For the first time, physicists have filmed the thing this publication is named for. Not a spectrum. Not a lifetime. Not a peak that shifts with temperature or field. The wavefunction itself: the spatial shape of an exciton, with its internal quantum-mechanical phase, caught in the first moments after its birth and watched as it pulled itself together.
The result, published August 28, 2026, in Physical Review X (announced September 1), comes from a team led by Peter Puschnig at the University of Graz, working with colleagues at Marburg University and Forschungszentrum Jülich. Using a technique called time-resolved photoemission orbital tomography — trPOT — the team reconstructed the wavefunction of an exciton in a thin film of the organic semiconductor alpha-sexithiophene, tracking its evolution over the first few hundred femtoseconds of its existence. The exciton begins spread coherently across roughly three molecules, about 1.5 nanometres, and then contracts by roughly a fifth to a quarter within 400 femtoseconds, as vibrations of the molecular lattice trap it in place.
This article is not breaking news. The paper appeared in late August and was missed by our regular scans; it reaches these pages as a landmark measurement that deserved a closer reading than it got. It is also not a device result. Alpha-sexithiophene is a model system, the film was measured in ultra-high vacuum and at cryogenic temperatures, and nothing in the work demonstrates a more efficient solar cell. What it demonstrates is something more fundamental: that the quantum-mechanical object at the heart of every organic solar cell and light-emitting diode can now be imaged directly, in space and in time, as it lives and changes.
Why a wavefunction was so hard to see
An exciton is a bound pair: an electron promoted to a higher energy state, still tethered by electrostatic attraction to the positively charged hole it left behind. It is neutral, it carries energy but no charge, and it is the primary way that light’s energy moves through an organic semiconductor before it becomes electricity or is re-emitted as light. Excitons have been known for decades, and their energies, lifetimes, and diffusion lengths have been measured exhaustively. But their internal structure — the actual quantum-mechanical wavefunction describing where the electron is likely to be found relative to the hole, and with what phase — stayed hidden.
The reason is a mismatch of scales. Excitons in organic materials decay within picoseconds, often faster. Standard spectroscopy measures energy with exquisite precision but gives up spatial information; imaging methods resolve space but are far too slow. As the Graz team put it, the exciton’s internal quantum-mechanical structure had “remained largely hidden until now” precisely because no single measurement could capture spatial structure and momentum at the same time, on the timescale of the exciton’s own life. A film of the wavefunction, rather than another spectrum, is a new kind of evidence — and a theory of exciton transport that only knows energies is working half-blind.
The camera: two pulses and a momentum microscope
The experiment is a pump-probe measurement, but with a decisive twist in the detection. First, an ultrashort laser pulse with an energy of 2.35 electron-volts strikes an ordered film of alpha-sexithiophene and creates excitons. Then, after a precisely controlled delay, a second, much higher-energy pulse — 21.7 electron-volts, produced by high-harmonic generation — ejects electrons from those excitons by photoemission. A time-of-flight momentum microscope records the energy and the three-dimensional direction of each ejected electron.
Varying the delay between the two pulses produces snapshots of the exciton at different ages. That much is standard pump-probe logic, and it is described in detail in our survey of ultrafast methods. The breakthrough is what the Graz team did with the snapshots. Photoemission orbital tomography, a technique developed in Puschnig’s group and extended with the Jülich team, turns a photoelectron’s momentum distribution into a picture of the orbital it came from. For an exciton, that requires a quantitative model of how photoemission works when the electron being ejected is part of a bound pair rather than a single-particle state — a model the Graz group built, together with ab initio many-body perturbation-theory calculations in the GW approximation, to interpret the data.
“It is possible to deduce the spatial shape and the internal quantum-mechanical phase of the exciton wave function directly from measured photoelectron images,” said Siegfried Kaidisch, a doctoral researcher in Puschnig’s group who made a key contribution to the work as part of his thesis. That sentence is the whole paper in miniature. The phase is the part that ordinary measurements cannot touch: the sign pattern of the wavefunction across the molecules it spans, which determines how its parts interfere. The reconstruction recovers it, and the recovered phase modulation matches the GW calculations. Theory and experiment are not merely consistent on an energy; they agree on a shape.
The experiment itself was a logistical feat. The Jülich group, led by Stefan Tautz, grew the ordered films — rod-shaped six-ring thiophene molecules, aligned on a specially prepared copper surface and decoupled from the substrate so the exciton would survive long enough to be filmed — and transported them to Marburg under ultra-high vacuum, in what the team calls a vacuum suitcase. The Marburg group, led by Ulrich Höfer, performed the photoemission measurements and analyzed the data. The Graz group developed the theory, ran the simulations, and built the analytical model that turned momentum images into a wavefunction. The project was a major milestone of the European Union’s “Orbital Cinema” program, whose aim is to visualize electron dynamics in materials with unprecedented spatial and temporal resolution.
What the film shows
The film has two acts, and both are short. In the first, the newborn exciton is coherently delocalized: its wavefunction extends across approximately three molecular units, about 1.5 nanometres, with a characteristic phase pattern that the GW calculations reproduce. “Coherent” here is doing real work. The exciton is not merely spread out in the sense that it might be found on any of three molecules. Its phase is correlated across them, so the parts of the wavefunction interfere. That is the difference between a quantum superposition and a classical probability distribution, and it is exactly the distinction our article on coherence and dephasing insists on: population tells you the exciton exists, coherence tells you its phase survives.
