The SRAM memory generator in this thesis is divided into four parts: row decoder, storage cell, column decoder, and sense amplifier & write controller. The row decoder is designed using pass-transistors logic with better area and regularity compared with conventional NAND based decoders. Two different column decoders, tree structure and NOR based predecoder, are provided in current version. Although only SRAM is implemented in this thesis, the memory generator platform is complete with all the necessary models required in the embedded design. In the future, other memories, such as cache, shift register, FIFO, stacks, ROM, register files, and content addressable memory, can be integrated in this memory generator platform.

Memory wall is becoming a more and more serious bottleneck of the processing speed of microprocessors. The mismatch between CPUs and memories has been increasing since three decades ago. SRAM was introduced as the bridge between the main memory and the CPU. SRAM is designed to be on the same die with CPU and stores temporary data and instructions that are to be processed by the CPU. Thus, the performance of SRAMs has a direct impact on the performance of CPUs. With the application of mass amount data to be processed nowadays, there is a great need for high-performance CPUs. Three dimensional CPUs and CPUs that are specifically designed for machine learning are gaining popularity. The objective of this work is to design high-performance SRAM for these two emerging applications. Firstly, a novel delay cell based on dummy TSV is proposed to replace traditional delay cells for better timing control. Secondly, a unique SRAM with novel architecture is custom designed for a high-performance machine learning processor. Post-layout simulation shows that the SRAM works well with the processing core and its design is optimized to work well with machine learning processors based on convolutional neural networks. A prototype of the SRAM is also tapped out to further verify our design.

京都大学 / 新制・課程博士 / 博士(情報学) / 甲第24746号 / 情博第834号 / 新制||情||140(附属図書館) / 京都大学大学院情報学研究科通信情報システム専攻 / (主査)教授 佐藤 高史, 教授 橋本 昌宜, 教授 新津 葵一 / 学位規則第4条第1項該当 / Doctor of Informatics / Kyoto University / DFAM

The overall energy consumption of portable devices has been projected to triple over the next decade, growing to match the total power generated by the European Union and Canada by 2025. The rise of the internet-of-things (IoT) and ubiquitous and embedded computing has resulted in an exponential increase in such devices, wherein projections estimate that 50 billion smart devices will be connected and online by 2020. In order to alleviate the associated stresses placed on power generation and distribution networks, a holistic approach must be taken to conserve energy usage in electronic devices from the component to the circuit level. An effective approach to reduce power dissipation has been a continual reduction in operating voltage, thereby quadratically down-scaling active power dissipation. However, as state-of-the-art silicon (Si) complimentary metal-oxide-semiconductor (CMOS) field-effect transistors (FETs) enter sub-threshold operation in the ultra-low supply voltage regime, their drive current is noticeable degraded. Therefore, new energy-efficient MOSFETs and circuit architectures must be introduced. In this work, tunnel FETs (TFETs), which operate leveraging quantum mechanical tunneling, are investigated. A comprehensive investigation detailing electronic materials, to novel TFET device designs, to memory and logic digital circuits based upon those TFETs is provided in this work. Combined, these advances offer a computing platform that could save considerable energy and reduce power consumption in next-generation, ultra-low voltage applications. / Ph. D.

Physically unclonable functions are increasingly used as security primitives for device identification and anti-counterfeiting. However, PUFs are associated with noise and bias which in turn affects its property of reliability and predictability. The noise is corrected using fuzzy extractors, but the helper data generated during the process may cause leakage in min-entropy due to the bias observed in the response. This thesis offers two optimization techniques for PUF based protocols. The first part talks about the construction of a secure enrollment solution for PUFs on a low-end resource-constrained device using a microcontroller and a secure networked architecture. The second part deals with the combined optimization of min-entropy and error-rate using symbol clustering techniques to improve the reliability of SRAM PUFs. The results indicate an increase in min-entropy without much effect on the error rate but at the expense of PUF size. / Master of Science

With a view to reduce the number of components in a Static Random Access Memory (SRAM) cell, the feasibility of use of Lambda Bipolar Transistor (LBT)in the bistable element of the cell has been explored under the present study. The LBT under consideration here comprises of an enhancement mode MOSFET integrated with a parasitic bipolar transistor so as to perform as a negative resistance device. LBTs for the study have been fabricated and analysed. The devices have been shown to function at much lower voltage and current levels than those reported earlier/ and thus have been shown to be suitable for lower power applications. The issues of agreements and discrepancies of the experimental results with the original DC model of the device have been highlighted and discussed. The factors contributing to the drain current of the MOSFET in the LBT have been identified. It has also been shown that in the real case of an LBT in operation, the MOSFET in it does not function as a discrete device for the same conditions of voltages and current levels as in an LBT. As per the present study, it is assessed to be influenced by the presence of the BJT in operation and this effect is felt more at the lower current levels of operation. With a separate and tailored p-well implantation the possibility of fabrication of LBTs with a CMOS technology is established. Along with a couple of polysilicon resistors, the LBTs have been successfully made to perform in the common-collector configuration as the bistable storage element of SRAM cell (as proposed in the literature). The bistable element with the LBT in common-emitter mode also has been visualised and practically achieved with the fabricated devices. The WRITE transients for either case have been simulated for various levels of WRITE voltages and their time of hold.The speed of Writing achieved are found comparable with that of the standard SRAMs. The advantages and disadvantages of using the LBT in either mode have been highlighted and discussed. The power consumption of the bistable element with the LBT in either mode is however shown to be the same. A different approach of READING has been proposed to overcome the factors known to increase the cycle time. On the whole, under the present study, the proposal of using LBTs in the bistable storage element of the SRAM cell has been shown to be feasible. Such SRAM circuits can find possible applications in the fields where smaller circuit area is the major concern.

