Today, the use of embedded Dynamic Random Access Memory (eDRAM) is increasing in our electronics that require large memories, such as gaming consoles and computer network routers. Unlike external DRAMs, eDRAMs are embedded inside ASICs for faster read and write operations. Until recently, eDRAMs required high manufacturing cost. Present process technology developments enabled the manufacturing of eDRAM at competitive costs. Unlike SRAM, eDRAM exhibits retention time bit fails from defects and capacitor leakage current. This retention time fail causes memory bits to lose stored values before refresh. Also, a small portion of the memory bits are known to fail at a random retention time. At test conditions, more stringent than use conditions, if all possible retention time fail bits are detected and replaced, there will be no additional fail bits during use. However, detecting all the retention time fails requires long time and also rejects bits that do not fail at the use condition. This research seeks to maximize the detection of eDRAM fail bits during test by determining effective test conditions and model the failure rate of eDRAM retention time during use conditions.

abstract: Memory systems are becoming increasingly error-prone, and thus guaranteeing their reliability is a major challenge. In this dissertation, new techniques to improve the reliability of both 2D and 3D dynamic random access memory (DRAM) systems are presented. The proposed schemes have higher reliability than current systems but with lower power, better performance and lower hardware cost. First, a low overhead solution that improves the reliability of commodity DRAM systems with no change in the existing memory architecture is presented. Specifically, five erasure and error correction (E-ECC) schemes are proposed that provide at least Chipkill-Correct protection for x4 (Schemes 1, 2 and 3), x8 (Scheme 4) and x16 (Scheme 5) DRAM systems. All schemes have superior error correction performance due to the use of strong symbol-based codes. In addition, the use of erasure codes extends the lifetime of the 2D DRAM systems. Next, two error correction schemes are presented for 3D DRAM memory systems. The first scheme is a rate-adaptive, two-tiered error correction scheme (RATT-ECC) that provides strong reliability (10^10x) reduction in raw FIT rate) for an HBM-like 3D DRAM system that services CPU applications. The rate-adaptive feature of RATT-ECC enables permanent bank failures to be handled through sparing. It can also be used to significantly reduce the refresh power consumption without decreasing the reliability and timing performance. The second scheme is a two-tiered error correction scheme (Config-ECC) that supports different sized accesses in GPU applications with strong reliability. It addresses the mismatch between data access size and fixed sized ECC scheme by designing a product code based flexible scheme. Config-ECC is built around a core unit designed for 32B access with a simple extension to support 64B and 128B accesses. Compared to fixed 32B and 64B ECC schemes, Config-ECC reduces the failure in time (FIT) rate by 200x and 20x, respectively. It also reduces the memory energy by 17% (in the dynamic mode) and 21% (in the static mode) compared to a state-of-the-art fixed 64B ECC scheme. / Dissertation/Thesis / Doctoral Dissertation Electrical Engineering 2017

