Dissertations / Theses on the topic 'Single Photon Avalanche Diode (SPAD)'

Cette thèse fait progresser la modélisation, la simulation et l'optimisation des diodes à avalanche à photon unique (Single-Photon Avalanche Diodes, ou SPAD), capables de détecter des photons individuels avec une grande sensibilité. Les SPAD sont essentiels dans des applications telles que les communications quantiques, l'imagerie et les mesures de temps de vol, où la détection précise de photons uniques est cruciale. Cependant, les SPAD fonctionnent par des processus complexes et stochastiques, comme la multiplication par avalanche, la gigue temporelle et l'extinction, rendant leur modélisation précise difficile. Cette thèse aborde ces complexités en développant des modèles de simulation avancés, en appliquant des techniques d'optimisation et en explorant des méthodes pour améliorer les performances des SPAD. La thèse commence par une revue de la technologie SPAD et de ses principes. Les SPAD détectent les photons en déclenchant un processus d'avalanche dans un champ électrique élevé, qui amplifie le signal du photon en une impulsion mesurable. Cette thèse étend le modèle de McIntyre, traditionnellement utilisé pour les dispositifs à avalanche, en trois dimensions, permettant des simulations plus précises des géométries complexes et des champs électriques des SPAD. Une contribution majeure est l'introduction de l'optimisation par essaim de particules (Particle Swarm Optimization, ou PSO) couplée à un solveur de Poisson non linéaire pour optimiser des paramètres de SPAD comme la tension de claquage, la largeur de la zone de déplétion et la gigue temporelle. La PSO explore efficacement l'espace de conception, équilibrant les exigences de performance et permettant la création de SPADs adaptés à diverses applications, de l'imagerie en basse lumière à la détection rapide de photons. Pour améliorer la précision des simulations de SPAD, la thèse développe une méthode Monte Carlo par advection-diffusion (ADMC), qui combine les processus d'advection et de diffusion pour modéliser de manière réaliste le transport des porteurs, notamment dans les zones de champ élevé où les avalanches se produisent. Ce modèle surmonte les limites des méthodes Monte Carlo traditionnelles, permettant une représentation précise de la gigue temporelle, de la probabilité de claquage et du taux de comptage de bruit.La thèse culmine avec le développement d'un modèle auto-cohérent Monte Carlo-Poisson pour les simulations transitoires de SPAD. En combinant ADMC avec un solveur de Poisson en 3D, ce modèle capture en temps réel des comportements critiques des SPAD, tels que l'initiation de l'avalanche, la réduction de champ et l'extinction. Cette boucle de rétroaction est essentielle pour comprendre les comportements transitoires des SPAD, car les porteurs influencent le champ électrique lorsqu'ils s'accumulent. Ce modèle est particulièrement utile pour étudier la nature stochastique de l'extinction, qui affecte la fiabilité et la temporalité des SPAD. En résumé, cette thèse apporte des contributions significatives à la modélisation, la simulation et l'optimisation des SPAD. En développant un modèle auto-cohérent Monte Carlo-Poisson et en intégrant des techniques avancées d'optimisation, ce travail fournit un cadre complet pour améliorer les performances des SPAD et soutenir les avancées futures dans des applications de haute précision<br>This thesis advances the modeling, simulation, and optimization of Single-Photon Avalanche Diodes (SPADs), which detect individual photons with high sensitivity. SPADs are essential for applications in quantum communications, imaging, and time-of-flight measurements, where precise single-photon detection is crucial. However, SPADs operate through complex, stochastic processes such as avalanche multiplication, timing jitter, and quenching, making accurate modeling a challenge. This thesis addresses these complexities by developing advanced simulation models, applying optimization techniques, and exploring methods to enhance SPAD performance. The thesis starts by reviewing SPAD technology and principles. SPADs detect photons by triggering an avalanche process in a high electric field, which amplifies the photon's signal into a measurable pulse. This thesis extends the McIntyre model, traditionally used for avalanche devices, to three dimensions, allowing for more accurate simulations of complex SPAD geometries and electric fields. A core contribution is introducing Particle Swarm Optimization (PSO) coupled with a nonlinear Poisson solver to optimize SPAD parameters like breakdown voltage, depletion width, and timing jitter. PSO navigates the design space effectively, balancing competing performance requirements and enabling customized SPADs for different applications, from low-light imaging to fast photon counting. To improve SPAD simulation accuracy, the thesis develops an Advection-Diffusion Monte Carlo (ADMC) method, which combines advection and diffusion processes for a realistic model of carrier transport, especially in high-field regions where avalanches occur. This model overcomes limitations of traditional Monte Carlo methods, achieving accurate representations of timing jitter, breakdown probability, and dark count rate. The thesis culminates in a self-consistent Monte Carlo-Poisson model for transient SPAD simulations. By combining ADMC with a 3D Poisson solver, this model captures critical SPAD behaviors like avalanche initiation, field depletion, and quenching in real time. This feedback loop is essential for understanding transient SPAD behaviors, as carriers impact the electric field as they accumulate. The model is especially useful for studying the stochastic nature of quenching, which influences SPAD reliability and timing. In summary, this thesis makes significant contributions to SPAD modeling, simulation, and optimization. By creating a self-consistent Monte Carlo-Poisson model and integrating advanced optimization techniques, this work provides a comprehensive framework for improving SPAD performance and supports further advances in high-precision applications

Ce travail a pour objectif la conception, la simulation, la modélisation et la caractérisation électrique de diodes à avalanche à photon unique (Single Photon Avalanche Diodes - SPAD) intégrées dans une technologie CMOS Fully Depleted Silicon on Insulator - FDSOI. Les SPAD sont des diodes (jonctions PN) polarisées en inverse au-delà de la tension de claquage, fonctionnant dans le mode Geiger. Grace à leur haute sensibilité et rapidité, les SPAD sont utiles pour plusieurs applications, telles que les mesures de temps de vol (Time of Flight - ToF), l’imagerie médicale (Fluorescence Lifetime Imaging Microscopy - FLIM), ainsi que la détection de particules chargées, dans le domaine de la physique de haute énergie. L’intégration de SPAD dans une technologie CMOS FDSOI permet d’obtenir une intégration 3D monolithique intrinsèque avec la diode sous l’oxyde enterré, et l’électronique associée dans le film silicium, en optimisant ainsi le facteur de remplissage. Afin d’analyser le comportement des SPAD FDSOI, plusieurs cellules ont été conçues, respectant les principales règles de dessin imposées par la fonderie, mais présentant des variantes structurelles telles que la zone d'intégration, la géométrie, la distance de garde et le circuit d’étouffement. Des simulations TCAD et des calculs analytiques ont été effectués afin d'estimer les principales figures du mérite de SPAD. Plusieurs modèles d'avalanche et de génération de porteurs ont été étudiés pour une meilleure adaptation du modèle simulé aux dispositifs fabriqués. Des caractérisations électriques ont été réalisées pour estimer des paramètres importants tels que la tension de claquage, le taux de comptage dans l'obscurité (DCR) et la réponse en l'électroluminescence. Bien que les résultats obtenus restent inférieurs par rapport à l'état de l’art, la faisabilité d’intégration de SPAD dans une technologie FDSOI a été démontrée comme preuve de concept, mais des améliorations sont nécessaires et certaines pistes sont proposées<br>This work aims at the design, simulation, modelling and electrical characterization of Single Photon Avalanche Diodes (SPAD) in an advanced Fully Depleted Silicon on Insulator (FDSOI) technology. SPADs are PN junctions reversed bias above breakdown voltage, operating in the so-called Geiger mode. Such an implementation should provide an intrinsic monolithic integration of those devices, along with their mandatory associated electronics, thanks to the buried oxide layer present in that technology, optimizing fill factor. Due to its high sensitivity, SPAD are useful for several applications, such as Time of Flight (ToF) and Fluorescence Lifetime Imaging Microscopy (FLIM) measurements, as well as the detection of charged particles, in high-energy physics domain. The designed cells follow the main design rules imposed by the foundry and present variations in aspect as integration zone, geometry, guard distance and quenching circuit. TCAD simulations were performed in order to estimate some of the SPAD main Figures of Merit. Several avalanche and carrier generation models were studied for better adapting the simulated model to the actual fabricated devices. Electrical characterizations were realized for estimating important parameters such as breakdown voltage, Dark Count Rate (DCR) and electroluminescence response. Although the obtained results are still poor when compared to State-of-the-Art, its feasibility was demonstrated and can be used as a proof of concept, at the same time that improvements are proposed

