Degradation of transistor parameter values due to Negative Bias Temperature Instability (NBTI) has emerged as a major reliability problem in current and future technology generations. NBTI Aging of a Static Random Access Memory (SRAM) cell leads to a lower noise margin, thereby increasing the failure rate. The register file, which consists of an array of SRAM cells, can suffer from data loss, leading to a system failure. In this work, we study the source of NBTI stress in an architecture and physical register file. Based on our study, we modified the register file structure to reduce the NBTI degradation and improve the overall system reliability. Having evaluated new register file structures, we find that our techniques substantially improve reliability of the register files. The new register files have small overhead, while in some cases they provide saving in area and power.

In recent years, amorphous oxide semiconductors have been studied due to their superior characteristics, such as transparent property, high electron mobility exceeding 10 cm2/V¡Es, and can be fabricated on plastic substrates at low temperatures. According to these advantages, a-IGZO thin-film transistors are promising as next-generation electronic devices. Although a-IGZO TFTs have such unique properties, the electrical performances are strongly dependent on its environment such as oxygen, water and visible light. In this study, the electrical characteristics of a-IGZO TFTs under positive bias stress with different ambient gases have been discussed. In particular, the total duration of the negative gate bias applied on the switching transistor is larger than that of the positive gate bias in display application. Therefore, the electrical stability under negative bias stress is vital to investigate. Moreover, a-IGZO TFT regarded as a panel switch may be exposed to visible light for the application of liquid crystal display. The electrical stability under illumination of visible light is also important to study. Experiment results show that device characteristics are affected under water-containing oxygen ambience. We indicates that the existence of water molecules can assist more oxygen to adsorb on the a-IGZO surface than the case without water assisting. That cause the variation of transfer curve under positive bias stress. However, the degradations in subthreshold swing and threshold voltage are caused by the state-related adsorption of water molecules under negative bias stress. Furthermore, adsorbed oxygen on the surface of a-IGZO can be desorbed by illumination of visible light, leading to large variation in transfer curve.

Transistor aging through negative bias temperature instability (NBTI) has become a major lifetime constraint in VLSI circuits. We propose a technique that uses antifuses to widen PMOS transistors later in a circuit?s life cycle to combat aging. Using HSPICE and 70nm BPTM process numbers, we simulated the technique on four circuits (a ring oscillator, a fan-out four circuit, an ISCAS c432 and c2670). Over the lifetime of the circuit, our simulations predict a 8.89% and a 13% improvement in power in the c432 and c2670 circuits respectively when compared to similarly performing traditional circuits.

With the gradual advance of the state-of-the-art VLSI manufacturing technology into the sub-45nm regime, engineering a reliable, high performance VLSI chip with economically attractive yield in accordance with Moore's law of scaling and integration has become extremely difficult. Some of the most serious challenges that make this task difficult are: a) the delay of a transistor is strongly dependent on process induced mechanical stress around it, b) the reliability of devices is affected by several aging mechanisms like Negative Bias Temperature Instability (NBTI), hot carrier injection (HCI), etc and c) the delay and reliability of any device are strongly related to lithographically drawn geometry of various features on wafer. These three challenges are the main focus of this dissertation. High performance fabrication processes routinely use embedded silicon-germanium (eSiGe) technology that imparts compressive mechanical stress to PMOS devices. In this work, cell level timing models considering flexibility to modulate active area to change mechanical stress, were proposed and exploited to perform timing optimization during circuit placement phase. Analysis of key physical synthesis optimization steps such as gate sizing and repeater insertion was done to understand and exploit mechanical stress to significantly improve delay of interconnect and device dominated circuits. Regarding circuit reliability, the proposed work is focused on reducing the clock skew degradation due to NBTI effect specially due to the use of clock gating technique for achieving low power operation. In addition, we also target the detrimental impact of burn-in testing on NBTI. The problem is identified and a runtime technique to reduce clock skew increase was proposed. For designs with predictable clock gating activities, a zero overhead design time technique was proposed to reduce clock skew increase over time. The concept of using minimum degradation input vector during static burn-in testing is proposed to reduce the impact of burn-in testing on parametric yield. Delay and reliability strongly depend on dimension of various features on the wafer such as gate oxide thickness, channel length and contact position. Increased variability of these dimensions can severely restrict ability to analyze or optimize a design considering mechanical stress and circuit reliability. One key technique to control physical variability is to move towards regular fabrics. However, to make implementation on regular fabrics attractive, high quality physical design tools need to be developed. This dissertation proposes a new circuit placement algorithm to place a design on a structured ASIC platform with strict site and clock constraints and excellent overall wirelength. An algorithm for reducing the clock and leakage power dissipation of a structured ASIC by reducing spine usage is then proposed to allow lower power dissipation of designs implemented using structured ASICs. / text

