IBM researchers have made a major breakthrough in an entirely new form of electronic device, one that Apple--without spending a dime on the research behind it--might use in future decades to create an iPad that would make today’s model look like young Abe Lincoln’s homework shovel.
IBM is an old-school corporation that believes in basic scientific and technological research. Apple isn’t. Such efforts can sometimes take decades, but can benefit not only the sponsoring company but the world as a whole. Think of the development of the transistor by AT&T’s Bell Labs back in the late 1940s. Can you imagine a world without transistors now? I can’t.
IBM’s breakthrough has the potential of being nearly as disruptive. They’ve breathed new life into an entirely different way of creating electronic devices, a technology called “spintronics.”
While spintronics may sound like a doo-wop group opening for the Delphonics back in the early 1970s, it’s actually a portmanteau for “spin transfer electronics,” which uses an electron’s spin to represent digital data’s ones and zeros rather than assigning that chore to an electron’s charge as is done in conventional electronics.
An electron’s spin can theoretically be manipulated much faster using much less energy than can its charge. If researchers can create devices that reliably apply, maintain, and read quantum-mechanical electron spin rather than having to muck about with comparatively clunky charge-based, classical-physics electronics, we will enter an era of power and miniaturization as yet undreamed of.
How about a supercomputer in your glasses frame that will listen to your companion speak to you in, say, Urdu and whisper a translation into your ear in real time? Or how about a dime-sized flying nanobot that can store and manage detailed navigational info about the city in which you live? On second thought, scratch that last one--it’s a bit spooky.
The work of IBM scientists such as Matthias Walser could have a dramatic impact on technology in the future.
Spintronics has already made its presence felt in hard drives’ GMR (giant magnetoresistive) read/write heads--Google it for a geeky deep dive--and in research on high-density, high-speed nonvolatile memory such as might be used in the SSD (solid state disk) of a future MacBook Air. Getting coherent electron spin to hang around long enough to be useful in computational structures, however, has been tough: until IBM got into the act, electron spin could be reliably maintained for only around 100 trillionths of a second, not long enough to be useful for a computing device chugging along at a comparatively slow 1GHz, or one billionth of a second per clock cycle.
Enter the IBM researchers. Using a combination of a semiconductor material known as gallium arsenide, a laser pulsing at one quadrillionth of a second, and other micromagic they were able to lock a group of coherently spinning electrons into a “spin helix” with a relatively Methuselan lifecycle of 1.1 nanoseconds, long enough to fit within a processor’s 1GHz clock cycle.
But don’t start ordering spintronic chippery quite yet. There are still a boatload of problems to be solved in theory, design, materials, and manufacturing. There’s also another niggling detail: the IBM breakthrough was made on a device cooled to –378°F, significantly colder than even the chilliest day in Fargo, North Dakota. Or Mars.
Formidable challenges, to be sure, but just the kind to fire the imaginations of researchers at companies such as IBM that still believe in basic research. Here’s a stat you might find of interest: last time we looked, Apple had about $117 billion in the bank, and its R&D budget for its last quarter represented 0.75 percent of that amount. IBM had $11.2 billion in cash, and spent 14.6 percent of that number on research in its last three months. Which company is getting a free ride?
>> Rik Myslewski was editor-in-chief of MacAddict from 2001 until transformed into Mac|Life in 2007, and how writes for The Register, which is "biting the hand that feeds IT" daily at www.theregister.co.uk
