ABSTRACT
An incredibly rapid expansion in the demand of portable consumer electronics ranging from personal computers, cellular phones, MP3 players, digital cameras, PDAs, USB memory sticks to applications in the networking arena continues to drive the pursuit for the development of higher density, faster, more eµcient, and economical electronic memory devices (Scott 2004). Currently, most commercially available memory elements consist of dynamic random access memory (DRAM), static random access memory (SRAM), or °ash memory (nonvolatile memory). DRAM is considered to be very fast and cheap but its contents are lost when power is switched o˜ (volatility). SRAM is much faster and needs less power but is far more expensive and also su˜ers from volatility. Flash memory is nonvolatile but currently operates at low write speed and thus has a slow random access. In addition, it is power-hungry and very expensive. In the past, incremental improvements in memory capacity and capability were primarily achieved by the simple scaling of the physical dimensions of the devices. However, the semiconductor industry is quickly reaching the fundamental limits on the miniaturization, encountering, for example, extreme diµculties due to short channel e˜ects in the scaling of metal-oxide-semiconductor œeld-e˜ect transistor
(MOSFET). ›is diµculty has stimulated an intense growth in research and development in the area of new technologies and materials that can overcome these limitations and deliver unprecedented capabilities for memory storage and access. Some promising new types of memory-storage devices include magnetoresistive RAM (MRAM), ferroelectric RAM (FRAM), phase-change memory (PRAM), and novel high speed-density nonvolatile memory (these are primarily based on carbon nanotubes [CNTs]) (Waser 2003).