Avec la tendance continue vers la technologie nanométrique et l'augmentation des fonctions complexes intègres dans les électroniques systèmes embarqués, Assurant la compatibilité électromagnétique (CEM) des systèmes électroniques est un grand défi. CEM est devenu une cause majeure de redesign des Circuits intègres (CI). D’ailleurs, les performances des circuits pourraient être affectés par les mécanismes de dégradation tels que hot carrier injection (HCI), negative bias temperature instability (NBTI), gate oxide breakdown, qui sont accélérés par les conditions d'exploitation extrême (haute / basse température, surcharge électrique, le rayonnement). Ce vieillissement naturel peut donc affecter les performances CEM des circuits intégrés.Les travaux développés dans notre laboratoire vise à clarifier le lien entre les dégradations induites par le vieillissement et les dérives CEM, de développer les modèles de prédiction et de proposer des "insensibles au cours du temps" structures pour CEM protection, afin de fournir des méthodes et des guidelines aux concepteurs d'équipements et CI pour garantir la CEM au cours de durée de vie de leurs applications. Ce sujet de recherche est encore sous-exploré en tant que communautés de recherche sur la «fiabilité IC» et «compatibilité électromagnétique IC» n’a souvent pas de chevauchement.Ce manuscrit de thèse introduit une méthode pour quantifier l'effet du vieillissement sur les CEM des circuits intégrés par la mesure et la simulation. Le premier chapitre donne un aperçu du contexte général et le deuxième chapitre est dédié a l’état de l'art de CEM des circuits intégrés et de problèmes de fiabilité IC. Les résultats expérimentaux de circuits CEM évolution sont présentés dans le troisième chapitre. Ensuite, le quatrième chapitre est consacré à la caractérisation et la modélisation des mécanismes de dégradation du CI. Un EMR modèle qui inclut l'élément le vieillissement pour prédire la dérive du niveau CEM de notre puce de test après stress est proposé / With the continuous trend towards nanoscale technology and increased integration of complex electronic functions in embedded systems, ensuring the electromagnetic compatibility (EMC) of electronic systems is a great challenge. EMC has become a major cause of IC redesign. Meanwhile, ICs performance could be affected by the degradation mechanisms such as hot carrier injection (HCI), negative bias temperature instability(NBTI), gate oxide breakdown, which are accelerated by the harsh operation conditions (high/low temperature, electrical overstress, radiation). This natural aging can thus affect EMC performances of ICs. The work developed in our laboratory aims at clarifying the link between ageing induced IC degradations and related EMC drifts, developing prediction models and proposing “time insensitive” EMC protection structures, in order to provide methods and guidelines to IC and equipment designers to ensure EMC during lifetime of their applications. This research topic is still under-explored as research communities on “IC reliability” and “IC electromagnetic compatibility” has often no overlap. The PhD manuscript introduced a methodology to quantify the effect of ageing on EMC of ICs by measurement and simulation. The first chapter gives an overview of the general context and the second chapter states the EMC of ICs state of the art and IC reliability issues. The experimental results of ICs EMC evolution are presented in the third chapter. Then, the fourth chapter is dedicated to the characterization and modeling IC degradation mechanism. An EMR model which includes the ageing element to predict our test chip’s EMC level drift after stress is proposed