Flash memories are solid-state non-volatile memories. They play a vital role especially in information storage in a wide range of consumer electronic devices and applications including smart phones, digital cameras, laptop computers, and satellite navigators. The demand for high density flash has surged as a result of the proliferation of these consumer electronic portable gadgets and the more features they offer – wireless internet, touch screen, video capabilities. The increase in the density of flash memory devices over the years has come as a result of continuous memory cell-size reduction. This size scaling is however approaching a dead end and it is widely agreed that further reduction beyond the 20 nm technological node is going to be very difficult, as it would result to challenges such as cross-talk or cell-to-cell interference, a high statistical variation in the number of stored electrons in the floating gate and high leakage currents due to thinner tunnel oxides. Because of these challenges a wide range of solutions in form of materials and device architectures are being investigated. Among them is three-dimensional (3-D) flash, which is widely acclaimed as the ideal solution, as they promise the integration of long-time retention and ultra-high density cells without compromising device reliability. However, current high temperature (>600 °C) growth techniques of the Polycrystalline silicon floating gate material are incompatible with 3-D flash memory; with vertically stacked memory layers, which require process temperatures to be ≤ 400 °C. There already exist some low temperature techniques for producing polycrystalline silicon such as laser annealing, solid-phase crystallization of amorphous silicon and metal-induced crystallization. However, these have some short-comings which make them not suitable for use in 3-D flash memory, e.g. the high furnace annealing temperatures (700 °C) in solid-phase crystallization of amorphous silicon which could potentially damage underlying memory layers in 3-D flash, and the metal contaminants in metal-induced crystallization which is a potential source of high leakage currents. There is therefore a need for alternative low temperature techniques that would be most suitable for flash memory purposes. With reference to the above, the main objective of this research was to develop a novel low temperature method for growing silicon structures at ≤ 400 °C. This thesis thus describes the development of a low-temperature method for polycrystalline silicon growth and the application of the technique in a capacitor-like flash memory device. It has been demonstrated that silicon structures with polycrystalline silicon-like properties can be grown at ≤ 400 °C in a 13.56 MHz radio frequency (RF) plasma-enhanced chemical vapour deposition (PECVD) reactor with the aid of Nickel Formate Dihydrate (NFD). It is also shown that the NFD coated on the substrates, thermally decomposes in-situ during the deposition process forming Ni particles that act as nucleation and growth sites of polycrystalline silicon. Silicon films grown by this technique and without annealing, have exhibited optical band gaps of ~ 1.2 eV compared to 1.78 eV for films grown under identical conditions but without the substrate being coated. These values were determined from UV-Vis spectroscopy and Tauc plots. These optical band gaps correspond to polycrystalline silicon and amorphous silicon respectively, meaning that the films grown on NFD-coated substrates are polycrystalline silicon while those grown on uncoated substrates remain amorphous. Moreover, this novel technique has been used to fabricate a capacitor-like flash memory that has exhibited hysteresis width corresponding to charge storage density in the order of 1012 cm-2 with a retention time well above 20 days for a device with silicon films grown at 300 °C. Films grown on uncoated films have not exhibit any significant hysteresis, and thus no flash memory-like behaviour. Given that all process temperatures throughout the fabrication of the devices are less than 400 °C and that no annealing of any sort was done on the material and devices, this growth method is thermal budget efficient and meets the crucial process temperature requirements of 3-D flash memory. Furthermore, the technique is glass compatible, which could prove a major step towards the acquisition of flash memory-integrated systems on glass, as well as other applications requiring low temperature polycrystalline silicon.

The influence of depolarization and its role in causing data retention failure in ferroelectric memories is investigated. Ferroelectric Hf₀.₅Zr₀.₅O₂ thin films 8 nm thick incorporated into a metal-ferroelectric-metal capacitor are fabricated and characterized with varying thicknesses of an Al₂O₃ interfacial layer. The magnitude of the depolarization field is adjusted by controlling the thickness of the Al₂O₃ layer. The initial polarization and the change in polarization with electric field cycling is strongly impacted by the insertion of Al₂O₃ within the device stack. Transient polarization loss is shown to get worse with larger depolarization fields and data retention is evaluated up to 85 °C.

With the discovery of the ferroelectric (FE) properties within HfO₂, the scaling gap between state-of-the-art technology nodes and non-volatile memories based on FE materials can be bridged. In addition to non-volatility, new memory concepts should guarantee sufficient endurance and operation stability. However, in contrast to optimized perovskite based FEs, binary oxide based FE memories still show changes in the memory window (MW) followed by either hard breakdown or closure of the MW. Recently, we have shown that anti-FE (AFE) materials exhibit very stable and significantly higher endurance with respect to the FE counterparts. Inspired by the robustness and remarkable cycling performance of the AFE materials, we analyze the remaining reliability aspects of these devices. By characterizing the pure film properties of capacitor stacks and switching performance when integrated into devices, we compare and investigate temperature stability, imprint, retention, and variability of both FE and AFE memories. We investigate if the lower energetic barrier to be overcome together with partial switching and lower switching induced stress are responsible for the higher endurance of the AFE with respect to the FE based memories. By utilizing charge trapping and charge pumping tests together with leakage current spectroscopy in combination with comprehensive modeling we check that assumption. Moreover, we identify the interfacial buffer layer as the weakest link of these devices.

