6G terahertz devices require 3D electronics

Smartphones have scaling issues. Specifically, the radio frequency (RF) filters that all phones (and all wireless devices in general) use to extract information from isolated radio signals are too large, too flat, and too numerous. Masu. Without these filters, wireless communication will not work at all.

“They are literally the entire backbone of the wireless system,” says researcher Roosbe Tavrian of the University of Florida, Gainesville.

So Tabrizian and other researchers at the University of Florida developed an alternative three-dimensional RF filter that could save space on smartphones and IoT devices. If someday these his 3D filters replace his bulky stack of 2D filters, there will be even more room for other components such as batteries. It could also make it easier to push wireless communications into terahertz frequencies, a key spectrum range being explored for 6G mobile phone technology.

“Soon, we will have trillions of devices connected to wireless networks. We will need new bands. We just need full range frequencies and full range filters.” —Roozbeh Tabrizian, University of Florida

The filters currently used in wireless devices are called planar piezoelectric resonators. Each resonator has a different thickness. The specific thickness of the resonator is directly related to the radio frequency band to which the resonator responds. Wireless devices that rely on multiple spectral bands (which are becoming increasingly common today) increasingly require these planar resonators.

However, as wireless signals proliferate and the spectrum on which they depend expands, many weaknesses in planar cavity technology have become apparent. One is that it is becoming increasingly difficult to make filters thin enough for the new spectral bands that wireless researchers are interested in exploiting for next-generation communications. The other one is about space. It is proving increasingly difficult to incorporate all the necessary signal filters into a device.

The vertical fins of a ferroelectric gate fin resonator can be constructed in the same way as FinFET semiconductors.Faysal Hakim/Rouzbe Tavrian/University of Florida

“Soon, we’ll have trillions of devices connected to wireless networks. All we need is new bands, full range frequencies and full range filters,” Tabrian said. “When you open up your phone, you have five or six specific frequencies, and that’s it. Five or six frequencies can’t handle that. It’s as if you had five or six roads and the traffic in a city of 10 million people. It’s like we want to respond to that.”

To switch to 3D filters, Tavrian and his research colleagues took a page from another industry that has made the leap into the third dimension: semiconductors. When the industry finally reached an impasse in its continued pursuit of decreasing chip size, a new approach to pulling electronic channels onto semiconductor substrates breathed new life into Moore’s Law. This chip design he called a FinFET (short for “fin field-effect transistor,” where the “fin” refers to a shark-fin-shaped vertical electron channel).

“The fact that we can change the width of the fins plays a big role in further increasing the capabilities of the technology.” —Roozbeh Tabrizian, University of Florida

“We were definitely inspired [by FinFETS]” says Tabrian. “The fact that the planar transistor was converted to a fin type was simply to ensure that the effective size of the transistor was reduced while having the same active area.”

Despite taking inspiration from FinFETs, Tabrian says there are fundamental differences in the way vertical fins need to be implemented in RF filters compared to chips. “If you think about FinFET, all the fins are about the same width. People aren’t changing the dimensions of the fins.”

This is not the case for filters, which require fins of different widths. This allows each fin of the filter to be tuned to a different frequency, allowing a single 3D filter to handle multiple spectral bands. “The fact that we can change the width of the fins plays a big role in further increasing the capabilities of the technology,” says Tabrian.

Tabrizian’s group has already produced several three-dimensional filters called ferroelectric gate fin (FGF) resonators spanning frequencies from 3 to 28 GHz. He also built a processor with a spectrum consisting of 6 integrated FGF resonators, covering frequencies from 9 to 12 GHz (as a comparison, the coveted mid-band of 5G his spectrum consists of 1 to 12 GHz). 6 GHz). The researchers published their findings in January. Nature Electronics.

Tabrian admits that the development of 3D filters is still in its early stages and has a long way to go. But again, drawing inspiration from his FinFET, he found a clear path for the development of his FGF resonator. “The good news is that by looking at FinFET technology, we can already deduce what many of these challenges are,” he says.

To someday incorporate FGF resonators into commercial devices, several manufacturing issues will need to be resolved, including how to increase the density of the filter’s fins and improve the electrical contacts. “Fortunately, FinFET has already gone through a lot of these answers, so the manufacturing part is already solved,” he says Tabrizian.

One thing the research group is already working on is a process design kit (PDK) for FGF resonators. PDKs are common in the semiconductor industry and serve as a kind of guidebook for designers to manufacture chips based on components detailed by the chip foundry.

Tabrian also sees a lot of potential for future manufacturing that integrates FGF resonators and semiconductors into one component, given the similarities in design and manufacturing. “It’s human innovation and creativity that comes up with new types of architectures that could revolutionize the way we think about resonators, filters, and transistors.”