Actuellement, les circuits numériques nécessitent d'être de plus en plus performants. Aussi, les produits doivent être conçus le plus rapidement possible afin de gagner les précieuses parts de marché. Les méthodes rapides de conception et l'utilisation de MPSoC ont permis de satisfaire à ces exigences, mais sans tenir compte précisément de l'impact du vieillissement des circuits sur la conception. Or les MPSoC utilisent les technologies de fabrication les plus récentes et sont de plus en plus soumis aux défaillances matérielles. De nos jours, les principaux mécanismes de défaillance observés dans les transistors des MPSoC sont le HCI et le NBTI. Des marges sont alors ajoutées pour que le circuit soit fonctionnel pendant son utilisation, en considérant le cas le plus défavorable pour chaque mécanisme. Ces marges deviennent de plus en plus importantes et diminuent les performances attendues. C'est pourquoi les futures méthodes de conception nécessitent de tenir compte des dégradations matérielles en fonction de l’utilisation du circuit. Dans cette thèse, nous proposons une méthode originale pour simuler le vieillissement des MPSoC à haut niveau d'abstraction. Cette méthode s'applique lors de la conception du système c.-à-d. entre l'étape de définition des spécifications et la mise en production. Un modèle empirique permet d'estimer les dégradations temporelles en fin de vie d'un circuit. Un exemple d'application est donné pour un processeur embarqué et les résultats pour un ensemble d'applications sont reportés. La solution proposée permet d'explorer différentes configurations d'une architecture MPSoC pour comparer le vieillissement. Aussi, l'application la plus sévère pour le vieillissement peut être identifiée. / Nowadays, more and more performance is expected from digital circuits. What’s more, the market requires fast conception methods, in order to propose the newest technology available. Fast conception methods and the utilization of MPSoC have enabled high performance and short time-to-market while taking little attention to aging. However, MPSoC are more and more prone to hardware failures that occur in transistors. Today, the prevailing failure mechanisms in MPSoC are HCI and NBTI. Margins are usually added on new products to avoid failures during execution, by considering worst case scenario for each mechanism. For the newest technology, margins are becoming more and more important and products performance is getting lower and lower. That’s why the conception needs to take into account hardware failures according to the execution of software. This thesis propose a new methodology to simulate aging at high level of abstraction, which can be applied to MPSoC. The method can be applied during product conception, between the specification phase and the production. An empirical model is used to estimate slack time at circuit's end of life. A use case is conducted on an embedded processor and degradation results are reported for a set of applications. The solution enables architecture exploration and MPSoC aging can thus be compared. The software with most severe impact on aging can also be determined.

The higher mobility is needed for thin film transistor (TFT) mainly used to be applied in the larger size flat-panel displays (FPDs). The amorphous metal oxide TFT has mobility higher than 10 cm2/V¡Es and can substitute the poor mobility (<1 cm2/V¡Es) of traditional amorphous silicon TFT, which shows a great potential for the next generation. Due to the superior characteristics in amorphous metal oxide TFT, therefore, the amorphous metal oxide TFT has been studied extensively. Usually, the source/drain with island type device has a large overlapped/contact area that we cannot determine the exact electron path. That the sample of inverted stagger £\-IGZO TFTs with via type device has smaller contact area and can be estimated the electron path. In this thesis, the devices with different M1 overlaps etching stop layer (ESL) via distance, M2 £\-IGZO contact size and the fringe field effect are investigated. Although the characteristics of £\-IGZO TFTs have great performance, the electrical stability under illumination and long term bias stress are still a important issue to study before implement them into display. Thus, the devices with different structures that we mentioned previously are investigated the electrical reliability which are the negative bias stress of gate voltage, hot carrier stress effect and negative bias of illumination. The electron path of via type is extracted by contact resistance which is greater than the distance between S/D via. Experiment results show that the increased offset between M1 and ESL via generates the resistance-liked effect in electrical characteristics. The hot carrier stress effect is independent of M2 £\-IGZO contact size in short channel length devices and there are close depletion lengths in drain side. The negative bias stress of illumination is proceeded in the fringe field effect devices, which results a negative shift of threshold voltage due to the hole trapping.