La portabilité des mémoires Flash embarquées sur les nœuds CMOS technologiques avancés tel que le 28nm pose de nombreux problèmes de compatibilité avec les nouvelles étapes de fabrication telles que le diélectrique de grille haute permittivité, l’utilisation de grille métallique, les stresseurs et tenseurs utilisés pour piloter la performance du transistor élémentaire. L’ajout d’un dispositif à double grille classique tel que celui de la Flash apparait comme très couteux en termes de nombre de masques et d’étapes de fabrication additionnelles. De nombreuses alternatives ont vu le jour : les mémoires à changement de phase, les mémoires magnétiques et les mémoires resistives. Ce dernier type de mémoire est particulièrement attrayant pour une intégration en tant que mémoire « embarquée » sur technologie CMOS. Les matériaux utilisés (diélectrique à base d’oxyde métallique tel que le HfO₂ ou le Ta₂O₅) sont compatibles avec le procédé de fabrication CMOS comparés à ceux utilisés pour les mémoires magnétiques (risques de contamination). Les mémoires résistives sont par ailleurs basées sur une conduction filamentaire qui s’avère également particulièrement économe en énergie et adaptée aux faibles géométries quand elles sont comparées aux mémoires à changement de phase (changement d’état volumique du matériau). De nombreux industriels ont focalisé leurs efforts sur les matériaux de type HfO₂ et Ta₂O₅. Le sujet proposé fait suite à trois années de collaboration intensive entre ST Microelectronics et le CEA-LETI qui ont permis d’établir les bases d’un cellule mémoire de type Oxram fonctionnelle et facilement intégrable facilement sur une technologie CMOS. Il aura pour objectifs d’analyser les paramètres responsables des instabilités des états résistifs observés et de rechercher les différents moyens susceptibles de mieux contrôler la dispersion de ces états. Les études réalisées pourront porter sur les matériaux (diélectrique et électrodes), la technologie mise en œuvre, les conditions électriques de formation du filament [20]. La consolidation du choix du matériau et l’analyse des modes de défaillance et de la fiabilité du plan mémoire feront également partie du travail de cette première année. Ce travail sera orienté par les résultats statistiques obtenus par le biais de test à plus grande échelle (circuit de plusieurs Kbits). / Embedded Flash memories integration on advanced CMOS technological nodes such as the 28nm leads to serious compatibility problems with the new manufacturing steps such as the high-permittivity gate dielectric, the use of metal gate, etc. The addition of a conventional double-grid device such as the one for Flash appears to be very expensive in terms of number of masks and additional manufacturing steps. Many alternatives have emerged: phase change memories PCRAM, magnetic memories MRAM and resistive memories OxRAM. However, the high programming current of the PCRAM memories and the risks associated to the contamination of the materials used for the MRAM memories represent the weak points of these technologies. On the other hand, OxRAM memories are particularly attractive for integration as CMOS embedded memory. The materials used (metal oxide dielectric such as HfO₂ or Ta₂O₅) compatible with the CMOS manufacturing process and their low programming voltages due to filament conduction are an advantage for OxRAM memories.In this thesis, an in depth memory stack optimization is done to make up the OxRAM memory cell in order to be integrated into a matrix of memories. Thus, various top and bottom electrodes and various switching oxides have been studied in order to better control and improve the variability of the resistive states of the OxRAM memory cell. An evaluation of the reliability and the main memory performances in terms of Forming voltage, memory window, endurance and thermal stability were performed for each memory stack through electrical characterizations. These assessments highlighted efficient memory stacks which have been integrated into a 16Kb demonstrator. Finally, a study of the variability of the resistive states as well as their degradation mechanisms during the endurance and thermal stability were carried out through simples models and atomistic simulations (ab-initio calculations).