What this paper is about: The ATLAS collaboration studied how the distributions of quarks and gluons inside lead nuclei change depending on where a probing photon strikes the nucleus. These distributions are known as nuclear parton distribution functions (nPDFs). The experiment used photon-induced (γ+A) jet production in ultra-peripheral lead–lead collisions at the Large Hadron Collider. The data come from 2018 and correspond to 1.72 nb⁻¹ recorded with the ATLAS detector.

What the researchers did: They compared two classes of events distinguished by whether one or both nuclei emitted neutrons very close to the beam direction. Events with no forward neutrons on either side (called 0n0n) are expected to come from photon interactions that hit the outer edge of the struck nucleus and therefore probe a larger impact parameter — that is, a larger distance from the nucleus center. Events with at least one forward neutron (0nXn) tend to come from interactions that excite and break up the struck nucleus, corresponding to smaller impact parameters. The team reconstructed jets with the anti-k_t algorithm using a radius R = 0.4 and used jet kinematics to form photon and nuclear parton momentum fractions (called z₋ and x₊) that describe the hard scattering.

How the measurement works at a high level: In these ultra-peripheral collisions the heavy ions pass by each other at distances larger than twice the nuclear radius, so the hard interactions are usually caused by photons emitted by one nucleus hitting the other. By selecting events with large rapidity gaps and by vetoing forward neutrons, the analysis isolates non-diffractive photonuclear jet production in which the struck nucleus is left essentially intact. The shape of the measured γ+ A cross section as a function of x₊ was then compared between the 0n0n and 0nXn samples.

Why it matters: The study finds a statistically significant difference between the two samples. The shape of the cross section versus x₊ differs with an observed significance of 6.0σ. At large x₊ the results are consistent with the idea that peripheral collisions (large impact parameter) show little or no modification of the nPDFs compared with free nucleons, while more central photon hits (smaller impact parameter) do show the familiar nuclear modifications seen in other hard processes. This directly supports the long-standing expectation that nPDF modifications depend on where inside the nucleus the interaction occurs.