Galactic PeV neutrinos from dark matter annihilation

The IceCube Neutrino Observatory has observed highly energetic neutrinos in excess of the expected atmospheric neutrino background. It is intriguing to consider the possibility that such events are probing fundamental physics beyond the standard model of particle physics. In this context, $\mathcal{O}(\mathrm{PeV})$ dark matter particles decaying to neutrinos have been considered, while dark matter annihilation has been dismissed invoking the unitarity bound as a limiting factor for the annihilation rate. However, the latter claim was done ignoring the contribution from dark matter substructure, which in a PeV cold dark matter scenario would extend down to a free-streaming mass of $\mathcal{O}(10^{-18}{\mathrm{M}}_{\odot })$. Since the unitarity bound is less stringent at low velocities, $({\sigma }_{\text{ann}}v)\le 4\pi /m_{\chi }^{2}v$, then it is possible that these cold and dense subhalos would contribute dominantly to a dark-matter-induced neutrino flux and easily account for the events observed by IceCube. A dark matter model in which annihilations are enhanced by a Sommerfeld mechanism can naturally support such a scenario. Interestingly, the spatial distribution of the events shows features that would be expected in a dark matter interpretation. Although not conclusive, 9 of the 37 events appear to be clustered around an extended region near the Galactic center, while 6 others spatially coincide, within the reported angular errors, with 5 of 26 Milky Way satellites. However, a simple estimate of the probability of the latter occurring by chance is $\sim 35\%$. More events are needed to statistically test this hypothesis. PeV dark matter particles are massive enough that their abundance as standard thermal relics would overclose the Universe. This issue can be solved in alternative scenarios, for instance, if the decay of new massive unstable particles generates significant entropy reheating the Universe to a slightly lower temperature than the freeze-out temperature, $T_{\mathrm{RH}}\lesssim T_{\mathrm{f}}\sim 4\times 10^4\mathrm{GeV}$.

Figure 1

Average rate of neutrinos ($E_{\nu }=m_{\chi }=1.2\mathrm{PeV}$) expected in IceCube within an angle $\mathrm{\Psi }$ from the Galactic center (solid angle $\mathrm{\Delta }\mathrm{\Omega }$) produced by dark matter annihilating exclusively into neutrinos along the line of sight. The solid lines are for the smooth dark matter distribution, while the dashed and dotted lines are the contributions from substructure down to a mass of $10^{-18}{\mathrm{M}}_{\odot }$ using two different models to account for substructure. For the left panel, we assume a constant $⟨{\sigma }_{\text{ann}}v⟩$ equal to the unitarity limit set locally [Eq. (2) with ${\beta }_{\text{loc}}\sim 3.7\times 10^{-4}$]. For the right panel (red lines), $⟨{\sigma }_{\text{ann}}v⟩$ is normalized to the same local value, but it scales as $1/\beta$ (Sommerfeld enhancement) until it saturates at $S_{\max }=100$ times the local value. The blue dotted line is for $m_{\min }=10^{-12}{\mathrm{M}}_{\odot }$. The dotted-dashed line in both panels is the estimate according to Eq. (1).

Figure 2

Top panel: Aitoff projection in equatorial coordinates of 33 of the 37 high-energy neutrino events reported by IceCube (crosses). We removed from the sample the two events with the largest angular errors ($\gtrsim 42^{\circ }$) and also two more (numbers 28 and 32 in Ref. [3]) that are very likely produced in cosmic ray air showers. The median angular error in the location of each event is shown with a circle. The blue circles mark the locations of 26 MW satellites. The six red crosses are neutrinos coincident with the position of at least one satellite (excepting Sagittarius). The red square marks the position of SgrA*. The red line circles all eight events in the inner halo region (defined arbitrarily as 40° around Sagittarius). Bottom panel: Discrete cumulative probability distribution function of randomly drawing 25 points in the sky, excluding the inner halo region, and having $N>N_{\text{rand}}$ neutrinos coincident with at least one of these 25 points (within the angular errors). The dashed blue line marks the six actual events coincident with five MW satellites. The probability that this number (or higher) occurs randomly is $\sim 35\%$.

Figure 3

Ratio of the expected number of neutrino events within 0.5° of the center of 17 MW satellites to that within 40° from the Galactic center produced by dark matter annihilating exclusively into neutrinos. The $J$ factor [Eq. (10)] for each galaxy was taken from Ref. [29] (assuming a NFW profile; see their Table I) scaled up by the average Sommerfeld enhancement given by the corresponding velocity dispersion (taken from Ref. [31]): $⟨{\sigma }_{\text{ann}}v⟩\propto 1/{\overline{\sigma }}_{\text{vel}}$. In the left panel, the MW halo is assumed to have an Einasto profile as described in Sec. 2a, while in the right a constant density core of 8.5 kpc is assumed. In addition, the satellites have a subsubstructure boost with a Sommerfeld enhancement saturated at $S_{\max }=100$; see the text for details. The horizontal dashed line marks the $1\sigma$ region of the hypothetically observed satellite: Galactic center ratio of events, $1:8$.