The simulations evaluated models with twin spacings of 4a, 6a, 8a and 12a. The abstract reports that the 4a model accumulated more heat, attributing this result to its excessive interfacial density, while the 8a and 12a models displayed less stable plastic flow and poorer surface quality. The researchers assessed cutting force, friction coefficient, shear strain, stress distribution, temperature evolution, surface morphology, material removal, phase transformation and dislocation activity across the simulations.

Twin inclination also shaped the cutting response. Simulations at 30° and 45° were associated with more efficient material removal and lower surface roughness. By contrast, the abstract states that twin orientations of 0° and 90° acted as strong barriers to dislocation motion, increasing cutting resistance and surface damage in the models. Together, the reported results associate twin-boundary architecture with differences in plastic deformation, thermal evolution, material removal and the condition of the simulated surface.

The findings offer atom-scale guidance for microstructural design intended to improve the machinability and surface quality of CoCrFeMnNi high-entropy alloys. However, the available source provides only the study abstract and a graphical-abstract preview. The full methodology, model parameters and quantitative results therefore cannot be assessed from the supplied evidence. The reported outcomes arise from molecular dynamics simulations; on that basis, they do not demonstrate equivalent performance in physical machining experiments or industrial production.

Subject of Research: Simulated effects of twin architecture on the orthogonal cutting of CoCrFeMnNi high-entropy alloys

Article Title: Investigation on surface morphology and deformation mechanisms of twin lamellar CoCrFeMnNi high entropy alloys during orthogonal cutting

Article References: Vu, T.-N., Pham, V.-T., Nguyen, T.-P., Nguyen, V.-D., & Fang, T.-H. (2026). Investigation on surface morphology and deformation mechanisms of twin lamellar CoCrFeMnNi high entropy alloys during orthogonal cutting. Journal of Materials Science. https://doi.org/10.1007/s10853-026-13755-z

Image Credits: AI Generated

DOI: 10.1007/s10853-026-13755-z

Keywords: CoCrFeMnNi, high-entropy alloys, molecular dynamics, orthogonal cutting, twin boundaries, surface morphology, dislocation activity, machinability, plastic deformation, Investigation, surface, morphology

That gap may now be closing. Researchers at Zhejiang University have developed AtomDiT, a machine learning framework that transforms a single high-resolution electron micrograph into a complete digital twin of a dislocation system. The twin does not merely reproduce the picture; it simultaneously records the type and position of each dislocation, the surrounding stress field, and the distribution of light elements such as hydrogen, none of which are directly visible in the original image. The work, published in National Science Review, is demonstrated across pure metals and complex alloys, including dual-phase systems and grain boundaries decorated with segregated hydrogen, and it uncovers how stress and chemistry together govern dislocation motion. In effect, the framework turns one static snapshot of atoms into a multidimensional portrait of the mechanics hidden inside a crystal.

The key insight behind AtomDiT is that the information needed to reconstruct the invisible fields is already encoded in the visible structure. The framework relies on the local atomic environment, meaning the geometric arrangement of atoms in the immediate neighborhood of each atomic site. Dislocations distort the crystal lattice in highly specific and characteristic ways: the planes of atoms bend around the defect core, interplanar spacings stretch and compress, and the symmetry of the local neighborhood changes in patterns that depend on the dislocation’s type and position. Because these distortions are caused by the stress field itself, the local arrangements of atoms contain information not only about the dislocations but also about the mechanical and chemical fields surrounding them. AtomDiT learns these relationships from large sets of atomistic simulations and then applies that learned knowledge to experimental images, without requiring any prior information about loading conditions or the detailed composition of the material being studied.

The technical pipeline works in a strikingly modular way. To identify dislocations, AtomDiT first assigns each atom a descriptor based on its distances to nearby neighbor atoms, a compact numerical fingerprint of the local environment. A set of trained neural networks then predicts, for each atom, how far that atom sits from each possible dislocation type. The design resembles a bank of sensors, each one tuned to detect a different kind of defect signal buried in the lattice geometry. From these per-atom predictions, the framework builds probability maps over the entire imaged region and then locates dislocation cores using unsupervised clustering, letting the data itself reveal where the defects are rather than imposing a predefined template. Crucially, the same local descriptors that identify the dislocations are also used to reconstruct the stress and chemical distributions around them, so all components of the digital twin emerge from a single unified representation of the atomic environment.

