Brain-Inspired Molecular Memristor Has Outstanding Memory Reconfigurability | Computer Science, Materials Science

The new reconfigurable memristor, or an electronic memory device, is based on a molecular system that can go from an on state to an off state at several discrete sequential voltages, according to a paper published in the journal Nature.

The abundant dendritic-synaptic interconnections between neurons in the neocortex incorporate complex logical structures enabling sophisticated decision-making that far surpasses any artificial electronic analog. The physical complexity goes far beyond existing circuit fabrication technologies: moreover, the network in a brain is dynamically reconfigurable, which provides flexibility and adaptability to changing environments. In contrast, advanced solid-state logic circuits are based on threshold switches that are wired to perform predefined logic functions. To advance the performance of logic circuits, Goswami et al. reinvented the fundamental elements of electronic circuits by expressing a complex logic in the properties of materials at the nanoscale. Image credit: National University of Singapore.

“This work is a significant breakthrough in our quest to design low-power calculations,” said Dr. A. Ariando, researcher at national university of singapore.

“The idea of ​​using multiple commutations in a single element takes inspiration from the functioning of the brain and fundamentally reinvents the strategy for designing a logic circuit.”

Unlike standard hard-wired circuits, the new memristor can be reconfigured using voltage to accommodate different computational tasks.

“This new discovery may contribute to the developments of advanced computing as a sophisticated in-memory computing approach to overcome the von Neumann bottleneck, a delay in computer processing seen in many digital technologies due to the physical separation of storage memory from a device’s processor, ”said Dr. Ariando.

The new memristor also has the potential to help design next-generation processing chips with improved computing power and speed.

“Similar to the flexibility and adaptability of connections in the human brain, our memory device can be reconfigured on the fly for different computational tasks by simply changing the applied voltages,” said Dr. Sreetosh Goswami, also of the ‘National University of Singapore.

“In addition, like the way nerve cells can store memories, the same device can also store information for future recovery and treatment.”

In their research, scientists conceptualized and designed a molecular system belonging to the chemical family of phenyl azo pyridines which have a central metal atom bonded to organic molecules called ligands.

“These molecules are like electron sponges that can offer up to six electron transfers resulting in five different molecular states”,

“The interconnectivity between these states is the key to the reconfigurability of the device,” said Dr. Sreebrata Goswami, researcher at the Indian Association for the Culture of Science.

The authors created a small electrical circuit consisting of a 40nm layer of molecular film sandwiched between an upper layer of gold and a lower layer of nanodisk infused with gold and indium tin oxide. .

They observed an unprecedented current-voltage profile when applying negative voltage to the device.

Unlike conventional metal oxide memristors which are turned on and off at a single fixed voltage, these organic molecular devices could switch between on-off states at multiple discrete sequential voltages.

Building on their research, the team used molecular memory devices to run programs for various real-world computational tasks.

As a proof of concept, the researchers demonstrated that their technology could perform complex calculations in one step and could be reprogrammed to perform another task the next instant.

An individual molecular memory device could perform the same computational functions as thousands of transistors, making the technology a more powerful and energy efficient memory option.

“The technology could first be used in portable devices, such as cell phones and sensors, and other applications where power is limited,” Dr Ariando said.


S. Goswami et al. 2021. Decision trees within a molecular memristor. Nature 597, 51-56; doi: 10.1038 / s41586-021-03748-0

Source link

Comments are closed.