Published online by Cambridge University Press: 01 February 2011
It expected that for many years to come, the majority of the nonvolatile memories shipped will be based on current mainstream flash technologies, which utilize transistor based charge storage memory cells and multi-level-cell concepts, for storing more than one logic bit in a single physical cell. Moore's law will continue to drive transistor based memory technology scaling but technology complexity will be increasing. In order to meet technology scaling, the mainstream transistor based flash technologies will start evolving to incorporate material and structural innovations. This paper will review the current status and discuss the approaches being explored to provide scaling solutions for future transistor based non-volatile memory products. Based on the introduction of material innovations, it is expected that the planar transistor based flash memory cell can scale into the 32nm node. Further, more complex, structural innovations will be required to maintain further scaling. New memory concepts, not relying on transistors as a basis of the memory cell, provide new opportunities for future low cost memories. Several of these new concepts will be summarized and contrasted with the mainstream transistor based flash memory technologies.