Kneelike Structure in the Spectrum of the Heavy Component of Cosmic Rays Observed with KASCADE-Grande

We report the observation of a steepening in the cosmic ray energy spectrum of heavy primary particles at about $8\times {10}^{16}\mathrm{eV}$. This structure is also seen in the all-particle energy spectrum, but is less significant. Whereas the “knee” of the cosmic ray spectrum at $3–5\times {10}^{15}\mathrm{eV}$ was assigned to light primary masses by the KASCADE experiment, the new structure found by the KASCADE-Grande experiment is caused by heavy primaries. The result is obtained by independent measurements of the charged particle and muon components of the secondary particles of extensive air showers in the primary energy range of ${10}^{16}$ to ${10}^{18}\mathrm{eV}$. The data are analyzed on a single-event basis taking into account also the correlation of the two observables.

Figure 1

Layout of the KASCADE-Grande experiment: shown are the Grande array as well as the KASCADE array with its central detector (CD) and muon tracking detector (MTD). The shaded area marks the outer 12 clusters (16 detector stations each) of the KASCADE array consisting of shielded (muon array) and unshielded detectors. The inner 4 clusters consist of unshielded detectors only.

Figure 2

Two-dimensional distribution of the shower sizes: charged particle number and total muon number. All quality cuts are applied. In addition, a roughly estimated energy scale is indicated in the upper panel. The lower panel shows a zoom to higher energies of the same observables, which is now corrected for attenuation.

Figure 3

Evolution of the $k$ parameter as a function of the reconstructed energy for experimental data compared with simulations of primary masses for the angular range 0°–24°. The error bars assign statistical as well as reconstruction uncertainties of $k$. The line displays the chosen energy dependent $k$ values for separating the mass groups, where the dashed lines assign the uncertainty of the selection.

Figure 4

Reconstructed energy spectrum of the electron-poor and electron-rich components together with the all-particle spectrum for the angular range 0°–40°. The error bars show the statistical uncertainties; the bands assign systematic uncertainties due to the selection of the subsamples. Fits on the spectra and resulting slopes are also indicated.

Figure 5

Energy spectra of electron-poor (heavy) event samples obtained by different selection and reconstruction criteria. The original spectrum from Fig. 4 is compared with the spectrum from a more selective cut in the $k$ parameter and with the spectrum obtained by using the $Y_{\mathrm{CIC}}$ parameter for selecting the electron-poor events (see text).