Monolithic Active Pixel Sensors (MAPS) are good candidates to be used in High Energy Physics (HEP) experiments for charged particle detection. In the HEP applications, MAPS chips are placed very close to the interaction point and are directly exposed to harsh environmental radiation. This thesis focuses on the study and improvement of the MAPS radiation hardness. The main radiation effects and the research progress of MAPS are studied firstly. During the study, the SRAM IP cores built in MAPS are found limiting the radiation hardness of the whole MAPS chips. Consequently, in order to improve the radiation hardness of MAPS, three radiation hard memories are designed and evaluated for the HEP experiments. In order to replace the SRAM IP cores, a radiation hard SRAM is developed on a very limited area. For smaller feature size processes, in which the single event upset (SEU) effects get significant, a radiation hard SRAM with enhanced SEU tolerance is implemented by an error detection and correction algorithm and a bit-interleaving storage. In order to obtain higher radiation tolerance and higher circuitry density, a dual-port memory with an original 2-transistor cell is developed and evaluated for future MAPS chips. Finally, the radiation hardness of the MAPS chips using new available processes is studied, and the future works are prospected.

Les capteurs monolithiques actifs à pixels (Monolithic Active Pixel Sensors, MAPS) sont de bons candidats pour être utilisés dans des expériences en Physique des Hautes Énergies (PHE) pour la détection des particules chargées. Dans les applications en PHE, des puces MAPS sont placées dans le voisinage immédiat du point d’interaction et sont directement exposées au rayonnement intense de leur environnement. Dans cette thèse, nous avons étudié et amélioré la résistance aux radiations des MAPS. Les effets principaux de l’irradiation et le progrès de la recherche sur les MAPS sont étudiés tout d'abord. Nous avons constaté que les cœurs des SRAM IP incorporées dans la puce MAPS limitent sensiblement la tolérance aux radiations de la puce MAPS entière. Aussi, pour améliorer la radiorésistance des MAPS, trois mémoires radiorésistantes sont conçues et évaluées pour les expériences en PHE. Pour remplacer les cœurs des IP SRAM, une SRAM radiorésistante est développée sur une petite surface. Pour les procédés de plus petit taille de grille des transistors, dans lequel les effets SEU (Single Event Upset) deviennent significatifs, une SRAM radiorésistante avec une tolérance SEU accrue est réalisée à l’aide d’un algorithme de détection et de correction d'erreurs (Error Detection And Correction, EDAC) et un stockage entrelacé des bits. Afin d'obtenir une tolérance aux rayonnements et une densité de micro-circuits plus élevées, une mémoire à double accès avec une cellule à 2 transistors originale est développée et évaluée pour des puces MAPS futures. Enfin, la radiorésistance des puces MAPS avec des nouveaux procédés disponibles est étudiée, et les travaux futurs sont proposés. / Monolithic Active Pixel Sensors (MAPS) are good candidates to be used in High Energy Physics (HEP) experiments for charged particle detection. In the HEP applications, MAPS chips are placed very close to the interaction point and are directly exposed to harsh environmental radiation. This thesis focuses on the study and improvement of the MAPS radiation hardness. The main radiation effects and the research progress of MAPS are studied firstly. During the study, the SRAM IP cores built in MAPS are found limiting the radiation hardness of the whole MAPS chips. Consequently, in order to improve the radiation hardness of MAPS, three radiation hard memories are designed and evaluated for the HEP experiments. In order to replace the SRAM IP cores, a radiation hard SRAM is developed on a very limited area. For smaller feature size processes, in which the single event upset (SEU) effects get significant, a radiation hard SRAM with enhanced SEU tolerance is implemented by an error detection and correction algorithm and a bit-interleaving storage. In order to obtain higher radiation tolerance and higher circuitry density, a dual-port memory with an original 2-transistor cell is developed and evaluated for future MAPS chips. Finally, the radiation hardness of the MAPS chips using new available processes is studied, and the future works are prospected.

