A new study has captured something physicists have been trying to see directly for years: excitons trapped inside the repeating nanoscale landscape of a twisted two-dimensional semiconductor. Using room-temperature photocurrent atomic force microscopy, researchers imaged where different types of moiré excitons gather within individual moiré unit cells in a twisted bilayer of molybdenum disulfide, or MoS₂. The work provides the clearest real-space evidence yet that excitons do not simply spread across a moiré pattern, but instead occupy specific stacking sites according to their electronic character. The findings could help transform moiré materials from intriguing laboratory systems into engineered platforms for light emission, sensing and quantum information technologies.
Excitons are electrically neutral quasiparticles formed when light promotes an electron to a higher-energy state, leaving behind a positively charged vacancy known as a hole. The electron and hole remain bound together by electrostatic attraction, much like a miniature hydrogen atom embedded in a solid. In atomically thin semiconductors, this binding can be especially strong because the particles are confined to two dimensions and experience reduced dielectric screening. Their properties can be modified dramatically when two layers are stacked with a small rotational mismatch. The resulting interference pattern, known as a moiré superlattice, creates a periodic energy landscape that can act as an artificial crystal for excitons.
The researchers examined a bilayer made from two MoS₂ sheets rotated by approximately 2 degrees relative to one another. This slight twist shifts the atomic registries gradually across the sample, producing a much larger periodic pattern than the atomic lattice itself. Within each moiré unit cell, the two layers align in several distinct ways. Some registries favour the formation or confinement of particular excitonic states, while others raise their energy and act as less favourable locations. In principle, this arrangement allows excitons to be organized into programmable lattices without chemically altering the material. In practice, however, proving exactly where the excitons reside has been exceptionally difficult.
Most previous evidence for moiré exciton localization has come from optical measurements collected over areas containing many moiré cells. Photoluminescence and absorption spectroscopy can reveal the energies of excitonic states and indicate that a periodic potential exists, but these techniques generally average together signals from countless sites. They can show that excitons are affected by the moiré landscape without revealing the precise position of an exciton within a single unit cell. At nanometre scales, that distinction matters. Two states with similar energies may be localized at entirely different stackings, and the apparent intensity of an optical signal may reflect how excitons are generated rather than where they ultimately settle.
The new measurements overcome this limitation by combining photocurrent detection with atomic force microscopy. In this approach, a nanoscale probe scans across the surface while the material is illuminated, and the resulting photocurrent records how efficiently photoexcited carriers are generated, separated or collected at each position. Because the probe can map variations in the electronic response with nanometre-scale precision, the method links local structure to excitonic behaviour directly. Crucially, the measurements were performed at room temperature, demonstrating that the spatial signatures are not limited to fragile cryogenic conditions. The result is a direct image of site-selective exciton confinement under conditions more relevant to practical devices.
The study distinguishes between direct and indirect excitons, two related species that differ in how their electrons and holes are distributed between the layers and valleys of the semiconductor. In a direct exciton, the electron and hole occupy configurations that allow them to recombine efficiently while emitting light. An indirect exciton has its electron and hole separated in layer, momentum or both, which can reduce their recombination rate and give the pair a longer lifetime. That separation also changes how the exciton interacts with the moiré potential. Rather than following an identical confinement pattern, the two species were found to localize at different stacking registries within the same moiré unit cell.
One of the study’s most important insights is that the measured contrast does not simply provide a map of the deepest potential wells. Instead, it depends on the relationship between where excitons are created and where the moiré landscape confines them. A particular stacking site may generate excitons efficiently but not be their final low-energy destination. Conversely, excitons produced elsewhere may migrate toward a different registry before they are detected. The photocurrent therefore reflects a combination of local optical generation, exciton motion, interlayer charge distribution and trapping. Recognizing this distinction helps explain why different experiments can produce apparently different spatial patterns while still describing the same underlying moiré physics.
To interpret the images, the researchers used a Wannier-based effective moiré-exciton model. Wannier functions are localized wavefunctions constructed from extended quantum states, providing a useful language for describing particles moving through a periodic potential. In the model, excitons are treated as composite objects that occupy effective sites of the moiré lattice and can move between them through quantum tunnelling. The calculation incorporates the variations in energy produced by different local stackings and reproduces both the measured exciton energies and their spatial localization. Agreement between the model and the images offers a microscopic explanation for why direct and indirect excitons choose different sites.
The ability to observe these species separately could reshape the design of excitonic devices. Direct excitons are attractive for light-emitting applications because they can recombine radiatively, while indirect excitons may be useful when long lifetimes, dipole interactions or electrically controlled transport are desired. If each species can be directed to a chosen registry, researchers may be able to build excitonic circuits in which light generation, energy transfer and information storage occur at predetermined locations. Moiré potentials could also provide a route to studying collective phases, including excitonic order and other correlated states, because the periodic confinement brings many interacting excitons into a controlled geometry.
The work also establishes an important benchmark for theories of twisted van der Waals materials. A model that reproduces an averaged spectrum but fails to predict the correct real-space localization is incomplete. By resolving the exciton distribution inside a single moiré cell, the study places stricter constraints on calculations of interlayer coupling, dielectric screening, tunnelling and relaxation. It shows that the moiré pattern is not merely a passive backdrop that shifts optical resonances. It actively determines where quasiparticles are born, where they move and where they spend their time. As nanoscale photocurrent imaging becomes more widely applied, similar measurements could reveal how moiré landscapes control excitons, trions and other emergent particles across a broad range of two-dimensional materials.
The findings mark a significant step toward turning twisted semiconductors into programmable quantum materials. A small change in twist angle, layer composition or external electric field can reshape the moiré potential, potentially moving confinement sites or changing the balance between competing excitonic states. The new room-temperature imaging method makes it possible to test such possibilities directly rather than inferring them from spatially averaged signals. For a field built around structures too small to see with conventional optical microscopes, the ability to watch excitons choose their sites is more than a technical advance. It is a direct view of how an artificial lattice controls matter—and a vivid demonstration that the future of nanoscale light and information may be written into the geometry of two slightly twisted sheets.
Subject of Research: Real-space, site-selective confinement of direct and indirect moiré excitons in a 2° twisted bilayer MoS₂.
Article Title: Real-space imaging of moiré-confined excitons in twisted bilayer MoS₂
Article References: Westenberg, L.J.M., Eek, L., Verbakel, J.D. et al. Real-space imaging of moiré-confined excitons in twisted bilayer MoS₂. Nature Physics (2026). https://doi.org/10.1038/s41567-026-03425-x
Image Credits: AI Generated
DOI: https://doi.org/10.1038/s41567-026-03425-x
Keywords: moiré excitons, twisted bilayer MoS₂, two-dimensional semiconductors, exciton confinement, atomic force microscopy, photocurrent imaging, direct excitons, indirect excitons, moiré superlattices, van der Waals materials
Tags: engineered 2D semiconductor heterostructuresexciton behavior in layered materialsexciton trapping in 2D semiconductorslight emission in moiré materialslocalized exciton statesmoiré excitonsmoiré superlattice electronic propertiesmoiré-induced exciton site occupationnanoscale moiré pattern imagingquantum information applications of twisted bilayersroom-temperature photocurrent atomic force microscopytwisted bilayer MoS2
