Tel Aviv University Demonstrates Nanoscale Graphene Switch That Flips Stacking Order With Less Than One Femtojoule of Energy
Researchers achieved reversible switching between graphene polytypes at 30-nanometer scale using sub-nanonewton forces and sub-femtojoule energy, opening a path to slidetronic memory and neuromorphic computing.
Overview
A team at Tel Aviv University has demonstrated a method for switching the internal stacking arrangement of graphene layers at the nanometer scale, using energy inputs orders of magnitude lower than those required by existing memory technologies. The work, published in Nature Nanotechnology, achieves fully reversible structural transformations in graphene islands as small as 30 nanometers in diameter, consuming less than one femtojoule per switching event.
The technique, which the researchers call a superlubricant array of polytypes (SLAP), could lay the groundwork for a new class of ultra-efficient electronic memory, sensors, and neuromorphic computing devices built on the emerging principle of slidetronics — controlling material properties by physically sliding atomic layers rather than breaking and reforming chemical bonds.
What We Know
Graphene, a single-atom-thick sheet of carbon arranged in a hexagonal lattice, behaves differently depending on how its layers are stacked. The two main stacking arrangements, known as Bernal and rhombohedral polytypes, exhibit distinct electronic properties including differences in intrinsic polarization, orbital magnetism, and even superconductivity. Switching between these arrangements has long been proposed as a mechanism for data storage, but previous attempts required micrometer-scale domains and micronewton-level forces, making practical devices unrealistic.
The Tel Aviv team, led by Prof. Moshe Ben-Shalom of the School of Physics and Astronomy alongside Prof. Michael Urbakh and Prof. Oded Hod of the School of Chemistry, solved this scaling problem with a novel device architecture. They inserted an intentionally misaligned spacer layer, patterned with nanometer-scale cavities, between a pair of aligned graphene bilayers. Within each cavity, the bilayers sag into contact and form stable single-domain polytypes. Outside the cavities, the layers rest on superlubric — nearly frictionless — interfaces, allowing them to slide freely with ultralow friction.
The result is a system where boundary solitons nucleated at the edges of each island can glide across the cavity to convert the stacking configuration from Bernal to rhombohedral or vice versa. The researchers demonstrated this switching down to 30-nanometer-scale islands using lateral shear forces of less than one nanonewton and energy consumption below one femtojoule per event.
“We simply slide atomic layers over one another — a natural process that is much faster and more efficient” than breaking and rebuilding chemical bonds, Prof. Ben-Shalom said.
The experimental work was directed by Dr. Nirmal Roy and Dr. Pengua Ying, with contributions from Simon Salleh Atri, Yoav Sharaby, Noam Raab, and Dr. Youngki Yao. The collaboration also involved researchers from Japan’s National Institute for Materials Science, who contributed the hexagonal boron nitride substrates used in the devices.
What We Don’t Know
The research demonstrates the fundamental switching mechanism at laboratory scale, but the path from a proof-of-concept device to commercial slidetronic memory remains unclear. The team has not yet reported switching speeds, endurance over millions of cycles, or performance at the temperatures encountered in real electronic systems.
It is also uncertain how SLAP devices would integrate with existing semiconductor manufacturing processes. Graphene device fabrication still relies on mechanical exfoliation and precise manual alignment of layers, techniques that do not translate easily to the wafer-scale production needed for commercial electronics.
The energy advantage over conventional memory technologies is striking on a per-event basis, but the total system-level efficiency — including the circuitry needed to apply and detect the shear forces — has not been characterized.
Analysis
The significance of this work lies in resolving a longstanding mismatch between the theoretical promise and practical feasibility of slidetronic devices. Previous demonstrations of polytype switching operated at length scales and force levels that precluded integration into nanoscale electronics. By shrinking the switching domain to 30 nanometers and reducing the required energy to sub-femtojoule levels, the Tel Aviv team has brought slidetronics into a regime compatible with modern device dimensions.
The neuromorphic computing implications are particularly notable. Because the cavities in a SLAP device can mechanically interact with one another through the shared elastic medium, networks of such elements could in principle implement the kind of coupled, analog computation that characterizes biological neural networks. This mechanical coupling distinguishes slidetronics from purely electrical approaches to neuromorphic hardware.
The field of two-dimensional material electronics has produced many laboratory demonstrations that have struggled to cross the threshold into practical technology. Whether slidetronic memory follows that pattern or breaks it will depend on progress in scalable fabrication and the development of readout mechanisms that can detect stacking changes quickly and reliably at the nanometer scale.