Over the past several decades, memory technologies have exploited continual scaling of CMOS to drastically improve performance and cost. Unfortunately, charge-based memories become unreliable beyond 20 nm feature sizes. A promising alternative is Phase-Change-Memory (PCM) which leverages scalable resistive thermal mechanisms. To realize PCM's potential, a number of challenges, including the limited wear-endurance and costly writes, need to be addressed. This thesis introduces novel methodologies for encoding data on PCM which exploit asymmetries in read/write performance to minimize memory's wear/energy consumption. First, we map the problem to a distance-based graph clustering problem and prove it is NP-hard. Next, we propose two different approaches: an optimal solution based on Integer-Linear-Programming, and an approximately-optimal solution based on Dynamic-Programming. Our methods target both single-level and multi-level cell PCM and provide further optimizations for stochastically-distributed data. We devise a low overhead hardware architecture for the encoder. Evaluations demonstrate significant performance gains of our framework.

In addition, this project also investigated schemes that might enhance the phase change in the SPR sensor. The "double-pass" and "multi-pass" approaches through which the SPR phase can be amplified upon hitting the sensor surface more than once, have been experimentally studied and successfully demonstrated. A double-pass method can immediately offer two times of phase change as compared to the singlepass one. Accordingly the multi-pass scheme offers a higher then two times phase enhancement. Such improvement in phase detection is extremely important for biosensing applications involving small molecules, small proteins, DNA and etc. Another approach for detection performance improvement is to incorporate a multilayer configuration for the biosensing surface. In order to improve the dynamic measurement response, we proposed to use a multiple resonant angle measurement approach in conjunction with the single-beam self-referenced phase-sensitive SPR configuration. With the use of many multiple incident angles, the system provided sensing capability that covers a refractive index (RI) 1.33 to over 1.38. A 128-element array detector was employed to measure the resonance phase change over the range of the incident angles to ensure a reasonably continuous phase response curves achievable from the system. / This project is concerned with the development and optimization of optical sensors based on measuring the phase change of surface plasmon resonance (SPR) effect. The phase sensitive SPR technique provides very high sensitivity performance due to the fact that an abrupt phase jump occurs near the resonance dip, thus resulting in large phase shift with very small change in the sensing medium. A range of different measurement techniques for enhancing system sensitivity have been investigated. Moreover we also studied the phase change characteristics around the SPR dip region by means of simulation in order to explore various approaches for achieving further improvement in sensitivity and as well as wide dynamic range. Since SPR is caused by electron charge density oscillations in metal surface in which the wave momentum required for plasmon wave excitation is always larger than that for free space, an inverted prism-coupling scheme (prism-metal-dielectric) is commonly used and this configuration was also employed in our experimental setup, particularly for the SPR biosensor based on differential phase Mach-Zehnder interferometer configuration. This design primarily operates by taking advantage of the fact that SPR only affects the p-polarization while leaving the s-polarization unchanged. This means that differential phase measurement between the p- and s- polarizations will result in SPR signals that are completely free from any disturbances that are common to both channels. Experimental results obtained from glycerin/water mixtures indicate that the sensitivity limit of our scheme is 5.48 x 10 -8 refractive index unit per 0.01° phase change. To our knowledge, this is a significant improvement over previously obtained results when gold is used as the sensor surface. While acknowledging that accurate optical alignment is a crucial requirement for the Mach-Zehnder interferometer and it is often not easy to maintain high degree alignment accuracies in practical situations, we have developed a versatile and low cost single-beam self-referenced phase-sensitive surface SPR sensing system. The system exhibits a root-mean-square phase fluctuation of +/-0.0028° over a period of 45 minutes, i.e. a resolution of +/-5.2x10 -9 refractive index units. The enhanced performance has been achieved through the incorporation of three design elements: (i) a true single-beam configuration enabling complete self-referencing so that only the phase change associated with SPR gets detected; (ii) a differential measurement scheme to eliminate spurious signals not related to the sensor response; (iii) elimination of retardation drifts by incorporating temperature stabilization in the liquid crystal phase modulator. Our design should bring the detection sensitivity of non-labeling SPR biosensing closer to that achievable by conventional florescence-based techniques. / Wu, Shu Yuen. / Source: Dissertation Abstracts International, Volume: 73-06, Section: B, page: . / Thesis (Ph.D.)--Chinese University of Hong Kong, 2011. / Includes bibliographical references (leaves 132-147). / Electronic reproduction. Hong Kong : Chinese University of Hong Kong, [2012] System requirements: Adobe Acrobat Reader. Available via World Wide Web. / Electronic reproduction. [Ann Arbor, MI] : ProQuest Information and Learning, [201-] System requirements: Adobe Acrobat Reader. Available via World Wide Web. / Abstract also in Chinese.

