Nonvolatile memories

• Autor(a) principal: Rocha, Leandro S.R. [UNESP]
• Data de Publicação: 2018
• Outros Autores: Simões, Alexandre Z. [UNESP], Cortés, Johan A. [UNESP], Ramirez, Miguel A. [UNESP], Moura, Francisco, Foschini, Cesar R. [UNESP], Tararam, Ronald, Cilense, Mario [UNESP], Longo, Elson, Paiva-Santos, Carlos O. [UNESP]
• Tipo de documento: Capítulo de livro
• Idioma: eng
• Título da fonte: Repositório Institucional da UNESP
• Texto Completo: http://dx.doi.org/10.1016/B978-0-12-811180-2.00014-1; http://hdl.handle.net/11449/201685
• topic: CaCu3Ti4O12 (CCTO); Data storage; Nonvolatile memory; Random access memory; Semiconductor

The need for faster, smaller, cheaper and energy-efficient electronic devices has been growing continuously in the last decade, with the conventional data storage technologies (i.e., static random access memory and dynamic random access memory), which have been so far fulfilled by CMOS-charge storage-based circuits, approaching their fundamental limits, due to the lesser progress of technology in comparison to logic. To overcome this challenge, increasingly high storage density memories has become one of the crucial approaches, aiming to improve storage capacity and reading/writing speed. A semiconductor memory is an indispensable component of all modern electronic devices, with all recognizable computing platforms, from hand-held devices to large supercomputer storage systems being used for storing data, temporarily or permanently. The ability of a material to store information is defined as a solid-state memory effect, which requires at least two switchable memory states that can be addressed by an externally controlled parameter. Based on storing data volatility, memories are basically classified into two categories, volatile and nonvolatile, with the former immediately losing the stored data after turning off the power, whereas the latter being capable of retaining the stored data for a longer period, even after the power is turned off. To optimize the performance-cost trade-off, hierarchical systems made from devices with varying speed, density, and cost have been adopted and the novel nonvolatile memory (NVM) concepts, such as ferroelectric random access memory, phase-change RAM, magnetic RAM, spin-transfer-torque RAM (STT-RAM), and resistive RAM (RRAM) are fulfilling the changing market trend requirements, from electronics to high performance computing, due to encouraging recent experimental demonstrations of high density, excellent scalability, low power consumption, endurance, and low cost. An RRAM is normally referred as those NVM technologies built on the resistance changing mechanisms, which can be varied by applying a voltage pulse, other than phase-change memory and STT-RAM. Data is stored by changing the resistance across a dielectric solid-state material in the RRAM cell, which presents two resistance states: a high-resistance state and a low-resistance state, being respectively defined as R ON and R OFF. Thus, the device can be used as a Boolean logic switch returning (0) when the resistance is R OFF and (1) when R ON.