Multimessenger gamma-ray counterpart of the IceCube neutrino signal

A signal of high-energy extraterrestrial neutrinos from unknown source(s) was recently discovered by the IceCube experiment. Neutrinos are always produced together with $\gamma$-rays, but the $\gamma$-ray flux from extragalactic sources is suppressed due to attenuation in the intergalactic medium. We report the discovery of a $\gamma$-ray excess at high galactic latitudes starting at energies 300 GeV in the data of the Fermi telescope. We show that the multi-TeV $\gamma$-ray diffuse emission has spectral characteristics at both low and high galactic latitudes compatible with those of the IceCube high neutrino signal in the same sky regions. This suggests that these $\gamma$-rays are the counterpart of the IceCube neutrino signal, implying that a sizable part of the IceCube neutrino flux originates from the Milky Way. We argue that the diffuse neutrino and $\gamma$-ray signal at high galactic latitudes originates either from previously unknown nearby cosmic ray “PeVatron” source(s), an extended galactic cosmic ray halo or from decays of heavy dark matter particles.

Figure 1

Top: The multimessenger spectrum of the full sky. Fermi/LAT spectrum of $\gamma$-ray emission is shown by black data points; thick and thin error bars show statistical and systematic uncertainties. IceCube data are shown by blue data points and by the green bow-tie (from Ref. [21]). The dash-dotted curve shows the model of galactic diffuse emission component from ${\pi }^0$ decays [26]. Middle: The Fermi/LAT spectrum of the galactic plane $|b|<10°$ (black data points). The blue dash-dotted curve shows model-dependent upper limit on neutrino flux derived under the assumption about particular shape of the ${\pi }^{\pm }$ decay spectrum [23]. The thin blue curve is an envelope of the upper bounds on the power-law spectra [21] (see the Appendix pp2). The gray dotted curve shows the model of ${\pi }^0$ decay component of diffuse $\gamma$-ray flux [26]. Bottom: Fermi/LAT spectrum of $|b|>10°$ region, compared to the IceCube neutrino flux measurements. The dash-dotted curve shows the best-fit model of the isotropic diffuse $\gamma$-ray background (IGRB) [27].

Figure 2

High galactic latitude emission for the local PeVatron (top) and DM (bottom) models. Thick and thin error bars of Fermi/LAT data points (black) show statistical and systematic uncertainties, including the uncertainties of subtraction of IGRB and residual CR backgrounds. Vertical arrows show KASCADE upper limits on the $\gamma$-ray flux from Northern sky [28]. Solid thin lines show the gamma-ray emission from the additional hard component. Dashed lines show the neutrino emission. The dotted line shows a broken power-law fit to the sub-TeV $\gamma$-ray spectrum. The thick solid line shows the sum of the sub-TeV and additional hard $\gamma$-ray components.

Figure 3

Comparison of Fermi/LAT spectra of Crab extracted using unbinned likelihood (gray data points) and aperture photometry (red data points) methods. The gray thick line shows the Crab spectrum measured by ground-based $\gamma$-ray telescopes [50].

Figure 4

Stacked Fermi/LAT spectrum of isolated sources listed in Table 1. Gray thin data points show the spectrum extracted assuming effective area calculated with the gtexposure tool, black thick data points show the spectrum calculated with effective area renormalized by the cross-calibration factor $\kappa$ (black data points). The gray thick line shows the average over the sources model spectrum.

Figure 5

Combined Fermi/LAT, ARGO-YBJ [19] and MILARGO [20] spectrum of the $40°<l<100°$ stretch of the galactic plane. Gray and black data points show the measurements (notations are the same as in Figs. 1 and 2 of the main text). The green data point is from MILAGRO [20], blue data points are from ARGO-YBJ [19].

Figure 6

Fermi/LAT spectrum of diffuse $\gamma$-ray emission from high galactic latitude (black data points and gray data point above 3 TeV) compared to the level of residual CR background (dotted thin line) derived in Ref. [27] extended to the energy range above 1 TeV (thick dotted line). The red thin dotted line shows the increase of residual CR background under the assumption of hardening of the CR count rate power-law slope by 1. Red data points show the high galactic latitude diffuse emission spectrum calculated assuming this higher residual CR background. The green shaded band shows the range of uncertainty of IGRB derived in Ref. [27]. Gray data points below 3 TeV show total high galactic latitude flux without subtraction of catalog sources, isotropic diffuse $\gamma$-ray background and residual CR background contributions.

Figure 7

Comparison of the total count ($\gamma$-ray $+$ residual cosmic ray background) spectra of the sky regions $|b|>20°$ extracted using PASS7 IGRB and PASS8 ULTRACLEANVETO event selections. The top panel shows the spectral data, the bottom panel shows the difference between the two spectral measurements compared to the systematic error delimited by the dashed line. The PASS7 spectrum is from Ref. [27].

Figure 8

The same as in Fig. 6 but for the “maximal possible” residual cosmic ray background estimate. The gray solid curve shows the maximal possible residual cosmic ray background in the PASS8 ULTRACLEANVETO event selection.

Figure 9

Point source contribution to the flux of high galactic latitude emission. The rose shaded range shows the estimate from Ref. [27]. Red data points show the calculation based on the 3FGL catalog [60].