Spectrum of PeV Cosmic-Ray Protons and Helium Nuclei with IceCube

TL;DR

The paper addresses the gap between direct measurements and IceCube’s prior composition results by extracting the light-component spectrum (protons and helium) in the 0.3–1.4 PeV range using a hybrid IceTop–in-ice dataset and a neural-network predictor for energy and lightness. It introduces two calibration schemes (agnostic and GSF-informed) to infer the H+He fraction, validating results across hadronic models Sibyll 2.3d, QGSJet-II.04, and EPOS-LHC. The reconstructed spectrum follows a approximately $E^{-2.7}$ trend with a knee near $8\,\text{PeV}$, with light primaries dominating in the studied region and general agreement with indirect measurements. This work closes the gap between direct and indirect measurements of the H+He flux and highlights systematic avenues to improve future analyses, including atmospheric corrections, ice modeling, and updated hadronic-model MCs.

Abstract

The IceCube Observatory comprises a cubic-kilometer particle detector deep in the Antarctic ice and the cosmic-ray air-shower array IceTop at the surface above. Previous analyses of the cosmic-ray composition have used coincident events with IceTop detecting the electromagnetic shower footprint as well as GeV muons, while the sensors submerged in the ice measure the TeV muons from the same events. The energy range of previous composition analyses, however, has been limited to 3 PeV primary energy and above, whereas the IceTop all-particle energy spectrum has been extended down to 250 TeV. This contribution presents a method to reconstruct the combined spectrum of cosmic-ray protons and helium nuclei, starting at 200 TeV primary energy. The resulting H+He spectrum closes the gap in the measurements of light cosmic rays between IceCube as well as KASCADE and experiments measuring in the TeV energy range, such as DAMPE and HAWC.

Figures (4)

• Figure 1: Calibration of the H+He fraction for the energy bin $10^6$--$10^{6.15}$ GeV. Top left: distribution of sampled subsets, nominal fit to the median true fraction and definition of sampling uncertainty via percentiles, top right: composition uncertainty for the agnostic method (gray band) with extreme composition assumptions, bottom right: composition uncertainty for the GSF-informed method (red band) using 200 variations of the best-fit GSF model.
• Figure 2: Reconstructed H+He fraction using the calibrations that are based on Monte Carlo produced with Sibyll 2.3d, QGSJet-II.04 and EPOS-LHC. Error bars represent the combined systematic uncertainty due to sampling as well as unknown composition in the vertical direction and energy resolution horizontally with the markers positioned at $\langle\log_{10}(E/\mathrm{GeV})\rangle$. Top: agnostic calibration, bottom: GSF-informed method.
• Figure 3: Compilation of the reconstructed H+He flux using the agnostic (blue) and the GSF-informed method (yellow), based on Sibyll 2.3d Monte Carlo. Systematic uncertainties are represented by brackets while error bars describe statistical uncertainty and energy resolution. The combined proton and helium flux of the composition models H4a, GST and GSF (dashed lines) is included for comparison.
• Figure 4: The differential spectrum of cosmic-ray protons and helium nuclei with one year of IceCube data (black). The hadronic model used is Sibyll 2.3d and the GSF-informed calibration method has been applied. Error bars represent energy resolution and statistical uncertainty in the horizontal and vertical direction, respectively. The total systematic uncertainty is indicated by brackets. The combined proton and helium flux as measured in the past by IceCube and other experiments is shown in various colors. Boxes represent their systematic and error bars their statistical uncertainties with the exception of CREAM and KASCADE (2005) where only total uncertainties are available. Open markers represent results obtained via direct detection. For references, see the text.