Since the advent of the integrated circuit in 1958, the same year the Advanced Research Projects Agency was established, engineers have been jamming ever more microelectronic integration into ever less chip real estate. Now it has become routine to pack billions of transistors onto chips the size of fingernails.

DARPA (the D for Defense was first added in 1972) has played key roles in this ongoing miracle of miniaturization, giving rise to new and sometimes revolutionary military and civilian capabilities in domains as diverse as communication, intelligence gathering, and optical information processing. ?Now a DARPA-funded team has drastically miniaturized highly specialized electronic components called circulators and for the first time integrated them into standard silicon-based circuitry.

The feat could lead to a doubling of radiofrequency (RF) capacity for wireless communications-meaning even faster web-searching and downloads, for example-as well as the development of smaller, less expensive and more readily upgraded antenna arrays for radar, signals intelligence, and other applications.

The defining feature of circulators is that RF signals, in the form of electronic waves in the circuitry, travel only in a forward direction with reverse propagation of the wave forbidden by the physics of the circuit. That's what you need for minimizing on-chip interference and for keeping signals separated. Most materials can't play this role because RF traffic can flow both ways through them; these materials exhibit what engineers refer to as reciprocal behavior.

Nonreciprocal components like the new circulator, on the other hand, act like one-way highways for RF signals. Traditionally, circulators have relied on external, ferrite-based magnets to force RF signals into a one-way course through downstream circuitry.

Those magnets and ferrite materials have rendered the circulators bulky, expensive, and incompatible with the workhorse microcircuit technology, known by insiders as CMOS, which stands for complementary metal-oxide semiconductor. So it has been hard to miniaturize circulators for CMOS integrated circuits.

The Columbia researchers got around this roadblock to miniaturization by coming up with a path-breaking design that does away with the need for bulky ferrites and magnets. Their design achieves the one-way RF flow with a series of capacitors coordinated with a minuscule and precise clock, electronically emulating the direction-dictating magnetic "twist" that in conventional ferrite circulators is imposed on RF signals by an external magnetic field.

That novel design makes possible an unprecedented microelectronic assemblage: A receiver connected to one "on-ramp" (or port) of the new circulator structure; a transmitter connected to another port of that same circulator; and an antenna shared by those two tiny devices, itself coupled to the circulator via a third port situated between the other two. Since the RF propagation is one way (non-reciprocal) in the circulator, the transmitted and received signals smoothly traverse their respective paths without getting mixed up with one another.

That clean segregation of received and transmitted signals opens a powerful new capability. In most two-way RF systems, transmission and reception at a given frequency have to be staggered in time with a switching process, slowing communication speeds.

The way around this bottleneck has been to transmit and receive at two different frequencies, which requires twice as much spectrum-a limited resource. By contrast, the new pinky-nail-sized circulator opens the door to communications and radar systems operating in full duplex mode-that is, transmitting and receiving at the same frequency at the same time with a single shared antenna.

"This new circulator component could enable full-duplex systems that let you speak and listen all at once," said William Chappell, director of DARPA's Microsystems Technology Office.

In radar applications, this capability could put an end to brief but potentially deadly blind moments since the system would not have to toggle between separate transmission and reception modes. And by halving the frequency needs, Krishnaswamy said, "full-duplex communication has the potential to double a network's capacity" for voice, data, and other forms of information.

In powerful radar and other RF systems that require large arrays of transmitters and receivers, he continued, "a compact, efficient, high-performance circulator" makes it easier for RF engineers to make their systems smaller.

Finally, noted Chappell, the new circulator's CMOS-compatibility feature is critical because it should ease integration into existing chip-manufacturing methods, potentially making all the difference between a laboratory achievement that stays in the lab and one that transforms a raft of RF technologies.

The work, funded under DARPA's Arrays at Commercial Timescales (ACT) program, was led by Columbia University electrical engineers Harish Krishnaswamy and Negar Reiskarimian and described in the April 15, 2016 issue of the journal Nature Communications.