The central idea is deceptively simple. Instead of assembling flat graphene oxide sheets, the researchers first induced controlled folds on the surface of multilayer graphene oxide flakes, then stacked these folded units into a film. When the flakes are compressed during film formation, the folds on neighbouring sheets clasp onto one another, creating a lamellar architecture in which adjacent layers are mechanically engaged rather than merely stacked. The team calls this a fold-clasping lamellar structure, and it fundamentally changes how force travels through the material when it is stretched.

To make the folded flakes, the researchers modified the oxidation process used to prepare graphene oxide, carrying out reactions at reduced temperatures ranging from minus 5 to minus 25 degrees Celsius. Temperature proved to be a tuning knob for the fold geometry: colder reactions produced thicker flakes with denser and wider surface folds, which the team characterised in detail using atomic force microscopy. The folded flakes, which the researchers abbreviate SFGOs, retain the layered crystallinity of graphite while developing a heterogeneous core-shell character, with the folded surface regions differing in oxidation state from the flake interior, as revealed by Raman spectroscopy and aberration-corrected transmission electron microscopy.

Assembling these folded flakes into films was accomplished by blade coating, a scalable processing route that can produce sheets tens of centimetres across. During coating and subsequent drying, the surface folds of adjacent flakes interpenetrate and clasp, locking the lamellar stack together. Chemical reduction then removes most of the oxygen-containing groups, converting the folded graphene oxide film into a reduced graphene film with a high proportion of sp2-bonded carbon. X-ray diffraction and wide-angle X-ray scattering confirmed that the films are well aligned and densely packed, with the folded architecture preserved through processing.

The mechanical performance of the resulting films is the headline result. A tensile strength of 2.9 gigapascals places these films among the strongest macroscopic graphene assemblies reported, while the toughness of 84.5 megajoules per cubic metre indicates the material can absorb enormous amounts of energy before failing. Perhaps most striking is the elastic strain of 5 percent, meaning the film can stretch by that fraction and spring back without permanent deformation. For comparison, most lamellar graphene films fail at strains well below this level, and their stress-strain curves show early plastic yielding rather than clean elastic recovery.

The secret lies in how stress is transferred between layers. In a conventional stacked film, tension applied to one sheet is transmitted to its neighbours through shear at the interfaces, a mechanism described by the classical shear-lag model. That works well at low loads, but the interfaces eventually slip, and the film deforms permanently. In the fold-clasping architecture, the researchers identified a fundamentally different mode of force transfer. At small strains, the clasped folds are pulled against each other, but as deformation increases the force chains undergo a pull-to-push transition: the geometry of the interlocked folds converts the tensile loading into compressive contact between fold surfaces, which presses the layers together and increases frictional interlock rather than allowing slip.

The team formalised this insight as a shear-lock stress transfer strategy, supported by coarse-grained molecular dynamics simulations of a single folded sheet under tension. The simulations tracked three regimes during loading: an initial state, a maximum-stress state, and an interlock failure state. They showed that stress concentrates at the head-to-head and head-to-tail junctions of the folds, where the clasping geometry redistributes load and delays catastrophic failure. In-situ wide-angle X-ray scattering during tensile testing confirmed that the films recover their original layer orientation after unloading, direct evidence of the elastic reversibility that underpins the 5 percent elastic strain.

The concept of using topology and geometry to control mechanical behaviour draws on a growing body of work on knots, tangles and entangled structures, where the way elements wrap around each other determines how forces propagate. The Zhejiang and Tsinghua teams had previously shown that interlayer entanglement can strengthen nacre-inspired materials, and the new work extends that topological thinking to two-dimensional building blocks. Rather than changing the chemistry of the sheets or the bonding between them, the approach reshapes the building blocks themselves, treating each folded flake as a modular unit whose spatial conformation is designed before assembly.

This modular design principle is what distinguishes the work from earlier strategies that improved graphene films by refining stacking order, alignment or interfacial bonding. Those approaches, which include vacuum-assisted self-assembly, shear-flow alignment and intercalation-modulated plasticization, all operate on flat sheets and improve the interfaces between them. The fold-clasping strategy instead modifies the conformation of the individual units, adding a third design axis that had been exploited in other material classes but remained elusive for two-dimensional sheets. Because the folds are created during flake synthesis and preserved through scalable coating, the approach is compatible with large-area film production.

