question, demonstrating that this nonlinearity does not originate on the applied strain, but on the presence of trapped charges that induce a static electric field inside the waveguide. Based on this outcome, a way to induce larger electric fields in silicon waveguide is studied. Using lateral p-n junctions, strong electric fields are introduced in the waveguides, demonstrating both electro-optic effects and second-harmonic generation. These results, together with a detailed modeling of the system, pave the way through the demonstration of spontaneous parametric down-conversion in silicon.

In this thesis, second order optical nonlinearities in silicon waveguides are studied. At the beginning, the strained silicon platform is investigated in detail. In recent years, second order nonlinearities have been demonstrated on this platform. However, the origin of these nonlinearities was not clear. This thesis offers a clear answer to this question, demonstrating that this nonlinearity does not originate on the applied strain, but on the presence of trapped charges that induce a static electric field inside the waveguide. Based on this outcome, a way to induce larger electric fields in silicon waveguide is studied. Using lateral p-n junctions, strong electric fields are introduced in the waveguides, demonstrating both electro-optic effects and second-harmonic generation. These results, together with a detailed modeling of the system, pave the way through the demonstration of spontaneous parametric down-conversion in silicon.

Silicon photonic biosensors have the potential to transform medical diagnostics and healthcare delivery. Hundreds of these nano-scale sensors can be integrated onto a single millimeter-sized silicon substrate. They are fabricated in established CMOS foundries leveraging similar economies-of-scale achieved by electronic integrated circuits. This also enables their potential integration with electronic read out circuitry on a single chip. As near-infrared light propagates through nanoscale silicon wires, a portion of the light resides outside the waveguide interacting with biomolecules on the waveguide’s surface. While silicon photonic biosensors have demonstrated performances approaching today’s gold-standard diagnostic, the enzyme-linked immunosorbent assay (ELISA), improving their performance expands the potential use for applications requiring higher sensitivities and detection limits. To this end, this thesis describes efforts to optimize established biosensor configurations and develop novel structures with performance that exceeds commercially available silicon photonic biosensor platforms. This involves improving the bulk and surface sensitivity, detection limit, and quality factor of transverse electric (TE) and magnetic (TM) mode resonators in various waveguide topologies. Specifically, TM mode microring resonators, microdisk resonators, thin waveguide resonators, and the first of its kind sub-wavelength grating microring resonator with a 10X sensitivity improvement over today’s commercially available ring resonators are presented. Furthermore the use novel TE mode slot-waveguide and TM mode strip waveguide Bragg gratings which facilitate higher sensitivities (8X) and lower detection limits for biosensing applications are described. Finally, suspended Bragg grating structures are investigated to further improve sensitivity. To support the design and characterization efforts required to efficiently investigate many different sensors, a testing platform and process design kit (PDK) was developed. The test platform automatically tests hundreds of devices and orchestrates complex, multi-hour assays. The PDK reduces first-time design risk and expedites chip testing. Both have been open-sourced and are in use by more than a dozen academic and commercial research groups in various countries.
Applied Science, Faculty of
Graduate

Optical trapping enables the non-contact manipulation of micro and nanoparticles with extremely high precision. Recent research on integrated optical trapping using the evanescent fields of photonic devices has opened up new opportunities for the manipulation of nano- and microparticles in lab-on-a-chip devices. Considerable interest has emerged for the use of optical microcavities as “sensors-on-a-chip”, due to the possibility for the label-free detection of nanoparticles and molecules with high sensitivity. This dissertation focuses on the demonstration of an on-chip optical manipulation system with multiple functionalities, including trapping, buffering, sorting, and sensing. We demonstrate the optically trapping of polystyrene particles with diameters from 110 nm to 5.6 \(\mu m\) using silicon microrings and photonic crystal cavities. By integrating multiple microrings with different resonant wavelengths, we show that tuning the laser wavelength to the resonance wavelengths of different rings enables trapped particles to be transferred back and forth between the rings in a controllable manner. We term this functionality “buffering”. We furthermore demonstrate an integrated microparticle passive sorting system based on the near-field optical forces exerted by a 3-dB optical power splitter that consists of a slot waveguide and a conventional channel waveguide. In related work, we demonstrate an ultra-compact polarization splitter design leveraging the giant birefringence of silicon-on-insulator slot waveguides to achieve a high extinction ratio over the entire C band. We demonstrate trapping-assisted particle sensing, using the shift in the microcavity resonance induced by the trapped particle. We show that this permits the sensing of proteins via a binding assay approach, in which the presence of green fluorescent protein causes the particles to bind. By detecting the size distribution of particles clusters using the microcavity, we quantitatively detect the GFP concentration. In a complementary approach, we demonstrate a reusable and reconfigurable surface-enhanced Raman scattering (SERS) sensing platform. We use a photonic crystal cavity to trap silver nanoparticles in a controllable manner, and measure SERS from molecules on their surfaces. We anticipate that the on-chip sensing approaches we introduce could lead to various applications in nanotechnology and the environmental and life sciences.
Engineering and Applied Sciences

