Beyond millimeter wavelengths, the terahertz domain awaits
When developing the standards for 5G, the 3rd Generation Partnership Project (3GPP) was always about millimeter waves. 5G would inherently make use of millimeter wavelengths to accommodate specific applications and provide access to several gigahertz of spectrum unavailable at lower frequencies. But it didn’t stop there.
Researchers throughout the world are exploring frequencies well above 100 GHz and even into the terahertz region.
At the moment, wireless carriers are having significant difficulty implementing networks at frequencies of 24 GHz and 28 GHz, so it’s reasonable to wonder how operation at frequencies an order of magnitude higher will be achieved. Like all discussions of a technical nature, the truth lies in the details, especially those concerning propagation characteristics at various wavelengths.
To fully appreciate the terahertz region, it’s important to recognize where it falls in the electromagnetic spectrum, which is between where oscillators are used to generate radio signals and where photonics approaches generate light. While millimeter wavelengths are exceedingly small, a full wavelength at 100 GHz is smaller yet, about 3 mm at 100 GHz and a minuscule 0.3 mm at 10 THz. So, theoretically, if immense challenges can be overcome, entire systems could be constructed in fractions of an inch.
However, it’s important to understand why using these extremely high frequencies would be necessary when there is so much bandwidth available at much lower millimeter-wave frequencies. The answer is that the terahertz region, defined as 100 GHz to 10 THz, offers unique benefits.
These benefits include: resistance to interference so that devices could operate very close to each other; inherently high security because propagation distances are so short; and the ability to achieve data rates of 1 Tb/s (1 million Mb/s) — 1,000 times faster than 5G — over a single continuous band of frequencies. Some likely applications include high-definition holographic gaming, high-speed wireless data distribution in data centers, wireless cognition, sensing, imaging, and incredibly precise positioning and location.
It’s not a linear world
It might seem logical that the higher the frequency, the higher the loss through space, and while this is true, it’s not that simple. At the UHF and microwave frequencies, path loss results primarily from molecular absorption that is very low. But in higher reaches of the spectrum, other factors come into play as well, such as scattering from precipitation and foliage which can significantly impede range and reliability.
However, path loss is not linear because the resonant frequencies of oxygen, hydrogen and other gases in the atmosphere that absorb electromagnetic energy vary with frequency. So, looking at the figure below, atmospheric absorption at 100 GHz (0.1 THz) is just slightly higher at 300 GHz and continues to increase with wide variations through 10 THz, at which it becomes extremely high.
While logic dictates that as atmospheric attenuation increases with frequency, operation in the terahertz region should be fantasy. But as this figure shows, the progression of attenuation is not linear, and with the addition of extremely high gain antennas, it can be significantly mitigated. (Source: The Truth About Terahertz, Carter M. Armstrong, IEEE Spectrum, August 2012)
These characteristics have been pointed out by, among others, Dr. Ted Rappaport, professor of electrical and computer engineering at New York University’s Tandon School of Engineering and founding director of NYU Wireless. Rappaport notes that path loss through space decreases quadratically as frequency increases, as long as the antenna aperture is the same size at each end of the transmission path. So, for a given RF output power and identical antennas at each end of a link, signal strength at 140 GHz in free space is actually 5.7 dB greater than at 73 GHz and 14 dB greater than at 28 GHz.
In addition, the terahertz bands have surprisingly low loss when compared to sub-6 GHz frequencies, adding only about 10 dB/km up to about 300 GHz. In addition, although much of the spectrum between 600 and 800 GHz suffers massive attenuation of 100 to 200 dB/km, over a distance of 100 m this amounts to only 10 or 20 dB, which happens to be the typical coverage of a small cell.
When combined with the fact that antennas at these frequencies are so small that thousands of antenna elements could be used to generate an enormous amount of forward gain over very narrow beamwidths, considerable power can be directed to specific areas using active phased array antenna technology. That said, 300 GHz is far lower than 10 THz, where attenuation of 100,000 dB/km effectively makes propagating signals over long distances impossible. Still, it is possible to use mesh networks or repeaters to increase range.
These frequencies have many other uses beyond voice and data communications, perhaps the most obvious of which is inter-satellite communications, where the inherent security of narrow-beam signals has substantial benefits for ensuring security. In fact, the Department of Defense has been using millimeter-wave frequencies to achieve this for many years and is interested in moving higher up in the spectrum to make jamming and interception of signals far more difficult.
Difficult but not impossible
Realizing the potential of terahertz frequencies will require enormous advances in technologies from semiconductors through network design. Because they won’t be needed for many years — at least for communications purposes — researchers have at least a decade to work on these problems, probably longer.
The FCC recognized the technological challenges that terahertz operation presents. In 2019 the agency initiated a program called New Horizons that provides an almost rules-free pathway for experimentation at unlicensed frequencies between 95 GHz and 3 THz, creating the Spectrum Horizons Experimental Radio license. It provides broad flexibility in specifications such as frequency range, power and emissions, with the only caveat being that the experimenter must refrain from creating interference to existing services, of which there are obviously few. More than 21 GHz of spectrum is available at 116 to 123 GHz, 174.8 to 182 GHz, 185 to 190 GHz and 244 to 246 GHz.