In the second act, the exciton shrinks. Within 400 femtoseconds — four hundred quadrillionths of a second, or 0.4 picoseconds — its radius contracts. The preprint reports a contraction of about 20 percent; the University of Graz announcement and the Physics World coverage put it at about 25 percent. The discrepancy is a matter of how the radius is extracted from the reconstruction, and this article gives the range rather than pretending to a precision the sources do not agree on. Either way, the direction and the mechanism are unambiguous: exciton-phonon coupling drives partial self-trapping. The exciton’s own presence distorts the surrounding lattice, the lattice vibrations respond, and the exciton settles into a more localized configuration, pulled inward by the distortion it created.
Self-trapping is not a new idea — it has long been invoked to explain why excitons in organic materials move less freely than band theory predicts — but it has mostly been inferred from indirect signatures such as Stokes shifts or temperature-dependent mobility. Here it is watched directly, as a shrinking wavefunction, on the timescale on which it happens. A long-standing interpretation becomes an observation.
Why solar engineers should care — and how far the claim goes
The honest payoff of this result is indirect, and it should be stated as such. Organic solar cells work by a sequence: absorb a photon, form an exciton, move the exciton to a donor-acceptor interface, split it into free charges, collect the charges. Every step after absorption depends on how the exciton moves, and how it moves depends on how delocalized it is. A delocalized exciton samples more of its environment per unit time than a localized one; it can reach an interface faster, and it can outrun the trapping processes that would otherwise strand it. The standard design rules for organic photovoltaics — in particular the distances between donor and acceptor domains — are built on hopping models in which excitons move by incoherent jumps between localized sites. The Graz film shows that, for the first few hundred femtoseconds, the exciton is not hopping between sites at all. It is spread across them, coherently.
That window is brief, and it would be a mistake to extrapolate it into an efficiency claim. The film was made on a model compound, not a device-relevant blend; in vacuum and in the cold, not in a working cell at room temperature; and the exciton self-traps within 400 femtoseconds regardless. But design rules are built from assumptions about the earliest moments of exciton life, and this measurement replaces an assumption with a picture. If the delocalized window survives, even partially, in real blends — an open question, not a finding of this paper — then junction distances designed from hopping models may be leaving transport on the table. The next step the team names is the right one: to watch electron-hole separation in donor-acceptor systems. “This process determines how efficiently light can be converted into electrical current and is therefore central to future developments in organic photovoltaics,” Puschnig said.
What it is not
Three cautions, stated plainly because the press-release framing invites overreading. First, this is not a solar-cell result. Alpha-sexithiophene is employed in optoelectronic applications, and it is a legitimate model for organic semiconductors, but it is not a solar-cell absorber, and the measurement involved no device, no electrodes, and no power conversion. The University of Graz announcement frames the work as “filmed” energy generation from light; the more precise statement is that the team filmed the exciton, which is the first step on the way from light to electricity. The distinction matters, and this article keeps it.
Second, this is not a room-temperature result. The films were measured under ultra-high vacuum and at cryogenic temperatures, conditions chosen to keep the exciton alive and the molecules ordered long enough for the measurement to work. Whether the same delocalized-to-trapped trajectory plays out at room temperature, in a disordered blend, under illumination at solar intensities, is unknown. Readers of our coverage of coherence and dephasing will recognize the pattern: cryogenic measurements reveal what is possible in principle, and room temperature is where the phonons fight back. The self-trapping observed here is itself a phonon effect, which makes the temperature question doubly live.
Third, this is not the only exciton-tomography result in circulation, and it should not be confused with another one. A separate, Göttingen-led study used a related technique, photoemission exciton tomography, and was published in Nature Communications. That is a different paper, with a different author list, a different journal, and different claims — none of its figures or findings appear in this article. The confusion is understandable, since the techniques share a name and Graz appears in both stories, but the two studies should be kept apart.
What to watch
The paper itself points to the most interesting follow-up: trPOT as a general method. The authors present it as an approach for resolving exciton wavefunctions, with spatial, phase, and temporal sensitivity, “in a broad class of molecular and low-dimensional materials.” That is a claim about a method’s future, not a demonstrated result. The interesting tests will be concrete: a wavefunction film in a donor-acceptor blend, where the exciton meets an interface; a film in a two-dimensional semiconductor, where the exciton is far more tightly bound; a film at elevated temperature, where self-trapping competes with thermal motion.
There is also a quieter implication for theory. The GW calculations reproduced not just an energy but a phase pattern and a contraction trajectory — a demanding test of many-body perturbation theory in a molecular solid, and it passed. For a field in which theory and experiment often meet only at the level of a peak position, agreement on a wavefunction’s shape raises the bar for what counts as agreement in the next paper.
The deepest reason this result matters is simpler. Excitons are the entity this publication is named for, and until now they have been known the way a city is known from its skyline: by the light it gives off, from a distance. The Graz, Marburg, and Jülich teams have produced something closer to a street-level film: the exciton itself, spread across three molecules, coherent, and then — in less than half a picosecond — pulling itself together. It is the first time the protagonist of excitonic physics has been seen as it actually is, and that changes what the next measurement has to live up to.
Sources and further reading
Marcel Theilen, Siegfried Kaidisch, Monja Stettner, Sarah Zajusch, Eric Fackelman, Alexa Adamkiewicz, Robert Wallauer, Andreas Windischbacher, Christian S. Kern, Michael G. Ramsey, François C. Bocquet, Serguei Soubatch, F. Stefan Tautz, Ulrich Höfer, and Peter Puschnig, “Observing the Spatial and Temporal Evolution of Exciton Wave Functions in Organic Semiconductors,” Physical Review X (2026), DOI: 10.1103/3zmg-276c; preprint arXiv:2511.23001. University of Graz announcement, “Ultrafast snapshots reveal the first moments on the way from light to electricity,” September 1, 2026. Physics World, “Exciton wavefunctions come into view,” October 1, 2026. For the wider measurement context, see this site’s explainers on how excitons are seen and on coherence, dephasing, and why room temperature is hard.