Packaging is a critical component in bringing silicon photonics to application. Without low cost, high throughput packaging, the per‐unit cost of silicon photonic integrated circuits will be too high for mass markets. Passive alignment of optical fibres to silicon photonic waveguides would significantly reduce the assembly time currently required where active alignment is labour intensive and time consuming. In this work a design is presented that has the potential for high volume cost effective packaging. The design accomplishes this by supporting and positioning multiple fibres relative to a silicon photonic integrated circuit. The capping chip passively aligns the fibres to silicon nanowires via a grating couplers. V‐groove structures in the capping chip are used to support and position the fibres and the end facet of the v‐groove reflects the light down on to the grating coupler. Plugs on the capping chip align with holes on the photonic chip assuring accurate positioning and optimal coupling. The processing required has been detailed and demonstrated, including a hybrid lithography process, crystallographically aligned v‐grooves and highly accurate alignment structures. Once combined these processes will passively align fibre optic cables with silicon photonic waveguide gratings with a misalignment less than 2μm which in theory will produce an added loss less than 1dB, grating couplers have been produced based on a bespoke design fitting the demands of the packaging solution and show a loss of 2.7dB with room for improvement compared to a simulated result of 1.67dB.

Thesis (Ph. D.)--Massachusetts Institute of Technology, Dept. of Materials Science and Engineering, 2013.
This electronic version was submitted by the student author. The certified thesis is available in the Institute Archives and Special Collections.
Cataloged from student-submitted PDF version of thesis.
Includes bibliographical references (p. 253-263).
Ge-on-Si devices are explored for photonic integration. Importance of Ge in photonics has grown and through techniques developed in our group we demonstrated low density of dislocations (<1x109cm-2) and point defects Ge growth for photonic devices. The focus of this document will be exclusively on Ge light emitters. Ge is an indirect band gap material that has shown the ability to act like a pseudo direct band gap material. Through the use of tensile strain and heavy doping, Ge exhibits properties thought exclusive of direct band gap materials. Dependence on temperature suggests strong interaction between indirect bands, [Delta] and L, and the direct band gap at [Gamma]. The behavior is justified through increase in photoluminescence on Ge. The range of efficient emission is to 120° with the first band interaction, and above 400° on the second band interaction. Low defect concentration in Ge is achieved through chemical vapor deposition at high vacuum (~1x10-8 mbar) in a two-step process. The high temperature growth and low concentration of particles permits epitaxial growth with low defect concentration. Chemical selectivity forbids Ge growth on oxide. Oxide trenches permit the growth on Si for a variety of shapes, without detrimentally affecting the strain of the Ge devices. Dopant concentration above intrinsic growth concentration, ~1x1019cm-3 phosphorus, have been achieved through a series of methods non-CMOS, spin-on dopant; and CMOS, implantation and delta doping. All the techniques explored use enhanced dopant diffusion observed in Ge under heavy n-type doping. A dopant source, or well, is used to distribute the dopants in the Ge without increasing the defect concentration. The approach lead to the development of electrically injected devices, LEDs and LDs. Ge pnn double heterostructure diodes were made under low, ~1x1018cm-3, and heavy n-type doping, >1x1019cm-3. Both devices showed improved performance compared to pin Ge LED. Furthermore, heavy doped Ge diodes exhibit evidence of bleaching or transparency. The techniques described permitted the development of Ge-on-Si laser with a concentration ~1-2x1019cm-3. It is the first demonstration of a Ge laser optically pumped working under the direct band gap assumption like other semiconductors. It represents the evidence of carrier inversion on an indirect band gap semiconductor. With 50cm-1 gain, the material shows Fabry-Perot cavity behavior. Finally, we demonstrated a fully functioning laser diode monolithically integrated on Si. Ge pnn lasers were made exhibiting a gain >1000cm-1 and exhibiting a spectrum range of over 200nm, making Ge the ideal candidate for Si photonics.
by Rodolfo E. Camacho-Aguilera.
Ph.D.

