New Plasmonic Antenna Shines a Light on Terahertz Processors

Illustration: Christoph Hohmann/TUM

A laser (left) generates on-chip electric pulses in the terahertz frequency range. A second laser (right) reads the information transmitted by the first laser

https://spectrum.ieee.org/tech-talk/semiconductors/optoelectronics/plasmonic-antenna-development-could-shine-a-light-on-terahertz-frequency-processing

The antenna, just a few micrometers in size, uses ballistic electrons to transmit signals

By Michael Koziol

Imagine an antenna that could transmit at terahertz frequencies—generally defined at those between 300 gigahertz (GHz) and 3 terahertz (THz). Such an antenna could send and receive data at rates that would be orders of magnitude faster than any device we currently use. Even the 5G networks now being deployed will operate, at best, on frequencies well below 100 gigahertz.

Now, a team at the Technical University of Munich, headed by Alexander Holleitner and Reinhard Kienberger, has developed a terahertz antenna—but it won’t ever be used to send signals over the air. The TUM team’s antenna is designed to use electrons to transfer data across a minuscule gap on the surface of a chip. Their technique could open the door for much faster on-chip signal generation.

The antenna uses a quantum property of electrons to transmit very high-frequency signals across narrow gaps, and could generate signals for on-chip communication at rates far exceeding any technology currently available for moving bits of information across a chip.

The process begins with a phase-coherent laser, as the group describes in a recent Nature Communications report. “What is crucial is that it’s phase-stable,” Holleitner says, that is, each pulse has the exact same shape. Those laser pulses are used to excite delocalized electrons—that is, electrons that are “free-floating” rather than those attached to any atom in particular. These free-floating electrons, as the name suggests, are free to move between the atoms of the material they find themselves in—in this case, a tiny on-chip antenna.

When a photon from the laser pulse strikes one of these delocalized electrons, the electron produces a plasmon wave. These waves, associated with the frequency of a vibrating electron rather than the wavelength of a traveling photon, have a much shorter wavelength than electromagnetic waves. Shorter wavelengths mean smaller components—the upshot being, plasmonic antennas can transfer a lot of data in a very small space.

Image: TUM

The second key to this design is to use an asymmetric antenna, which means the emitter and the receiver, only nanometers apart, have different shapes, with the emitter side featuring more of a curve than the receiver side. It’s important to note that the Munich team is not the first to develop a plasmonic antenna, but past attempts used symmetric antennas, which Holleitner says makes it difficult to pick up the signal because ultimately they don’t generate a favorable current for the electrons to easily cross the gap. But Holleitner and the others discovered that by using asymmetric antennas (or, as they more accurately refer to them, “nanojunctions”), they could receive a stronger signal.

It’s important to stress how small this entire set-up is. The distance between the two nanojunctions that a signal will cross is 10 micrometers, about the thickness of a single cotton fiber. The data will be carried by the vibrating electron that is shot across that minuscule gap. “It’s just physics, you need to have very small gaps,” Holleitner says. Otherwise the electron wouldn’t make it.

The electron-exciting laser sends out 20 femtosecond-long pulses, which means, in one second, 50 trillion electrons would make the microscopic trip between nanojunctions. If each one was carrying one bit of data, that works out to a clean 10 terahertz transmission.

Holleitner says the team’s creation could be built on a silicon chip, and imagines that some day it could be used for extremely fast signal generation for communications.

For now, Holleitner is focused on finding a way to control the laser’s phase. “We tried controlling the phase, but the laser wasn’t strong enough,” he says, and explains that controlling the phase would enable them to create even higher frequency plasmonic antennas. And those higher frequencies will mean even faster processors.