FLIM is branch of microscopy mainly used in biology which is quickly improving thanks to a rapid enhancement of instrumentation and techniques enabled by new sensors. In FLIM, the most precise method of measuring fluorescent decays is called TCSPC. High voltage PMT detection devices together with costly and bulky optical setups which scan the sample are usually required in TCSPC instrumentation. SPADs have enabled a big improvement in TCSPC measurement setup, providing a CMOS compatible device which can be designed in wide arrays format. However, sensors providing in-pixel TCSPC do not scale in size and in large array like the time-gated SPAD pixel sensors do. Time-gated pixels offer a less precise lifetime estimation, discarding any photon information outside a given time window, but this loss in photon-efficiency is offset by gains in pixel size. This work is aimed at the development of a wide field TCSPC sensor with a pixel size and fill factor able to reduce the cost of such devices and to obtain a high resolution time-resolved fluorescence image in the shortest time possible. The study focuses on SPAD and pixel design required to maximise the fill factor in sub 10 μm pixel pitch. Multiple pixel designs are proposed in order to reduce pixel area and so enable affordable wide array TCSPC systems. The first proposed pixel performs the CMM lifetime estimation in order to reduce the frame rate needed to stream the data out of the SPAD array. This pixel is designed in a 10 μm pitch and attains with the most aggressive design a fill factor of 10:17 %. A second design proposes an analogue TCSPC which consists in a S/H TAC circuitry. This simpler pixel can achieve a higher fill factor of 19:63% as well as smaller pitch of 8 μm thanks to the adoption of SPAD n-well and electronics area sharing. This last design is implemented in a 320 x 256 SPAD array in which is included part of a novel ADC aimed at reduction of the processing time required to build a TCSPC histogram. A more conventional analogue readout is used to evaluate the pixel performance as well as a more fine TCSPC histogram. The device was used to measure the fluorescence lifetime of green micro-spheres while the 2b flash ADC is used to demonstrate rapid resolution and separation of two different fluorescence decays.

Fluorescence based analysis is a fundamental research technique used in the life sciences. However, conventional fluorescence intensity measurements are prone to misinterpretation due to illumination and fluorophore concentration non-uniformities. Thus, there is a growing interest in time-resolved fluorescence detection, whereby the characteristic fluorescence decay time-constant (or lifetime) in response to an impulse excitation source is measured. The sensitivity of a sample’s lifetime properties to the micro-environment provides an extremely powerful analysis tool. However, current fluorescence lifetime analysis equipment tends to be bulky, delicate and expensive, thereby restricting its use to research laboratories. Progress in miniaturisation of biological and chemical analysis instrumentation is creating low-cost, robust and portable diagnostic tools capable of high-throughput, with reduced reagent quantities and analysis times. Such devices will enable point-of-care or in-the-field diagnostics. It was the ultimate aim of this project to produce an integrated fluorescence lifetime analysis system capable of sub-nano second precision with an instrument measuring less than 1cm3, something hitherto impossible with existing approaches. To accomplish this, advances in the development of AlInGaN micro-LEDs and high sensitivity CMOS detectors have been exploited. CMOS allows electronic circuitry to be integrated alongside the photodetectors and LED drivers to produce a highly integrated system capable of processing detector data directly without the need for additional external hardware. In this work, a 16x4 array of single-photon avalanche diodes (SPADs) integrated in a 0.35μm high-voltage CMOS technology has been implemented which incorporates two 9-bit, in-pixel time-gated counter circuits, with a resolution of 400ps and on-chip timing generation, in order to directly process fluorescence decay data. The SPAD detector can accurately capture fluorescence lifetime data for samples with concentrations down to 10nM, demonstrated using colloidal quantum dot and conventional fluorophores. The lifetimes captured using the on-chip time gated counters are shown to be equivalent to those processed using commercially available external time-correlated single-photon counting (TCSPC) hardware. A compact excitation source, capable of producing sub-nano second optical pulses, was designed using AlInGaN micro-LEDs bump-bonded to a CMOS driver backplane. A series of driver array designs are presented which are electrically contacted to an equivalent array of micro-LEDs emitting at a wavelength of 370nm. The final micro-LED driver design is capable of producing optical pulses of 300ps in width (full width half maximum, FWHM) and a maximum DC optical output power of 550μW, this is, to the best of our knowledge, the shortest reported optical pulse from a CMOS driven micro-LED device. By integrating an array of CMOS SPAD detectors and an array of CMOS driven AlInGaN micro-LEDs, a complete micro-system for time-resolved fluorescence analysis has been realised. Two different system configurations are evaluated and the ability of both topologies to accurately capture lifetime data is demonstrated. By making use of standard CMOS foundry technologies, this work opens up the possibility of a low-cost, portable chemical/bio-diagnostic device. These first-generation prototypes described herein demonstrate the first time-resolved fluorescence lifetime analysis using an integrated micro-system approach. A number of possible design improvements have been identified which could significantly enhance future device performance resulting in increased detector and micro-LED array density, improved time-gate resolution, shorter excitation pulse widths with increased optical output power and improved excitation light filtering. The integration of sample handling elements has also been proposed, allowing the sample of interest to be accurately manipulated within the micro-environment during investigation.

Time resolved imaging is concerned with the measurement of photon arrival time. It has a wealth of emerging applications including biomedical uses such as fluorescence lifetime microscopy and positron emission tomography, as well as laser ranging and imaging in three dimensions. The impact of time resolved imaging on human life is significant: it can be used to identify cancerous cells in-vivo, how well new drugs may perform, or to guide a robot around a factory or hospital. Two essential building blocks of a time resolved imaging system are a photon detector capable of sensing single photons, and fast time resolvers that can measure the time of flight of light to picosecond resolution. In order to address these emerging applications, miniaturised, single-chip, integrated arrays of photon detectors and time resolvers must be developed with state of the art performance and low cost. The goal of this research is therefore the design, layout and verification of arrays of low noise Single Photon Avalanche Diodes (SPADs) together with high resolution Time-Digital Converters (TDCs) using an advanced silicon fabrication process. The research reported in this Thesis was carried out as part of the E.U. funded Megaframe FP6 Project. A 32x32 pixel, one million frames per second, time correlated imaging device has been designed, simulated and fabricated using a 130nm CMOS Imaging process from ST Microelectronics. The imager array has been implemented together with required support cells in order to transmit data off chip at high speed as well as providing a means of device control, test and calibration. The fabricated imaging device successfully demonstrates the research objectives. The Thesis presents details of design, simulation and characterisation results of the elements of the Megaframe device which were the author’s own work. Highlights of the results include the smallest and lowest noise SPAD devices yet published for this class of fabrication process and an imaging array capable of recording single photon arrivals every microsecond, with a minimum time resolution of fifty picoseconds and single bit linearity.

