This Microscopic Beam Splitter Could Enable Light-Based Computers

The word "electronics" may become a misnomer faster than anyone expected.

Computer engineers at the University of Utah have made a major breakthrough in the world of optical computing, creating a light beam splitter 50 times thinner than a human hair. This innovation—one that various engineering teams have been chasing for years—paves the way for a new kind of computing altogether, where computers send signals within the machine using light instead of electrons.

In a normal computer, information is processed using electrons that are sent down wires, which is a terribly inefficient way of doing things. As Motherboard's Michael Byrne wrote last year, it's like "building subway tunnels for cockroaches." Light-based processing offers the potential for computers that are many thousands of times faster, owing to the speed at which light travels. In theory, light-based processors would also use less energy and heat.

With the invention of this tiny beamsplitter (which looks strangely like a QR code), light can be divided into two channels of information and transmitted on a much smaller scale. The beamsplitter is so small—2.4 square microns—that millions of these light splitters could be fit on to one single computer chip. The splitters could then be used to funnel specific bits of information around the machine.

The efforts in Utah were spearheaded by computer engineering associate professor Rajesh Menon and a team of his colleagues, who published their findings in Nature Photonics. Menon believes his team's photonic technology could be implemented by supercomputers and large data processing centers in as few as three years.

"Light is the fastest thing you can use to transmit information," Menon said in a statement. "But that information has to be converted to electrons when it comes into your laptop. In that conversion, you're slowing things down. The vision is to do everything in light."

Like all innovation, optical computing has its skeptics, with some wondering if it could really ever compete with electronic computing in size- or cost-efficiency. Menon and his team have amply addressed the first concern, with 2.4 square microns being close to the limit of physical possibility. As for the second, the photonic chip designed by Menon and his team could be produced using the same techniques currently used for making silicon chips, saving on production cost.

Faster, more compact, and more energy-efficient, optical computing is a dizzying new frontier with endless applications. This isn't totally a pipe dream or some far-off idea, either: Companies such as Intel and IBM have been doing early work on optical computing. Menon's finding won't be the thing that makes optical computing a reality, but it's an important step on the journey there.