Pierre Auger Observatory

Evidence for a mixed mass composition at the 'ankle' in the cosmic ray spectrum

The highest energy cosmic rays remain elusive and mysterious, and their study requires extraordinary efforts. At the Pierre Auger Observatory in Argentina, the giant air showers of particles created by these cosmic rays are detected when they slam into the ground by a large array of water tanks equipped with electronic detectors. But on dark nights they are simultaneously detected by telescopes sensitive to the faint sky glow left by the air showers. The new report by the Pierre Auger Collaboration correlates in detail the signals in the water tanks with those from the telescopes. The correlation is uniquely sensitive to the presence in the primary beam of nuclei with different masses, and is used in particular to help resolve how many types of atomic nuclei contribute to the cosmic ray flux.

Evidence for a mixed mass composition at the ‘ankle’ in the cosmic ray spectrum

Cosmic rays are energetic particles (atomic nuclei) impinging upon the Earth from the vast reaches of the cosmos. They can have tremendous energies and thus must originate in remarkable but still mysterious astrophysical sources. If we understand the nature of these particles, we may be able to find the extragalactic sources of the highest energy cosmic rays.
The distribution of particle numbers with energy, or 'spectrum', shows a striking and very rapid reduction in numbers with increasing energy. At an energy of E ≈ 5 × 1018 eV a flattening in the spectrum (slightly slower rate of reduction with energy) is observed. This feature is called the 'ankle' in the cosmic ray spectrum.
The transition from cosmic rays from a Galactic origin to an extragalactic component may cause such a flattening. Alternatively, the 'dip' model in which highly energetic extragalactic protons interact with photons from the cosmic microwave background also creates a flattening of the spectrum. The different models can be distinguished by the predicted cosmic ray composition in the ankle region.
In this paper the important characteristics of the mass composition of the cosmic rays — the spread of their masses — is measured using a method relatively insensitive to either experimental uncertainties or uncertainties in the particle interaction models. The method takes advantage of the hybrid design of the Pierre Auger Observatory, which measures both the cascade of particles when it reaches the ground and the development of the shower through the atmosphere using specialized fluorescence telescopes on dark clear nights.
The method uses the independent information on the depth of shower maximum (Xmax) from the fluorescence telescopes and the signal at 1000 m from the shower axis, S(1000), from the surface detector. The original idea of the method is that a direct and robust estimation of the spread of masses in the primary beam can be obtained via a measurement of the correlation between Xmax and S(1000). For any given type of cosmic ray nucleus this correlation is close to or larger than zero as shown in the first figure (left) for proton and iron air showers generated with the particle interaction model EPOS-LHC (the correlation coefficient is denoted as rG). For mixed compositions a negative correlation emerges due to a very general characteristic of air showers: showers from heavier nuclei have smaller Xmax and larger S(1000) (due to a larger number of muons). Thus in a mixed composition shallower showers have on average larger signals. The correlation becomes more negative the larger the spread of masses.
For events successfully reconstructed with both Auger fluorescence and surface detectors in the energy range around the ankle E = 1018.5 - 1019.0 eV a significant negative correlation was found rG = -0.125 ± 0.024 (stat), as illustrated in the first figure (right). This value is at least 5 standard deviations away from predictions for pure compositions or any composition made up of protons and helium only.

2016-11 paper xmax1
Correlation between X*max (Xmax scaled to 1019 eV) and S*38 (S(1000) scaled to 1019 eV,
38° of zenith angle).      Left: simulations for protons and iron nuclei with EPOS-LHC. Right: Auger data.


To estimate the spread of the primary mass numbers σ(ln A) the value of the correlation found in the data is compared to values from simulations for compositions with varying fractions of protons, helium, oxygen and iron. As illustrated in the second figure, for different particle interaction models (EPOS-LHC or QGSJetII-04) the data can be described with compositions having a spread in atomic mass numbers within the range 1.0 ≲ σ(ln A)≲ 1.7. The results are practically unaffected by systematic uncertainties on Xmax and S(1000), or by modifications of the key parameters of the particle interaction models.


2016-11 paper xmax2
Correlation coefficient as a function of spread of cosmic ray mass number, in simulated (with two separate particle interaction models, EPOS-LHC and QGSJetII-04) mixtures with different fractions of protons, helium, oxygen and iron nuclei (points) compared to the correlation value found in Auger data (shaded area). The ranges of cosmic ray mass spread compatible with the data are marked by vertical lines.

The conclusion that the mass composition around the cosmic ray ankle energy is not pure but mixed has important consequences for theoretical source models. Proposals of almost pure compositions, such as the dip scenario, are disfavored as a sole explanation of the ultra-high energy cosmic rays. These findings, together with other observations made at the Pierre Auger Observatory, indicate that nuclei with mass number A > 4 are accelerated to ultra-high energies E ≥ 1018.5 eV and are able to escape the source environment. The search for the final resolution of the ankle puzzle is continuing and new astrophysical models are already emerging, e.g., including modifications of the nuclear composition in the environment of the acceleration sites.
Related paper:
Evidence for a mixed mass composition  at the 'ankle' in the cosmic-ray spectrum
A. Aab et al. (Pierre Auger Collaboration), Phys.Lett. B762 (2016) 288-295