The practical implications reach into lightweight structural components, flexible electronics and any application where thin films must survive repeated mechanical loading. A film that is simultaneously strong, tough and elastically resilient can absorb impacts, tolerate bending cycles and carry sustained loads without accumulating damage. The researchers suggest that the fold-clasping principle should generalise to other two-dimensional materials beyond graphene, opening a route to a broader family of architectured films in which the geometry of the nanoscale building blocks is engineered as deliberately as their chemistry. As two-dimensional materials move from laboratory curiosities toward structural applications, controlling how they clasp together may prove as important as controlling how they are made.

Subject of Research: Fold-clasping lamellar architecture in graphene oxide films for superior mechanical performance

Article Title: Clasping folds enable superior elastic strength of graphene films

Article References: Wang, L., Liao, Y., Huang, J., Shen, C., Chen, L., Li, X., Ding, D., Zhang, M., Peng, L., Li, K., Li, P., Liu, Y., Xu, Z., Pang, K., Xu, Z., & Gao, C. (2026). Clasping folds enable superior elastic strength of graphene films. Nature, 658(8136), 643-649. https://doi.org/10.1038/s41586-026-11109-y

Image Credits: AI Generated

DOI: 10.1038/s41586-026-11109-y

Keywords: graphene, graphene oxide, mechanical properties, lamellar structures, toughness, elastic strain, fold-clasping, two-dimensional materials, molecular dynamics, nacre-inspired materials, stress transfer, structural films

Chirality, the property of an object that cannot be superimposed on its mirror image, is usually associated with the handedness of molecules such as amino acids or DNA. Pasteur famously separated the left- and right-handed crystals of tartaric acid in 1848, and chemists have long known that achiral molecules can occasionally crystallize into chiral forms. But deliberate, controllable twisting in inorganic crystals remains rare, and in layered materials it is rarer still. Previous examples have relied on screw dislocations, which impose an Eshelby twist on growing whiskers, or on growth on curved, non-Euclidean surfaces that force layered crystals into supertwisted spirals. The new work demonstrates a different mechanism altogether, one rooted in the internal architecture of the crystal lattice itself.

The materials at the heart of the study are two-dimensional organic-inorganic hybrid perovskites of the general form (organic)2PbBr4, in which sheets of corner-sharing lead-bromide octahedra are separated by layers of organic cations. These cations are not passive spacers. In the compound (FCA3)2PbBr4, where FCA3 is a fluorinated aromatic cation, the organic layers form hydrogen-bonded networks that the authors describe as hydrogen-bonded organic frameworks embedded within the inorganic lattice. Using single-crystal X-ray diffraction, the team determined the structures at 150 K and 294 K, with crystallographic data deposited at the Cambridge Crystallographic Data Centre under deposition numbers 2485415, 2485416 and 2485417. The structures revealed that the hydrogen-bonded organic networks and the inorganic lead-bromide sheets want to deform in incompatible ways when the crystal grows.

That incompatibility is the essence of geometric frustration. As a ribbon grows from a supersaturated solution, mismatched strains accumulate between the organic and inorganic sublattices. When the strain is released, the ribbon relieves it by bending and twisting, exactly as a seed pod, made of two layers with different shrinkage directions, coils as it dries. The researchers combined brightfield and cross-polarized optical microscopy, scanning electron microscopy, transmission electron microscopy with selected-area electron diffraction, and molecular dynamics simulations to connect the atomic-scale mismatch to the macroscopic shape. The simulations, carried out by Zhichen Nian under the supervision of Brett M. Savoie, reproduced the spontaneous deformation of the ribbons and confirmed that incompatible elasticity within the lattice, rather than any external influence, drives the twisting.

The scale of the phenomenon is remarkable. The team documented ribbons roughly one micrometer wide and about a millimeter long carrying as many as 86 complete twists along their length, grown simply by cooling a supersaturated solution at 2.5 degrees Celsius per minute from 75 to 55 degrees Celsius. Crucially, microscopy under brightfield, photoluminescence and cross-polarized modes showed that the twisted microribbons are free of internal boundaries, meaning each ribbon is a single crystal that has deformed elastically rather than a composite of misoriented domains. Statistical and mechanical analyses of hundreds of ribbons showed that the shape is governed by intrinsic elastic strain and geometric constraints, and that tuning those parameters allows a continuous transition from twisted ribbons to fully helical ones.

