72 Years After Its Prediction, Physicists Finally Took a Picture of Superradiance. It Was a Ring.
In 1954, Robert Dicke had an idea so elegant it sounds almost obvious today: a group of excited atoms packed close together shouldn’t glow independently. They should coordinate. Like an orchestra section where each musician can hear the others, the atoms would sync up and release their energy in a single, blazing burst: brighter, faster, and more directional than if each atom acted alone. He called it superradiance.
The theory was beautiful. Testing it took decades. And now, 72 years later, physicists in France have done something nobody had managed before: they took a picture of the exact spatial pattern of that collective emission and found it looked nothing like what the textbooks predicted.
Background & Context
Dicke’s superradiance paper imagined an idealized scenario: atoms so close together you couldn’t tell them apart. In that limit, they share a single channel for emitting light: one collective voice. But real atomic clouds are never that small. They have size, shape, and spatial structure. When the sample is larger than the wavelength of the emitted light, things get messy. The atoms’ positions break their indistinguishability, and multiple collective channels start competing.
This messiness is actually physics, not noise. In 2000, Howard Carmichael and collaborators introduced the concept of “collective jump operators”: mathematical objects that describe distinct electromagnetic modes through which a finite-sized atomic ensemble could emit light. Some modes would be superradiant (faster than a single atom), others subradiant (slower). But these operators remained theoretical tools for two decades. No one had directly measured one.
Meanwhile, the standard working theory for superradiance in pencil-shaped samples said the emission should peak straight along the cloud’s axis, like a flashlight beam. This “spin-wave” picture worked reasonably well for explaining experiments dating back to the 1970s. Nobody had a reason to doubt it.
What the Researchers Did
The team at Laboratoire Charles Fabry, led by Antoine Browaeys and Igor Ferrier-Barbut, built an experiment to settle the question directly. They trapped clouds of up to 8,000 cold rubidium-87 atoms in an optical dipole trap shaped like a tiny cigar, roughly 78 micrometers long and just a few micrometers wide. Each cloud sat in a glass cell surrounded by three microscope objectives.
Here’s where it gets clever: instead of just measuring how much light came out and when, they pointed a camera at the Fourier plane of their imaging system. In plain terms, this let them photograph the angular distribution of the emitted light. Where the photons were going. It’s the difference between measuring how loud an orchestra is versus mapping out which direction the sound is coming from and what note each section is playing.
The sequence was simple. They turned off the trap, hit the atoms with a 9-nanosecond laser pulse to flip them into their excited state (achieving 88% excitation), and then let them decay while the camera recorded the light. They repeated this 200 times on the same cloud to build up statistics.
What They Found
The image that came back was striking. Instead of a bright spot at the center (the along-the-axis peak that standard theory predicted), they saw a ring. The light emerged in a cone around the cloud axis, with a dark hole in the middle. This was the fingerprint of the most superradiant collective mode.
The ring only appeared above a critical atom number of about 1,700. Below that threshold, the emission was diffuse and spread across all angles. Above it, the directional ring grew in brightness, scaling roughly as N², exactly what Dicke superradiance demands. A second ring, from the next-most-superradiant mode, was theoretically present but buried in the noise.
| Measurement | Value |
|---|---|
| Cloud length (σx) | 23–80 µm |
| Cloud radius (σr) | 1.3–2.2 µm |
| Atom number (N) | Up to 8,000 |
| Critical N for superradiance | ~1,700 |
| Enhancement factor (η) | ~6 × 10⁻⁴ |
| π-pulse duration | 9 ns (0.34/Γ0) |
| Excited population after pulse | 88% |
| Collection angle | Up to 30° |
When they compared their image to the theoretical emission pattern calculated from the most superradiant collective jump operator, the match was near-perfect. They had, for the first time, directly measured a mathematical entity that had existed only on paper for 26 years.
The team also performed time-resolved measurements, swapping the camera for single-photon detectors. When they collected light only from within the ring (spatially filtering to isolate the most superradiant mode), they saw the classic superradiant burst: a delayed pulse that peaked after the driving laser turned off. The peak intensity scaled as N² above the threshold. When they blocked the ring and collected light from outside it, the emission was non-exponential but weaker. Multiple modes still above threshold, but none dominant. And when they collected perpendicular to the cloud axis, they saw plain exponential decay at the single-atom rate. No collective enhancement at all.