In this dissertation, we investigates the electrical stress induced degradation in low-temperature polycrystalline-silicon thin film transistors (LTPS TFTs) applied for system-on-panel (SOP), including the electrical degradations of device for switch operation in active matrix flat-panel displays, driving circuit and nonvolatile memory. Finally, we also present the reliability of LTPS TFTs applied for flexible displays. In first part, electrical degradation of conventional and pattered metal-shielding LTPS TFTs under darkened and illuminated dynamic AC stresses are investigated. Experimental results reveal that competitive mechanisms are generated in conventional LTPS TFTs during illuminated stress, namely, carrier increase and electric field weakening. This phenomenon is verified by stressing the patterned source/drain open metal-shielding LTPS TFTs, which determines that the electric field weakening dominates; conversely, the carrier increase is dominated the electrical degradation in channel open metal-shielding device under illuminated stress. In addition, an improvement in anomalous on-current and subthreshold swing (S.S.) in n-channel LTPS TFTs after positive gate bias stress are studied. These improved electric properties are due to the hole trapping at SiO2 above the lightly doped drain regions, which causes a strong electric field at the gate corners. The effect of the hole trapping is to reduce the effective channel length and the S.S.. Besides, the stress-related electric field was also simulated by TCAD software to verify the mechanism above. Secondly, a mechanism of anomalous capacitance in p-channel LTPS TFTs was investigated. In general, the effective capacitance of the LTPS TFTs was only dependent with the overlap area between gate and source/drain under the off-state. However, the experimental results reveal that the off-state capacitance was increased with decreasing measurement frequency and/or with increasing measurement temperature. Besides, by fitting the curve of drain current versus electric field under off-state region, it was verified that the TAGIDL is consisted of the Pool-Frenkel emission and Thermal-Field emission. In addition, the charge density calculated from the Cch-Vg measurement also the same dependence with electric field. This result demonstrates that the anomalous capacitance is mainly due to the trap-assisted-gate-induced-drain-leakage (TAGIDL). In order to suppress the anomalous capacitance, a band-to-band hot electron (BTBHE) stress was utilized to reduce the vertical electric field between the gate and the drain. In third part, in order to realize the reliability in p-channel TFTs under illuminated environment operation, the degradation of negative bias temperature instability (NBTI) with illumination effect is investigated. The generations of interface state density (Nit) are identical under various illuminated intensity DC NBTI stresses. Nevertheless, the degradation of the grain boundary trap (Ntrap) under illumination was more significant than for the darkened environment, with degradation increasing as illumination intensity increases. This phenomenon is mainly caused by the extra number of holes generated during the illuminated NBTI stress. The increased Ntrap degradation leads to an increase in the darkened environment leakage current. This indicates that more traps are generated in the drain junction region that from carrier tunneling via the trap, resulting in leakage current. Conversely, an increase of Ntrap degradation results in a decrease in the photoleakage current. This indicates that the number of recombination centers increases in poly-Si bulk, affecting photosensitivity in LTPS TFTs. Besides, the transient effect assisted NBTI degradation in p-channel LTPS TFTs under dynamic stress is also presented, in which the degradation of the Ntrap becomes more significant as rise time decreases to 1 £gs. Because the surface inversion layer cannot form during the short rise time, transient bulk voltage will cause excess holes to diffuse into the poly-Si bulk. Therefore, the significant Ntrap increase is assisted by this transient effect. Fourthly, we study the electric properties of n- and p-channel LTPS TFTs under the mechanical tensile strain. The improved on-current for tensile strained n-channel TFTs is originated form an increase in energy difference between 2- and 4-fold valleys, reducing the inter-valley scattering and further improving the carrier mobility. On the contrary, the hole mobility decreases in p-channel, suggesting the split between the light hole and heavy hole energy bands and an increase in hole population on the heavy hole energy band of poly-Si when the uniaxial tensile strain is parallel to the channel direction. In addition, the Nit and Ntrap degradations induced by NBTI for tensile strained LTPS TFTs are more pronounced than in the unstrained. Extracted density-of-states (DOS) and conduction activation energy (EA) both show increases due to the strained Si-Si bonds, which implies that strained Si-Si bonds are able to react with dissociated H during the NBTI stress. Therefore, the NBTI degradation is more significant after tensile strain than in an unstrained condition. Finally, the SONOS-TFT applied to nonvolatile memory is prepared and studied. In the gate disturb stress, a parasitic capacitance and resistance in off-state region are identified as electrons trapped in the gate-insulator (GI) near the defined gate region. Meanwhile, these trapping electrons induced depletions in source/drain also degraded the I-V characteristic when the gate bias is larger than the threshold voltage. However, these degradations slightly recover when the trapped electrons are removed after negative bias stress. The electric field in the undefined gate region is also verified by TCAD simulation software.