Perhaps the most consequential discovery concerns handedness. Because the underlying molecules are achiral, one might expect equal numbers of left- and right-handed ribbons, yielding no net optical signature. Instead, the researchers found that the direction of twisting is controlled by crystal symmetry and by growth kinetics. By adjusting the cooling rate, for example to a slow 0.044 degrees Celsius per minute, they could bias nearly all ribbons in a bulk sample to twist in the same direction, producing centimeter-sized forests of ribbons with a single handedness. Ensemble-averaged circular dichroism measurements on these chirality-pure samples showed genuine chiroptical activity. The team took particular care to isolate the true signal: by rotating the sample about the light axis and flipping it, they removed artifacts arising from linear birefringence and linear dichroism, leaving the circular dichroism that arises from the coupling between electronic and magnetic transition dipoles.

The symmetry analysis explains why the effect appears in some compositions and not others. The inorganic [PbBr4]2- lattices contain specific symmetry elements, such as 21 screw axes or inversion centers, positioned at precise crystallographic sites, and the morphology of the growing crystal reflects these elements. When the team doped a second cation, BrCA3, into the (FCA3)2PbBr4 lattice, the doped crystals adopted the triclinic space group P-1 and their hydrogen-bonded organic networks aligned in a configuration resembling that of the pure (BrCA3)2PbBr4 compound. Partial occupancy refinement, made possible by the strong scattering contrast between fluorine and bromine, established a doping level of about 18 percent in one crystal, corresponding to the composition (FCA3)1.821(BrCA3)0.179PbBr4. The doped crystals showed markedly different twisting behavior, demonstrating that molecular doping offers a practical handle for tuning the frustration and hence the final morphology.

The work situates layered perovskites within a broader family of twisting crystals that includes benzamide, hippuric acid, aspirin and cholesterol monohydrate, all of which form helicoidal or twisted habits through growth-actuated bending. It also echoes recent advances in van der Waals materials, where chiral twisted nanowires and helical crystals with discretized Eshelby twist have been reported, and in nanoplatelet systems where ligand-induced curvature mismatches control polymorphism and chirality. What distinguishes the new result is the mechanism: rather than dislocations, external templates or curved substrates, the twist emerges from the interplay of two chemically distinct, hydrogen-bonded and ionic sublattices within a single crystal. The authors suggest this provides a general strategy for programming chiral shapes into layered two-dimensional materials through molecular design.

The practical implications could extend well beyond crystallography. Chiral perovskites are already of intense interest for spintronics, where chiral-induced spin selectivity has enabled room-temperature spin light-emitting diodes, and for photovoltaics, where chiral-structured heterointerfaces have been shown to improve the durability of solar cells. A material that can be grown in bulk with controlled handedness, using nothing more than cooling rate and molecular doping, would be an attractive platform for such technologies, as well as for circularly polarized light detection and chiral optoelectronics. Because the twisting is an elastic response baked into the lattice, it also raises the possibility of dynamic behavior, analogous to the reversible twisting motions observed in other molecular crystals and to light-powered flagella-like motion in photomechanical microwires.

For now, the study stands as an elegant demonstration that macroscopic chirality can be encoded in the geometry of interactions at the molecular scale. A crystal built from mirror-symmetric parts, given the right internal mismatch and the right growth conditions, will coil itself into a helix with a preferred hand, and that hand can be read out optically from an ensemble of thousands of ribbons. The researchers, whose work was supported by the US Department of Energy Office of Basic Energy Sciences and the US National Science Foundation, describe their results as a route to controlling chiral morphologies in layered two-dimensional materials. If the principle generalizes across the rich family of hybrid perovskites and related layered compounds, the humble twisted ribbon may become a standard building block in the designer materials of the coming decade.

Subject of Research: Geometric frustration driving spontaneous chiral twisting in layered halide perovskite nanoribbons

Article Title: Geometric frustration in morphologically chiral nanoribbons of layered perovskites

Article References: Shao, W., Nian, Z., Lu, Y., Yang, H., Yu, Y., Savoie, B. M., & Dou, L. (2026). Geometric frustration in morphologically chiral nanoribbons of layered perovskites. Nature Synthesis. https://doi.org/10.1038/s44160-026-01167-5

Image Credits: AI Generated

DOI: 10.1038/s44160-026-01167-5

Keywords: layered perovskites, geometric frustration, chirality, nanoribbons, two-dimensional materials, hydrogen-bonded organic frameworks, circular dichroism, molecular doping, crystal growth, elastic strain, chiroptical activity, metal-halide perovskites