Why It Matters
The result matters for two reasons. The first is conceptual: it falsifies the spin-wave picture that has guided superradiance theory for half a century. In an elongated cloud, the most superradiant mode is not the one where all atoms radiate in phase along the axis. Geometric constraints shift the winner to a mode with a nonzero transverse wavevector, hence the ring. The atoms aren’t all singing the same note; they’re choosing the most efficient arrangement given their physical layout.
The second is practical. Collective jump operators are more than a conceptual curiosity. They are the natural communication channels between light and matter in free space. If you can measure them and isolate them (which this experiment proves you can), you can manipulate them. A spatial light modulator could selectively excite one mode. Reference light could be interfered with it. This is exactly the toolkit you need to build light-matter interfaces for quantum networks, where information encoded in atoms needs to be reliably transferred to photons and back again.
The experiment also uncovered an unexpected requirement: for a single mode to dominate, the eigenvalue spectrum of the decay matrix needs a spectral gap. The most superradiant mode must be separated from the pack by a measurable margin, or the emission gets diluted across many competing channels. The team found this gap exists for a wide range of cloud geometries but not all of them, and it peaks along a curve of constant Fresnel number.
How It Could Change Our Lives
The most immediate applications are in quantum technology. Free-space quantum interfaces (devices that convert stationary qubits in atoms into flying qubits in photons) have been a bottleneck in quantum networking. Current approaches often use optical cavities to enhance the coupling between a single atom and a single photon mode. But cavities add complexity, loss, and bandwidth limitations.
If you can instead engineer an atomic ensemble to emit into exactly one well-defined free-space mode, you get the enhancement without the cavity. The ring-shaped emission pattern this team observed is essentially a highly directional antenna made of atoms, and the direction and shape can be tuned by changing the cloud geometry. Future arrays of individually trapped atoms, arranged in specific patterns, could be designed to emit into precisely chosen modes. Your quantum repeater node might look less like a mirror-lined cavity and more like a carefully arranged grid of atoms.
Further out, the ability to image collective emission modes opens a new window for studying many-body quantum physics in a controlled setting. Watching how modes compete, how correlations build up, and how geometry shapes collective behavior is the kind of experimental playground that theorists have wanted for decades.
The Bigger Picture
Dicke’s 1954 paper was less than a page long and contained no experimental data. It was a thought experiment that happened to be right. But the story of superradiance since then has been one of gradual refinement: first the realization that real samples are finite, then the development of mean-field theories, then the collective jump operator framework, and now, finally, direct experimental access to those operators.
What makes this moment satisfying is the convergence. The theoretical tools Carmichael built in 2000 weren’t validated by some indirect signature or statistical inference. They were photographed. The agreement between the computed and measured emission patterns leaves little room for interpretation. The collective jump operators are real, they can be measured, and they can be controlled. After 72 years and a series of approximations, superradiance finally has its ground-truth image.
Limitations & What’s Next
The experiment used clouds of up to 8,000 atoms, which is large enough to see collective effects but far from the regime where quantum correlations between individual atoms could be tracked. The bosonic model they used to explain mode competition makes significant approximations and doesn’t capture the full quantum dynamics. Resonant dipole-dipole interactions (the Hamiltonian part of the atom-atom coupling) were mostly ignored.
The natural next step is to move from random clouds to ordered arrays. If you can place atoms in a precise lattice and image their collective emission, you can design the modes from scratch. The team has already suggested using a spatial light modulator to selectively excite or interfere with specific collective jump operators. A full quantum simulation, perhaps using truncated Wigner or cumulant expansion methods, could reveal the role of dipole-dipole interactions that their simplified models couldn’t handle.
For now, the ring of light in that Fourier image is enough. It’s the kind of result where the picture tells the story before the equations do.
📄 Source: Gavalda, Tremblier, Poitrinal, Pancaldi, Browaeys & Ferrier-Barbut, “Fourier imaging of collective spontaneous emission modes in superradiant cold atomic clouds,” arXiv:2607.08421v1 (2026).