image: Two ultrathin crystal layers, shown as blue and red atoms, are rotated relative to each other. When the layers bond strongly, their atoms rearrange into a square grid of defects in the crystal structure, highlighted here by golden lines. The spacing between the lines can be controlled by changing the twist angle between the layers. Illustration: John Mangeri.
Credit: Illustration: John Mangeri.
Researchers have fabricated two ultrathin crystal sheets that, when stacked and rotated relative to each other, create a square lattice of defects at their interface. These defects change how electrical charges are arranged, and by bonding and twisting the 10-nanometre-thick sheets, tiny electrical patterns emerge at the interface. These controllable arrangements could eventually prove useful in the electronics industry, although using them in a working device will require further research.
The study, led by the Technical University of Denmark (DTU) and described in a new paper in Nature, shows why strong bonding between the sheets matters for understanding their behaviour.
“We have spent years learning how to make high-quality sheets, put them together in the right position, bond them and twist them. Once you can do all of that, you can start looking at what happens where these two sheets meet,” says Professor Nini Pryds of DTU.
“So, the bonding is essential to understanding what happens between the layers.”
The starting point is a familiar pattern. Lay two mesh screens crosswise and rotate one slightly, and a larger repeating pattern will appear. This is a moiré pattern, and the same geometry appears when the regular atomic structures of two crystal sheets overlap at an angle.
Here, the researchers used sheets of strontium titanate (SrTiO3) – a hard, chemically stable material often used as a model oxide in electronics. Strong bonding during the twisting allowed the atoms at the interfaces between the sheets to adopt arrangements that fit together. The remaining mismatch gathered along narrow lines of distorted crystal, known as dislocations, forming a square grid.
The first indication of this rearrangement came from simulations by William Sandholt during his work as a Ph.D. student at DTU. Using a machine learning model, he calculated how atoms would move as the sheets bonded and twisted. The resulting grid gave the team a structure to look for in the laboratory.
Further modelling showed that distortions around the grid could slightly separate positive and negative charges. This separation, called polarisation, formed swirling patterns. Since controlling polarisation is relevant to memory and energy storage, there is a possibility that future devices could be designed based on these patterns.
“Changing the twist angle changes the spacing between the dislocations. So now, we have a way to control how small the structure becomes,” says joint first author William Sandholt.
“These dislocations could be used for memory applications, for example, and the grid could be tuned through the twist angle.”
Finding the buried grid
The first microscope images appeared to show an ordinary moiré pattern and offered no clear view of the grid predicted by the simulations. The team therefore modelled how the electron beam would pass through the material, then cut samples in different directions to examine the buried structure in cross-section.
“The electron beam goes through about ten nanometres of material, reaches the interface, and then goes through another ten nanometres. So, the image is influenced by the material on both sides of what we want to see,” explains John Mangeri, a postdoc at DTU and joint first author of the study.
“Right where the sheets met, we could see the repeating pattern, and it agreed almost perfectly with the theory. That was amazing.”
Electron microscopy by collaborators at the University of Antwerp in Belgium and the SuperSTEM Laboratory in the UK was central to identifying and characterising the grid. Further microscopy, interpreted alongside the simulations, supported the link between the rearranged atoms and the electrical patterns.
The predictions were also essential to deciding how to examine the samples, explains Nini Pryds:
“We would never have discovered the dislocations if the calculations had not predicted their existence. The finding highlights the power of combining theory and experiment, bringing together researchers who synthesized the ultrathin sheets, assembled them with a controlled twist, simulated their behavior, and examined the resulting material.”
The team now wants to explore whether the approach can produce useful electrical behaviour in SrTiO3 or other oxides. For possible memory technologies, the ability to control the spacing between the grid lines could be an advantage. Whether the grid lines can conduct electricity is a question that future experiments will need to answer.
-
Fact Box
The science behind the grid
Oxide membranes are freestanding crystalline sheets composed of oxygen and other elements. The team used strontium titanate (SrTiO₃), with each sheet approximately 10 nanometres thick. One nanometre is one billionth of a metre.
A moiré pattern is the larger repeating pattern produced when two regular grids overlap at an angle. Strong bonding allows atoms at the contact between the sheets to rearrange, creating an ordered network of line defects called screw dislocations.
Polarisation means a small separation of positive and negative charges. Uneven distortion around the grid can create polarisation via a phenomenon called flexoelectricity. Models predict swirling polarisation patterns, known as polar vortices. The microscopy signal supports the reconstructed structure; interpreting its polar character also relies on modelling.
The team observed dislocation networks at twist angles of 5, 7.5 and 10 degrees. The pronounced network was absent at 20 degrees, because the grid spacing becomes too small. Calculations also predict electronic states concentrated near the grid. Electrical conduction along its lines has not been demonstrated, yet.
Study and contacts
“Designing dislocation-driven polar vortex networks in twisted perovskites” by W. Sandholt, N. Gauquelin, J. Mangeri and colleagues. Sandholt, Gauquelin and Mangeri contributed equally.
Participating institutions: Technical University of Denmark; University of Antwerp; SuperSTEM Laboratory; University of York; Politecnico di Bari; University of Leeds; University of Minnesota Twin Cities.
Funding: This work was supported by the European Research Council (ERC) Advanced Grant NEXUS and the Villum Investigator Grant EUREKA.
Journal
Nature
Article Title
Designing dislocation-driven polar vortex networks in twisted perovskites
Article Publication Date
18-Nov-2026
COI Statement
The authors do not declare any competing interests.