The main objective of this research is to provide an efficient and reliable method for using multi-level cell (MLC) phase-change memory (PCM) as a main memory. As DRAM scaling approaches the physical limit, alternative memory technologies are being explored for future computing systems. Among them, PCM is the most mature with announced commercial products for NOR flash replacement. Its fast access latency and scalability have led researchers to investigate PCM as a feasible candidate for DRAM replacement. Moreover, the multi-level potential of PCM cells can enhance the scalability by increasing the number of bits stored in a cell. However, the two major challenges for adopting MLC PCM are the limited write endurance cycle and the resistance drift issue. To alleviate the negative impact of the limited write endurance cycle, this thesis first introduces a secure wear-leveling scheme called Security Refresh. In the study, this thesis argues that a PCM design not only has to consider normal wear-out under normal application behavior, most importantly, it must take the worst-case scenario into account with the presence of malicious exploits and a compromised OS to address the durability and security issues simultaneously. Security Refresh can avoid information leak by constantly migrating their physical locations inside the PCM, obfuscating the actual data placement from users and system software. In addition to the secure wear-leveling scheme, this thesis also proposes SAFER, a hardware-efficient multi-bit stuck-at-fault error recovery scheme which can function in conjunction with existing wear-leveling techniques. The limited write endurance leads to wear-out related permanent failures, and furthermore, technology scaling increases the variation in cell lifetime resulting in early failures of many cells. SAFER exploits the key attribute that a failed cell with a stuck-at value is still readable, making it possible to continue to use the failed cell to store data; thereby reducing the hardware overhead for error recovery. Another approach that this thesis proposes to address the lower write endurance is a hybrid phase-change memory architecture that can dynamically classify, detect, and isolate frequent writes from accessing the phase-change memory. This proposed architecture employs a small SRAM-based Isolation Cache with a detection mechanism based on a multi-dimensional Bloom filter and a binary classifier. The techniques are orthogonal to and can be combined with other wear-out management schemes to obtain a synergistic result. Lastly, this thesis quantitatively studies the current art for MLC PCM in dealing with the resistance drift problem and shows that the previous techniques such as scrubbing or error correction schemes are incapable of providing sufficient level of reliability. Then, this thesis proposes tri-level-cell (3LC) PCM and demonstrates that 3LC PCM can be a viable solution to achieve the soft error rate of DRAM and the performance of single-level-cell PCM.

No description available.