The Silicon Photomultiplier (SiPM) is a novel solid state photon counting detector consisting of a parallel array of avalanche photodiodes biased beyond their breakdown voltage. It has known a fast development in the last few years as a possible alternative to vacuum photomultiplier tubes (PMTs) and conventional avalanche photodiodes (APDs). Indeed, current research in photodetectors is directed toward an increasing miniaturization of the pixel size, thus both improving the spatial resolution and reducing the device dimensions. SiPMs show high photon detection efficiency in the visible and near infrared range, low power consumption, high gain, ruggedness, compact size, excellent single-photon response, fast rise time and reduced sensitivity with temperature, voltage fluctuations, and magnetic fields. Furthermore, solid-state technology owns the typical advantages of the planar integration process, therefore, they can be manufactured at low costs and with high reproducibility. SiPMs performances in photon counting regime have been deeply investigated in literature, using picosecond pulsed lasers. In this regime, they can be used in applications like positron emission tomography, magnetic resonance imaging, nuclear physics instrumentation, high energy physics. An optical characterization performed via continuous wave (CW) sources has seldom been reported even though this kind of excitation seems to be very useful in several fields such as low power measurements, near-infrared spectroscopy and immunoassay tests. In this Thesis, I perform an electrical and optical analysis of two novel classes of SiPMs in the CW regime. After a brief introduction about the SiPM operating principle, parameters and properties (Chapter 1), I describe my responsivity measurements made with an incident optical power down to tenths of picowatts, monitoring the temperature of the device packages, and on a spectrum ranging from ultraviolet to near infrared (Chapter 2). These measurements allowed to define an innovative criterion to establish the conditions necessary for the device to be usable in CW regime. Chapter 3 continues with an investigation of the SiPM signal-to-noise ratio. Measurements employed a 10 Hz equivalent noise bandwidth, around a tunable reference frequency in the range 1 - 100 kHz, and were performed varying the applied bias and the temperature of the SiPM package. These results were compared with similar measurements performed on a PMT. Once the SiPM is characterized, Chapter 4 reports an innovative application: an optical characterization of a class of photonic crystals infiltrated with a new ethanol responsive hydrogel employing the SiPM as a reference photodetector. This activity shows innovative developments for the ethanol sensing to be applied into inexpensive and minimally invasive breathalyzers. Finally, Appendix A shows an electro-optical characterization of a novel class of Silicon Carbide (SiC) vertical Schottky UV detectors. I performed responsivity measurements as a function of the wavelength and the applied bias, varying the temperature of the SiC package, in the 200 - 400 nm range. The results of this work show a new approach to investigate the SiPM capabilities, the CW regime, demonstrating its outstanding performances and innovative applications. This Thesis was made in collaboration with the "Advanced Sensors Development Group" of STMicroelectronics and partially supported by the Project HIGH PROFILE (HIGH-throughput PROduction of FunctIonaL 3D imagEs of the brain), which is funded by the European Community under the ARTEMIS Joint Undertaking scheme.<br>The Silicon Photomultiplier (SiPM) is a novel solid state photon counting detector consisting of a parallel array of avalanche photodiodes biased beyond their breakdown voltage. It has known a fast development in the last few years as a possible alternative to vacuum photomultiplier tubes (PMTs) and conventional avalanche photodiodes (APDs). Indeed, current research in photodetectors is directed toward an increasing miniaturization of the pixel size, thus both improving the spatial resolution and reducing the device dimensions. SiPMs show high photon detection efficiency in the visible and near infrared range, low power consumption, high gain, ruggedness, compact size, excellent single-photon response, fast rise time and reduced sensitivity with temperature, voltage fluctuations, and magnetic fields. Furthermore, solid-state technology owns the typical advantages of the planar integration process, therefore, they can be manufactured at low costs and with high reproducibility. SiPMs performances in photon counting regime have been deeply investigated in literature, using picosecond pulsed lasers. In this regime, they can be used in applications like positron emission tomography, magnetic resonance imaging, nuclear physics instrumentation, high energy physics. An optical characterization performed via continuous wave (CW) sources has seldom been reported even though this kind of excitation seems to be very useful in several fields such as low power measurements, near-infrared spectroscopy and immunoassay tests. In this Thesis, I perform an electrical and optical analysis of two novel classes of SiPMs in the CW regime. After a brief introduction about the SiPM operating principle, parameters and properties (Chapter 1), I describe my responsivity measurements made with an incident optical power down to tenths of picowatts, monitoring the temperature of the device packages, and on a spectrum ranging from ultraviolet to near infrared (Chapter 2). These measurements allowed to define an innovative criterion to establish the conditions necessary for the device to be usable in CW regime. Chapter 3 continues with an investigation of the SiPM signal-to-noise ratio. Measurements employed a 10 Hz equivalent noise bandwidth, around a tunable reference frequency in the range 1 - 100 kHz, and were performed varying the applied bias and the temperature of the SiPM package. These results were compared with similar measurements performed on a PMT. Once the SiPM is characterized, Chapter 4 reports an innovative application: an optical characterization of a class of photonic crystals infiltrated with a new ethanol responsive hydrogel employing the SiPM as a reference photodetector. This activity shows innovative developments for the ethanol sensing to be applied into inexpensive and minimally invasive breathalyzers. Finally, Appendix A shows an electro-optical characterization of a novel class of Silicon Carbide (SiC) vertical Schottky UV detectors. I performed responsivity measurements as a function of the wavelength and the applied bias, varying the temperature of the SiC package, in the 200 - 400 nm range. The results of this work show a new approach to investigate the SiPM capabilities, the CW regime, demonstrating its outstanding performances and innovative applications. This Thesis was made in collaboration with the "Advanced Sensors Development Group" of STMicroelectronics and partially supported by the Project HIGH PROFILE (HIGH-throughput PROduction of FunctIonaL 3D imagEs of the brain), which is funded by the European Community under the ARTEMIS Joint Undertaking scheme.

Les diodes à avalanche à photon unique (SPAD) sont utilisées pour les capteurs à temps de vol afin de déterminer la distance d'une cible. Cependant, ils sont sujets à des déclenchements parasites par des porteurs de charge générés de manière parasitaire, quantifiés en tant que taux de comptage dans l’obscurité (DCR), ce qui peut compromettre la précision de la distance mesurée. Pour résoudre ce problème, une méthodologie de simulation a été mise en place pour évaluer le DCR. Cela est réalisé en simulant la probabilité de claquage d'avalanche, intégrée avec le taux de génération de porteurs de charge à partir de défauts. Cette méthodologie permet d'identifier les sources potentielles de DCR avant stress. Pour garantir l'intégrité des mesures de distance sur une longue période, il est nécessaire de prédire le niveau de DCR dans diverses conditions d'exploitation. La méthodologie de simulation susmentionnée est utilisée pour identifier les sources potentielles de DCR après stress. Pour un modèle cinétique précis de dégradation de type porteurs chauds (HCD), il est essentiel de considérer non seulement la distribution d'énergie des porteurs, mais également la distribution de l'énergie de dissociation de la liaison Si-H à l'interface Si/SiO2. La probabilité de dissociation d'ionisation d'impact est utilisée pour modéliser le processus de création de défauts, qui présente une dépendance temporelle sous-linéaire en raison de l'épuisement progressif des précurseurs de défauts. Une mesure précise de la distance nécessite de distinguer le signal du bruit ambiant et du plancher de DCR. L'impact de DCR peut être estimé en considérant la réflectance de la cible et les conditions d'éclairage ambiant. En résumé, ce travail utilise une méthodologie de caractérisation et de simulation approfondie pour prédire le DCR dans les dispositifs de type SPAD le long de sa durée de vie, permettant ainsi d'évaluer son impact sur les mesures de distance<br>Single-Photon Avalanche Diode (SPAD) are used for Time-of-Flight (ToF) sensors to determine distance from a target by measuring the travel time of an emitted pulsed signal. These photodetectors work by triggering an avalanche of charge carriers upon photon absorption, resulting in a substantial amplification which can be detected. However, they are subject to spurious triggering by parasitic generated charge carriers, quantified as Dark Count Rate (DCR), which can compromise the accuracy of the measured distance. Therefore, it is crucial to identify and eliminate the potential source of DCR. To tackle this issue, a simulation methodology has been implemented to assess the DCR. This is achieved by simulating the avalanche breakdown probability, integrated with the carrier generation rate from defects. The breakdown probability can be simulated either in a deterministically, based on electric-field streamlines, or stochastically, by means of drift-diffusion simulation of the random carrier path. This methodology allows for the identification of the potential sources of pre-stress DCR by comparing simulation results to experimental data over a wide range of voltage and temperature. To ensure the accuracy of distance range measurements over time, it is necessary to predict the DCR level under various operating conditions. The aforementioned simulation methodology is used to identify the potential sources of post-stress DCR by comparing simulation results to stress experiments that evaluate the principal stress factors, namely temperature, voltage and irradiance. Furthermore, a Monte-Carlo study has been conducted to examine the device-to-device variation along stress duration. For an accurate Hot-Carrier Degradation (HCD) kinetics model, it is essential to consider not only the carrier energy distribution function but also the distribution of Si−H bond dissociation energy distribution at the Si/SiO2 interface. The number of available hot carriers is estimated from the carrier current density according to the carrier energy distribution simulated by means of a full-band Monte-Carlo method. The impact-ionization dissociation probability is employed to model the defect creation process, which exhibits sub-linear time dependence due to the gradual exhaustion of defect precursors. Accurate distance ranging requires distinguishing the signal from ambient noise and the DCR floor, and ensuring the target’s accumulated photon signal dominates over other random noise sources. An analytical formula allows to estimate the maximum distance ranging using the maximum signal strength, ambient noise level, and confidence levels. The impact of DCR can be estimated by considering the target’s reflectance and the ambient light conditions. In a nutshell, this work makes use of a in-depth characterization and simulation methodology to predict DCR in SPAD devices along stress duration, thereby allowing the assessment of its impact on distance range measurements