Thesis (Ph. D.)--Massachusetts Institute of Technology, Dept. of Materials Science and Engineering, 2004.
Includes bibliographical references (p. 199-206).
This thesis focuses on silicon-based high index contrast (HIC) photonics. In addition to mature fiber optics or low index contrast (LIC) platform, which is often referred to as Planar Lightwave Cirrcuit (PLC) or Silica Optical Bench (SiOB), the use of HIC platform has been attracting considerable attention recently for the purpose of dense integration of optical components on chip. There are two ultimate solutions to mold of the flow of light. One is high index contrast HIC optics, where the index difference ([delta]n) of core and cladding is more than 0.5 and light is strongly confined in the core, which enables us to integrate optical circuits in m order. Another technique is the introduction of photonic crystal, with which the flow of light is controlled by its photonc bandgap (PBG) and the defect. The concept of photonic crystal can be applied to optical wavgeuides by placing the defect, which is surrounded with photonic crystal structures. In addition to wavgeuide applications, there are lots of unexplored attractive applications for photonic crystal, especially for high index contrast photonic crystal (HIC-PC or HIC-PBG), such as Si/SiO₂ or Si/Si₃N₄ materials systems, due to the wide stop-band. In this thesis, the various applications based on HIC-PBG platform are proposed and investigated. All of the works in this thesis are based on Silicon CMOS-compatible techniques for practical applications. In first three chapters (chapter 2,3 and 4), waveguide applications are mainly focused based on HIC or HIC-PBG platform. In the latter chapters (chapter 5, 6 and 7), the applications of HIC-PBG are explored such as visible-light reflector, semiconductor saturable absorber (SESAM) and thermophotovoltaic (TPV) applications.
by Shoji Akiyama.
Ph.D.

Thesis (S.M.)--Massachusetts Institute of Technology, Dept. of Materials Science and Engineering, 2004.
This electronic version was submitted by the student author. The certified thesis is available in the Institute Archives and Special Collections.
Includes bibliographical references (p. 75-78).
As silicon photonics enters mainstream technology, we find ourselves in need of methods to seamlessly transfer light between the optical fibers of global scale telecommunications networks and the on-chip waveguides used for signal routing and processing in local computing networks. Connecting these components directly results in high loss from their unequal sizes. Therefore, we employ a coupler, which acts as an intermediary device to reduce loss through mode and index matching, and provide alignment tolerance. This thesis presents a potential fiber-to-waveguide coupler design for use in integrating such networks. A quadratic index stack focuses incident light from a fiber in one plane, while a planar lens and linear taper do likewise in the perpendicular plane. Once the mode is sufficiently compressed, the light then enters and propagates through the waveguide. We performed simulations using the beam propagation method and finite difference time domain, among other modeling techniques, to optimize coupling efficiency and gain an understanding of how varying certain parameters affects coupler performance. The simulation results were then incorporated into a mask layout for fabrication and measurement.
by Trisha M. Montalbo.
S.M.

Thesis (Ph. D. in Electrical Engineering)--Massachusetts Institute of Technology, Dept. of Electrical Engineering and Computer Science, 2013.
Cataloged from PDF version of thesis.
Includes bibliographical references (p. 155-163).
Silicon photonics, emerging from the interface of silicon technology and photonic technology, is expected to inherit the incredible integration ability of silicon technology that has boomed the microelectronic industry for half a century, as well as the unparalleled communication capability of photonic technology that has revolutionized the information industry for decades. Being a prevailing research topic in the past decade, silicon photonics has seen tremendous progresses with the successful demonstrations and commercializations of almost all of the key components, including on-chip light source, low-loss silicon waveguide, and ultrafast silicon modulators and detectors. It seems silicon photonics is ready to take off by following the successful path the microelectronic industry has been traveling through to achieve a large-scale integration of millions of photonic devices on the silicon chip with the aide of the well-established complementary metal-oxide-semiconductor (CMOS) technology. However, there remain some substantial challenges in silicon photonics, including the reliable design and fabrication of silicon photonic devices with unprecedented accuracy, and the large-scale integration of otherwise discrete silicon photonic devices. To this end, this thesis explored several examples as possible means of addressing these two challenges in silicon photonics. Two different ways of improving silicon photonic device accuracy were presented from perspectives of fabrication and device design respectively, followed by a successful integration demonstration where more than 4,000 components worked together on a silicon chip to form a functional large-scale silicon photonic system, representing the largest silicon photonic integration demonstrated to date.
by Jie Sun.
Ph.D.in Electrical Engineering

Silicon photonics is nowadays a mature technology and is on the verge of becoming a blossoming industry. Silicon photonics has also been pursued as a platform for integrated nonlinear optics based on Raman and Kerr effects. In recent years, more futuristic directions have been pursued by various groups. For instance, the realm of silicon photonics has been expanded beyond the well-established near-infrared wavelengths and into the mid-infrared (3 – 5 ?m). In this wavelength range, the omnipresent hurdle of nonlinear silicon photonics in the telecommunication band, i.e., nonlinear losses due to two-photon absorption, is inherently nonexistent. With the lack of efficient light-emission capability and second-order optical nonlinearity in silicon, heterogeneous integration with other material systems has been another direction pursued. Finally, several approaches have been proposed and demonstrated to address the energy efficiency of silicon photonic devices in the near-infrared wavelength range. In this dissertation, theoretical and experimental works are conducted to extend applications of integrated photonics into mid-infrared wavelengths based on silicon, demonstrate heterogeneous integration of tantalum pentoxide and lithium niobate photonics on silicon substrates, and study two-photon photovoltaic effect in gallium arsenide and plasmonic-enhanced structures. Specifically, performance and noise properties of nonlinear silicon photonic devices, such as Raman lasers and optical parametric amplifiers, based on novel and reliable waveguide technologies are studied. Both near-infrared and mid-infrared nonlinear silicon devices have been studied for comparison. Novel tantalum-pentoxide- and lithium-niobate-on-silicon platforms are developed for compact microring resonators and Mach-Zehnder modulators. Third- and second-harmonic generations are theoretical studied based on these two platforms, respectively. Also, the two-photon photovoltaic effect is studied in gallium arsenide waveguides for the first time. The effect, which was first demonstrated in silicon, is the nonlinear equivalent of the photovoltaic effect of solar cells and offers a viable solution for achieving energy-efficient photonic devices. The measured power efficiency achieved in gallium arsenide is higher than that in silicon and even higher efficiency is theoretically predicted with optimized designs. Finally, plasmonic-enhanced photovoltaic power converters, based on the two-photon photovoltaic effect in silicon using subwavelength apertures in metallic films, are proposed and theoretically studied.
Ph.D.
Doctorate
Optics and Photonics
Optics and Photonics
Optics and Photonics