Silicon based 3D fin structure is believed to be the potential future of current semiconductor technology. However, there are significant challenges still exist in realizing a manufacturable fin based process. In this work, we have studied the effects of hydrogen anneal on the structural and electrical characteristics of silicon fin based devices: tri-gate, finFET to name a few. H₂ anneal is shown to play a major role in structural integrity and manufacturability of 3D fin structure which is the most critical feature for these types of devices. Both the temperature and the pressure of H₂ anneal can result in significant alteration of fin height and shape as well as electrical characteristics. Optimum H₂ anneal is required in order to improve carrier mobility and device reliability as shown in this work. A new hard-mask based process was developed to retain H₂ anneal related benefit while eliminating detrimental effects such as reduction of device drive current due to fin height reduction. We have also demonstrated a novel 1T-1C pseudo Static Random Access Memory (1T-1C pseudo SRAM) memory cell using low cost conventional tri-gate process by utilizing selective H₂ anneal and other clever process techniques. TCAD-based simulation was also provided to show its competitive advantage over other types of static and dynamic memories in 45nm and beyond technologies. A high gain bipolar based on silicon fin process flow was proposed for the first time that can be used in BiCMOS technology suitable for low cost mixed signal and RF products. TCAD-based simulation results proved the concept with gain as high 100 for a NPN device using single additional mask. Overall, this work has shown that several novel process techniques and selective use of optimum H₂ anneal can lead to various high performance and low cost devices and memory cells those are much better than the devices using current conventional 3D fin based process techniques. / text

L’émergence des circuits complexes numériques, ou System-On-Chip (SOC), pose notamment la problématique de la consommation énergétique. Parmi les blocs fonctionnels significatifs à ce titre, apparaissent les mémoires et en particulier les mémoires statiques (SRAM). La maîtrise de la consommation énergétique d’une mémoire SRAM inclue la capacité à rendre la mémoire fonctionnelle sous très faible tension d’alimentation, avec un objectif agressif de 300 mV (inférieur à la tension de seuil des transistors standard CMOS). Dans ce contexte, les travaux de thèse ont concerné la proposition d’un point mémoire SRAM suffisamment performant sous très faible tension d’alimentation et pour les nœuds technologiques avancés (CMOS bulk 28nm et FDSOI 28nm). Une analyse comparative des architectures proposées dans l’état de l’art a permis d’élaborer deux points mémoire à 10 transistors avec de très faibles impacts de courant de fuite. Outre une segmentation des ports de lecture, les propositions reposent sur l’utilisation de périphéries adaptées synchrones avec notamment une solution nouvelle de réplication, un amplificateur de lecture de données en mode tension et l’utilisation d’une polarisation dynamique arrière du caisson SOI (Body Bias). Des validations expérimentales s’appuient sur des circuits en technologies avancées. Enfin, une mémoire complète de 32kb (1024x32) a été soumise à fabrication en 28 FDSOI. Ce circuit embarque une solution de test (BIST) capable de fonctionner sous 300mV d’alimentation. Après une introduction générale, le 2ème chapitre du manuscrit décrit l’état de l’art. Le chapitre 3 présente les nouveaux points mémoire. Le 4ème chapitre décrit l’amplificateur de lecture avec la solution de réplication. Le chapitre 5 présente l’architecture d’une mémoire ultra basse tension ainsi que le circuit de test embarqué. Les travaux ont donné lieu au dépôt de 4 propositions de brevet, deux conférences internationales, un article de journal international est accepté et un autre vient d’être soumis. / Emergence of large Systems-On-Chip introduces the challenge of power management. Of the various embedded blocks, static random access memories (SRAM) constitute the angrier contributors to power consumption. Scaling down the power supply is one way to act positively on power consumption. One aggressive target is to enable the operation of SRAMs at Ultra-Low-Voltage, i.e. as low as 300 mV (lower than the threshold voltage of standard CMOS transistors). The present work concerned the proposal of SRAM bitcells able to operate at ULV and for advanced technology nodes (either CMOS bulk 28 nm or FDSOI 28 nm). The benchmarking of published architectures as state-of-the-art has led to propose two flavors of 10-transitor bitcells, solving the limitations due to leakage current and parasitic power consumption. Segmented read-ports have been used along with the required synchronous peripheral circuitry including original replica assistance, a dedicated unbalanced sense amplifier for ULV operation and dynamic forward back-biasing of SOI boxes. Experimental test chips are provided in previously mentioned technologies. A complete memory cut of 32 kbits (1024x32) has been designed with an embedded BIST block, able to operate at ULV. After a general introduction, the manuscript proposes the state-of-the-art in chapter two. The new 10T bitcells are presented in chapter 3. The sense amplifier along with the replica assistance is the core of chapter 4. The memory cut in FDSOI 28 nm is detailed in chapter 5. Results of the PhD have been disseminated with 4 patent proposals, 2 papers in international conferences, a first paper accepted in an international journal and a second but only submitted paper in an international journal.