As transistors have been made smaller and packed more densely onto computer chips, chips have consumed more and more power, quickly draining batteries and threatening to make laptops unbearably hot. This has many looking ahead to a day when something other than a transistor might serve as the workhorse of the computer processor.
One candidate for such an alternative technology recently took another step toward practical reality. As reported last week in the journal Science, researchers at Notre Dame have combined magnetic nanoparticles into a logic gate that theoretically could be used to perform all the operations of today’s computers. Instead of electricity, as in transistors, the technology uses the particles’ magnetic fields for processing information, leading the researchers to estimate that a computer based on this technology could run on a thousand times less power.
It’s experimental evidence for a theoretical approach that “could very well be the most efficient way of computing,” says Stan Williams, director of quantum science research at Hewlett-Packard, who calls the Notre Dame research “first rate.” While it’s unlikely to appear commercially in computers within the next decade, he says, “what it has done is inject a note of optimism that there are physical processes that can be used for computing that can be very, very low power consumption.”
Furthermore, since the process does not require power to maintain its settings, it could be the basis of instant-on computing, as well as surviving power outages, says Williams. “Somebody can pull the plug on you, and you can plug it in maybe five years later and the thing’s going to take up exactly where it left off and keep on going.”
At the heart of the new technology are magnetic nanoparticles that “flip” in response to the orientation of similar nearby particles – as refrigerator magnets sometimes flip over if they’re brought close together. In a row of such particles, flipping the first particle can cause the rest of the magnets to flip, like a row of dominoes falling. This, in effect, transmits the information about the first magnet’s position to the end of the row.