Progress toward the Holy Grail of magnetic-free and semiconductor-compatible non-reciprocal devices

By Jean-Jacques DeLisle, contributing writer

For the majority of transceiver and radar systems, there is a compromise between the typology RF front end and isolation between the transmitter and receiver circuits. Many cellphones and radar leverage switches on the front end to ensure a high level of isolation as the other options require relatively bulky and expensive magnetic circulators, isolators, and gyrators. This is certainly an issue in defense and aerospace applications, which is why there’s an active DARPA project in the works called Magnetic Miniaturized and Monolithically INtegrated Components (M3IC). The goal of this project is to achieve circulators, isolators, and gyrators that can be integrated into standard semiconductor processes and replace discrete devices.

M3IC. Image source: DARPA.

It could be that researchers at Columbia University in New York have developed a potential solution to this problem without any magnetic materials. Not only is the research reported in Nature Communications non-magnetic, it can supposedly also be integrated into a standard CMOS process. Moreover, the proposed technology can also operate well into the millimeter-wave bands, possibly enabling two-way communication for upcoming 5G telecommunication services.

According to Harish Krishnaswamy, lead of the research project, the mm-wave circulator enables mm-wave wireless full-duplex communications. This could revolutionize emerging 5G cellular networks, wireless links for virtual reality, and automotive radar.

The tone of the research paper and other articles on the topic are very optimistic. The reason behind this is that the ability to have non-reciprocity for outgoing and incoming signals can enable fully bi-directional communications and radar operation. This means that while a device is transmitting, it could also be receiving with minimal impact on the receiver circuitry performance. Current wireless communication systems must either operate with two separate frequency bands for uplink and downlink, frequency division duplex (FDD), or separate the uplink and downlink signals in time, time division duplex. With fully bi-directional communication, neither of these systems of duplexing would be necessary, which could enable reduced spectrum requirements for communication and higher throughput data transfer with the same amount of spectrum use.

It appears that the researchers are using switch-capacitor filter concepts that exhibit nonreciprocal phase shift properties as a base for how the conductivity of transmission line segments is modulated. By controlling the modulation depth at the different sections of the transmission lines, the realization of the proposed non-reciprocal relationships is demonstrated. By varying the structure of the transmission lines, different configurations of nonreciprocal devices are possible. For example, fabricating a transmission line as a ring, and a circulator is the result.

The research presents an experimental 25-Ghz 45-nm SOI CMOS circulator, which is only 1.2 x 1.8 mm in size. At 25 GHz, the clockwise transmission responses (S 21, S 32, and S 13) were found to be −3.3 dB, −3.2 dB, and −8.7 dB, respectively, and the counter-clockwise isolation responses (S 12, S 23, and S 31) were measured as −10.3 dB, −9 dB, and −18.9 dB, respectively. Without any port impedance matching, a broadband isolation of −18.3 dB to −20.2 dB over 4.6 GHz of bandwidth. The results could be demonstrated with higher isolation if the experiment also had an impedance matching circuit between the 50-Ω millimeter-wave probes and the ports of the circulator.

The evolution of this technology could eventually provide millimeter-wave isolators, circulators, and gyrators as integrated components within complex 5G and radar systems-on-chip (SOCs). As this is a new approach, there are many factors to consider before this technology can be considered viable. Evaluation of environmental behavior, power handling, frequency behavior, and device characterization are all necessary steps before this technology can be leveraged in devices. Also, vendors would have to help mature this technology from R&D to production grade, which isn’t always possible while meeting cost and compatibility constraints. Time will tell if this new approach to nonreciprocal devices can really be a herald to a new era of communications and radar technology.