GeTe and Ge2Sb2Te5 alloys are potential candidates for non-volatile phase change random access memories (PCRAM). For electrical data storage applications the materials should have stable amorphous and crystalline phases, fast crystallization time, low power to switch, and high crystallization activation energy (to be stable at normal operating temperatures). Phase change memories can be tuned through compositional variations to achieve sufficient phase change contrast and thermal stability for data retention. Selenium is one of the attractive choices to use as an additive material owing to its flexible amorphous structure and a variety of possible applications in optoelectronics and solar cells. GeSb2Te3Se alloy, in which 25 at.% of Se substituted for Te, show a higher room temperature resistance with respect to parent GeSb2Te4 alloy, but the transition temperature is lowered which will affect the thermal stability. The RESET current observed for Sb65Se35 alloys were reduced and the crystallization speed increased 25 % faster with respect to Ge2Sb2Te5. Alloys of Ga-Sb-Se possess advantages such as higher crystallization temperatures, better data retention, higher switching speed, lower thermal conductivity and lower melting point than the GST, but the resistance ratio is limited to about two orders of magnitude. This affects the resistance contrast and data readability. It is with this background a study has been carried out in GeTe and GeSbTe system with Se doping. Studies on structural, thermal and optical properties of these materials all through the phase transition temperatures would be helpful to explore the feasibility of phase change memory uses. Thin films along with their bulk counterparts such as (GeTe)1-x Sex ( 0 < x ≤ 0.50) and (GST)1-xSex (0 < x ≤ 0.50), including GeTe and GST alloys, have been prepared. The results are presented in four chapters apart from the Introduction and Experimental techniques chapters. The final chapter summarizes the results. Chapter 1 provides an introduction to chalcogenide glasses, phase change memory materials and their applications. The fundamental properties of amorphous solids, basic phase change properties of Ge2Sb2Te5 and GeTe alloys and their applications are presented in detail. Various doping studies on GeTe and Ge2Sb2Te5 reported in literatures are reviewed. The limitations, challenges, future and scope of the present work are presented. In chapter 2, the experimental techniques used for thin film preparation, electrical characterizations, optical characterization and surface characterizations etc. are explained. Chapter 3 deals entirely on Ge2Sb2Te5 films studied throughout the phase transition, by annealing at different temperatures. Changes in sheet resistance, optical transmission, morphology and surface bonding characteristics are analyzed. The crystallization leads to an increase of roughness and the resistance changes to three orders of magnitude at 125 oC. Optical studies show distinct changes in transmittance during phase transitions and the optical parameters are calculated. Band gap contrast and disorder variation with annealing temperatures are explained. The surface bonding characteristics studied by XPS show Ge-Te, Sb-Te bonds are present in both amorphous and crystalline phases. The temperature dependent modifications of the band structure of amorphous GST films at low temperatures have been little explored. The band gap increment of around 0.2 eV is observed at low temperature (4.2 K) compared to room temperature 300 K. Other optical parameters like Urbach energy and B1/2 are studied at different temperatures and are evaluated. The observed changes in optical band gap (Eopt) are fitted to Fan’s one phonon approximation, from which a phonon energy (ћω) corresponding to a frequency of 3.59 THz resulted. The frequency of 3.66 THz optical phonons has already been reported by coherent phonon spectroscopy experiment in amorphous GST. This opens up an indirect method of calculating the phonon frequency of the amorphous phase change materials. Chapter 4 constitutes comparison of optical, electrical and structural investigation of GST and (GST)1-xSex films. It is well known that GST alloys have vacancy in their structure, which leads to the possibility of switching between the amorphous and crystalline states with minimum damage. Added Se may occupy the vacancy or change the bonding characteristics which intern may manifest in the possibility of change in optical and electrical parameters. The structural studies show a direct amorphous to hexagonal transition in (GST)1-xSex, where x ≥ 0.10 at.