Magnetic random access memory (MRAM) may soon see widespread market adoption due to the work of IBM and TDK Corporation, as these two have spawned a program to develop it in high capacity.

If you’re unfamiliar with MRAM, you only need to know that it is a far superior memory. But, assuming you won’t take my word for it, MRAM’s advantages over other memory types include:

1. Power usage - MRAM requires no refresh of its cells to maintain memory; whereas dynamic random access memory (DRAM) requires its capacitors to be refreshed approximately 1,000 times per second. This allows the MRAM to not constantly draw power, consuming 99% less power than DRAM.

2. Quick speeds - MRAM operation is based on measuring voltages rather than currents, thus allowing it to operate quicker than the most advanced DRAM chips in production. IBM researchers have developed MRAM with access times of 2 nanoseconds.

3. Unlimited read and write cycles - Unlike Flash RAM, MRAM cells do not degrade, allowing a device to read and write to the memory an unlimited amount of times. Flash cells can only be written to a fixed number of times before it simply won’t work.

4. Inherent non-volatility - Similar to Flash memory, MRAM will retain the contents of its memory even after the power is switched off.

But if MRAM is so great, why isn’t it used everywhere? It seems that market adoption of MRAM has always been limited by the inability to cost effectively increase capacity of the chip. In other words, it’s too expensive. 

MRAM has been plagued with capacity since it went under development in the 90s. Memory cells within the MRAM were limited to 180nm in size until the arrival of toggle-mode MRAM. But toggle-mode could still only be developed to 90nm, the size of most DRAM cells.

It has been assumed that for MRAM to be worth putting into large production, it will have to exceed to 65nm.

Using a technology known as the “spin momentum transfer effect” (SMT), IBM, together with TDK, believe this method will allow a much more compact memory cell than is currently possible.

Spin momentum transfer technology enables MRAM to significantly reduce cell size and thereby increase capacity cost-effectively while maintaining the low power, endurance and non-volatility advantages.

“This joint research and development will broaden the application of magnetic materials which has been TDK’s core technology since 1935,” said Mr. Minoru Takahashi, CTO, TDK Corporation.

Perhaps the most exciting aspect of this co-operation is that both companies have a certain expertise in the fields needed to develop a high density, high capacity MRAM module based on the spin momentum transfer effect.

IBM is not only a leader in the development of MRAM technology, but is on the leading edge of research in magnetic tunnel junction (MTJ) and spin momentum transfer effect, upon which its method for creating MRAM is based.

TDK is the leader in applying MTJ technology into recording heads for hard drives (a head nearly identical to this will be used in the proposed MRAM).

“This collaborative initiative reinforces IBM’s commitment to explore new phenomena for memory applications,” said Dr. T.C. Chen, vice president, Science and Technology, IBM Research. “The project will focus on creating and demonstrating advanced magnetic materials in demanding memory chip designs.”

The research will be conducted at IBM’s research center in New York.

Related:

Entry was posted on Monday, August 20th, 2007 at 4:01 pm under Uncategorized.  Print This Post

Read more entries by George Gardner.
Stumble It!