Science2 publishers2 min readPublished
A few days of oxygen beams at CERN gave air-shower models their first laboratory test
Reading a cosmic-ray shower has always meant trusting a computer model of what happens in air. Last July the LHC made the collision those models start from, under conditions physicists chose.
The Scientist · Science desk
What happened
- CERN collided oxygen with protons for the first time on July 1, 2025, holding the configuration at the Large Hadron Collider for a few days of running.
- Interpreting cosmic-ray data depends on computer simulations, and different codes predict different things about how the resulting particle showers form in the sky.
- The run followed several years of work by the account's author and physicist Lydia Beresford, who argued for reconfiguring existing CERN instruments to study cosmic rays.
- Ground-based cosmic-ray work normally uses hundreds of cameras spread across miles of land, such as the Telescope Array in Utah, which sees the highest-energy particles arriving from space.
- Scientists have studied these atmospheric showers for more than a century without a precise understanding of how they form.
Compiled by The ScientistSomething wrong?How this is made
Why it matters
- capability Codes that turn ground-level shower measurements into statements about what the incoming cosmic ray was can now be checked against a measured proton-oxygen collision instead of only against one another.
- decision Groups holding decades of archived shower events have to decide whether retuned simulations justify reprocessing that archive, and whether earlier composition estimates need restating.
- precedent A years-long request to repurpose collider instruments for an atmospheric-physics question ended with real beam time. Requests from fields outside collider physics now have a case to point to.
In the sky the projectile is a fast-moving nucleus, usually hydrogen, and the target is an atom of air, high above the ground [12][20]. At CERN that was mapped onto two beams: the proton beam played the cosmic ray, the oxygen beam played the atmosphere [3]. The collider adds control. The species and the energy are set by the machine, and the collision happens inside a detector at a known moment.
The account frames the goal as finding out which model, if any, describes real collisions [5]. That phrasing allows for the outcome in which every existing code misses the data and all of them need retuning.
ATLAS recorded the run. The detector is the size of a football field and thousands of physicists operate it, and the high-speed camera at its centre can take more than 200 million photographs in a day, which is roughly 2,300 a second [7][8][16].
As published, the write-up describes the design and leaves out the measured distributions, the collision energy and the number of collisions recorded, so it never names the shower model the data favour [18].
That leaves the rest of the cascade. The collision recreates the first moments of a shower [14]; what a ground-based array measures is what reaches the surface after everything downstream of that first step, and that part is still computed on a machine [11][9]. Fixing the first interaction narrows the inputs those computations begin from.
Both publishers ran the same first-person account by a physicist who works on ATLAS, so the description of the run comes from one participant [15][7]. The measurements are in Physical Review Letters [2].
What to watch
- Whether the Physical Review Letters distributions separate the competing shower codes, and by how much.
- Whether CERN schedules further oxygen or other light-ion running after the few days in July 2025.
- Whether array groups such as Telescope Array refit archived events with codes retuned to the collider data.