%. Raman spectra of the as deposited and annealed (GST)1-xSex films show structural modifications. The infrared transmission spectra indicate a shift in absorption edges from low to high photon energy when Se concentration increases in GST. Band gap values calculated from Tauc plot show the band gap increment with Se doping. It is noted that a small amount of Se doping increases the resistance of the amorphous and crystalline phases and maintains the same orders of resistance contrast. This will be beneficial as it improves the thermal stability and reduces the write current in a device. Switching studies show an increasing threshold voltage as the Se doping concentration increases. Chapter 5 comprises compositional dependent investigations of the bulk GeTe chalcogenides alloys added with different selenium concentrations. The XRD investigations on bulk (GeTe)1-xSex (x = 0.0, 0.02, 0.10, 0.20 and 0.50 at.%) alloys show that the crystalline structure of GeTe alloys does not affect ≤ 0.20 at.% of Se concentration. With increasing amount of Se concentration the alloys gets modified in to a homogeneous amorphous structure. This result has been verified from the XRD, Raman, XPS, SEM and DSC measurements. The possibility that Se occupying the Ge vacancy sites in GeTe structure is explained. Since Se is an easy glass former, the amorphousness increases in the alloys due to new amorphous phases formed by the Se with other elements. It is shown from Raman and XPS analysis that the Ge-Te bonds exists up to Se 0.20 at.% alloys. Ge-Se and GeTe2 bonds are increasing with increasing Se at.%. Melting temperature has found decreases and the reduction in melting point may reduces the RESET current. Further studies on switching behavior may bring out its usefulness. Chapter 6 deals with studies on (GeTe)1-xSex films for phase change memory applications based on the insight received from their bulk study. Even at low at.% addition of Se makes the as prepared (GeTe)1-xSex film amorphous. At 200 oC, GeTe crystalline structure is evolved and the intensity of the peaks reduces in the alloys with increase of Se content. At 300 oC, more evolved GeTe crystalline structure is seen compared to 200 oC annealed films whereas 0.20 at.% Se alloy remain amorphous. Resistance and thermal studies shows increase in crystallization temperature. It is expected that Se sits in the vacancies of the GeTe crystalline structural formation. This may also account for the increased threshold voltages with increasing Se doping. The band gap increase with increase of Se at.% signifying the possibility of band gap tuning in the material. Possible explanation for the increased order in GeTe due to Se doping is presented. The modifications in the alloy with Se addition can be explained with the help of chemical bond energy approach. Those bonds having higher energy leads to increased average bond energy of the system and hence the band gap. The XPS core level spectra and Raman spectra investigation clearly shows the GeTe bonds are replaced by Ge-Se bonds and GeTe2 bonds. The 0.10 at.% Se alloy is found to have a higher thermal stability in the amorphous state and maintains a gigantic resistance contrast compared to other Se concentration alloys. This alloy can be considered as an ideal candidate for multilevel PCM applications. Chapter 7 summarizes the major findings from this work and the scope for future work.

The Science behind amorphous Chalcogenide materials opened up new technologies in the arena of Phase Change Memories. The Ovonic universal phase change memory is called universal because it can replace flash memory, DRAM and SRAM. These are not only basic computer memory devices but also are becoming the driving force for the ongoing revolutionary growth of cell phones and other mobile devices, which are in desperate need of memory providing higher density, faster speed and lower power consumption. In this thesis, compositional dependence of various properties of different chalcogenide glasses are investigated, to explore the possibility of their application in Phase Change Memories. Efforts are also made to understand the effect of rigidity and extended rigidity transition on the composition dependence of properties investigated. This thesis comprises of 9 chapters; a brief summary is given below. Chapter 1 deals with fundamental aspects of amorphous semiconductors with a particular reference to chalcogenide glasses. The advantages and applications of chalcogenide glasses are also described. Chapter 2 outlines preparation and characterization of the glasses investigated. The sample preparation and various experimental setup used in the present thesis work like Raman Scattering, Nanoindentation, Alternating Differential Scanning Calorimetry (ADSC), Photo-thermal Deflection Spectroscopy (PDS), Electrical Switching are summarized here. Chapter 3 deals with Micro-Raman studies in Ge15Te85-x Inx Glasses. Micro-Raman studies reveal that as-quenched