Building a model that generalizes across real materials required an unusually broad training database. The researchers trained AtomDiT on atomistic configurations spanning several crystal structures, the common dislocation types found in engineering metals, grain boundaries, cracks, triple junctions, and full polycrystalline networks, all sampled under a wide range of deformation states. This diversity ensures that when the model confronts an experimental image, it has already seen the relevant lattice distortions in simulation. For complex alloys with severe lattice distortion, where the local environments can deviate strongly from those of pure metals, the framework employs transfer learning, allowing the model to adapt to a new material system with only a small amount of additional data. In tests across pure metals and multicomponent alloys, the resulting digital twin located dislocation cores with an accuracy well below a single atomic spacing, meaning the predicted positions of the defects are reliable at the finest scale that atomic imaging can resolve.

The real test of any digital twin is whether it reveals physics that was previously out of reach, and the researchers demonstrated this in two demanding scenarios. The first involved a dual-phase alloy, a material composed of two distinct crystalline phases separated by an interface. Dislocations moving through such a material must eventually confront a phase boundary, and whether they pass through it or pile up against it has enormous consequences for the alloy’s strength and ductility. Using AtomDiT, the digital twin tracked a dislocation as it approached and crossed a phase boundary, simultaneously recording the transformation of its dislocation type and the evolution of the stress field around it. The twin showed that a smooth, broadly distributed stress field allows the dislocation to transmit through the interface, while a sharply fluctuating stress field blocks its passage. This connection between the spatial character of the stress field and the fate of a moving defect provides a direct, atomically resolved explanation of how interfaces control plastic flow in multiphase metals.

The second scenario addressed one of the most consequential problems in materials science: hydrogen embrittlement. Hydrogen atoms are so small that they slip into the spaces between metal atoms and segregate at grain boundaries, the interfaces between individual crystalline grains, where they are notoriously difficult to detect and even harder to correlate with mechanical behavior. AtomDiT’s digital twin mapped hydrogen concentration, dislocation character, and stress together in hydrogen-segregated grain boundaries, producing a combined chemical and mechanical picture of the same nanoscale region. The analysis revealed that hydrogen changes dislocations from highly mobile types into less mobile ones and creates a tensile stress peak at the boundary itself. Atomistic simulations confirmed that the energy barrier for a dislocation to transmit across the boundary increases with hydrogen content, directly connecting the chemical picture reconstructed from the micrograph to the mechanical degradation that hydrogen is known to cause.

What makes this approach conceptually powerful is the direction of the information flow. Traditional multiscale modeling struggles to connect the atomistic scale, where defects live and interact, with the continuum scale, where engineers describe plasticity in terms of stresses and strains. The authors note that the digital twin can be built from real atomic structures obtained from experimental measurements, which opens a path to feed experimental data directly into larger-scale plasticity models. Instead of relying on idealized simulated structures as the starting point for multiscale simulations, researchers can now begin from the actual atomic configuration of a real material, complete with its measured defect content and reconstructed fields. This bridges atomic-scale defect dynamics with the behavior of engineering materials, grounding continuum-level predictions in experimentally verified atomic reality.

The implications extend across the materials economy. Alloys designed for aircraft, pipelines, nuclear reactors, and hydrogen-powered vehicles all fail, when they fail, through the motion and interaction of dislocations, often accelerated by environmental species like hydrogen. A tool that can extract defect identity, stress, and chemistry from a single micrograph turns routine atomic-resolution imaging into a quantitative probe of the mechanics that were previously invisible. By turning a single picture into a multidimensional portrait of structure, stress, and composition, AtomDiT offers a new way to see the invisible machinery inside metals, and with it, a new route to understanding, predicting, and ultimately preventing crystal deformation and damage.

Subject of Research: Machine learning reconstruction of dislocation structure, stress, and chemical fields from atomic-resolution electron micrographs

Article Title: How to see the atomic-scale invisibles via digital twinning?

