Study of the mass composition of cosmic rays with the Underground Muon Detector of AMIGA

Despite several decades of study, the origin of ultra-high-energy cosmic rays (UHECRs)—ionized nuclei from outer space bombarding the Earth's atmosphere at energies above $10^{17}$ eV—remains a key open question in astroparticle physics.
Their enormous energies, exceeding by orders of magnitude those achievable in the most powerful human-made particle accelerators, are indicative of the most violent and extreme processes in our universe.
Since their flux is extremely low, UHECRs can only be detected indirectly by measuring the extensive air showers (EAS) of secondary particles they produce when interacting with an air nucleus.
Consequently, the energy, arrival direction and mass composition must be inferred from the signals that EAS produce in large ground-based detectors such as the Pierre Auger Observatory.
In particular, the mass composition is of utmost importance, as it would help to constrain astrophysical scenarios explaining the origin and sources of these extremely energetic particles.
One of the most mass-sensitive EAS observables is the number of muons produced in the shower, which can be directly sampled by underground segmented scintillator detectors, such as the Underground Muon Detector (UMD) of the Pierre Auger Observatory.
... mehr

In this work, we present a measurement of the muon content of air showers within the energy range $10^{17.5}$ eV to $10^{18.9}$ eV using the Underground Muon Detector (UMD) of the Pierre Auger Observatory.
We revisited and improved the existing procedure for reconstructing the lateral distribution function (LDF) of muons at the individual event level.
The likelihood model used to fit the LDF was extended to account for detector effects—such as detector noise, inefficiency, and corner-clipping muons—by introducing probabilistic models with physical motivation.
In addition, we developed a novel data-driven method to quantify and correct for corner-clipping muons—inclined muons that can generate signals in adjacent segments, leading to overcounting. Previously, this effect was addressed solely through simulations; in this work, we demonstrate how it can be studied and corrected using data.
The framework developed here, both for the likelihood model and the corner-clipping correction, is generally applicable to any type of segmented detector.

The muon content measured in this work was compared to air shower simulations to infer the mass composition.
The energy-dependent trend of the inferred mass is consistent with that observed using another mass-sensitive observable—the depth of shower maximum, $X_\text{max}$—indicating a transition from heavy to light elements from $10^{17.5}$ eV up to $\sim 10^{18.4}$ eV, where a break toward heavier elements is observed. However, the absolute mass is systematically heavier than that predicted from $X_\text{max}$.
This discrepancy between the inferred mass from these two observables has been observed in other experiments and is interpreted as a muon deficit in simulations. This suggests that current hadronic models used to simulate air showers—tuned to LHC data and extrapolated to higher energies—do not consistently reproduce all aspects of EAS.
We find that the discrepancy between the inferred mass of the two observables remains relatively constant with energy up to $10^{18.4}$ eV but gradually increases beyond that point.
Although the number of events above $10^{18.4}$ eV is still low, our results suggest that the muon deficit may become more pronounced at these energies, providing valuable insights for model builders seeking to improve hadronic interaction models.