Computers may soon be a thousand times faster than they are now. Scientists at UT Dallas have designed an all-carbon computing system that exploits electromagnetic waves to significantly improve on silicon transistors used today.

Transistors use the flow of electrons through a wire to generate an electric current and behave like switches that turn on and off to create binary logic. The new design takes advantage of magnetic fields created when electrons move through a circuit, according to Joseph Friedman, lead author of the study published in Nature Communications and assistant professor at UT Dallas.

Electrons move through a carbon nanotube, Friedman said, creating a magnetic field in a nearby graphene nanoribbon. The nanoribbon is the unzipped form of the nanotube and has a very high resistance, so it also functions as a switch.

“The most exciting aspect is the possibility of extremely fast computing because the mechanism of switching is based on electromagnetic wave propagation rather than conventional charge transfer, possibly on the order of terahertz,” Friedman said.

Currently, computers have processing speeds in gigahertz, which is a thousand times slower than terahertz.

Traditionally, silicon transistors have to be physically connected by wires for information to be saved and used by other computing elements, but this is not necessary for the carbon logic gate, which uses the magnetic field to directly influence the resistance of nearby nanotubes. Friedman said this is why this new cascaded design is so fast.

The team includes Jean-Pierre Leburton, professor at University of Illinois at Urbana-Champaign, who worked on the basic physics and simulations of the device.

“The real challenge,” Leburton said, “was during the simulation to consider the right structure that would work for such a device, the right parameters for the carbon nanotube and ribbon, the distance between them, et cetera.”

Alan Sahakian, professor at Northwestern University, was Friedman’s PhD advisor at Northwestern, where they worked on a number of alternative circuit and device concepts for realizing logic circuits.

Sahakian said that other researchers have been trying to use carbon transistors for a long time, but they used traditional electric circuits and fields.

“In several years, the standard circuit approach will no longer be scalable, so many academic and industrial researchers are looking at unconventional alternatives,” Sahakian said. “Our approach, or others based on our ideas, may well take over in many applications.”

For now, Sahakian said the team is working with others in the field to build a prototype and develop a large-scale, fully functional system incorporating millions or billions of logic gates.