Power Splitter for Terahertz Waves Developed to Improve Data Capacity in Cellular and Wi-Fi networks

A power splitter is one of the most fundamental parts of any communications network. It allows a signal to be transmitted to numerous devices and users. A team of researchers at Brown University have created such a device meant for terahertz radiation — a range of frequencies that probably will facilitate data transfer up to 100 times quicker than current Wi-Fi and cellular networks.

One of the most basic components of any communications network is a power splitter that allows a signal to be sent to multiple users and devices. Researchers from Brown University have now developed just such a device for terahertz radiation -- a range of frequencies that may one day enable data transfer up to 100 times faster than current cellular and Wi-Fi networks. (Photo Credit: Mittleman lab / Brown University)

“One of the big thrusts in terahertz technology is wireless communications,” said Kimberly Reichel, a post-doctoral researcher in Brown’s School of Engineering who led the device’s development. “We believe this is the first demonstration of a variable broadbrand power splitter for terahertz, which would be a fundamental device for use in a terahertz network.”

The device could have a number of uses, including as a part in terahertz routers that would transmit data packets to several computers, similar to routers in current Wi-Fi networks.

This innovative new device is illustrated in the Nature journal Scientific Reports.

The current Wi-Fi and cellular networks depend on microwaves, but the quantity of data that can travel on microwaves is restricted by frequency. Terahertz waves span between 100 and 10,000GHz on the electromagnetic spectrum, and possess greater frequency and thus have the potential to transmit plenty more data. Until recent times, however, terahertz has not received a lot of attention from researchers and scientists, therefore several of the standard parts for a terahertz communications network just does not exist.

Daniel Mittleman, a professor in Brown’s School of Engineering, has been involved in building some of those standard parts. Recently, his lab built the first system for terahertz multiplexing and demultiplexing, a technique of transmitting many signals via a single medium and then splitting them back out on the other side. Mittleman’s lab has also been involved in creating a unique type of lens for focusing terahertz waves.

All the parts created by Mittleman make use of parallel-plate waveguides — metal sheets capable of restraining terahertz waves and guiding them in specific directions.

“We’re developing a family of waveguide tools that could be integrated to create the appropriate signal processing that one would need to do networking,” said Mittleman, who was a co-author on the new paper along with Reichel and Brown research professor Rajind Mendis. “The power splitter is another member of that family.”

The novel device comprises many waveguides set to form a T-junction. Signal passing into the leg of the T is divided by a triangular septum at the junction, transmitting a fraction of the signal down the two arms. The triangular shape of the septum reduces the quantity of radiation that reflects back down the leg of the T, decreasing signal loss. The septum can be shifted left or right so as to vary the quantity of power that is transmitted down either arm.

We can go from an equal 50/50 split up to a 95/5 split, which is quite a range.


The septum is controlled manually in this proof-of-concept device.  Mittleman states that the process could easily be automated to facilitate dynamic switching of power output to each channel. Then it may be possible for the device to be integrated in a terahertz router.

It’s reasonable to think that we could operate this at sub-millisecond timescales, which would be fast enough to do data packet switching. So this is a component that could be used to enable routing in the manner of the microwave routers we use today.


The Brown University team aim to further tweak the new device. The subsequent step would be to initiate testing error rates in data streams transmitted via the device.

The goal of this work was to demonstrate that you can do variable power switching with a parallel-plate waveguide architecture. We wanted to demonstrate the basic physics and then refine the design.


The National Science Foundation and the W. M. Keck Foundation funded part of this project.

Tell Us What You Think

Do you have a review, update or anything you would like to add to this news story?

Leave your feedback