La tomographie d'émission par positrons (TEP) se distingue des autres modalités d'imagerie par sa capacité à localiser et quantifier la présence de molécules marquées, appelées radiotraceurs, au sein d'un organisme. Cette capacité à mesurer l'activité biologique des différents tissus d'un sujet apporte des informations uniques et essentielles à l'étude de tumeurs cancéreuses, au fonctionnement du cerveau et de ses maladies neurodégénératives et de la pharmacodynamique de nouveaux médicaments. Depuis les tout débuts de la TEP, les scientifiques rêvent de pouvoir utiliser l'information de temps de vol des photons pour améliorer la qualité de l'image TEP. L'arrivée des photodiodes avalanche monophotoniques (PAMP), rend maintenant ce rêve possible. Ces dispositifs détectent la faible émission de lumière des scintillateurs et présentent une réponse grandement amplifiée avec une faible incertitude temporelle. Mais le potentiel des PAMP n'est pas encore entièrement exploré. Plutôt que de faire la somme des courants d'une matrice de PAMP, il est possible d'utiliser leur nature intrinsèquement binaire afin de réaliser un photodétecteur numérique capable de déterminer avec précision le temps d'arrivée de chaque photon de scintillation. Toutefois, la conception de matrices de PAMP numériques en est encore à ses débuts, et les outils de conception se font rares. Ce projet de doctorat propose un simulateur facilitant la conception de matrices de PAMP, que celles-ci soient analogiques ou numériques. Avec cet outil, l'optimisation d'une matrice de PAMP numérique basée dans une technologie Teledyne DALSA HV CMOS \SI{0,8}{\micro\metre} est proposée. En plus de guider les choix de conception de l'équipe, cette optimisation permet de mieux comprendre quels paramètres influencent les performances du détecteur. De plus, puisque le photodétecteur n'est pas l'unique acteur des performances d'un détecteur TEP, une étude sur l'impact des scintillateurs est aussi présentée. Cette étude vérifie l'amélioration apportée par l'intégration de photons prompts dans des scintillateurs LYSO. Enfin, une approche novatrice pour discriminer l'énergie des évènements TEP basée sur l'information temporelle des photons de scintillation a été développée et vérifiée à l'aide du simulateur. Bien que ce simulateur et les études réalisées dans le cadre de cette thèse soient concentrés sur des détecteurs TEP, l'utilité des PAMP et du simulateur ne se limite pas à cette application. Les matrices de PAMP sont prisées pour le développement de détecteur en physique des particules, physique nucléaire, informatique quantique, LIDAR et bien d'autres.<br>Abstract : Positron emission tomography (PET) stands out among other imaging modalities by its ability to locate and quantify the presence of marked molecules, called radiotracers, within an organism. The capacity to measure biological activity of various organic tissues provides unique information, essential to the study of cancerous tumors, brain functions and the pharmacodynamics of new medications. Since the very beginings of PET, scientists dreamed of using the photon's time-of-flight information to improve PET images. With the recent progress of Single Photon Avalanche Diodes (SPAD), this dream is now possible. These photodetectors detect the scintillators' low light emission and offers a greatly amplified response with only a small time uncertainty. However the potential of SPAD has not yet been entirely explored. Instead of summing the currents of a SPAD array, it is possible to use their intrinsically binary operation to build a digital photodetector, able to establish with precision the time of arrival of each scintillation photon. With this information, the time-of-flight measurements will be much more precise. Yet the design of digital SPAD arrays is in its infancy and design tools for this purpose are rare. This project proposes a simulator to aid the design of SPAD arrays, both analog and digital. With this tool, we propose an optimised design for a digital SPAD array fabricated in Teledyne Dalsa HV CMOS \SI{0.8}{\micro\metre} technology. In addition to guiding the design choices of our team, this optimisation led to a better understanding which parameters influence the performance of a PET detector. In addition, since the photodetector is not the sole actor in the performance of a PET detector, a study on the effect of scintillators is also presented. This study evaluates the improvement brought by incorporating a prompt photon emission mechanism in LYSO crystals. Finally, we describe a novel approach to energy discrimination based on the timing information of scintillation photons was developped and tested using the simulator. While this simulator and the studies presented in this thesis focus on PET detectors, SPAD are not limited to this sole application. SPAD arrays are promising for a wide variety of fields, including particle physics, high energy physics, quantum computing, LIDAR and many more.

Cette thèse porte sur une famille de photo-détecteurs appelés SPAD pour Single Photon Avalanche Diodes qui sont des jonctions PN polarisées en inverse au-delà de la tension de claquage. Les diodes SPADs sont reconnues pour présenter de très bonnes performances en détection de faibles flux lumineux avec une réponse extrême rapide. Afin d’améliorer l’efficacité de détection dans le proche infrarouge de diodes SPAD sur silicium, les objectifs de la thèse sont de concevoir, fabriquer et caractériser une nouvelle génération de photodiodes SPADs dans une technologie CMOS 40nm en intégrant une cavité de germanium. Les travaux menés comportent i) un volet conception en utilisant des outils de simulation TCAD pour proposer une architecture originale optimisée, ii) le développement du flot complet du procédé technologique avec la création de nouvelles briques telles que la gravure de la cavité et l’épitaxie de germanium dopé in-situ, 3) la caractérisation électro-optique des composants issus des premiers lots fabriqués. Les analyses et interprétations des résultats obtenus révèlent la difficulté technologique pour réaliser une hétérojonction silicium-germanium sans défauts et une couche germanium de qualité. Néanmoins, les mesures réalisées ont démontré la capacité de cette nouvelle famille de SPAD à cavité de germanium sur plateforme silicium pour détecter les flux jusqu’à 1300nm, démontrant un fort potentiel applicatif pour les applications d’aide à la navigation<br>This thesis deals with a family of photo-detectors called SPAD for Single Photon Avalanche Diodes which are a PN junctions reverse biased beyond the breakdown voltage. SPADs diodes are known to have very good performance in detecting low light fluxes with an extremely fast response. In order to improve the near infrared detection efficiency of SPAD diodes on silicon, the objectives of the thesis are to design, manufacture and characterize a new generation of SPAD photodiodes in 40nm CMOS technology by integrating a germanium cavity. The work carried out includes i) design and simulation using TCAD tools to propose an optimized original architecture, ii) development of the process flow in industrial imager technological with the creation of new bricks such as etch of the cavity and epitaxy of germanium in-situ doped 3) the electro-optical characterization of the manufactured devices. The results obtained reveal technological difficulty to produce a silicon-germanium heterojunction without defects. Nevertheless, the measurements carried out demonstrated the ability of this new family of germanium cavity SPADs on a silicon platform to detect wavelengths up to 1300nm, demonstrating a strong potential for time of light applications