Article References: How to see the atomic-scale invisibles via digital twinning?. (n.d.). Original publication

Image Credits: AI Generated

DOI: Not provided

Keywords: dislocations, digital twin, machine learning, electron microscopy, hydrogen embrittlement, stress fields, plastic deformation, grain boundaries, dual-phase alloys, atomistic simulation, Zhejiang University, National Science Review

The material at the heart of the study is Ni3Fe, an intermetallic alloy that adopts the L12 superstructure, an ordered arrangement of atoms in which nickel and iron occupy distinct sublattices on a face-centered cubic framework. Ordered alloys of this kind are scientifically precious because their atomic ordering changes how dislocations move and how grain boundaries behave, making them model systems for understanding the interplay between chemistry, crystallography, and mechanics. In the undeformed state, the alloy’s grain boundary ensemble is dominated by a particular family of boundaries known as Σ3 twins, which correspond to a 60-degree misorientation between neighboring crystals and are among the most benign, low-energy interfaces a metal can possess. These coherent twin boundaries resist corrosion, block crack propagation, and generally improve material performance, which is why engineers have long sought to maximize their fraction through a strategy called grain boundary engineering.

To see what happens to this carefully ordered boundary landscape under load, the researchers turned to electron backscatter diffraction, or EBSD, a scanning electron microscopy technique that maps the crystallographic orientation of every point on a polished surface. The team examined four states of the alloy: undeformed, and three levels of compressive strain corresponding to roughly 0.1, 5, and 28 percent. These strain levels were chosen deliberately, because they span two distinct regimes of the alloy’s stress–strain curve: the transitional stage at low strain and the so-called stage III, the well-known strain-hardening regime that begins around 18 percent strain. For each state, the authors mapped four scan areas of 820 by 640 micrometers with a one-micrometer step, capturing between 3,500 and 4,000 grains and up to 400,000 individual boundary segments per condition. That enormous statistical base is what separates this work from many earlier studies, which often relied on small fields of view where a handful of unusual boundaries could skew the picture.

The headline result is striking. At 28 percent strain, within stage III of the deformation curve, the total length of high-angle grain boundaries—interfaces with large misorientation angles that act as strong obstacles to dislocations—increased by 93 percent relative to the undeformed state. At the same time, the length of boundaries classified as special type, meaning those that approximate low-energy coincidence site lattice configurations, dropped by 15 percent. The spacing between neighboring boundaries, and between general-type boundaries in particular, shrank markedly, reflecting the intense fragmentation of the original grains as deformation progressed. In other words, the alloy’s boundary network became simultaneously denser and less special, a double shift that reshapes how the material stores strain energy, conducts heat, and resists chemical attack.

The changes in the misorientation spectrum are equally telling. The fraction of boundaries sitting near the 60-degree twin orientation plummeted from 0.75 in the undeformed alloy to 0.23 at 28 percent strain. Meanwhile, the mean deviation from ideal coincidence site lattice parameters, averaged across the special-boundary spectrum, grew from 1.1 degrees to 3.8 degrees, and the kernel average misorientation—a local measure of lattice distortion—rose near the boundaries themselves. Taken together, these numbers describe a boundary population that is drifting away from its ideal, relaxed configurations and accumulating the kind of elastic distortion that accompanies dislocation pile-ups and strain incompatibilities between neighboring grains.

One of the most conceptually important parts of the study concerns energy. Classification schemes such as the widely used Brandon criterion label a boundary as special if its misorientation deviates from the ideal coincidence lattice angle by less than a tolerance that scales with the inverse square root of the coincidence index. But geometric proximity to an ideal boundary does not guarantee low energy, particularly in a deformed, non-equilibrium material. To probe this, the authors estimated the relative grain boundary energy at individual triple junctions using the Herring relation, a force-balance condition that links the energies of three boundaries meeting at a point to the angles between them. The analysis revealed that a portion of the boundaries that still satisfied the Brandon criterion at 28 percent strain were nevertheless not low-energy boundaries. This is a cautionary result for the field: a boundary can look special on paper while carrying the energetic baggage of deformation-induced distortion.

