Search for a diffuse flux of astrophysical muon neutrinos with the IceCube 59-string configuration

TL;DR

This study leverages IceCube's 59-string data to search for a diffuse astrophysical muon-neutrino flux via a global likelihood fit to two-dimensional energy-zenith PDFs, separating astrophysical, prompt atmospheric, and conventional atmospheric components while rigorously treating systematic uncertainties. The best-fit astrophysical flux is $E^2\Phi(E) = 0.25 \times 10^{-8}$ with a zero prompt contribution, and a 90% CL upper limit of $1.44 \times 10^{-8}$; the result shows a mild excess (~1.8σ) that prevents excluding the Waxman-Bahcall bound. The analysis excludes several models for diffuse astrophysical flux at 90% CL and places meaningful limits on prompt atmospheric neutrinos, illustrating the improved sensitivity attainable with multi-parameter likelihoods and careful treatment of systematics. These findings demonstrate IceCube’s growing capability to probe high-energy extragalactic neutrinos and guide future multi-channel investigations with the full detector.

Abstract

A search for high-energy neutrinos was performed using data collected by the IceCube Neutrino Observatory from May 2009 to May 2010, when the array was running in its 59-string configuration. The data sample was optimized to contain muon neutrino induced events with a background contamination of atmospheric muons of less than 1%. These data, which are dominated by atmospheric neutrinos, are analyzed with a global likelihood fit to search for possible contributions of prompt atmospheric and astrophysical neutrinos, neither of which have yet been identified. Such signals are expected to follow a harder energy spectrum than conventional atmospheric neutrinos. In addition, the zenith angle distribution differs for astrophysical and atmospheric signals. A global fit of the reconstructed energies and directions of observed events is performed, including possible neutrino flux contributions for an astrophysical signal and atmospheric backgrounds as well as systematic uncertainties of the experiment and theoretical predictions. The best fit yields an astrophysical signal flux for $ν_μ+ \barν_μ$ of $E^2 \cdot Φ(E) = 0.25 \cdot 10^{-8} \textrm{GeV} \textrm{cm}^{-2} \textrm{s}^{-1} \textrm{sr}^{-1}$, and a zero prompt component. Although the sensitivity of this analysis for astrophysical neutrinos surpasses the Waxman and Bahcall upper bound, the experimental limit at 90% confidence level is a factor of 1.5 above at a flux of $E^2 \cdot Φ(E) = 1.44 \cdot 10^{-8} \textrm{GeV} \textrm{cm}^{-2} \textrm{s}^{-1} \textrm{sr}^{-1}$.

Figures (13)

• Figure 1: Distribution of primary neutrino energies and zenith angles for conventional atmospheric Honda:2006qjGaisser:2012zz, prompt atmospheric Enberg:2008te and astrophysical $\nu_{\mu} + \overline \nu_{\mu}$ expected in the runtime of $348.1$ days, folded with the detection efficiency of this analysis. Note, the flux normalization of the latter two are multiplied by factors of 200 and 500 for better visibility in the right figure.
• Figure 2: Event view of the highest-energy neutrino event observed in this analysis. The grey dots mark IceCube DOMs. DOMs hit by photons are shown in color. The color code indicates the photon arrival time with red colors marking early times and blue colors standing for late times. The radius of the DOMs correlates with the observed charge. The reconstructed zenith angle of this event is $91.2^{\circ}\pm 0.1^{\circ}$ and the reconstructed, truncated muon energy loss is $\log(dE/dx\,\textrm{[GeV/m]}) = 1.37$ within the detection volume. Assuming the best-fit energy spectrum from this analysis (see Fig. \ref{['img:EnergyResolution']}), this event most likely originated from a neutrino of energy 500TeV-1PeV, producing a muon that passed through the detector with an energy of about 400 TeV.
• Figure 3: The cumulative distribution of the angular resolution for astrophysical $E^{-2}$ and conventional atmospheric events reconstructed by the MPE fit Ahrens:2003fg obtained from Monte Carlo simulation.
• Figure 4: Correlation between the truncated energy loss of the muon reconstructed with the algorithm Truncated Energy Abbasi:2012wht and the true energy loss of the muon (left), the muon energy when entering the detector (middle) and the primary neutrino energy (right) obtained from Monte Carlo. The spectral shape assumed in these plots is the best-fit superposition of atmospheric and astrophysical neutrino fluxes from this analysis. Each column has been normalized individually to $1$ for a better visibility of the reconstruction uncertainties.
• Figure 5: Neutrino effective area averaged for $\nu_{\mu}$ and $\overline \nu_{\mu}$ of the final event selection for different zenith bands.