First search for atmospheric and extraterrestrial neutrino-induced cascades with the IceCube detector

We report on the first search for atmospheric and for diffuse astrophysical neutrino-induced showers (cascades) in the IceCube detector using 257 days of data collected in the year 2007–2008 with 22 strings active. A total of 14 events with energies above 16 TeV remained after event selections in the diffuse analysis, with an expected total background contribution of $8.3\pm 3.6$. At 90% confidence we set an upper limit of $E^2{\Phi }_{90\%\mathrm{CL}}<3.6\times {10}^{-7}\mathrm{GeV}·{\mathrm{cm}}^{-2}·{\mathrm{s}}^{-1}·s{r}^{-1}$ on the diffuse flux of neutrinos of all flavors in the energy range between 24 TeV and 6.6 PeV assuming that $\Phi \propto E^{-2}$ and the flavor composition of the ${\nu }_e:{\nu }_{\mu }:{\nu }_{\tau }$ flux is $1:1:1$ at the Earth. The atmospheric neutrino analysis was optimized for lower energies. A total of 12 events were observed with energies above 5 TeV. The observed number of events is consistent with the expected background, within the uncertainties.

Figure 1

Flux predictions for atmospheric muon and electron neutrinos, and measurements of the atmospheric muon neutrino spectrum by Fréjus [4], SuperK [5], AMANDA [6, 7], and IceCube [8, 9], and of the atmospheric electron neutrino spectrum by Fréjus [4]. Predictions for conventional [27] and prompt [12] atmospheric neutrinos are shown separately. All fluxes are averaged over all arrival directions, include both neutrinos and antineutrinos, and are weighted by $E_{\nu }^2$ for clarity.

Figure 2

a) The filled circles show the positions of the strings in the $x-y$ (horizontal) plane for the 22-string detector configuration. The lines show the boundaries of the fiducial volume and are described in the text. b) The reconstructed center-of-gravity position $x$ (COGX) versus $y$ (COGY) for events passing the cascade online filter. The right axis shows the rate [Hz].

Figure 3

Normalized distributions of (top) zenith angle and (bottom) reduced log-likelihood ratios after Level-5 event selections for the data (filled circles), muon background Monte Carlo (continuous histogram) and signal neutrino (dashed histogram).

Figure 4

Normalized distributions of cascade reconstructed vertex position: $x$ component (left panel), $y$ component (middle panel) and $z$ component (right panel) after level-6 event selections for the data (filled circles), standard muon background Monte Carlo (continuous histogram).

Figure 5

(top) Normalized distribution of the reconstructed cascade energy after level-6 event selections for the data (filled circles), muon background Monte Carlo (continuous histogram) and signal neutrino (dashed histogram). (bottom) Absolute event rate versus primary cosmic ray energy after level-6 event selections with an additional selection criterium on the reconstructed cascade energy of 6.3 TeV for the standard CORSIKA Monte Carlo (continuous histogram) and high-energy optimized CORSIKA Monte Carlo (open triangles).

Figure 6

The fill-ratio, $F$, versus the distance, $D$, as defined in the text for (top-left panel) the Monte Carlo simulated signal sample, (top-right panel) the Monte Carlo simulated muon background sample, and the data at the bottom part of the detector (bottom-left panel) and at the top part of the detector (bottom-right panel). The right axis shows the rate [Hz]. The dashed lines show the selection boundaries applied at Level-7 in the analysis.

Figure 7

Normalized distributions of the Level-4 neural network classification score.

Figure 8

Normalized distributions (after Level-4 atmospheric selections) of (left panel) reconstructed energy (middle panel) number of direct hits (right panel) total number of hits.

Figure 9

Normalized distributions (after Level-4 atmospheric selections) of (left panel) distance from the cascade vertex to the first hit in the event (middle panel) difference in $z$ vertex positions for the two split cascade reconstructions (right panel) fill-ratio from the mean hit distance.

Figure 10

Normalized distributions of the neural network classification scores (top) ${\mathrm{Q}}_{\mathrm{B}}$ and (bottom) ${\mathrm{Q}}_{\mathrm{C}}$.

Figure 11

The effective area, $A_{\mathrm{eff}}$, for (left panel) ${\nu }_e$, (middle panel) ${\nu }_{\mu }$, and (right panel) ${\nu }_{\tau }$ versus neutrino energy at trigger level (dashed curves) and at analysis level after all selections for both the extraterrestrial and the atmospheric analysis (continuous curves).

Figure 12

Distribution of the multivariate selection parameter ${\mathrm{Q}}^{\star }$ for data and for a simulated cascade signal from ${\nu }_e$ and ${\nu }_{\mu }$ with reconstructed energies above 5 TeV in the atmospheric analysis.

Figure 13

Experimental events reconstructed with energies above 5 TeV in the atmospheric analysis compared to an expected cascade signal from atmospheric neutrinos. (top) The cumulative number of events above a threshold in the multivariate selection parameter ${\mathrm{Q}}^{\star }$ for data and simulated signal. (bottom) The ratio of data to expected signal as a function of selection threshold.

Figure 14

Cascade reconstructed energy distribution after all selections in the extraterrestrial analysis.

Figure 15

Experimental upper limits on the diffuse flux of neutrinos from sources with $\Phi \propto E^{-2}$ energy spectrum and theoretical predictions for neutrino fluxes from astrophysical neutrino sources.

Figure 16

The MC muon background event which passed all selections in extraterrestrial analyses and which has the reconstructed energy of 18 TeV is displayed from the side. Different colors of the circles represent different DOMs hit times, with early hits in red and late hits in blue. The size of the circles represents the amplitude.

Figure 17

The data event which passed all selections in atmospheric and extraterrestrial analyses and which has the highest reconstructed energy of 134 TeV is displayed from the side. Different colors of the circles represent different DOMs hit times, with early hits in red and late hits in blue. The size of the circles represents the amplitude.