Fujitsu Makes a Terahertz Receiver Small Enough for a Smartphone

http://spectrum.ieee.org/tech-talk/telecom/wireless/fujitsu-makes-a-terahertz-receiver-small-enough-for-a-smartphone
By John Boyd

It’s a good time to be alive for pixel peepers. TV makers are pushing 4K-resolution sets to replace our present 1080p screens; Apple’s iMacs sport a 5K resolution; and NHK, Japan’s national broadcaster, is testing 8K broadcasting equipment, targeting 2020 and the Tokyo Olympics for its introduction.

To help wireless devices cope with the higher speeds demanded by such applications, Fujitsu has developed a 300-GHz prototype receiver compact enough to fit into a cellphone. Though limited to about 1 meter in range, the company says the device can download 4K and 8K video almost instantly.

Today’s cellphones operate in frequency ranges between 0.8 to 2.5-GHz, and are capable of download speeds of around 230 megabits per second, while the top speed for 802.11n Wi-Fi operating in the same frequency range can reach speeds as high as 600 Mb/s. Fujitsu touts its new receiver as operating in theterahertz band—frequencies of over 300 GHz—where terminals can communicate at speeds hundreds of times faster than today’s mobile handsets.

Devices to enable such high speeds have been developed, but because terahertz-band waves quickly attenuate, receiver-amplifier chips need to be sensitive enough to deal with a weak signal. Present designs rely on a separate antenna, which in turn requires a waveguide component to transport the incoming signal from the antenna to the chip. This makes the overall combination far too bulky for cellphone use, says Fujitsu.

The goal, then, is to create a receiver-amplifier module with a built-in antenna  to increase miniaturization. This has been achieved for devices employed in millimeter wave-band equipment operating at 60-GHz to 80-GHz frequencies, for instance, and used in applications such as collision-avoidance radar. These modules connect the antenna to the receiver-amplifier  chip through an internal printed-circuit substrate making a waveguide unnecessary.

Illustration: Fujitsu

“Typical printed-circuit-substrate materials used in these higher frequency ranges are ceramics, quartz ,and Teflon,” says Yasuhiro Nakasha, a research manager at Fujitsu’s Devices & Materials Lab. “But when these are used in terahertz-band communications, there is significant signal attenuation and loss of receiving sensitivity.”

To get round this, Fujitsu has micro-fabricated a printed-circuit substrate using a polyimide (a heat-resistant synthetic polymer) material.  Signals from the antenna are transmitted to the receiver-amplifier chip through a connecting circuit on the substrate.

In order to ensure stable signal transmission with low loss, the top and bottom faces of the printed circuit substrate are grounded and connected using through-hole metalized vias. This and the connecting circuit together form a grounded coplanar-waveguide structure: a transmission pathway designed to enhance high frequency signal propagation. To reduce signal interference from the printed circuit substrate, the vias need to be spaced apart less than one-tenth of the signal’s wavelength—in this case less than a few tens of micrometers.

Though the polyimide material experiences a signal loss ten-percent greater than quartz, Fujitsu says the material’s processing accuracy is more than four times higher than the latter. This makes it possible to space the vias closer together, thereby halving the overall signal loss compared to using a quartz substrate.

To facilitate a strong connection between the antenna connecting-circuit on the printed-circuit substrate and the receiver-amplifier chip, Fujitsu adapted a millimeter-mounting technology to handle terahertz transmission. This method let the receiver-amplifier circuitry directly face the printed circuit substrate.

The outcome is a module with an overall volume of just 0.75 cubic centimeters—not including output terminals—small enough to be incorporated into a mobile phone. Download speeds obtained so far in the lab reached 20 Gb/s.

Fujitsu will begin field-testing by the end of March 2016, and aims to launch the technology in 2020. The application the engineers envision include instant downloading of large volumes of data from servers and terminals, electronic versions of printed guides and brochures used at events, and downloading video and music from kiosks.

Nakasha isn’t looking beyond 2020 at the moment, but he believes the technology has the potential to one day achieve speeds of 100 Gb/s.

Part of the research used was obtained from an R&D project on expanding radio spectrum resources commissioned by Japan’s Ministry of Internal Affairs and Communications.