Search for Galactic PeV gamma rays with the IceCube Neutrino Observatory

Gamma-ray induced air showers are notable for their lack of muons, compared to hadronic showers. Hence, air shower arrays with large underground muon detectors can select a sample greatly enriched in photon showers by rejecting showers containing muons. IceCube is sensitive to muons with energies above $\sim 500\mathrm{GeV}$ at the surface, which provides an efficient veto system for hadronic air showers with energies above 1 PeV. One year of data from the 40-string IceCube configuration was used to perform a search for point sources and a Galactic diffuse signal. No sources were found, resulting in a 90% C.L. upper limit on the ratio of gamma rays to cosmic rays of $1.2\times {10}^{-3}$ for the flux coming from the Galactic plane region ($-80°\lesssim l\lesssim -30°$; $-10°\lesssim b\lesssim 5°$) in the energy range 1.2–6.0 PeV. In the same energy range, point source fluxes with $E^{-2}$ spectra have been excluded at a level of $(E/\mathrm{TeV}{)}^2\mathrm{d}\Phi /\mathrm{d}E\sim {10}^{-12}–{10}^{-11}{\mathrm{cm}}^{-2}{\mathrm{s}}^{-1}{\mathrm{TeV}}^{-1}$ depending on source declination. The complete IceCube detector will have a better sensitivity (due to the larger detector size), improved reconstruction, and vetoing techniques. Preliminary data from the nearly final IceCube detector configuration have been used to estimate the 5-yr sensitivity of the full detector. It is found to be more than an order of magnitude better, allowing the search for PeV extensions of known TeV gamma-ray emitters.

Figure 1

IceCube consists of a $\sim {\mathrm{km}}^2$ surface air-shower array and 86 strings holding 60 optical modules each, filling a physical volume of ${\mathrm{km}}^3$. The region in the center of the buried detector is more densely instrumented. See text for details.

Figure 2

Map of the locations of all 86 strings of the completed IceCube detector. The blue dots represent the 40-string configuration that is used for this analysis. At the surface level, each of these 40 strings is complemented by an IceTop station consisting of two tanks. The large (red) circles indicate the “inner strings” of the IceCube40 configuration.

Figure 3

Contours of HI column densities [14] in Galactic coordinates (flat projection). The blue circle indicates the gamma-ray FOV for IceCube in the present IC40 analysis. The red rectangles indicate the regions for which CASA-MIA [6] has set an upper limit on the Galactic diffuse photon flux in the 100 TeV–1 PeV energy range. IceCube’s FOV is smaller but covers a different part of the Galactic plane, close to the Galactic center.

Figure 4

Distribution of the number of muons with $E>500\mathrm{GeV}$ at 3 km altitude in samples of simulated showers of 0.6 PeV (red, dashed), 1 PeV (black, solid), and 2 PeV (blue, dotted). Each of the samples consists of 10,000 showers simulated with CORSIKA (using QGSJETII) with zenith angle $<30°$. The 1-PeV sample contains only 17 showers that contain no muons with energy above 500 GeV.

Figure 5

Ratio between reconstructed and true energy of simulated gamma-ray showers as a function of their true energy. At low energies, the overestimation of the gamma-ray energy is largely due to a bias effect of the eight-station filter. At higher energies, this overestimation decreases.

Figure 6

Distribution of true energy of gamma-ray (red, solid) and proton (blue, dotted) showers for an energy cut at $E_{\mathrm{reco}}>1.4\mathrm{PeV}$, weighted to an $E^{-2.7}$ spectrum.

Figure 7

Optimization scans for cut parameters $E_{\min }$, $d_{\max }$, and $C_{\max }$. The fraction of scrambled data sets that perform equal or better than the real data set is plotted against cut value. For each plot, the other 2 parameters are kept constant at their optimal value. The actual scan was done three dimensionally. For each plot, the shaded region indicates the parameter space that is excluded by the optimized cuts.

Figure 8

Equatorial map of the 268 candidate gamma-ray events of the IC40 data set superimposed on HI column densities based on Ref. [14]. The dotted black curve encloses the source region, defined as within 10° of the Galactic plane.

Figure 9

Equatorial map of the part of the sky for which an unbinned point source search is performed. The contours indicate the value of $\lambda$, and the black dots are the candidate events.

Figure 10

Distribution of the largest value of $\lambda$ observed in each scrambled data set. The red dotted line indicates the value for $\lambda$ that corresponds to the hottest spot in the data.

Figure 11

Sky map of 90% C.L. upper limits on point source flux $(E/\mathrm{TeV}{)}^2\mathrm{d}\Phi /\mathrm{d}E$ in ${\mathrm{cm}}^{-2}{\mathrm{s}}^{-1}{\mathrm{TeV}}^{-1}$ for $E^{-2}$ source spectra in the energy range $E=1.2–6.0\mathrm{PeV}$. The limit is typically more constraining at low declinations where the effective area is largest.

Figure 12

Observed time of isolated hits (SLC) relative to the arrival time of the air-shower front measured by IceTop. The left plot shows the distribution of SLC hits for all events; the right plot is the same but restricted to the subset of events which have only SLC hits and no HLC hits. There is an excess of SLC hits in the region were a muon signal associated to the shower is expected. This allows an additional cut to separate gamma-ray showers from hadronic showers. The black box indicates the region in which an SLC hit counts as a veto (see text for details). The variation in the number of hits as a function of depth in the left plot is due to variations in the optical properties of the ice.

Figure 13

The number of L3 (red) and L4 (blue) events in the data as a function of the three main cut parameters. The ratio of the number of L4 to L3 events is given by the black dotted line, for which the corresponding axis is drawn on the right-hand side of the plot.

Figure 14

Acceptance (effective area integrated over solid angle) for showers with an axis through both IceTop and IceCube for IC40, $C_{\mathrm{IT}}=0.8$ (black/lower line), and IC86, $C_{\mathrm{IT}}=1.0$ (blue/upper line).

Figure 15

Existing limits (red triangles for CASA-MIA and purple line for the present IC40 analysis) and IceCube sensitivity to a diffuse gamma-ray flux from a region within 10° from the Galactic plane. The blue dashed line indicates the 5-yr sensitivity of the completed detector, while the blue dots represent the sensitivity in smaller energy bins.

Figure 16

IceCube 5-yr sensitivity to point sources as a function of declination. The solid (dashed) black line indicates the sensitivity to an $E^{-2}(E^{-2.5})$ flux. The dashed red line indicates the lowest declination reached by the Galactic plane. The blue points indicate the flux at 1 PeV with extrapolated uncertainties of the sources listed in Table 1 in the absence of a cutoff.