Flash memory technology - a review

Abstract

Since the first flash memory patent issued on October 6, 1987, flash memory’s multiple useful characteristics, including nonvolatility, in-circuit reprogrammability, low power consumption, and high density, have led it to become the fastest growing memory segment in recent years. Mobile computing and communication have driven the demand for flash memories. Nascent applications, such as digital cameras, personal digital assistants (PDA), digital telephone answering devices @TAD) are expected to further add to growth in flash memory demand. Typically, flash memories are descendants of EPROM or EEPROM technologies and therefore many similarities in the architecture and operation can be noticed. As in the case of EPROMs, flash devices based on NOR architecture and the hot-electron injection mechanism for programming still dominate the marketplace. To provide erasability, the Fowler-Nordheim tunneling mechanism is typically employed. However, booming demand for flash memory has set-off a ferment of new architectures, cell structures, and manufacturing processes. For mass storage applications, NAND architecture is emerging as a contender. AND and DiNOR (Divided bit-line NOR) architectures have also emerged, each having their own advantages and disadvantages. Requirements on low power, low voltages, and higher densities have led to the development of devices which use Fowler-Nordheim tunneling for both programming and erasing. Multilevel 1087-4852/96 $5.00