Le Groupe de Recherche en Appareillage Médical de Sherbrooke (GRAMS) travaille actuellement sur un programme de recherche portant sur des photodiodes à avalanche monophotoniques (PAMP) opérées en mode Geiger en vue d'une application à la tomographie d’émission par positrons (TEP). Pour opérer dans ce mode, la PAMP, ou SPAD selon l’acronyme anglais (Single Photon Avalanche Diode), requiert un circuit d'étouffement (CE) pour, d’une part, arrêter l’avalanche pouvant causer sa destruction et, d’autre part, la réinitialiser en mode d’attente d’un nouveau photon. Le rôle de ce CE comprend également une électronique de communication vers les étages de traitement avancé de signaux. La performance temporelle optimale du CE est réalisée lorsqu’il est juxtaposé à la PAMP. Cependant, cela entraîne une réduction de la surface photosensible ; un élément crucial en imagerie. L’intégration 3D, à base d'interconnexions verticales, offr une solution élégante et performante à cette problématique par l’empilement de circuits intégrés possédant différentes fonctions (PAMP, CE et traitement avancé de signaux). Dans l’approche proposée, des circuits d’étouffement de 50 [mu]m x 50 [mu]m réalisés sur une technologie CMOS 130 nm 3D Tezzaron, contenant chacun 112 transistors, sont matricés afin de correspondre à une matrice de PAMP localisée sur une couche électronique supérieure. Chaque circuit d'étouffement possède une gigue temporelle de 7,47 ps RMS selon des simulations faites avec le logiciel Cadence. Le CE a la flexibilité d'ajuster les temps d'étouffement et de recharge pour la PAMP tout en présentant une faible consommation de puissance ( ~ 0,33 mW à 33 Mcps). La conception du PAMP nécessite de supporter des tensions supérieures aux 3,3 V de la technologie. Pour répondre à ce problème, des transistors à drain étendu (DEMOS) ont été réalisés. En raison de retards de production par les fabricants, les circuits n’ont pu être testés physiquement par des mesures. Les résultats de ce mémoire sont par conséquent basés sur des résultats de simulations avec le logiciel Cadence.

Ce travail porte sur la modélisation pour la simulation et la prédiction desparamètres de performance des photodiodes à avalanche polarisée en mode Geiger en technologieCMOS, ou SPADs CMOS, pour Lidars spatiaux. Ce travail de thèse vise à développerune méthodologie basée sur : des modèles de la physique du semi-conducteur, des mesuresfournissant des informations sur le procédé technologique visé et des outils commerciaux desimulation. Ceci, dans le but de simuler les paramètres de performance des SPADs en serapprochant autant que possible de la réalité du procédé technologique afin d’améliorer lesprédictions. Des SPADs ont été conçues et caractérisées de manière à acquérir les paramètresde performances et les confronter aux résultats de simulation pour valider notre approche.De plus, la conception des SPADs s’est faite en regard des spécifications Lidar du CNESet d’Airbus Defence and Space en vue d’obtenir des structures permettant d’améliorer nosconnaissances en matière de : compréhension des mécanismes physiques, conception et méthodede caractérisation des SPADs CMOS. Ceci, dans l’intention d’étudier la possibilitéd’intégrer ces détecteurs dans leurs futurs systèmes Lidars spatiaux<br>This work focuses on modelling for simulation and prediction purposes ofCMOS SPADs performance parameters used in spaceborne Lidars. The innovative side ofthis work lies in a new methodology based on physical models for semiconductor devices,measurements performed on the targeted CMOS process and commercial simulation tools topredict CMOS SPADs performances. This method allows to get as close as possible to theprocess reality and to improve predictions. A set of SPAD has been designed and fabricated,and is used for measurements and model validation. SPAD design has been done with respectto CNES and Airbus Defence Space Lidar specification, in order to produce devices that willimprove our knowledge in terms of understanding of the involved physical mechanisms, SPADsdesign and test method, for a possible integration within their future spaceborne Lidars

We have studied a new optical sensor characterized by performances that will extend the capabilities of several new physical investigation techniques. Our imaging device is based on a two-dimensional array of Single Photon Avalanche Diode (SPAD), sensitive to the single photon with a subnanosecond timing precision. It is able to perform a continuous photon acquisition without the necessity to break to perform the readout process. Moreover it is not damageable by intense light sources. The proposed solution constitutes a step forward for all Time Correlated Single Photon Counting analysis, as Fluorescence Lifetime Imaging Microscopy, Dynamic Light Scattering, 3D Camera, Particle Imaging Velocimetry and Adaptive Optics. An electric characterization of the single SPAD has been carried out to perform multiple readout strategies, and an electric model has been used to perform the simulation of different two-dimensional electric array configurations. We have also deeply studied the source of the counting distortion of the single passive quenched SPAD and have been able to extend the dynamic range of four order of magnitude and to use the dead time saturation as a compression feature for data produced by our imaging sensor. The dead time compensation laws established in Literature have been extended over the steady state analysis to include the time dependent source and any type of dead time. The acquisition electronics, the sensor calibration and the imaging reconstruction algorithm have been performed on a working prototype. The device has been tested with many experimental setups, developed to evaluate the features and the limits of our technological solutions.<br>Abbiamo studiato un nuovo sensore ottico caratterizzato da prestazioni che estenderanno le funzionalita' di molte nuove tecniche di indagine fisica. Il nostro dispositivo si basa su una matrice bidimensionale di Single Photon Avalanche Diode (SPAD), in grado di fornire il tempo di arrivo di ogni singolo fotone con una precisione del decimo di nanosecondo. Il nostro apparato e' in grado di acquisire là ¢ arrivo dei fotoni con continuita', senza interruzioni dovute al processo di lettura, ed e' inoltre resistente a fonti di luce eccessiva che costituiscono una limitazione per i normali dispositivi a singolo fotone. La soluzione proposta costituisce un passo in avanti per tutte le analisi basate sulla correlazione temporale a singolo fotone, come la Fluorescence Lifetime Imaging Microscopy, Dynamic Light Scattering, 3D Camera, Particle Imaging Velocimetry e Adaptive Optics. Grazie allo studio delle caratteristiche elettriche del singolo SPAD e' stato possibile individuare varie strategie di lettura. Il modello elettrico sviluppato e' stato inoltre utilizzato per simulare diverse configurazioni elettriche della matrici bidimensionali di sensori. Abbiamo studiato le caratteristiche funzionali del singolo SPAD ponendo l'attenzione sui fenomeni che alterano la linearita' di ri-sposta, siamo stati cosi' in grado di estendere di quattro ordini di grandezza il suo intervallo di utilizzo, e di utilizzare la saturazione come una funzione di compressione dei dati prodotti dal sensore. Le equazioni presentate estendono la correzione degli effetti del tempo morto, gia' presenti in letteratura, dallà ¢ analisi del caso stazionario a quello delle sorgenti variabili nel tempo, e sono inoltre estendibili a qualunque configurazione di tempo morto. La produzione di un prototipo funzionante ha compreso inoltre la realizzazione dell'elettronica di acquisizione, dell'algoritmo di calibrazione del sensore e di ricostruzione delle immagini. Il dispositivo e' stato testato realizzando diversi esperimenti, che hanno permesso di valutare le caratteristiche e i limiti delle soluzioni tecnologiche adottate.

