If research in silicon chips were a sport, the IEEE International Solid-State Circuits Conference would be its Olympics, and the Lewis Winner Award for Outstanding Paper the equivalent of the 100 meter gold medal. As announced during the plenary session of ISSCC 2018 today, IBM Research has won this coveted award for the 2017 conference for work in the field of silicon-based millimeter wave phased arrays. This is the fourth time in the last 15 years that IBM Research has received this recognition, and the third time for the work of the IBM Research RF Circuits and Systems Group, continuing a tradition of leadership contributions to the worldwide solid-state circuits community.
The 2017 ISSCC paper, which resulted from a collaboration between IBM Research and Ericsson, describes a chip for 5G communications. Silicon chips continue to be the backbone of computation and communications technology, and underneath the hood of every new generation of cellular technology are game-changing silicon chips. Last year, IBM Research and Ericsson unveiled a chip that will revolutionize mobile communications technology by using narrow pencil-like beams at very high frequencies for high-data-rate communications. Such electronically steerable narrow beams reduce energy wastage and increase data rates by focusing the energy towards only the intended user rather than broadcasting the signal in all directions, as is done today.
Millimeter-wave technology breakthroughs from the first Si-based 60GHz circuits to the first phased array for 5G
Our RF Circuits and Systems group at IBM has been a pioneer in silicon-based mmWave research; the mmWave frequency band lies above the radio frequency and microwave bands. Prior to the 2017 award, the IBM Research RF Circuits and Systems group won the 2004 Lewis Winner Award for a paper entitled “60-GHz Transceiver Circuits in SiGe Bipolar Technology,” demonstrating, for the first time, silicon-based circuits operating at frequencies as high as 60 GHz. The group followed this up with a second award for the ISSCC 2006 paper, “A Silicon 60GHz Receiver and Transmitter Chipset for Broadband Communications,” describing the first demonstration of silicon-based fully integrated radio transmitters and receivers at 60 GHz. These two remarkable breakthroughs were able to convince skeptics that mmWave circuits, traditionally implemented using discrete III-V blocks, could function reliably when fully integrated into silicon-based processes. Subsequent advancements by the RF Circuits and Systems group included the demonstration of a scaled 94GHz phased array in 2013 and a low-power CMOS transceiver for 60GHz communications in 2016. With respect to III-V-based discrete designs, silicon integration has enabled a stunning volume reduction of ~1000X alongside a ~1000X increase in integration complexity and has already enabled the first generation of mmWave commercial automotive radar and data communication products. Moreover, these advances placed mmWave as a key component of the next global mobile communications standard – 5G.
In the ISSCC 2017 award-winning paper, “A 28GHz 32-element phased-array transceiver IC with concurrent dual polarized beams and 1.4 degree beam-steering resolution for 5G communication,” the IBM Research-Ericsson team presented the world’s first reported silicon-based 5G mmWave phased array antenna module operating at 28GHz. Towards the end of last year, the ﬁrst 5G specifications were released by 3GPP, supporting mmWave communications. Today, 5G trials are being carried out and mobile operators around the world are preparing their networks to support mmWave communications. With this new technology, these operators are hoping to usher in the next revolution in wireless technology, and allow high-data rate applications such as virtual reality over cellular networks to finally take flight.
Beyond 5G: Vertically Integrated Antennas-to-AI Systems
Having been at the forefront of silicon-based mmWave and phased array technology. since its inception, our group continues to grapple with the next set of challenges for 5G technology and beyond. Electronic beam steering brings a new level of complexity hitherto unexplored in wireless communication.
One path we are exploring to meet the next generation of phased array challenges is work on vertical stack integration to more tightly couple software and hardware. Specifically, we believe that coupling complex multi-antenna systems such as our phased arrays to machine learning-driven back ends will enable us to take these systems to an even higher level, driving the creation of new adaptive communication systems and of the next generation of portable imaging systems.