In this thesis, we demonstrate long transverse magnetic (TM) Bragg gratings wrapped compactly using spiral waveguides on the silicon-on-insulator (SOI) platform. We developed three types of TM spiral Bragg grating waveguides (SBGWs) including uniform spiral Bragg grating (U-SBGWs), phaseshifted spiral Bragg grating (P-SBGWs), and chirped spiral Bragg gratings (C-SBGWs). Our spiral waveguides are space-efficient, requiring an area of only 189 x 189 μm² to accommodate 1 cm long Bragg grating waveguides and, thus, are less susceptible to fabrication non-uniformities. Due to these factors, the TM U-SBGWs are able to successfully obtain narrow bandwidths and high extinction ratios (ERs), as narrow as 0.09 nm and as large as 52 dB respectively. Also, the TM P-SBGWs can obtain sharp resonance peaks with high quality factors of 78790. Finally, we demonstrate the TM C-SBGWs, which exhibit group delays that are linear functions of the wavelengths over their passbands. Traditionally, due to the large coupling coefficients and the flexibility for achieving desired spectral characteristics, short Bragg grating waveguides for transverse electric (TE) modes on the SOI platform have been developed for applications in optical communication and sensing systems. In contrast, TM modes Bragg gratings on SOI platform have small coupling coefficients and, therefore, the grating lengths need to be much longer than TE mode devices, in order to obtain large ERs. However, such TM mode Bragg gratings can achieve very narrow bandwidths. Creating long gratings in regular straight waveguides suffer from the fabrication non-uniformity effects caused by the wafer thickness. As is shown here, spiral-shaped waveguides can be used to increase the grating length, while still being made using little on chip real estate, thus reducing the effects of fabrication non-uniformity.
Applied Science, Faculty of
Electrical and Computer Engineering, Department of
Graduate

Les progrès dans la fabrication des systèmes de calcul reconfigurables de type « Field Programmable Gate Arrays » (FPGA) s’appuient sur la technologie CMOS, ce qui engendre une consommation des puces élevée. Des nouveaux paradigmes de calcul sont désormais nécessaires pour remplacer les architectures de calcul traditionnel ayant une faible performance et une haute consommation énergétique. En particulier, optique intégré pourrait offrir des solutions intéressantes. Beaucoup de travail sont déjà adressées à l’utilisation d’interconnexion optique pour relaxer les contraintes intrinsèques d’interconnexion électronique. Dans ce contexte, nous proposons une nouvelle architecture de calcul reconfigurable optique, la « optical lookup table » (OLUT), qui est une implémentation optique de la lookup table (LUT). Elle améliore significativement la latence et la consommation énergétique par rapport aux architectures de calcul d’optique actuelles tel que RDL (« reconfigurable directed logic »), en utilisant le spectre de la lumière au travers de la technologie WDM. Nous proposons une méthodologie de conception multi-niveaux permettant l'explorer l’espace de conception et ainsi de réduire la consommation énergétique tout en garantissant une fiabilité élevée des calculs (BER~10-18). Les résultats indiquent que l’OLUT permet une consommation inférieure à 100fJ/opération logique, ce qui répondait en partie aux besoins d’un FPGA tout-optique à l’avenir
Advances in the design of high performance silicon chips for reconfigurable computing, i.e. Field Programmable Gate Arrays (FPGAs), rely on CMOS technology and are essentially limited by energy dissipation. New design paradigms are mandatory to replace traditional, slow and power consuming, electronic computing architectures. Integrated optics, in particular, could offer attractive solutions. Many related works already addressed the use of optical on-chip interconnects to help overcome the technology limitations of electrical interconnects. Integrated silicon photonics also has the potential for realizing high performance computing architectures. In this context, we present an energy-efficient on-chip reconfigurable photonic logic architecture, the so-called OLUT, which is an optical core implementation of a lookup table. It offers significant improvement in latency and power consumption with respect to optical directed logic architectures, through allowing the use of wavelength division multiplexing (WDM) for computation parallelism. We proposed a multi-level modeling approach based on the design space exploration that elucidates the optical device characteristics needed to produce a computing architecture with high computation reliability (BER~10-18) and low energy dissipation. Analytical results demonstrate the potential of the resulting OLUT implementation to reach <100 fJ/bit per logic operation, which may meet future demands for on-chip optical FPGAs