This thesis explores the use of single-photon avalanche diodes (SPADs) for visible light communications (VLC). The high sensitivity of SPADs can potentially enhance the performance of VLC receivers. However, a SPAD-based system has challenges that need to be addressed before it can be considered as a viable option for VLC. The first challenge is the susceptibility of SPAD-based receivers to variations in ambient light. The high sensitivity of SPADs is advantageous for signal detection, but also makes SPADs vulnerable to variations in ambient light. In this thesis, the performance of a SPAD-based receiver is investigated under changing lighting conditions. Analytical expressions to quantify performance are derived, and an experiment is conducted to gain further understanding of system performance. It is shown that a SPAD-based receiver is highly sensitive to illumination changes when on-off keying (OOK) is employed, and that pulse-position modulation (PPM) is a preferred modulation scheme as it is more robust. The second challenge is broadcasting to SPAD-based receivers with different capabilities. A traditional broadcasting scheme is time-sharing, whereby a transmitter sends data to receivers in an alternating manner. Broadcasting to SPAD-based receivers is challenging as receivers may have diverse capabilities. In this thesis, a new multiresolution modulation scheme is proposed, which can potentially improve system performance over the traditional timesharing approach. The performance of the proposed scheme is analyzed, and a proof-of-concept experiment is performed to demonstrate its viability.

In order to improve the sensitivity of an optical receiver, the gain and the collection area of the photo-detectors within the receiver should be increased. Detectors with internal gain such as avalanche photodiodes (APD) are usually used to increase the sensitivity of the receiver. One problem with APDs is the sensitivity of their gain to their bias voltage, which makes them challenging to be fabricated in a standard CMOS process due to variations in their gain. However, when an APD is biased over its breakdown voltage, it is sensitive to a single photon, hence, referred to as a single photon avalanche diodes (SPAD). The SPADs are photon-counting detectors, which are less sensitive to their bias voltage, and can be integrated with rest of the electronic circuitry that form an optical receiver. An avalanche diode requires dedicated circuits to be operated in the SPAD mode. These circuits make the diode insensitive to an incident photon for a duration that is known as deadtime. Unfortunately, The collection area of the PD, APD, and SPADs are limited to their capacitance. Hence, a large photo-detector leads to a larger capacitance, which reduces the bandwidth of the receiver. In this thesis, a photon counting optical receiver based on an array of SPADs is proposed which increases the collection area with a low output capacitance. The avalanche diode and peripheral circuits which operate and readout-out the SPAD array are fabricated in the commercially available UMC 0.18 μm CMOS process. Initially, the avalanche diode is tested and characterised. A high performance circuit is then designed and tested which is able to achieve short deadtimes up to 4 ns. Once the photon counting operation of the SPAD is verified, a numerical model is developed to investigate the influence of several factors, including the deadtime, on the performance of the photon-counting detector in a communication link. Based on the simulation results, which show the advantages of an array over a single detector, a prototype detector array of 64 asynchronous SPADs is designed and tested. This array uses a high-speed readout mechanism which is inspired by the current steering digital-to-analogue converters. Bit error ratio tests (BERT) verify the photon counting capability of the proposed detector, and a bit error rate of 1E-3 has been achieved at data rate of 100 Mbps. In addition, the array of SPAD is compatible with a front-end of conventional optical receiver which uses a photodiode as a photo detector.

Nous présentons une caméra à balayage de fente intégrée basée sur un système de comptage de photon unique résolu en temps (TCSPC-SC) employant l'architecture linéaire « streak » pour surmonter la limitation de l'espace inhérent aux systèmes TCSPC bidimensionnels. Cette solution permet l'intégration de fonctionnalités électroniques complexes dans les pixels sans l'inconvénient d'un faible facteur de remplissage conduisant à une faible efficacité de détection. Le TCSPC-SC se compose de deux blocs principaux: une photodiode à avalanche (SPAD) et un bloc de mesure de temps, les deux blocs sont intégrés en technologie 180 nm CMOS standard. La structure de la SPAD utilisée a été sélectionnée parmi 6 structures différentes après un processus de caractérisation précise et approfondie. Le bloc de mesure du temps se compose d'un TOC hybride capable d'atteindre des résolutions de temps élevées et ajustables avec une large dynamique de mesure grâce à un système de conversion de temps (TOC) hybride qui combine l'approche analogique basée sur un convertisseur de temps vers amplitude(TAC), et les approches numériques utilisant une boucle à verrouillage de retard (DLL) et un compteur numérique. Le TOC hybride a été spécialement conçu pour être utilisé dans un système TCSPC qui intègre une ligne de TOC nécessitant ainsi une conception appropriée pour limiter la consommation d'énergie et la surface d'occupation et parvenir à une architecture flexible et facilement extensible. Suite à la conception et la réalisation de ces deux blocs dans une technologie180 nm CMOS standard, une structure de test de la caméra à balayage de fente (TCSPC-SC) qui englobe 8 unités a été réalisée dans le but final de mettre en œuvre un modèle TCSPC-SC complet et plus large<br>In this work we present a TCSPC Streak Camera (TCSPC-SC) that takes advantage of the streak mode imaging ta overcome the space limitation inherent ta 20 TCSPC sensor arrays. This cost-effective solution allows the integration of complex functionalities in the pixel without the inconvenience of low fill factor that leads ta low detection efficiency. The TCSPC~SC consists of two main building blacks: a SPAD and a time measurement black bath integrated in 180 nm Standard CMOS technology. The SPAD was selected among 6 different SPAD structures following a thorough characterization process ta fully determine its performance figures. The time measurement black consists of a hybrid TOC capable of achieving high adjustable time resolutions with large dynamic range owing ta a time conversion scheme that combines traditional Analog Time to Amplitude Converter (TAC), Digital DLL-based and counter-based TOC. Furthermore, thehybrid TOC was especially designed ta be used in a TCSPC system that incorporates an array of TDCs which required a careful design ta limit power consumption and occupation area in order to achieve a flexible and easily scalable architecture. These two building blacks were bath fabricated in a 180 nm standard CMOS technology and employed ta demonstrate a TCSPC Streak Camera(TCSPC-SC) test structure that englobes 8 units in order ta demonstrate the system's operation principle with the final aim of implementing a complete and bigger TCSPC-SC model in the near future