Ge15Te85-xInx samples exhibit two prominent peaks, at 123 and 155 cm-1. In thermally annealed samples, the peaks at 120 cm-1 and 140 cm-1, which are due to crystalline Te, emerge as the strongest peaks. The Raman spectra of polished samples are similar to those of annealed samples, with strong peaks at 123 cm-1 and 141 cm-1. The spectra of lightly polished samples outside the thermally reversing window resemble those of thermally annealed samples; however, the spectra of glasses with compositions in the thermally reversing window resemble those of as-quenched samples. This observation confirms the earlier idea that compositions in the thermally reversing window are non-ageing and are more stable. Chapter 4 explains nanoindentation studies undertaken on Ge15Te85-xInx glasse (1 ≤ x ≤ 11). Nanoindentation studies on Ge15Te85-xInx glasses indicate that the hardness and elastic modulus of these glasses increase with indium concentration. While a pronounced plateau is seen in the elastic modulus in the composition range 3 ≤ x ≤ 7, the hardness exhibits a change in slope at compositions x = 3 and x = 7. Also, the density exhibits a broad maximum in this composition range. The observed changes in the mechanical properties and density are clearly associated with the thermally reversing window in Ge15Te85-xInx glasses in the composition range 3 ≤ x ≤ 7. In addition, a local minimum is seen in density and hardness around x = 9, the chemical threshold of the system. Chapter 5 deals with crystallization kinetics of Ge15Te85-xInx glasses. The crystallization kinetics of Ge15Te85Inx glasses have been studied by non-isothermal method. The composition dependence of Tg and Tc at different heating rates, is investigated. The activation energy of crystallization is calculated using the Kissinger’s plot. It is found that the composition dependence of the glass transition temperature, Tg and the crystallization temperature, Tc, the activation energy of crystallization, Ec, and the stability factor, (ΔT= Tc-Tg) exhibit specific signatures of intermediate phase in the composition rang 3 ≤ x ≤ 7 and Chemical Threshold at x = 9. Chapter 6 explains Alternating Differential Scanning Calorimetric and XRD studies on silver doped Ge15Te80In5 glasses. X-ray diffraction studies on quaternary Ge15Te80-xIn5Agx glasses (2 ≤ x ≤ 24) reveal the presence of Te, GeTe, Ag8GeTe6, AgTe, In2Te3 and In4Te3. Thermal studies on quaternary Ge15Te80-xIn5Agx glasses exhibit signatures of Intermediate Phase (IP) in the variation of Tg, ∆HNR and ∆Cp with Ag addition in the composition range 8 ≤ x ≤ 16. The composition x = 16 has been identified to be the Chemical Threshold (CT) based on the saturation of flexible Ag-Te bonds. Micro-Raman, molar volume, thermal diffusivity studies on Ge15Te80-xIn5Agx glasses reveal a clear evidence of intermediate phase in the composition range 8 ≤ x ≤ 16 as depicted in the ADSC studies. Chapter 7 deals with Micro-Raman studies on as-quenched Ge15Te80-xIn5Agx glasses reveal the presence of tetrahedral structural units. Further, the Raman peak positions are found to shift with silver addition. In addition, specific signatures of the Intermediate Phase (IP) are observed in the composition dependence of Raman frequencies and corresponding intensities of different modes in the composition range, 8 ≤ x ≤ 16. In thermally annealed samples, the observed Raman peaks can be attributed to crystalline tellurium and silver lattice vibrational modes; significant increase in intensity is observed at 93 and 141cm-1 with silver addition in annealed samples, suggesting an increase in silver lattice vibrational modes. Also, the compositional dependence of density, molar volume and thermal diffusivity confirms the presence of the intermediate phase. Chapter 8 contains the current-voltage characteristics and electrical switching behavior of Ge15Te80-xIn5Agx glasses. The glasses are found to exhibit memory type switching for 3mA current in the voltage range 70 -120 V, for a sample thickness 0.3 mm. But when the current is lowered to 1mA the samples exhibit threshold switching. The compositional studies indicate the presence of an intermediate phase in the composition range 8 ≤ x ≤ 16. SET-RESET studies have been carried out using a triangular pulse of 6 mA amplitude for SET and 21 mA amplitude for RESET for a sample thickness 0.3 mm. Raman studies on SET and RESET indicates SET state resemble annealed samples and RESET state resemble as-quenched samples. It is interesting to note that the samples in the intermediate phase, especially compositions at x =10, 12, 14 withstand more set-reset cycles. This indicates compositions in the intermediate phase are suitable for PCM devices. Chapter 9 summarizes the significant results obtained and explains the scope of this thesis.