After Feynman's proposal, in 1982, to simulate quantum systems using quantum computers, much effort has been focused on the study and realisation of machines capable of harnessing the power of quantum mechanics for simulation and computation. Many difFerent implementations have been proposed for the realisation of quantum technologies, all with their advantages and disadvantages. Integrated silicon photonics recently emerged as a promising approach: in fact all the necessary components for quantum computation can be integrated together on a silicon chip. In addition, the information carriers (photons) have very long coherence times and can be manipulated in an intrinsically phase-stable manner. The realisation of quantum photonic technologies is tied to the existence of a high efficiency single photon source (ideally on-demand). One of the possible solutions is in the multiplexing of many probabilistic photon pair sources. In this thesis we present four different quantum photonics experiments. We show the integration in a silicon quantum photonics platform of fundamental components for the implementation of any quantum information processing. We show that with our approach we can obtain high fidelity quantum states and high levels of entanglement. Furthermore, we also demonstrate the implementation of a hybrid (time and space) multiplexed single photon source in bulk optics.

Thesis: S.M., Massachusetts Institute of Technology, Department of Electrical Engineering and Computer Science, 2017.
Cataloged from PDF version of thesis.
Includes bibliographical references (pages 121-128).
Microwave photonics is broadly defined as the study of optical devices operating in the microwave to millimeter wave spectrum, and interest in this field is continually driven by the need for higher electrical frequencies and larger signal bandwidths. However, electronic systems designed to meet these specifications are becoming increasingly challenging to create, while their microwave photonic counterparts offer a larger bandwidth, lower power consumption, higher linearity, and smaller footprint. Thus, microwave photonic systems are an attractive solution for challenging problems in the domain of high-speed electronics. This thesis will examine a specific component of a microwave photonic system, a frequency down-converter. This device takes two electronic input signals and outputs the difference between the two input frequencies. To date, all demonstrations of a photonic microwave frequency down-converter have been based on bulk optical devices that limit its deployability to a real-world application due to a large footprint. However, recent advances to silicon photonic fabrication processes stand to improve the performance and enable the mass-production of many different microwave photonic systems including a frequency down-converter. In this thesis, a library of passive CMOS-compatible silicon photonic components for microwave photonic systems was developed. Additionally, a high-saturation power germanium-on-silicon photodetector showing a 70% improvement in photocurrent generation under high incident powers, and a depletion mode optical phase shifter with a V[subscript [pi]]L of 0.9 V xcm and an electro-optical bandwidth of several gigahertz were designed and tested. Finally, all of these components were assembled in a novel architecture to create an integrated silicon photonic microwave frequency down-converter. Initial measurements of this structure showed an electrical-to-electrical conversion efficiency of -42 dB and a suppression of spurious frequencies of approximately 15 dB.
by Matthew Byrd.
S.M.

The design, simulation, and initial fabrication of a novel ultra-compact 2x2 silicon multimode-interference device evanescently coupled to a dual germanium metal-semiconductor-metal (MSM) photodetector is presented. For operation at the standard telecom wavelength of 1.5 µm, the simulations demonstrate high-speed operation at 30 GHz, low dark current in the nanoamp range, and external quantum efficiency of 80%. Error analysis was performed for possible tilt error introduced by hybrid integration of the MSM layer on top of the MMI waveguides by use of surface mount technology (SMT) and direct wafer bonding.

In this paper, the performance simulations of a novel ultra-compact balanced coherent photodetector for operation at a wavelength of 1.5 mu m are presented and design proposals for future fabrication processes are provided. It consists of a compact 2x2 MMI that is evanescently coupled into a germanium MSM photodetection layer. The simulations demonstrate dark current less than 10 nA, capacitance less than 20 fF, and optical bandwidth in the 10-30 GHz range. We propose utilizing the simplicity of direct wafer bonding to bond the detection layer to the output waveguides to avoid complicated epitaxial growth issues. This ultra-compact device shows promise as a high-speed, low-cost integrated silicon photonics solution for the telecommunications infrastructure.