With data throughput increasing exponentially in wireless communication networks, the limited radio frequency (RF) spectrum is unable to meet the future data rate demand. As a promising complementary approach, optical wireless communication (OWC) has gained significant attention since its licence-free light spectrum provides a considerable amount of communication bandwidth. In conventional OWC systems, the information-carried signal has to be real-valued and non-negative due to the incoherent light output of the conventional optical transmitter, light emitting diode (LED). Therefore, an intensity modulation and direct detection (IM/DD) system is used for establishing the OWC link. Some modified orthogonal frequency division multiplexing (OFDM) schemes have been proposed to achieve suitable optical signals. In previous research, three OFDM-based schemes have been presented, including DC-biased optical orthogonal frequency division multiplexing (DCO-OFDM), asymmetrically clipped optical orthogonal frequency division multiplexing (ACO-OFDM) and unipolar orthogonal frequency division multiplexing (U-OFDM). Basic concepts of SPAD receivers are studied and a novel application in OWC is proposed for a permanent downhole monitoring (PDM) system in the gas and oil industry. In this thesis, a complete model of the SPAD-based OWC system is presented, including some related SPAD metrics, the photon counting process in SPAD and a specific nonlinear distortion caused by passive quenching (PQ) and active quenching (AQ) recharged circuits. Moreover, a practical SPAD-based visible light communication (VLC) system and its theoretical analysis are presented in a long-distance gas pipe with a battery-powered LED and a basic on-off keying (OOK) modulation scheme. In this thesis, two novel optical orthogonal frequency division multiplexing (O-OFDM) technologies are proposed: non-DC-biased orthogonal frequency division multiplexing (NDCOFDM) and OFDM with single-photon avalanche diode (SPAD). The former is designed for optical multiple-input multiple-output (O-MIMO) systems based on the optical spatial modulation (OSM) technique. In NDC-OFDM, signs of modulated O-OFDM symbols and absolute values of the symbols are separately transmitted by different information carrying units. This scheme can eliminate clipping distortion in DCO-OFDM and achieve high power efficiency. Furthermore, as the indices of transmitters carry extra information bits, NDC-OFDM gives a significant improvement in spectral efficiency over ACO-OFDM and U-OFDM. In this thesis, SPAD-based OFDM systems with DCO-OFDM and ACO-OFDM are presented and analysed by considering the nonlinear distortion effect of PQ SPAD and AQ SPAD. A comprehensive digital signal processing of SPAD-based OFDM is shown and theoretical functions of the photon counting distribution in PQ SPAD and AQ SPAD are given. Moreover, based on Bussgang theorem, a conventional method for analysing memoryless distortion, close-formed bit-error rate (BER) expressions of SPAD-based OFDM are derived. Furthermore, SPAD-based OFDM is compared with conventional photo-diode (PD) based OFDM systems, and a gain of 40 dB in power efficiency is observed.

De nombreuses applications en sciences nucléaires bénéficieraient d’un détecteur possédant une précision temporelle de 10 ps largeur à mi-hauteur à la mesure d’un photon unique. Par exemple, le projet de Time-Imaging Calorimeter en cours de conception au CERN requiert un détecteur possédant une telle précision temporelle afin de mesurer le temps de vol (TDV) et la trajectoire des particules émises lors des collisions dans les expériences du Large Hadron Collider (LHC), ce qui permet d’identifier ces dites particules. De plus, un détecteur possédant une précision temporelle de l’ordre de 10 ps permettra la mitigation de l’empilement des événements. Un second exemple est la tomographie d’émission par positrons (TEP), une modalité d’imagerie médicale non-invasive qui mesure la distribution d’un traceur radioactif afin d’étudier et détecter le cancer. Dans le but de développer un scanner TEP temps réel, le groupe de recherche en appareillage médical de Sherbrooke (GRAMS) travaille sur l’intégration de la mesure du TDV de l’interaction TEP. Les meilleures performances actuelles des détecteurs TEP se situent aux alentours de 150 ps, ce qui n’est pas suffisant pour intégrer le TDV dans un scanner TEP préclinique. Cette mesure exige une résolution temporelle TEP de l’ordre de 10 ps. La solution proposée par le GRAMS est de développer un module de comptage monophotonique (MCMP) 3D qui est composé d’une matrice de photodiodes avalanches monophotoniques (PAMP) reliée par des interconnexions verticales (TSV) à une matrice de circuits de lecture composée d’un circuit d’étouffement et d’un convertisseur temps-numérique. Ce détecteur permet donc de mesurer précisément le temps d’arrivée de chaque photon détecté. Ce document présente la conception du circuit d’étouffement réalisé en technologie CMOS 65 nm de TSMC (Taiwan Semiconductor Manufacturing Company) intégré à chaque pixel de 50 × 50 µm2 dans un MCMP 3D. Afin de répondre au besoin de précision temporelle de 10 ps dans un détecteur 3D, le circuit proposé est un circuit d’étouffement passif avec une recharge active possédant un amplificateur opérationnel en boucle ouverte à titre de comparateur de tension. L’amplificateur opérationnel utilisé possède un seuil ajustable de 0 à 2,5 V afin d’être en mesure d’évaluer le seuil optimal pour la mesure de gigue temporelle avec une PAMP. La taille finale du circuit d’étouffement est de 18 × 30 µm2 incluant l’amplificateur qui est d’une taille de 13 × 8 µm2, ce qui représente respectivement environ 22% et 4% de la taille totale du pixel. Le circuit d’étouffement possède une gigue temporelle de 4 ps largeur à mi-hauteur (LMH). Les résultats obtenus prouvent qu’il est possible d’intégrer de l’électronique de lecture de PAMP dans un MCMP 3D possédant des performances temporelles sous les 10 ps.

碩士<br>國立交通大學<br>電子工程學系 電子研究所<br>101<br>In this work, the photon detection performance of single-photon avalanche diodes (SPADs) fabricated in the high-voltage (HV) 0.25-μm CMOS technology is studied and discussed in details, including photon detection probability (PDE) and jitter. The devices measured in this work exhibited a very low dark count rate in a previous study. The wavelength-dependent PDEs are measured under various excess voltages. The maximun PDE of about 14.2% at 510 nm is obtained. By squeezing the incident light spot into about 1-μm, the 2-D spatial distribution of photo-counts in the circular active area are mapped automatically. The 2-D mappings of photo-counts reveal a clear ring-like non-uniformity. The non-uniform distribution becomes more significant with a shorter wavelength and a higher bias voltage. Simulations with TCAD are performed to understand the spatial distributions of electric field inside the active region. It is found that the arrangement of contact pad and connection metal line affects the electric field underneath, which results the non-uniformity of photo-counts. In addition, by using pulsed laser diodes at 405 and 782 nm and a time-correlated photon-counting card, the jitter distributions of the devices under various bias voltages are measured and analyzed. At last, by using the same technology, two new structures of SPADs with opposite doping types are designed and simulated to study their transient photo response, which would be helpful for achieving low jitter SPADs in the future.

碩士<br>國立交通大學<br>電子研究所<br>106<br>To enhance detection efficiency and image resolution, high fill-factor array of single-photon avalanche diodes (SPADs) is applied by shortening the distance between devices. However, crosstalk between adjacent SPADs becomes an issue in this decade. In this thesis, an extensive study on various aspects of crosstalk in a CMOS SPAD array, including device structure, device-to-device distance, bias voltage, and its timing characteristics, is presented. According to the experimental results and a simple model fitting, we deduce that the crosstalk between CMOS SPADs is dominated by direct optical coupling. Timing characteristics of crosstalk support our explanation and provide an additional angle of observation on its mechanism. For single SPADs, by applying a proper doping concentration profile in TCAD simulation, experimental temperature-dependent breakdown voltages and C-V curves are fitted in good consistence. The size effect of photon-detection probability (PDP) is investigated by using 2-D photo-count mapping. The same simulation method and doping profiles tells that the PDP non-uniformity is caused by the guard-ring induced electric field lowering. A doping profile for guard-ring region is proposed to ease this problem. This work is highly valuable for SPAD array design for various imaging applications in the future.

碩士<br>國立交通大學<br>電子工程學系 電子研究所<br>104<br>In this work, the afterpulsing effect in single photon avalanche diodes (SPADs) fabricated by TSMC 25HV (high voltage) CMOS process are studied. A new method for evaluating afterpulsing effect has been proposed and demonstrated. Different from conventional method requiring photon correlation measurement and short-pulsed light source, the proposed scheme is simply a measurement of dark count rate (DCR) distribution. Because the afterpulsing events correlate with their parent breakdowns, the DCR distribution deviates from the original Poisson one, which can be used to evaluate afterpulsing probability (APP). To demonstrate the validity of our method, we established a system to measure the temperature-dependent DCRs of a SPAD and analyzed their distribution. At low temperature, as the afterpulsing effect worsens, a clear non-Poisson distribution of DCRs is observed. A quantitative simulation has been performed to find out the relation between the DCR distribution and the APP. Our method is useful for evaluating APPs either in single SPADs or in circuit-integrated SPAD arrays.

