The LHC Smashed Oxygen Nuclei Into Protons to Recreate the First Instant of a Cosmic-Ray Air Shower. ATLAS Measured the Debris 10 Times More Precisely Than Any Model Predicts It.

Every second, particles born high in the atmosphere pass through your body at nearly the speed of light. They are the remnants of cosmic rays, atomic nuclei flung across the galaxy by exploding stars and supermassive black holes, which slam into nitrogen and oxygen nuclei roughly 20 to 30 kilometers up and burst into cascades of secondary particles that rain down over square kilometers of ground. Physicists have studied these air showers for more than a century, but they have never had a precise picture of the very first collision that starts one. A new measurement from the ATLAS experiment at CERN's Large Hadron Collider, published in Physical Review Letters, changes that.

The problem is one of interpretation. Observatories such as the Telescope Array in Utah's high desert and the Pierre Auger Observatory in Argentina detect the shower, not the primary cosmic ray. To work backward to the energy and identity of the original nucleus, and to decide whether it was a bare proton or something as heavy as iron, they run computer simulations of the shower's development. Different simulation codes disagree about how many particles the first collision makes and how fast they emerge, and those disagreements propagate into the biggest unresolved question in the field: what the highest-energy cosmic rays are made of, and therefore where they come from.

On July 1, 2025, the LHC collided oxygen nuclei with protons for the first time, in a short special run that physicists Jesse Liu and Lydia Beresford had spent years arguing for. The configuration was chosen deliberately. In the atmosphere, a cosmic-ray proton hits an oxygen or nitrogen nucleus; at CERN, the proton beam played the cosmic ray and the oxygen beam played the atmosphere, at a center-of-mass energy of 9.62 teraelectronvolts per nucleon pair. The ATLAS detector, a football-field-sized instrument whose silicon tracking layers can record more than 200 million collision images a day, photographed the sprays of charged particles that flew out of each impact.

Liu's team measured how many charged particles each collision produced and the distribution of their energies, the two quantities air-shower simulations most need. The results pin those quantities down with more than 10 times the precision of the spread among existing model predictions. That means the models can now be calibrated against real proton-oxygen collisions rather than extrapolated from proton-proton and lead-lead data taken at other energies, which is what they have relied on until now.

The payoff is a sharper answer to the composition question. If the highest-energy cosmic rays are mostly hydrogen, they can be traced back through the galaxy's magnetic fields to specific sources; if they are heavy nuclei, they bend too much to point anywhere. Better shower models also feed directly into the muon puzzle, the long-standing observation that air showers produce more muons than any simulation predicts. Cosmic rays have practical uses too: they date archaeological finds through carbon-14, they have imaged hidden chambers in Egypt's pyramids and the interiors of active volcanoes, and each of those techniques depends on knowing the shower physics.

Liu writes that the study "renews the links between particle physics and high-energy astrophysics," two fields with shared roots that have drifted apart. The oxygen run lasted only days. Its data will be used for years.