Quantum information sciences offer novel capabilities in the fields of computation, data security and sensing. In order to deliver these improvements, information needs to be encoded in a physical quantum system the realization of which is technologically challenging. Even though several platforms have been proposed for computation, quantum information encoded in nonclassical states of light is going to play a central role since the classical fiber technology already provides an almost ideal transmitting medium. Furthermore, a hybrid approach formed together with solid state emitters could meet the non-trivial requirement needed for quantum computation. In the past ten years, efforts have been placed in integrating the bulky optical components required to manipulate nonclassical states of light, giving birth to the field of quantum photonics. Between all the different materials proposed, 3C silicon carbide (SiC) meets all the complex requirements needed for photonic quantum technologies and the development of essential components to this scope is the main subject of this thesis. The design, fabrication and characterization of small modal area waveguides, grating couplers and ring resonators made in SiC are reported. Four wave mixing was demonstrated thanks to the small modal volume achieved in the ring resonator and the Kerr coefficient of SiC was retrieved. The realization of photonic crystal cavities is also investigated with the aim to harness quantum emitters. Thanks to the demonstration of coupling between confined and propagating surface waves, SiC is a potential platform for quantum applications in the mid-infrared. Finally, the generation of photon pairs in the near-infrared by means of third order nonlinear process is reported using ring resonators fabricated in silicon nitride.