碩士<br>國立交通大學<br>電子研究所<br>100<br>In this work, we study single-photon-avalanche-diodes (SPADs) with an active quenching circuit (AQC) by using the standard 0.18 µm CMOS technology. Because the dead time of SPAD with passive quenching circuit is too long, the main target of the AQC design is to shorten it. On the other hand, the large chip area of AQC will reduce the spatial resolution of the fabricated SPAD array so to minimize the AQC is also important. Our SPADs with well-designed AQC show the fastest dead time of 4 ns, and use the smallest chip area of 15.7×15.2 〖µm〗^2, comparing with others using 0.18 µm CMOS technology. In addition, we demonstrate the function of tunable hold-off time which can be adjusted, according to the characteristics of SPAD, to reduce afterpulsing effect and to lower dark counts. The measured dead time can be extended to more than 280 ns, and the reduced equivalent dark counts are also obtained. Finally, we observe that dark counts and photon detection efficiency (PDE) both increase with increasing excess bias voltages. The highest PDE of 7.5% is achieved. The noise equivalent power (NEP) is about 10-14 WHz-1/2.

碩士<br>國立交通大學<br>電子研究所<br>107<br>In this work, we realize single photon avalanche diode(SPAD) by using EPISIL 0.8μm CMOS technology. In addition, we implement two SPADs device by using standard CMOS technology which is P+/NWELL SPAD, N+/PWELL SPAD and PWELL/NBL SPAD. Particularly, P+/NWELL SPAD with 27.8V breakdown voltage has a better performance with 10% excess, for example, low dark count rate(DCR) is 56Hz, high photon detect probability(PDP) is 15.38%@495nm, timing jitter FWHM is 157ps. To improve N+/PWELL SPAD and PWELL/NBL SPAD’s characteristics, we introduce the customized CMOS technology for new SPAD(SPAD-A, SPAD-B). From TCAD simulation result, SPAD-A’s breakdown voltage is identical but the depletion width is wider compared with P+/NWELL. The other way, SPAD-B’s breakdown voltage is decreased by using retrograded well. In the end, the customized SPAD’s characteristics are abnormal, whose breakdown voltage is not excepted by simulation. We found out the problem which is the simulation model for ion implant and doping diffuse is not accurate. Although the experiment is not work successfully, it provide the appropriate design flow for future research.

碩士<br>國立交通大學<br>電子工程學系 電子研究所<br>104<br>In this work, we investigate single-photon avalanche diodes (SPADs) in standard 0.18-m high-voltage CMOS technology provided by TSMC. The SPADs with various kinds of P-N junctions have been designed, fabricated, and characterized. Device simulation and afterward analysis are performed with Sentaurus-TCAD tool. Among the studied devices, 20-m-diameter SPADs formed with deep p-typed well (DPW) and n-typed buried layer (NBL) have the best performance including low dark count rate (DCR), high photon detection efficiency (PDE), low jitter and reduced breakdown voltage comparing with the previous ones in our group. Possible reasons for the improvement are discussed and explained by the simulation on revised doping profiles of the layers. However, the dependence of jitter on the photon wavelength exhibits unusual behavior around 720 nm and calls for further studies in the future.

碩士<br>國立交通大學<br>電子研究所<br>100<br>In this thesis, by using standard 0.18μm CMOS process, we study the vertical and lateral single-photon avalanche diodes (SPADs). Simulation results show that the lower p-typed and n-typed doping concentration in lateral SPADs can reduce the band-to-band tunneling rate so as their dark count rate. Fifteen devices are fabricated with various parameters such as with/without grating, operation voltages, with/without deep n-well (DNW). The measured breakdown voltages of the vertical and the lateral device are about 10 V and 15 V, respectively, which is consistent with the simulated ones. The characteristics of the SPADs biased below and above the breakdown voltages are measured and discussed. The lateral structure has a higher responsivity and fast transient time, comparing with the vertical structure. It is also found that the grating above the device shows no improvement on its responsivity. For the devices performance above breakdown voltages, different from the simulation results, the dark count rate of the lateral structures is much higher than that of the vertical ones. We suspect that much higher dark count rates are caused by the unwanted shallow trench isolation (STI) in the active region, whose existence is observed with optical and secondary electron microscope. By using gated-mode measurement, we have obtained the de-trapping times of the STI-induced traps under various exceed biases. The dark count rates, photon detection efficiency (PDE) and noise equivalent power (NEP) at 400 nm are measured with long enough dead time to avoid afterpulsing effect.

碩士<br>國立交通大學<br>電子工程學系 電子研究所<br>103<br>An active reset circuit with excess bias voltage tracking capability for single-photon avalanche diode (SPAD) achieving uniform dark count rate (DCR) and photon detection efficiency (PDE) has been presented. By using sample-and-hold circuit to detect the quenching voltage level of SPAD, the circuit adjusts the reset voltage level of SPAD and keeps the device at constant excess bias voltage of 1.6V. The circuit can compensate the variation of breakdown voltage and biasing voltage up to a range beyond 1V and achieve uniform DCR and PDE within this interval. The chip is fabricated in standard 0.18 m CMOS technology and operates under 3.3V supply voltage. This thesis also demonstrates a time-of-flight (TOF) rangefinding system operating by detecting the phase difference of light pulse. The system consists of SPAD circuits in this thesis as detector, a laser source modulated by pulse train and a gated-mode photon counter implemented by FPGA. The system is verified through introducing different time delay by varying the length of coaxial cable.

碩士<br>國立交通大學<br>電子研究所<br>101<br>In this thesis, we propose and demonstrate a device structure of low dark-count-rate (DCR) single photon detector. To avoid the breakdown events triggered by the trap of shallow trench isolation (STI) in the active region, we design a guard-ring structure to keep the STI in distance or relocate the active region from the top region to the deeper one. TCAD simulation tool is used to calculate the spatial distributions of electric field and impact ionization to confirm the feasibility of our design. With the 0.25-µm high-voltage standard CMOS technology, we have fabricated the designed devices successfully. The DCRs of devices under various excess voltages have been characterized with the setup in our lab and with the passive quenching circuit. The results show that the DCR of designed structure is lowered by more than two orders comparing with that of the conventional one. The lowest DCR less than 10 Hz is obtained. With a precise calibration of incident power, we have also measured the photon detection efficiency (PDE) of the devices under various excess voltage and incident wavelengths. The highest PDE reaches 15.4 % at 650 nm. At last, we discuss the DCR mechanism of the best device and suggest the direction for further improvement in the future.

碩士<br>國立交通大學<br>電子研究所<br>107<br>We demonstrate two simple designs to enhance the dynamic ranges for CMOS single-photon avalanche diodes. The one is the time gated active reset scheme, it can be reset by clock signal. The dynamic range of time gated active reset is the best, which the maximum count rate is 225 MHz, it can lift six times the maximum count rate with it closed to the area of the conventional passive reset scheme, and it is the highest SPAD count rate to the best of the author’s knowledge. The other is mutual reset scheme, the system can be quickly reset by SPADs’ mutual effect. It has the ability to dynamically adjust the reset frequency with the intensity of the background light. Under the same conditions, the intervention of mutual reset can enhance about twice the dynamic range. In the ranging histogram, it has no obvious peak of afterpulsing, the afterpulsing probability of mutual reset is 1-3 %, it can reduce misjudgment of mult-echo. Furthermore, we make 4×4 SPAD arrays by time gated active reset and mutual reset schematic, which can be measurement laser under the SPADs’ background count rate of 1 GHz. In this work, we will analyze the dynamic range, ranging experiment, afterpulsing, crosstalk, power, etc. for these systems, and comprehensively analyze their advantages and disadvantages.