Le cerveau humain est composé d’un grand nombre de réseaux neuraux interconnectés, dont les neurones et les synapses en sont les briques constitutives. Caractérisé par une faible consommation de puissance, de quelques Watts seulement, le cerveau humain est capable d’accomplir des tâches qui sont inaccessibles aux systèmes de calcul actuels, basés sur une architecture de type Von Neumann. La conception de systèmes neuromorphiques vise à réaliser une nouvelle génération de systèmes de calcul qui ne soit pas de type Von Neumann. L’utilisation de mémoire non-volatile innovantes en tant que synapses artificielles, pour application aux systèmes neuromorphiques, est donc étudiée dans cette thèse. Deux types de technologies de mémoires sont examinés : les mémoires à changement de phase (Phase-Change Memory, PCM) et les mémoires résistives à base d’oxyde (Oxide-based resistive Random Access Memory, OxRAM). L’utilisation des dispositifs PCM en tant que synapses de type binaire et probabiliste est étudiée pour l’extraction de motifs visuels complexes, en évaluant l’impact des conditions de programmation sur la consommation de puissance au niveau du système. Une nouvelle stratégie de programmation, qui permet de réduire l’impact du problème de la dérive de la résistance des dispositifs PCM est ensuite proposée. Il est démontré qu’en utilisant des dispositifs de tailles réduites, il est possible de diminuer la consommation énergétique du système. La variabilité des dispositifs OxRAM est ensuite évaluée expérimentalement par caractérisation électrique, en utilisant des méthodes statistiques, à la fois sur des dispositifs isolés et dans une matrice complète de mémoire. Un modèle qui permets de reproduire la variabilité depuis le niveau faiblement résistif jusqu’au niveau hautement résistif est ainsi développé. Une architecture de réseau de neurones de type convolutionnel est ensuite proposée sur la base de ces travaux éxperimentaux. La tolérance du circuit neuromorphique à la variabilité des OxRAM est enfin démontrée grâce à des tâches de reconnaissance de motifs visuels complexes, comme par exemple des caractères manuscrits ou des panneaux de signalisations routières. / The human brain is made of a large number of interconnected neural networks which are composed of neurons and synapses. With a low power consumption of only few Watts, the human brain is able to perform computational tasks that are out of reach for today’s computers, which are based on the Von Neumann architecture. Neuromorphic hardware design, taking inspiration from the human brain, aims to implement the next generation, non-Von Neumann computing systems. In this thesis, emerging non-volatile memory devices, specifically Phase-Change Memory (PCM) and Oxide-based resistive memory (OxRAM) devices, are studied as artificial synapses in neuromorphic systems. The use of PCM devices as binary probabilistic synapses is studied for complex visual pattern extraction applications, evaluating the impact of the PCM programming conditions on the system-level power consumption.A programming strategy is proposed to mitigate the impact of PCM resistance drift. It is shown that, using scaled devices, it is possible to reduce the synaptic power consumption. The OxRAM resistance variability is evaluated experimentally through electrical characterization, gathering statistics on both single memory cells and at array level. A model that allows to reproduce OxRAM variability from low to high resistance state is developed. An OxRAM-based convolutional neural network architecture is then proposed on the basis of this experimental work. By implementing the computation of convolution directly in memory, the Von Neumann bottleneck is avoided. Robustness to OxRAM variability is demonstrated with complex visual pattern recognition tasks such as handwritten characters and traffic signs recognition.

Chalcogenide based Phase Change Materials are currently of great technological interest in the growing field of optoelectronics. Ge₂Sb₂Te₅ (GST) is the most widely studied phase change material, and it has been commercially used in both optical and electronic data storage applications, due to its ability to switch between two different atomic configurations, at high speed and with low power consumption, as well as its high optical and electrical contrast between amorphous and crystalline states. Despite its well-known optical and electrical properties, the operation in combination of optical and electrical domains has not yet been fully investigated. This work studies the operation of GST nano-devices exposed to a combination of optical and electrical stimuli or mixed mode by asking, is it possible to electrically measure an optically induced phase change, or vice versa? If so, how do the optical and electrical responses relate to each other, and is it possible to operate GST with a combination of optical and electrical signals? What are the technical constraints that need to be considered in order to fabricate GST devices that could be operated either optically or electrically? In order to answer these questions, experiments that characterized the optical and electrical responses of GST based nano-devices were performed. It was found that different crystallization mechanisms may have influence in the response, and that the thermal and optical design characteristics of the device play a key role in its operation. Finally a proof of principle, of an opto-electonic memory device that can be read electrically, reset optically and write electrically, is presented. This opens up possibilities for the development of new opto-eloectronic applications such as non-volatile interfaces between future photonics and electronics, high speed optical communication detectors, high speed cameras, artificial retinas and many more.

