Evidence for Astrophysical Muon Neutrinos from the Northern Sky with IceCube

Results from the IceCube Neutrino Observatory have recently provided compelling evidence for the existence of a high energy astrophysical neutrino flux utilizing a dominantly Southern Hemisphere data set consisting primarily of ${\nu }_e$ and ${\nu }_{\tau }$ charged-current and neutral-current (cascade) neutrino interactions. In the analysis presented here, a data sample of approximately 35 000 muon neutrinos from the Northern sky is extracted from data taken during 659.5 days of live time recorded between May 2010 and May 2012. While this sample is composed primarily of neutrinos produced by cosmic ray interactions in Earth’s atmosphere, the highest energy events are inconsistent with a hypothesis of solely terrestrial origin at $3.7\sigma$ significance. These neutrinos can, however, be explained by an astrophysical flux per neutrino flavor at a level of $\mathrm{\Phi }(E_{\nu })={9.9}_{-3.4}^{+3.9}\times {10}^{-19}{\mathrm{GeV}}^{-1}{\mathrm{cm}}^{-2}{\mathrm{sr}}^{-1}{\mathrm{s}}^{-1}{(E_{\nu }/100\mathrm{TeV})}^{-2}$, consistent with IceCube’s Southern-Hemisphere-dominated result. Additionally, a fit for an astrophysical flux with an arbitrary spectral index is performed. We find a spectral index of ${2.2}_{-0.2}^{+0.2}$, which is also in good agreement with the Southern Hemisphere result.

Figure 1

The distribution of reconstructed zenith angles of events in the final sample, compared to the expected distributions for the fit of an $E^{-2}$ astrophysical neutrino spectrum. Only statistical errors are shown, though in almost all bins they are small enough to be hidden by the data markers.

Figure 2

The distribution of reconstructed muon energy proxy for events in the final sample, compared to the expected distributions for the fit of an $E^{-2}$ astrophysical neutrino spectrum. Only statistical errors are shown. The energy proxy does not have a linear relationship to actual muon energy, but values $\sim 3\times {10}^3$ are roughly equivalent to the same quantity in GeV. Larger proxy values increasingly tend to underestimate muon energies, while smaller values tend to overestimate.

Figure 3

Likelihood profile of the astrophysical flux power-law index and the flux normalization at 100 TeV in units of ${10}^{-18}{\mathrm{GeV}}^{-1}{\mathrm{cm}}^{-2}{\mathrm{sr}}^{-1}{\mathrm{s}}^{-1}$. While the $E^{-2}$ result is well within the 68% contour, it is not the overall best fit. Also shown are the best fits from various IceCube analyses of starting events, which generally have good agreement: Starting Events (HE) [4], Starting Events (LE 1) [32], Starting Events (LE 2) [33].

Figure 4

Comparison of the best-fit per-flavor astrophysical flux spectrum of $E^{-2}$ from this work, assuming a flavor ratio of $1:1:1$ (shown in dark green with the 68% error range in lighter green) to other selected IceCube measurements (heavy lines) [4, 12] and theoretical model predictions (thin dashed lines) [5, 6, 7, 17, 20, 28]. The sensitivity of this analysis is also shown as the thin green line.