Perhaps the most elegant piece of the analysis is a quantitative length balance. The team asked a simple question with a complicated answer: where did all the new general-type boundary length come from? By comparing the boundary length budgets across the four strain states, they determined that roughly 90 percent of the additional general-type boundary length is supplied by the formation of entirely new boundaries, while only about 10 percent can be attributed to pre-existing special boundaries losing their special character and being reclassified as general. This partitioning suggests that the transformation of the boundary ensemble under deformation proceeds along two parallel tracks: the birth of new, disordered interfaces as grains fragment and subdivide, and a slower, progressive degradation of the twin population that once dominated the microstructure.

The authors are careful to flag the limits of their method. Because the analysis compares different samples deformed to different degrees, rather than tracking individual boundaries in situ as deformation proceeds, the 90-to-10 split should be regarded as an estimate rather than a direct measurement. In situ EBSD experiments, which follow the same region of a sample through successive loading steps, could in principle confirm the partitioning and reveal the microscopic mechanisms—dislocation absorption, boundary migration, twin decomposition—behind the loss of special character. Still, the statistical weight of the present data, with hundreds of thousands of boundary segments per state and experimental scatter reported across scan areas, makes the conclusion difficult to dismiss.

Why does this matter beyond the laboratory? Grain boundary character distributions are not academic curiosities; they are levers that materials engineers pull to control real-world performance. Twin-rich microstructures resist intergranular corrosion and cracking in stainless steels and nickel-based superalloys, and thermomechanical processing routes have been designed specifically to boost special boundary fractions. The new results suggest that heavy plastic deformation, the very process used to shape and strengthen components, works against those efforts in ordered alloys, converting a benign, twin-dominated boundary network into a dense web of general, higher-energy interfaces. For applications in which deformed ordered alloys must subsequently withstand corrosive environments or cyclic loading, this deformation-induced boundary degradation could be a hidden liability. At the same time, the work provides a quantitative baseline for modeling how boundary ensembles evolve during stage III hardening, feeding directly into the mesoscale simulation frameworks that next-generation alloy design increasingly relies on.

The study also reinforces a broader lesson that has been accumulating across decades of grain boundary research: the boundary ensemble of a polycrystal is a dynamic, evolving system, not a fixed backdrop against which deformation plays out. From the faceting and roughening transitions documented in classical studies to the twin-boundary engineering of austenitic steels and the deformation behavior of faceted Σ3 boundaries, the evidence consistently shows that interfaces respond to their environment as actively as the grains they separate. By quantifying that response in an ordered L12 alloy with unprecedented statistical rigor, Perevalova, Konovalova, and Ivanov have added a crucial data point—and a useful warning—to the growing picture of materials whose most important features live at the edges of their grains.

Subject of Research: Grain boundary ensemble evolution in plastically deformed ordered Ni3Fe alloy with L12 superstructure

Article Title: Grain boundary ensemble transformation in ordered Ni3Fe alloy with L12 superstructure under plastic deformation

Article References: Perevalova, O. B., Konovalova, E. V., & Ivanov, K. V. (2026). Grain boundary ensemble transformation in ordered Ni3Fe alloy with L12 superstructure under plastic deformation. Journal of Materials Science. https://doi.org/10.1007/s10853-026-13876-5

Image Credits: AI Generated

DOI: 10.1007/s10853-026-13876-5

Keywords: grain boundaries, Ni3Fe, L12 superstructure, plastic deformation, EBSD, Σ3 twins, coincidence site lattice, Herring relation, strain hardening, grain boundary engineering, ordered alloys, materials science

Moiré patterns arise whenever two periodic lattices are overlaid with a slight rotational misalignment, producing a long-wavelength interference superlattice that can profoundly alter electronic behavior. The phenomenon became world-famous with magic-angle twisted bilayer graphene, where twisting two graphene sheets by about 1.1 degrees unleashed unconventional superconductivity. Until now, however, moiré engineering has been pursued almost exclusively as a route to emergent quantum states in atomically thin heterostructures, typically assembled layer by layer under demanding laboratory conditions. The new work demonstrates that the same rotational degree of freedom, when distributed throughout the interior of a bulk crystal, can serve an entirely different purpose: dissipating mechanical stress and preventing catastrophic fracture.

