By colliding ions of different species with each other and with protons, the LHC enables studies of fundamental properties of matter across an extraordinary range of energy scales, from MeV to ~10 TeV.

At the lower end of this spectrum, nuclear binding energies are of the order of a few MeV, and modifications of parton distribution functions for partons bound in nuclei (nuclear PDFs, or nPDFs) have been observed for decades. These modifications are expected to depend on the nucleon’s position within the nucleus, yet experimental constraints on this spatial dependence remain limited. Dijet production induced by electromagnetic probes in ultra-peripheral Pb+Pb collisions (UPCs), with impact parameters constrained via forward neutron detection in Zero Degree Calorimeters, provides a novel avenue for investigating the spatial dependence of nPDFs.

At the opposite extreme, proton–oxygen collisions at the LHC delivered in 2025 reproduce key features of PeV cosmic-ray interactions with the Earth’s atmosphere. Interpreting data from ground-based cosmic-ray detectors relies on detailed simulations of hadronic interactions at TeV-scale center-of-mass energies. These simulations are based on phenomenological models that require tuning to collider data, particularly in the forward and high-energy regimes.

This seminar presents two recent ATLAS measurements that probe nuclear cross sections at complementary frontiers. The first is a measurement of dijet production in Pb+Pb UPCs at √s_NN=5.02 TeV with no nuclear break-up on either side, directly testing the spatial dependence of nPDFs. The second is the first measurement of prompt charged-particle production in proton–oxygen collisions at √s_NN=9.62 TeV, delivering key constraints for tuning the hadronic models and describing the development of cosmic-ray air showers.