Sunday, September 12, 2004

Storage space

Nanotechnology-Based Data Storage on Rise (UPI)

Industry buzz is rising around MRAM, or magnetic RAM, where data are stored not electrically but magnetically. The advantage of MRAM is it preserves data after systems are switched off, which means computers, digital cameras, PDAs and other devices can be built with nonvolatile memory, the kind immediately ready for use after switching on.

    millipedeNanotechnology could yield billions of dollars of new data storage devices, based on exotic technologies, in just the next few years, with vastly larger memory and faster response times, analysts said.

    Advances customers could see from such devices include cell phones Latest News about cell phones with enough memory to download movies, suggested Lawrence Gasman, principal analyst for NanoMarkets, an industry research firm in Sterling, Va. The global market for such nanobased storage -- engineered at the scale of nanometers, or billionths of a meter, which is shorter than a wavelength of visible light -- is expected to increase dramatically.

    Experts predict a growth from $97 million in 2004 to $17.9 billion by 2008 and $65.7 billion by 2011, large enough to suggest future disruptions in existing markets and potentially the rise of new industry giants. More here
Related News
Nano Memory Scheme Handles Defects (Technology Review, via Techno-News Blog)

NanoBot Backgrounder
From the Dark Tower of my memory
'Terabyte territory'
Money for Memory


Jason Koulouras said...

Would there be a need to contain the magnetic fields used in MRAM or would the fields be so weak as to be inconsequential to other devices/magnetic field using instruments (for example a compass) etc.


Howard Lovy said...

It's an excellent question, Jason, and one that had the MRAM experts stumped for years. In fact, this magnetic field issue was a major barrier blocking MRAM's path from the laboratory to the marketplace until a Motorola scientist discovered the way out.

First, here's the Blogger's Digest explanation of magnetoresistance random access memory. Scientists and engineers, go talk amongst yourselves. You'll probably be horrified by my simplifications for the casual reader.

The way MRAM works is not in the storage of data inside particles, rather in the way particles – in this case, electrons – spin. They do this inside "tiny magnetic sandwiches" that sit appetizingly at the intersection of two perpendicular electrodes that run above and below it. Use these electrodes to zap the magnets from above and below, and the electrons inside will dervishly spin to your command.

Hit them with parallel currents, and the electrons will do an Esther Williams number, spinning in unison in the same direction, and in the process writing a "1" in binary code. But, then, break out the topspin you've been secretly practicing in the basement ping-pong table and you'll get the top layer of electrons to move in the opposite direction from the bottom. You've just written a "zero."

The trouble was, the magnetic field Jason mentioned had an annoying tendency to disturb the neighbors, forcing nearby bits to flip when they really should hold still.

Here's where the late great Leonid Savtchenko stepped in as the lead inventor in Motorola's Patent 6,545,906.

Simply put, it's a "toggling" solution that zaps the memory bit with two overlapping pulses, each guiding the rotation in proportion to how much the little bit decides to resist. The result is that just enough force is applied to make the bit say "1" or "0" without so much as rattling the neighboring bits' dishes.

That brings Motorola/Freescale to the front of the line, and it's going to license the technology to others. That's good for the spread of MRAM, and good for Freescale.

Sadly, Savtchenko died of a brain aneurysm at the age of 35 and did not see the final fruits of his labor.


Jason Koulouras said...

Thanks Howard for taking the time for such a concise answer laid out in layman terms.

I would like to indulge in a follow up question

Would a terrestial pole reversal altering the Earth's magnetic field be a potential risk to MRAM stability?


Howard Lovy said...

Well, this one's way too hard to wrap my liberal arts brain around, but I'm guessing that the folks at Motorola failed to account for the once-every-10,000-year event. If you want that question answered, you'll have to call them -- if you can ...


Jason Koulouras said...

Fair enough Howard...thanks for the details and your blogging of this complex subject