Actuellement, les circuits numériques nécessitent d'être de plus en plus performants. Aussi, les produits doivent être conçus le plus rapidement possible afin de gagner les précieuses parts de marché. Les méthodes rapides de conception et l'utilisation de MPSoC ont permis de satisfaire à ces exigences, mais sans tenir compte précisément de l'impact du vieillissement des circuits sur la conception. Or les MPSoC utilisent les technologies de fabrication les plus récentes et sont de plus en plus soumis aux défaillances matérielles. De nos jours, les principaux mécanismes de défaillance observés dans les transistors des MPSoC sont le HCI et le NBTI. Des marges sont alors ajoutées pour que le circuit soit fonctionnel pendant son utilisation, en considérant le cas le plus défavorable pour chaque mécanisme. Ces marges deviennent de plus en plus importantes et diminuent les performances attendues. C'est pourquoi les futures méthodes de conception nécessitent de tenir compte des dégradations matérielles en fonction de l'utilisation du circuit. Dans cette thèse, nous proposons une méthode originale pour simuler le vieillissement des MPSoC à haut niveau d'abstraction. Cette méthode s'applique lors de la conception du système c.-à-d. entre l'étape de définition des spécifications et la mise en production. Un modèle empirique permet d'estimer les dégradations temporelles en fin de vie d'un circuit. Un exemple d'application est donné pour un processeur embarqué et les résultats pour un ensemble d'applications sont reportés. La solution proposée permet d'explorer différentes configurations d'une architecture MPSoC pour comparer le vieillissement. Aussi, l'application la plus sévère pour le vieillissement peut être identifiée.

abstract: The aging process due to Bias Temperature Instability (both NBTI and PBTI) and Channel Hot Carrier (CHC) is a key limiting factor of circuit lifetime in CMOS design. Threshold voltage shift due to BTI is a strong function of stress voltage and temperature complicating stress and recovery prediction. This poses a unique challenge for long-term aging prediction for wide range of stress patterns. Traditional approaches usually resort to an average stress waveform to simplify the lifetime prediction. They are efficient, but fail to capture circuit operation, especially under dynamic voltage scaling (DVS) or in analog/mixed signal designs where the stress waveform is much more random. This work presents a suite of modelling solutions for BTI that enable aging simulation under all possible stress conditions. Key features of this work are compact models to predict BTI aging based on Reaction-Diffusion theory when the stress voltage is varying. The results to both reaction-diffusion (RD) and trapping-detrapping (TD) mechanisms are presented to cover underlying physics. Silicon validation of these models is performed at 28nm, 45nm and 65nm technology nodes, at both device and circuit levels. Efficient simulation leveraging the BTI models under DVS and random input waveform is applied to both digital and analog representative circuits such as ring oscillators and LNA. Both physical mechanisms are combined into a unified model which improves prediction accuracy at 45nm and 65nm nodes. Critical failure condition is also illustrated based on NBTI and PBTI at 28nm. A comprehensive picture for duty cycle shift is shown. DC stress under clock gating schemes results in monotonic shift in duty cycle which an AC stress causes duty cycle to converge close to 50% value. Proposed work provides a general and comprehensive solution to aging analysis under random stress patterns under BTI. Channel hot carrier (CHC) is another dominant degradation mechanism which affects analog and mixed signal circuits (AMS) as transistor operates continuously in saturation condition. New model is proposed to account for e-e scattering in advanced technology nodes due to high gate electric field. The model is validated with 28nm and 65nm thick oxide data for different stress voltages. It demonstrates shift in worst case CHC condition to Vgs=Vds from Vgs=0.5Vds. A novel iteration based aging simulation framework for AMS designs is proposed which eliminates limitation for conventional reliability tools. This approach helps us identify a unique positive feedback mechanism termed as Bias Runaway. Bias runaway, is rapid increase of the bias voltage in AMS circuits which occurs when the feedback between the bias current and the effect of channel hot carrier turns into positive. The degradation of CHC is a gradual process but under specific circumstances, the degradation rate can be dramatically accelerated. Such a catastrophic phenomenon is highly sensitive to the initial operation condition, as well as transistor gate length. Based on 65nm silicon data, our work investigates the critical condition that triggers bias runaway, and the impact of gate length tuning. We develop new compact models as well as the simulation methodology for circuit diagnosis, and propose design solutions and the trade-offs to avoid bias runaway, which is vitally important to reliable AMS designs. / Dissertation/Thesis / Doctoral Dissertation Electrical Engineering 2014