Les communications optiques basées sur la photonique sur silicium (SiP) sont au centre des récents efforts de recherche pour le développement des futures technologies de réseaux optiques à haut débit. Dans cette thèse, nous étudions le traitement numérique du signal (DSP) pour pallier aux limites physiques des modulateurs Mach-Zehnder sur silicium (MZM) opérés à haut débit et exploitant des formats de modulation avancés utilisant la détection cohérente. Dans le premier chapitre, nous présentons une nouvelle méthode de précompensation adaptative appelée contrôle d’apprentissage itératif par gain (G-ILC, aussi utilisé en linéarisation d’amplificateurs RF) permettant de compenser les distorsions non-linéaires. L’adaptation de la méthode G-ILC et la précompensation numérique linéaire sont accomplies par une procédure « hardware-in-the-loop » en quasi-temps réel. Nous examinons différents ordres de modulation d’amplitude en quadrature (QAM) de 16QAM à 256QAM avec des taux de symboles de 20 à 60 Gbaud. De plus, nous combinons les précompensations numériques et optiques pour contrevenir surmonter les limitations de bande-passante du système en régime de transmission haut débit. Dans le second chapitre, inspiré par les faibles taux de symbole du G-ILC, nous augmentons la vitesse de transmission au-delà de la limite de bande-passante du système SiP. Pour la première fois, nous démontrons expérimentalement un record de 100 Gbaud par 16QAM et 32QAM en transmission consécutive avec polarisation mixte. L’optimisation est réalisée sur le point d’opération du MZM et sur la DSP. Les performances du G-ILC sont améliorées par égalisation linéaire à entrées/sorties multiples (MIMO). Nous combinons aussi notre précompensation non-linéaire innovante avec une post-compensation. Par émulation de la polarisation mixte, nous réalisons un taux net de 833 Gb/s avec 32QAM au seuil de correction d’erreur (FEC) pour une expansion en largeur de bande de 20% et 747 Gb/s avec 16QAM (une expansion en largeur de bande de 7% du FEC). Dans le troisième chapitre, nous démontrons expérimentalement un algorithme de précompensation numérique basé sur une table de consultation (LUT) unidimensionnelle pour compenser les non-linéarités introduites à l’émetteur, e.g. réponse en fréquence non-linéaire du MZM en silicium, conversion numérique-analogique et amplificateur RF. L’évaluation est réalisée sur un QAM d’ordre élevé, i.e. 128QAM et 256QAM. Nous examinons la diminution en complexité de la LUT et son impact sur la performance. Finalement, nous examinons la généralisation de la méthode de précompensation proposée pour des jeux de données différents des données d’apprentissage de la table de consultation.
Les communications optiques basées sur la photonique sur silicium (SiP) sont au centre des récents efforts de recherche pour le développement des futures technologies de réseaux optiques à haut débit. Dans cette thèse, nous étudions le traitement numérique du signal (DSP) pour pallier aux limites physiques des modulateurs Mach-Zehnder sur silicium (MZM) opérés à haut débit et exploitant des formats de modulation avancés utilisant la détection cohérente. Dans le premier chapitre, nous présentons une nouvelle méthode de précompensation adaptative appelée contrôle d’apprentissage itératif par gain (G-ILC, aussi utilisé en linéarisation d’amplificateurs RF) permettant de compenser les distorsions non-linéaires. L’adaptation de la méthode G-ILC et la précompensation numérique linéaire sont accomplies par une procédure « hardware-in-the-loop » en quasi-temps réel. Nous examinons différents ordres de modulation d’amplitude en quadrature (QAM) de 16QAM à 256QAM avec des taux de symboles de 20 à 60 Gbaud. De plus, nous combinons les précompensations numériques et optiques pour contrevenir surmonter les limitations de bande-passante du système en régime de transmission haut débit. Dans le second chapitre, inspiré par les faibles taux de symbole du G-ILC, nous augmentons la vitesse de transmission au-delà de la limite de bande-passante du système SiP. Pour la première fois, nous démontrons expérimentalement un record de 100 Gbaud par 16QAM et 32QAM en transmission consécutive avec polarisation mixte. L’optimisation est réalisée sur le point d’opération du MZM et sur la DSP. Les performances du G-ILC sont améliorées par égalisation linéaire à entrées/sorties multiples (MIMO). Nous combinons aussi notre précompensation non-linéaire innovante avec une post-compensation. Par émulation de la polarisation mixte, nous réalisons un taux net de 833 Gb/s avec 32QAM au seuil de correction d’erreur (FEC) pour une expansion en largeur de bande de 20% et 747 Gb/s avec 16QAM (une expansion en largeur de bande de 7% du FEC). Dans le troisième chapitre, nous démontrons expérimentalement un algorithme de précompensation numérique basé sur une table de consultation (LUT) unidimensionnelle pour compenser les non-linéarités introduites à l’émetteur, e.g. réponse en fréquence non-linéaire du MZM en silicium, conversion numérique-analogique et amplificateur RF. L’évaluation est réalisée sur un QAM d’ordre élevé, i.e. 128QAM et 256QAM. Nous examinons la diminution en complexité de la LUT et son impact sur la performance. Finalement, nous examinons la généralisation de la méthode de précompensation proposée pour des jeux de données différents des données d’apprentissage de la table de consultation.
Optical communications based on silicon photonics (SiP) have become a focus of the recent research for future high speed optical network technologies. In this thesis, we investigate digital signal processing (DSP) approaches to combat the physical limits of SiP Mach-Zehnder modulators (MZM) driven at high baud rates and exploiting advanced modulation formats with coherent detection. In the first section, we present a novel adaptive pre-compensation method known as gain based iterative learning control (G-ILC, previously used in RF amplifier linearization) to overcome nonlinear distortions. We experimentally evaluate the G-ILC technique. Adaptation of the G-ILC, in combination with linear digital pre-compensation, is accomplished with a quasireal- time hardware-in-the-loop procedure. We examine various orders of quadrature amplitude modulation (QAM), i.e., 16QAM to 256QAM, and symbol rates, i.e., 20 to 60 Gbaud. Furthermore, we exploit joint digital and optical linear pre-compensation to overcome the bandwidth limitation of the system in the higher baud rate regime. In the second section, inspired by lower symbol rate G-ILC results, we push the baud rate beyond the bandwidth limit of the SiP system. For the first time, we experimentally report record-breaking 16QAM and 32QAM at 100 Gbaud in dual polarization back-to-back transmission. The optimization is performed on both MZM operating point and DSP. The G-ILC performance is improved by employing linear multiple input multiple output (MIMO) equalization during the adaptation. We combine our innovative nonlinear pre-compensation with post-compensation as well. Via dual polarization emulation, we achieve a net rate of 833 Gb/s with 32QAM at the forward error correction (FEC) threshold for 20% overhead and 747 Gb/s with 16QAM (7% FEC overhead). In the third section, we experimentally present a digital pre-compensation algorithm based on a one-dimensional lookup table (LUT) to compensate the nonlinearity introduced at the transmitter, e.g., nonlinear frequency response of the SiP MZM, digital to analog converter and RF amplifier. The evaluation is performed on higher order QAM, i.e., 128QAM and 256QAM. We examine reduction of LUT complexity and its impact on performance. Finally, we examine the generalization of the proposed pre-compensation method to data sets other than the original training set for the LUT.
Optical communications based on silicon photonics (SiP) have become a focus of the recent research for future high speed optical network technologies. In this thesis, we investigate digital signal processing (DSP) approaches to combat the physical limits of SiP Mach-Zehnder modulators (MZM) driven at high baud rates and exploiting advanced modulation formats with coherent detection. In the first section, we present a novel adaptive pre-compensation method known as gain based iterative learning control (G-ILC, previously used in RF amplifier linearization) to overcome nonlinear distortions. We experimentally evaluate the G-ILC technique. Adaptation of the G-ILC, in combination with linear digital pre-compensation, is accomplished with a quasireal- time hardware-in-the-loop procedure. We examine various orders of quadrature amplitude modulation (QAM), i.e., 16QAM to 256QAM, and symbol rates, i.e., 20 to 60 Gbaud. Furthermore, we exploit joint digital and optical linear pre-compensation to overcome the bandwidth limitation of the system in the higher baud rate regime. In the second section, inspired by lower symbol rate G-ILC results, we push the baud rate beyond the bandwidth limit of the SiP system. For the first time, we experimentally report record-breaking 16QAM and 32QAM at 100 Gbaud in dual polarization back-to-back transmission. The optimization is performed on both MZM operating point and DSP. The G-ILC performance is improved by employing linear multiple input multiple output (MIMO) equalization during the adaptation. We combine our innovative nonlinear pre-compensation with post-compensation as well. Via dual polarization emulation, we achieve a net rate of 833 Gb/s with 32QAM at the forward error correction (FEC) threshold for 20% overhead and 747 Gb/s with 16QAM (7% FEC overhead). In the third section, we experimentally present a digital pre-compensation algorithm based on a one-dimensional lookup table (LUT) to compensate the nonlinearity introduced at the transmitter, e.g., nonlinear frequency response of the SiP MZM, digital to analog converter and RF amplifier. The evaluation is performed on higher order QAM, i.e., 128QAM and 256QAM. We examine reduction of LUT complexity and its impact on performance. Finally, we examine the generalization of the proposed pre-compensation method to data sets other than the original training set for the LUT.

