Abstract
A nonconventional way of producing nonvolatile memories is to use ferroelectrics, a class of electroceramic materials. These materials have a remanent polarization. The direction of this polarization can be changed by an electric field. Ferroelectric materials possess a “natural memory,” so to speak. Ferroelectrics have been known for a long time, and the idea to use them for binary data storage originates in the 1950s. The basic element of this type of memory is formed by a ferroelectric capacitor—a ferroelectric layer sandwiched between electrodes. Early prototypes were unsuccessful because rather high voltages were needed to switch the ferroelectric capacitor (200–300 V) and the memories suffered from crosstalk. (Programming one particular cell influenced neighboring cells.) The revival of ferroelectric memories was driven by the development of thin-film deposition techniques that allowed the formation of capacitors with ferroelectric thin films of submicron thicknesses. These capacitors can be switched with normal intergrated-circuit (IC) voltages. The crosstalk problem is circumvented by isolating each memory cell by a transistor (similar to a dynamic random-access memory [DRAM]). Compared to “standard” nonvolatile memories, ferroelectric memories offer the advantage of very fast access times (both for reading and writing), low-voltage operation, and good write/read endurance. A ferroelectric material that is already being used in commercially available memories is lead zirconate titanate, PbZrxTi 1−xO3. To combine a ferroelectric material with IC technology is a challenge, and many problems have been (and will be) encountered.
Talk to us
Join us for a 30 min session where you can share your feedback and ask us any queries you have