Abstract

At present, the main driving forces for rapidly growing the market of nonvolatile memory devices and nonvolatile memory technology are portable electronics. Technical innovations will continue to drive the increase of memory density and speed in the future. For traditional nonvolatile memory devices like NOR flash, scaling could stop even before the 32nm node, while scaling of mass storage devices like NAND flash, is going to be very difficult beyond the 20nm node, too. In recent years, more research efforts have been devoted to finding a new memory technology able to overcome performance and scalability limits of currently memory devices. However as lithographic scaling becomes more challenging. Novel architectures are needed to improve memory device performances in embedded as well as stand-alone memory applications. It is now generally recognized that 3–D integration and vertically stacking are possible alternatives to scaling. In order to realize stacked memory cells a novel physical principle operation of a memory element is needed. Many new ideas have been proposed to find out the solution of such problems. Recently, it were developed a rewriteable 3–D NAND flash memory chip called Bit Cost Memory (BiCS) in which it stacked memory arrays vertically (Fukuzumi & et al., 2007) and 3–D doublestacked multi-level NAND flash memory (Park & et al., 2009). However, there is a cross talk between neighbouring memory cells in the same plane and between vertical neighbours in adjacent planes, too. It is a new challenge in addition to scaling issues. The major limitation, however, is likely to be power dissipation where 3–D designs are not efficient at dissipating heat. More over 3–D architectures are required to overcome a tyranny of interconnects. From physical point of view, a memory device based on ferroelectric phenomena is more suitable for creating 3–D architectures of universal memory devices. Ferroelectric materials can be electrically polarized in certain direction. Then, the polarization is retained after the polarizing field is removed. Therefore, these materials have intrinsic memory. Ferroelectric random access memory (FeRAM) is widely considered as an ideal nonvolatile memory with high write speeds, low-power operation and high endurance (Scott, 2000; Pinnow & Mikolajick, 2004). FeRAM also have an added significant advantage, which is high radiation hardness. FeRAM are inherently radiation-hard making it very suitable for space applications. The FeRAM with lowest active power is a voltage-controlled device as

Full Text
Published version (Free)

Talk to us

Join us for a 30 min session where you can share your feedback and ask us any queries you have

Schedule a call