Les Mémoires ont de plus en plus importance à l'époque actuelle, et sont fondamentales pour la définition de tous les systèmes électroniques avec lesquels nous entrons en contact dans notre vie quotidienne. Les mémoires non-volatiles (NVM), représentées par la technologie Flash, ont pu suivre jusqu'à présent l'effort à la miniaturisation pour satisfaire la demande croissante de densité de mémoire exigée par le marché. Cependant, la réduction de la taille du dispositif de mémoire est de plus en plus difficile et la complexité technologique demandé a augmenté le coût par octet. Dans ce contexte, les technologies de mémoire innovantes deviennent non seulement une alternative, mais la seule solution possible pour fournir une densité plus élevée à moindre coût, une meilleure fonctionnalité et une faible consommation d'énergie. Les Mémoires à Changement de Phase (PCM) sont considérées comme la solution de pointe pour la future génération de mémoires non-volatiles, grâce à leur non-volatilité , scalabilité, "bit-alterability", grande vitesse de lecture et d'écriture, et cyclabilité élevée. Néanmoins, certains problèmes de fiabilité restent à surmonter afin de rendre cette technologie un remplacement valable de la technologie Flash dans toutes les applications. Plus en détail, la conservation des données à haute température, est l'une des principales exigences des applications embarquées industrielles et automobiles. Cette thèse se concentre sur l'étude des mémoires à changement de phase pour des applications embarquées, dans le but d'optimiser le dispositif de mémoire et enfin de proposer des solutions pour surmonter les principaux obstacles de cette technologie, en abordant notamment les applications automobiles. Nous avons conçu, fabriqué et testé des dispositifs PCM basés sur des structures reconnues et innovantes, en analysant leurs avantages et inconvénients, et en évaluant l'impact de la réduction de la taille. Notre analyse de fiabilité a conduit au développement d'un système de caractérisation dédié à caractériser nos cellules PCM avec des impulsions de l'ordre de la nanoseconde, et à la mise en oeuvre d'un outil de simulation basé sur un solveur thermoélectrique et sur l'approche numérique "Level Set", pour comprendre les différentes mécanismes qui ont lieu dans nos cellules pendant les opérations de programmation. Afin de répondre aux spécifications du marché des mémoires non-volatiles embarquées, nous avons conçu le matériau à changement de phase intégré dans le dispositif PCM avec deux principales approches: la variation de la stoechiométrie et l'ajout de dopants. Nous avons démontré et expliqué comment la rétention des données dans les dispositifs PCM à base de GeTe peut être améliorée avec l'augmentation de la concentration de Te, et comment les inclusions de SiO2 peuvent réduire les défauts causés par la tension de lecture à températures de fonctionnement élevées. En outre, nous avons présenté les avantages sur la réduction de la puissance de programmation du dopage de carbone dans les dispositifs à base de GST. Enfin, nous avons étudié les effets de l'enrichissement en Ge dans le GST, combiné avec le dopage N et C, intégré dans des cellules PCM à l'état de l'art. Grâce à l'introduction d'une nouvelle technique de programmation, nous avons démontré la possibilité d'augmenter la vitesse de programmation de ces dispositifs, caractérisés par des performances de rétention des données parmi les meilleurs rapportés dans la littérature, et de réduire le phénomène de la dérive de la résistance qui affecte la stabilité de l'état programmé des cellules PCM. Nous avons donc prouvé, avec ces derniers résultats, la validité de la technologie PCM pour les applications embarquées. / Memories are getting an exponential importance in our present era, and are fundamental in the definition of all the electronic systems with which we interact in our daily life. Non-volatile memory technology (NVM), represented by Flash technology, have been able to follow till now the miniaturization trend to fulfill the increasing memory density demanded by the market. However, the scaling is becoming increasingly difficult, rising their cost per byte due to the incoming technological complexity. In this context, innovative memory technologies are becoming not just an alternative, but the only possible solution to provide higher density at lower cost, better functionality and low power consumption. Phase-Change Memory (PCM) technology is considered the leading solution for the next NVM generation, combining non-volatility, scalability, bit-alterability, high