Imagine a future where your phone, computer or even a tiny wearable device can think and learn like the human brain – processing information faster, smarter and using less energy.
A single ferroelectric domain wall memristive device: Left - interelectrode gap straddled by a single domain wall, which is pinned strategically at a few locations (shown by arrows) at the film's surface. Right – Electronic transport characteristics of this wall and the device. Image Credit: Dr Pankaj Sharma, Flinders University
A breakthrough approach developed at Flinders University and UNSW Sydney brings this vision closer to reality by electrically ‘twisting’ a single nanoscale ferroelectric domain wall.
The domain walls are almost invisible, extremely tiny (1-10 nm) boundaries that naturally arise or can even be injected or erased inside special insulating crystals called ferroelectrics. The domain walls inside these crystals separate regions with different bound charge orientations.
More importantly, these tiny boundaries despite being embedded in insulating crystals, can acts as channels for regulating electron flow, and thus are capable of storing and processing information like in a human brain, says Flinders University senior lecturer in physics Dr Pankaj Sharma, lead and corresponding author in a new American Chemical Society (ACS) article.
Why does this matter? Devices mimicking the human brain allow for faster processing of vast amounts of information while using far less energy compared to existing digital computers, in particular, for tasks such as image and voice recognition, the researchers say.
“With this new design, these ferroelectric domain walls in crystalline ferroelectric materials are poised to power a new generation of adaptable memory devices, bringing us closer to faster, greener and smarter electronics,” says Dr Sharma. “Our results reaffirm the promise of ferroelectric domain walls for brain-inspired neuromorphic and in-memory computing applications based on integrated ferroelectric devices.”
“In our research, a single ferroelectric domain wall has been controllably injected and engineered to mimic memristor behaviour. By applying electric fields, we carefully manipulate the shape and position of this single wall, causing it to bend and warp.”
“This controlled movement leads to changes in the wall’s electronic properties, unlocking its ability to store and process data at different levels.”
The new study reveals how ferroelectric domain walls straddling two terminal devices (see image below) can function as "memristors" – devices that can store information at varying levels and remember the history of its electrical activity – similar to synapses in a human brain.
Coauthor UNSW Professor Jan Seidel, says “the key lies in the interplay between the wall’s surface pinning (where it’s fixed) and its freedom to twist or warp deeper within the material.
“These controlled twists create a spectrum of electronic states, enabling multi-level data storage, and eliminates the need for repetitive wall injection or erasure, making the devices more stable and reliable,” he says.
Using advanced microscopy and theoretical phase field modelling, this research uncovers the physics behind these warping-induced electronic transitions at the domain walls.
Coauthor UNSW Professor Valanoor Nagarajan adds: “These new highly reproducible and energy-efficient domain wall devices could revolutionise neuromorphic computing, the brain-inspired systems that promise to reshape artificial intelligence and data processing.”
The article, Ferroelectric Domain Wall Warp Memristor (2024) by Pankaj Sharma, Chi-Hou Lei, Yunya Liu, Daniel Sando, Qi Zhang, Nagarajan Valanoor and Jan Seidel, has been published in journal ACS Applied Materials & Interfaces DOI: 10.1021/acsami.4c16347.
Acknowledgements: The study was supported by funding from Australian Research Council Discovery Projects (DP240102137, DP240100238) and Flinders University grants. The nanoscale device patterning is supported by the Australian National Fabrication Facility (ANFF, UNSW). This research was also partially supported by the ARC Centre of Excellence in Future Low Energy Electronics Technologies (FLEET). Liu acknowledges the support from the National Natural Science Foundation of China (12172318) and the Science and Technology Innovation Program of Hunan Province (2022RC3069).
Potential conflict of interest: Dr Pankaj Sharma, Professor Jan Seidel and Professor Valanoor Nagarajan declare the filing of the provisional patent (priority date 31/07/2024) related to this research.