This dissertation deals with the design and the characterization of novel reconfigurable silicon-on-insulator (SOI) devices to filter and route optical signals on-chip. Design is carried out through circuit simulations based on basic circuit elements (Building Blocks, BBs) in order to prove the feasibility of an approach allowing to move the design of Photonic Integrated Circuits (PICs) toward the system level. CMOS compatibility and large integration scale make SOI one of the most promising material to realize PICs. The concepts of generic foundry and BB based circuit simulations for the design are emerging as a solution to reduce the costs and increase the circuit complexity. To validate the BB based approach, the development of some of the most important BBs is performed first. A novel tunable coupler is also presented and it is demonstrated to be a valuable alternative to the known solutions. Two novel multi-element PICs are then analysed: a narrow linewidth single mode resonator and a passband filter with widely tunable bandwidth. Extensive circuit simulations are carried out to determine their performance, taking into account fabrication tolerances. The first PIC is based on two Grating Assisted Couplers in a ring resonator (RR) configuration. It is shown that a trade-off between performance, resonance bandwidth and device footprint has to be performed. The device could be employed to realize reconfigurable add-drop de/multiplexers. Sensitivity with respect to fabrication tolerances and spurious effects is however observed. The second PIC is based on an unbalanced Mach-Zehnder interferometer loaded with two RRs. Overall good performance and robustness to fabrication tolerances and nonlinear effects have confirmed its applicability for the realization of flexible optical systems. Simulated and measured devices behaviour is shown to be in agreement thus demonstrating the viability of a BB based approach to the design of complex PICs.

This dissertation deals with the design and the characterization of novel reconfigurable silicon-on-insulator (SOI) devices to filter and route optical signals on-chip. Design is carried out through circuit simulations based on basic circuit elements (Building Blocks, BBs) in order to prove the feasibility of an approach allowing to move the design of Photonic Integrated Circuits (PICs) toward the system level. CMOS compatibility and large integration scale make SOI one of the most promising material to realize PICs. The concepts of generic foundry and BB based circuit simulations for the design are emerging as a solution to reduce the costs and increase the circuit complexity. To validate the BB based approach, the development of some of the most important BBs is performed first. A novel tunable coupler is also presented and it is demonstrated to be a valuable alternative to the known solutions. Two novel multi-element PICs are then analysed: a narrow linewidth single mode resonator and a passband filter with widely tunable bandwidth. Extensive circuit simulations are carried out to determine their performance, taking into account fabrication tolerances. The first PIC is based on two Grating Assisted Couplers in a ring resonator (RR) configuration. It is shown that a trade-off between performance, resonance bandwidth and device footprint has to be performed. The device could be employed to realize reconfigurable add-drop de/multiplexers. Sensitivity with respect to fabrication tolerances and spurious effects is however observed. The second PIC is based on an unbalanced Mach-Zehnder interferometer loaded with two RRs. Overall good performance and robustness to fabrication tolerances and nonlinear effects have confirmed its applicability for the realization of flexible optical systems. Simulated and measured devices behaviour is shown to be in agreement thus demonstrating the viability of a BB based approach to the design of complex PICs.

This thesis work covers both classical and quantum aspects of nonlinear propagation of photons in nanophotonic Silicon waveguides. The work has been carried out within the framework of the project SIQURO, which aims to bring the quantum world into integrated photonics by using the Silicon platform and, therefore, permitting in a natural way the integration of quantum photonics with electronics. The research towards on chip bright quantum sources of photon pairs has been done by investigating Multi Modal Four Wave Mixing in micrometer-size waveguides, thus exploiting the large third order nonlinearity of Silicon. The possibility to induce second order nonlinearities by straining its unit cell has been also analyzed through the study of the electro-optic effect. This has been done with the aim to promote Silicon as a platform for the integration of quantum sources of entangled photons based on Spontaneous Parametric Down Conversion. New quantum interference effects have been reported in a free space unbalanced Mach Zehnder interferometer asymmetrically excited by colour entangled photon pairs. Innovative designs of integrated quantum circuits have been proposed, which extend the capabilities of the quantum circuits demonstrated so far and provide additional functionalities. This work represents a step forward to the realization of self subsistent integrated devices for quantum enhanced measurement, quantum computation and quantum crypthography.