A Georgia Tech-led research team with funding from the U.S. Defense Advanced Research Projects Agency (DARPA) is developing a new approach to computer logic that could dramatically reduce the amount of energy required to process information.
The project explores whether nanoscale mechanical strain can be used alongside electrical signals to perform computing operations. The researchers say the approach could ultimately deliver energy efficiency improvements of as much as 100 times compared with today’s state-of-the-art technologies, although that target remains a research objective rather than a demonstrated commercial result.
The work is being led by Asif Khan, an associate professor in Georgia Tech’s School of Electrical and Computer Engineering. The research is being conducted through DARPA’s Fast and Curious program to the tune of $10.6 million in funding.
The research is targeting new computing devices, materials and architectures capable of dramatically reducing energy consumption while remaining compatible with advanced semiconductor manufacturing.
The technology, called Mechanically Amplified Ferroic-Actuated (MEFA) Logic, takes a different approach from conventional CMOS logic. Modern processors rely primarily on the movement and control of electrical charge through billions of transistors. MEFA instead incorporates mechanical movement at the nanoscale.
The system uses ferroic materials that expand or contract slightly when an electrical voltage is applied. That mechanical movement can then be amplified and transferred to a nearby semiconductor channel, changing the flow of electrical current and effectively controlling a logic state.
That makes the materials a potential alternative to CMOS. CMOS has been around as long as computers have, but it is prone to power leakage and heat buildup even as transistors get smaller. Reducing power consumption is especially important for artificial intelligence, data centers, edge computing and defense systems.
DARPA’s Fast and Curious program is specifically investigating computing approaches that could benefit and run more efficiently under new technologies rather than conventional CMOS while still having a path toward integration with modern semiconductor manufacturing.
The research team plans to progress from individual MEFA devices toward increasingly complex logic circuits. The ultimate goal is to demonstrate a practical computing technology that combines substantially lower energy consumption with the speed and functionality needed for real-world computing.
