Imagine a future where your smartphone, computer, or even a compact wearable gadget can think and learn like the human brain — processing information more quickly, intelligently, and with reduced energy consumption. A groundbreaking method is bridging this vision with reality by electrically ‘twisting’ a single nanoscale ferroelectric domain wall.
Imagine a future where your smartphone, computer, or even a compact wearable gadget can think and learn like the human brain — processing information more quickly, intelligently, and with reduced energy consumption.
A new innovative technique developed at Flinders University and UNSW Sydney is bringing this dream closer to life by electrically ‘twisting’ an individual nanoscale ferroelectric domain wall.
These domain walls are nearly imperceptible, extremely small (1-10 nm) boundaries that naturally form or can be introduced or removed within specific insulating crystals known as ferroelectrics. The domain walls within these materials divide areas that have different orientations of bound charges.
What’s critical is that these minute boundaries, even though they are situated in insulating materials, can function as channels that regulate the flow of electrons. This gives them the ability to store and process information much like the human brain, explains Dr. Pankaj Sharma, a senior lecturer in physics at Flinders University and the lead and corresponding author of a recent article published by the American Chemical Society (ACS).
Why is this significant? Brain-like devices can process massive volumes of information much more quickly while consuming significantly less energy than current digital computers, particularly for tasks such as image and voice recognition, according to the researchers.
“With this innovative design, these ferroelectric domain walls in crystalline ferroelectric materials are set to power a new wave of adaptable memory devices, bringing us nearer to faster, greener, and smarter electronics,” Dr. Sharma states. “Our findings support the potential of ferroelectric domain walls for neuromorphic computing and in-memory computing applications based on integrated ferroelectric devices.”
“In our study, we have successfully injected and engineered a single ferroelectric domain wall to replicate memristor behavior. By applying electric fields, we can systematically adjust the shape and position of this single wall, causing it to bend and distort.”
“This controlled movement alters the electronic properties of the wall, enhancing its capacity to store and process information at various levels.”
The latest research illustrates how ferroelectric domain walls spanning two terminal devices (see image below) can operate as “memristors” — devices capable of storing information at multiple levels and retaining the history of electrical activity, similar to synapses in the human brain.
Coauthor Professor Jan Seidel from UNSW states, “The critical factor lies in the interaction between the wall’s surface pinning (where it’s anchored) and its ability to twist or warp more deeply within the material.”
“These managed twists create a range of electronic states, allowing for multi-level data storage, and eliminating the need for repetitive wall injection or erasure, thus enhancing the stability and reliability of the devices,” he explains.
Through advanced microscopy and theoretical phase field modeling, this study unveils the physics behind the electronic transitions driven by warping at the domain walls.
Coauthor Professor Valanoor Nagarajan from UNSW adds: “These new, highly reproducible, and energy-efficient domain wall devices have the potential to revolutionize neuromorphic computing, the brain-inspired systems that aim to transform artificial intelligence and data processing.”
