News

Brain-Inspired Electronics Offer Energy and Cost Benefits to AI and IoT Applications

April 27, 2020 by Luke James

Organic memory devices developed by American researchers reportedly show promise for flexible, wearable, and personalized computing devices and other electronics.

These so-called “brain-inspired” electronics, which use organic memristors as opposed to inorganic ones, could offer a functional and cost-effective platform for flexible, wearable, and personalized computing technologies.

Memristors made from organic materials are favored for their theoretical characteristics, such as spatially uniform switching, tuneable molecular functionalities, and ultra-low switching energies. 

However, despite research efforts that go back almost two decades, the performance and general understanding of organic memristors leave a lot to be desired, and one that is suitable for industrial applications is yet to emerge due to problems with stability, scalability, speed, reproducibility, and more. 

 

Functionally Promising and Cost-Effective Memristors

In a paper published in Applied Physics Reviews, American researchers discuss a new family of organic memristors that could be industrially competitive thanks to characteristics that include exhibit cyclability, prolonged retention, low switching energy, sufficient switching voltage, cyclability, and rise time of below 30 nanoseconds. 

"It is important for us to understand that the computing platforms of today will not be able to sustain at-scale implementations of AI algorithms on massive datasets," said Thirumalai Venkatesan, one of the authors of a paper published in Applied Physics Reviews.

He went on to add that so-called “brain-inspired electronics” that feature organic memristors capable of both storing data and performing computation could offer a functionally promising and cost-effective platform. In terms of functionality, memristors are analogous to neurons and so this makes them prime candidates for these brain-inspired electronics such as computing platforms

 

A graphic depicting device structure at a molecular level.

A graphical depiction of a molecular device structure at a molecular level. The gold particles on the bottom electrode enhance the field, enabling low energy operation of the device. Image credited to Sreetosh Goswami, Sreebrata Goswami, and Thirumalai Venky Venkatesan

 

A ‘New Generation’ of Organic Memristors

The organic memory devices, which the researchers refer to as the “new generation of organic memristors,” have been developed based on metal azo complex devices, pioneered by Sreebata Goswami, an author on the team’s paper and a professor at the Indian Association for the Cultivation of Science in Kolkata. 

For a long time, oxides were the leading candidate as the optimum material for memristors and while lots of different ones have been proposed, none have thus far been successful.

 

The Challenges of Developing Organic Memristors

Goswami notes that over the last 2 decades there have been many attempts to create organic memristors, but none have shown promise due to problems with stability and reproducibility. “At a device level, we are now able to solve most of these problems,” he added. 

The team’s next challenge is producing these organic memristors at scale. Goswami foresees them as suitable for a wide range of wearable and implantable technologies, going so far as to claim that eventually they could be used to build a “body net”—a series of wireless sensors that stick to the skin for health monitoring.