The research team, led by scientists at the Shanghai Institute of Ceramics of the Chinese Academy Sciences together with collaborators at Zhejiang University, concentrated on gallium germanium telluride, GaGeTe, a ternary layered van der Waals semiconductor. Like other van der Waals crystals, GaGeTe consists of robust covalently bonded sheets held together by weak interlayer forces, which allows individual layers to slip relative to one another. In a pristine, perfectly aligned crystal, the layers share a common crystallographic orientation, and under tension the material tends to fail through the initiation and rapid propagation of cracks. The researchers reasoned that if adjacent layers were rotated with respect to each other, the mismatch between their lattices would create moiré superlattices that fundamentally change how stress is accommodated during deformation.

Introducing such twists into a bulk crystal might seem to require assembling it layer by layer, but the team found a remarkably simple route. By applying compression along an axis slightly off the crystal’s principal orientation, they controllably generated interlayer moiré twisting spanning a broad range of angles commensurate with the crystal’s translational symmetry. The magnitude of the twist could be tuned through the degree of pre-compression: samples compressed by 5, 10, and 20 percent developed distinct distributions of twist angles, as revealed by transmission electron microscopy and synchrotron-based diffraction techniques. Three-dimensional precession electron diffraction tomography and Rietveld refinement of synchrotron powder diffraction data confirmed the structural integrity and composition of the treated crystals, while in situ TEM tensile experiments captured the moiré patterns evolving in real time as the material was stretched.

The mechanical consequences were striking. When the pre-compressed, twisted crystals were subsequently pulled in tension at room temperature, their stress–strain curves revealed a transformation from brittle to genuinely ductile behavior. Where untreated GaGeTe fractured after limited elastic and plastic elongation, the twisted crystals sustained tensile strains approaching 30 percent, an increase of up to 360 percent in macroscopic ductility. Critically, the enhancement scaled with the twist angles introduced by compression: higher pre-compression levels produced larger interlayer rotations and correspondingly greater stretchability. The team also verified that the effect persisted across variations in sample thickness and strain rate, underscoring that the mechanism is intrinsic to the twisted architecture rather than an artifact of a particular testing configuration.

The origin of this extraordinary plasticity lies in how moiré superlattices facilitate stress relaxation. Under mechanical loading, the twist angles between neighboring layers continue to increase, allowing the crystal to accommodate deformation through widespread interlayer slipping rather than through the nucleation of fatal cracks. In an untwisted crystal, slipping is constrained by the periodic registry between layers, and once a micro-crack forms, the covalent bonds within a layer offer little resistance to its propagation. In the twisted configuration, the moiré modulation effectively homogenizes and redistributes the interlayer potential energy landscape, lowering the barriers to slip along the basal planes. Density functional theory calculations of generalized stacking fault energies and crystal orbital Hamilton populations supported this picture, showing reduced slip energies and modified Te–Te and Te–Ga bonding along slipping pathways in twisted structures.

Direct atomic-scale imaging told a complementary story. High-angle annular dark-field scanning transmission electron microscopy of deformed samples revealed intralayer bending, ripplocations, and micro-cracks that remained arrested rather than propagating through the entire crystal. In situ TEM tensile videos documented moiré patterns with small twist angles of roughly 1.5 degrees at early deformation stages, growing to approximately 6 degrees as tensile strain accumulated, with the progressive twisting acting as a built-in reservoir for plastic accommodation. The researchers also constructed idealized twist-stacked models, denoted θ-GaGeTe, with alternating layers rotated by specific angles ranging from about 1.5 to nearly 22 degrees; simulated moiré patterns for these structures matched the experimentally observed contrast, validating the structural interpretation of the deformation mechanism.

Perhaps the most important practical feature of the approach is its selectivity. Because moiré twisting occurs only between layers while the atomic arrangement within each layer remains intact, the in-plane electronic properties that make these materials attractive for devices are essentially untouched. This stands in contrast to many toughening strategies that rely on defects, grain boundaries, or alloying, all of which can severely degrade charge transport, optical response, or thermoelectric performance. The authors also demonstrated the generality of the concept beyond GaGeTe: off-axis compression produced comparable enhancements in tensile plasticity across a family of bulk van der Waals crystals including GaS, GaSe, InSe, CrSiTe3, InSiTe3, and SnBi2Te4, suggesting the strategy is broadly applicable to the entire class of layered semiconductors.