write speed and read bandwidth and high cycle life endurance. However, some reliability issues remain to overcome, in order to be a valid Flash replacement in all the possible applications. In particular, retention of data at high temperature, is one of the main requirements of industrial and automotive embedded applications. This work focuses on the study of embedded Phase-Change Memories, in order to optimize the memory device and finally propose some solutions to overcome the main bottlenecks of this technology, in particular addressing automotive applications. We designed, fabricated, and tested PCM devices based on recognized and innovative structures, analyzing their advantages and disadvantages, and evaluating the scaling impact. Our reliability analysis led to the development of a characterization setup dedicated to characterize our PCM cells with pulses in the order of nanoseconds, and to the implementation of a simulation tool based on a thermoelectrical solver and on the Level Set numerical approach, to understand the different mechanisms taking place in our cells during the programming operations. In order to fulfill embedded NVM requirements, we engineered the phase-change material integrated in the PCM device with two main approaches: the stoichiometry variation and the dopants addition. We showed and explained how the data retention in GeTe based PCM devices can be enhanced increasing Te content, and how SiO2 inclusions can reduce the read voltage disturbs at high operating temperatures. Moreover, we reported the advantages on the programming power reduction of carbon doping in GST based devices. Finally, we studied the effects of Ge enrichment in GST, combined with N or C doping, integrated in state of the art PCM cells. Through the introduction of a new programming technique, we demonstrated the possibility to improve the programming speed of these devices, characterized by data retention performance among the best reported in the literature, and to reduce the drift phenomenon that affects the resistance state stability of PCM technology. We then proved, with these last results, the suitability of PCM for embedded applications.

Ces travaux de thèse portent sur la caractérisation thermique à l’échelle micrométrique d’un alliage à base de tellure lorsque ce matériau se trouve à l’état fondu, à haute température. À cette fin, une cellule innovante d’emprisonnement du matériau fondu a été conçue, et mise en place. Des structures de tellure au volume du microlitre ont été déposées sur un substrat de silicium et recouverts par la suite d’une couche de protection capable de les emprisonner dans une matrice : silice amorphe et alumine amorphe. La technique de la Radiométrie Photothermique Modulée a été utilisée pour étudier les propriétés thermiques de ce type de cellules et de ces constituants. La résistance thermique de dépôt a été ainsi estimée en utilisant un modèle d’étude des transferts de la chaleur utilisant le formalisme des impédances thermiques. Ceci nous a permit dans le cas de l’alumine amorphe de déterminer sa conductivité thermique et la résistance thermique de contact avec le substrat jusqu’à 600°C. Un long processus de conception, de mesure et d’analyse a été nécessaire afin d’obtenir une cellule capable de résister aux contraintes des hautes températures. À l’heure actuelle seule la caractérisation thermique jusqu’à 300°C a été possible à cause de l’instabilité mécanique de ce dépôt hétérogène. Ceci a été confirmé par des caractérisations physico-chimiques par techniques XRR, XRD et SEM. / This thesis is devoted to the thermal characterization of molten materials, namely chalcogenide glass-type tellurium alloys, at the micrometer scale. An experimental setup of Photothermal Radiometry (PTR), formerly developed for solid state measurements, has been adapted for this purpose. Using MOCVD technique, a random lattice of sub-micrometric tellurium alloy structures is grown on a thermally oxidized silicon substrate. These structures are then embedded in a protective layer (silica or alumina) to prevent evaporation during melting. Measurements are then performed from room temperature up to 650°C. SEM and XRD measurements performed after annealing show that these samples withstand thermal stress only up to 300°C. The coating’s thermal boundary resistance is estimated by a heat transfer model based on the thermal impedance formalism. Moreover, the thermal conductivity and thermal boundary resistance of thin amorphous alumina by low temperature ALD are measured from the room temperature to 600°C.