Science

A 100-Year-Old "Nuisance" is Now the Future of Smartphone Chips

You'll like the sound of this.

by Alasdair Wilkins
Getty Images / Dean Mouhtaropoulos

In 1922, a young French physicist named Léon Brillouin predicted sound waves could interfere with and scatter much faster-moving light waves. That’s bad news if you’re trying to send information on light pulses — which is exactly how fiber-optic communications networks connect the globe.

“That Brillouin scattering effect is really regarded as a nuisance in the telecommunications industry,” University of Sydney Professor Ben Eggleton tells Inverse. But he and his fellow researchers have found something potentially revolutionary in that interaction between light and sound waves. “What we are harnessing, they see as detrimental.”

As the researchers describe in Monday’s Nature Communications, they have taken microchip data carried by light waves and instead stored them inside sound waves. And while they are the first ones to pull off this feat, they have done it on existing tech, meaning the benefits of this technology could be felt sooner rather than later.

“Doing it on a chip is really a big part of the breakthrough,” says Eggleton, who observes that everything ends up in a smartphone one day. “We can build this onto a chip that’s compatible with your smartphone, we can build this onto a silicon chip, we can build all the other functions onto the chip as well.”

There’s a big, if counterintuitive advantage to holding data in the acoustic rather than the optical domain: The speed of sound is about 100,000 times slower than the speed of light.

Slowing down the flow of information is crucial for the much same reason that it’s important to be able to slow down when driving on a highway. When cars are in the middle of their long journey, they want to zip along at 70 miles per hour, but they can’t pull off to refuel or arrive at their destination without slowing down to exit. It’s the same principle with data inside a communications network: It’s simply not possible to process and manage that information if it’s moving at light speed.

A diagram of light and acoustic waves headed toward each other on a data chip.

University of Sydney

The University of Sydney team’s chip works by intentionally making use of the once dreaded Brillouin scattering, as the light wave first induces a sound wave. “Then it couples that data packet onto that sound wave,” says Eggleton. “And then it sits in the sound wave for as long as the sound wave sits there for, propagating 100,000 times slower, and then you read the information out off that sound wave.”

Not that the microchips of the future will feature a constant barrage of audible screams as data moves from light to sound and back again. The acoustic waves in question are what’s known as hypersound, roughly a million times higher in frequency than those of sounds you hear every day. Nor will any of those sounds hang around for very long.

“This is not meant to be like a USB stick,” say Eggleton. “You’re not storing these sound waves for a long time at all. You’re storing these sound waves for long enough that you can basically delay that [data] packet. It might be only tens of nanoseconds that you need to store that information.”

While a general disdain for Brillouin scattering may have kept previous researchers from considering the possibility of sound for storing data, scientists have long known there needed to be some way to slow down the speed of information in optical networks. The simplest solution is just to move those data packets to traditional electronic chips, but we have reached the upper limit of what those can do. The world’s most advanced data centers already consume too much energy and produce too much heat with electronic chips for those to be a realistic option.

All this groundbreaking research is happening inside a chip barely the size of a coin.

University of Sydney

It says a lot about how serious a problem this is that many researchers have spent more than a decade trying to figure out how to slow down the speed of light. It’s not impossible to do that, as Harvard University researchers slowed light to less than 40 miles per hour in the late 1990s, but that was only possible in temperatures barely above absolute zero. Repeating the feat at room temperature has proved more or less impossible, especially over anything but the most narrow band of frequencies.

“There seems to be this fundamental tradeoff between the speed of light, how slow you can make it, and the bandwidth over which you can slow it down,” says Eggleton. “And that’s hardwired in by the laws of physics. So that just tells you up front, you’re only going to be able to go so far. You’re never going to be able to massively slow light down for very, very high-speed data signals. I think it took a while for the community to get their heads around that.”

While moving data from light to sound waves has its own challenges, Eggleton says this is an elegant sidestep of what would otherwise be a fundamental barrier thrown up by the laws of physics.

The first and most important application will be in communications data centers, relieving the otherwise unmanageable strain. Eliminating the use of costly electronic processers could represent a pretty massive energy savings — and we live in a time where any energy we don’t need to use is worth paying attention to — but Eggleton stresses that hasn’t been demonstrated by their experiments so far.

For now, the real excitement for him and his team lies in opening up an entirely new way of handling all the world’s ever-growing mass of data.

“Up until now, we’d only considered the possibility of manipulating information in the electrical domain, then in the optical domain, and now there’s a new possibility that we can also imagine manipulating information in the acoustic, sound wave domain,” says Eggleton. “I guess I’d like to say it’s the new wave. It’s the next wave.”

Related Tags