The implications reach into several technology areas where mechanical fragility has long limited deployment. Flexible thermoelectrics, which convert waste heat into electricity in conformable formats, depend on semiconductors that can bend and stretch without failing; earlier work by some of the same groups established ductile inorganic semiconductors and exceptional plasticity in materials such as InSe and Bi2Te3-based crystals. The moiré-twisting strategy adds a new, geometrically driven lever to that toolkit, one that is compatible with the demanding electronic requirements of thermoelectric and optoelectronic devices. It may likewise inform the design of two-dimensional-material-based transistors, photodetectors, and silicon photonics components, where GaGeTe itself has recently been integrated into electro-optic devices, by improving the mechanical robustness of bulk crystals used as feedstock or substrates.

From a fundamental standpoint, the study expands the conceptual scope of moiré physics from quantum emergent phenomena to classical mechanics of deformation. It shows that a structural parameter long treated as an electronic tuning knob can double as a mechanical design variable, decoupled from in-plane transport. The work was supported by the National Natural Science Foundation of China, the Chinese Academy of Sciences, and Shanghai municipal programs, with synchrotron experiments performed at the RIKEN BL44B2 beamline of SPring-8 in Japan. As research groups worldwide continue to explore twistronics for superconductivity, magnetism, and correlated electron physics, this result suggests a parallel track: twisting crystals to make them tough. If the strategy can be scaled from laboratory single crystals to manufactured components, the humble rotational misalignment between atomic layers may become one of the most versatile tools in the materials engineer’s repertoire, turning inherently brittle semiconductors into materials that bend, stretch, and survive the demands of real-world flexible technology.

The result also reframes a long-standing dichotomy in materials science. Metals derive their ductility from dislocations, dense arrays of line defects that glide through the crystal lattice and dissipate strain; conventional semiconductors, with their stiff covalent bonding and scarcity of mobile defects, have almost always been relegated to the brittle side of that divide. The moiré-twisting mechanism offers a third route: rather than relying on point or line defects introduced during growth or processing, it exploits a rotational degree of freedom that is deliberately programmed into the stacking sequence itself, effectively designing the deformation pathway in advance.

The approach resonates with recent independent demonstrations that twist can strengthen otherwise fragile solids. Twisted-layer boron nitride ceramics have shown high deformability and strength, and twist-assisted toughening has been reported in two-dimensional transition metal dichalcogenides, indicating that rotational misalignment is emerging as a general design principle spanning scales from atomically thin films to bulk ceramics. The present study extends this principle to ternary van der Waals semiconductors and, crucially, shows that the twist can be introduced after synthesis by a scalable mechanical step rather than during crystal growth.

For practical adoption, the compatibility with existing characterization and processing workflows matters. Off-axis compression requires no epitaxial assembly, no layer-by-layer transfer, and no exotic chemistry, which means the treatment could in principle be applied to crystals grown by conventional bulk methods. Because the twist angles are commensurate with the crystal’s translational symmetry, the resulting superlattices are periodic and well defined, making the deformed crystals amenable to standard diffraction-based quality control before device fabrication.

Open questions remain, including how the twist distributions evolve under cyclic loading, elevated temperature, and long-term operation, and whether the ductility gains persist when crystals are integrated into multilayer device stacks. Answering these will determine whether moiré-twisted semiconductors move from laboratory demonstrations into flexible thermoelectric modules, wearable sensors, and compliant photonic platforms.

Subject of Research: Enhancing tensile plasticity of bulk van der Waals crystals through tunable interlayer moiré twist angles

Article Title: Tuning moiré twist angles enhances tensile plasticity in bulk van der Waals crystals

Article References: Zhou, J., Zou, J., Gao, Z., Zhang, J., Zhang, L., Li, Z., Zhou, Z., Qiu, P., Yang, Y., Yu, Q., Chen, L., & Shi, X. (2026). Tuning moiré twist angles enhances tensile plasticity in bulk van der Waals crystals. Nature Materials. https://doi.org/10.1038/s41563-026-02731-2

Image Credits: AI Generated

DOI: 10.1038/s41563-026-02731-2

Keywords: moiré twisting, van der Waals crystals, tensile ductility, GaGeTe, off-axis compression, plastic deformation, two-dimensional materials, twistronics, semiconductors, flexible electronics, interlayer slipping, Nature Materials