Prospects of establishing the origin of cosmic neutrinos using source catalogs

The cosmic neutrino flux recently discovered by IceCube will be instrumental in probing the highest-energy astrophysical processes. Nevertheless, the origin of these neutrinos is still unknown. While it would be more straightforward to identify a transient, or galactic source, class, finding a population of distant, continuous sources is challenging. We introduce a source-type classification technique that incorporates all available information from catalogs of source candidates. We show that IceCube-Gen2 can statistically establish the origin of cosmic neutrinos, even for the most challenging source populations–starburst galaxies, AGN, or galaxy clusters–if neutrino track directions can be reconstructed with a precision $\sim {0.3}^{°}$. We further show that the source catalog out to $\sim 100\mathrm{Mpc}$ can be sufficient for the most challenging source types, allowing for more straightforward source surveys. We also characterize the role of detector properties, namely angular resolution, size, and veto power in order to understand the effects of IceCube-Gen2’s design specifics.

Figure 1

Fraction of astrophysical neutrinos detected from sources within a luminosity distance, as a function of the luminosity distance, for different cosmic evolution models (see legend and Sec. 2e). For comparison, the top axis shows the corresponding redshift.

Figure 2

Simulated distribution of neutrino flux ${\mathcal{F}}_{\nu }$ corresponding to signal and background neutrinos. This example shows the densities for $\psi =0.3°$, $d_{\mathrm{th}}=200\mathrm{Mpc}$, for starburst galaxies. The dashed red line shows a power-law fit on the background density’s tail. The black solid lines show power-law slopes with ${{\mathcal{F}}_{\nu }}^{-3/2}$ and ${{\mathcal{F}}_{\nu }}^{-5/2}$, the theoretical expectations for the signal and neutrino models, respectively, for the case in which the expected number of sources within $d_{\mathrm{th}}$ coincident with a neutrino is $\ll 1$.

Figure 3

Simulated distributions of likelihood ratio $\lambda$ for the $\mathrm{signal}+\mathrm{background}$ and background-only models. This example shows the distributions for $d_{\mathrm{th}}=200\mathrm{Mpc}$, $T_{\mathrm{obs}}=10\mathrm{yr}$, for starburst galaxies, and with angular uncertainties specified in the legend. The vertical dashed line shows the likelihood ratio corresponding to the $p=3\sigma$ p-value from the background distribution.

Figure 4

Probability of a $\mathrm{signal}+\mathrm{background}$ observation having a p-value $\le p_0=3\sigma =0.00135$ as a function of threshold distance $d_{\mathrm{th}}$, for different angular uncertainties (see the legend). The results are obtained using the starburst-galaxy model, assuming $T_{\mathrm{obs}}=10\mathrm{yr}$ with IceCube-Gen2 for the highest energy ($\ge 30\mathrm{TeV}$) neutrinos.

Figure 5

Probability of a $\mathrm{signal}+\mathrm{background}$ observation having a p-value $\le p_0=0.00135$ ($3\sigma$) as a function of neutrino angular uncertainty, for different source assumptions (see the legend). For AGN, the spread corresponds to the range ${10}^{-5}–{10}^{-4}{\mathrm{Mpc}}^{-3}$ source density. The results are obtained assuming $T_{\mathrm{obs}}=10\mathrm{yr}$ with IceCube-Gen2 in the Sunflower 240 configuration, for the highest energies with $>24\mathrm{TeV}$ neutrinos, corresponding to a detection rate of 36 astrophysical and 36 atmospheric neutrinos per year.

Figure 6

(Same as Fig. 5, but for the low-threshold neutrino search.) Probability of a $\mathrm{signal}+\mathrm{background}$ model having a p-value $\le p_0=0.00135$ ($3\sigma$) as a function of neutrino angular uncertainty for a low-threshold ($>1\mathrm{TeV}$) neutrino search, for different source models (see the legend). The results are obtained assuming $T_{\mathrm{obs}}=10\mathrm{yr}$ with IceCube-Gen2.

Figure 7

Probability of a $\mathrm{signal}+\mathrm{background}$ observation having a p-value $\le p_0=0.00135$ ($3\sigma$) as a function of the number of detected neutrinos. The x axis is normalized to the case of IceCube-Gen2 in the Sunflower 240 configuration, for the highest energies with $>24\mathrm{TeV}$ neutrinos, which corresponds to a detection rate of 36 astrophysical and 36 atmospheric neutrinos per year. We show results for different source assumptions (see the legend). The number of both astrophysical and atmospheric neutrinos is scaled identically. For AGN, the spread corresponds to the range ${10}^{-5}–{10}^{-4}{\mathrm{Mpc}}^{-3}$ source density. The results are obtained assuming an angular uncertainty of 0.3° and $T_{\mathrm{obs}}=10\mathrm{yr}$.

Figure 8

Probability of a $\mathrm{signal}+\mathrm{background}$ observation having a p-value $\le p_0=0.00135$ ($3\sigma$) as a function of the number of detected atmospheric neutrinos. The x axis is normalized to the case of IceCube-Gen2 in the Sunflower 240 configuration, for the highest energies with $>24\mathrm{TeV}$ neutrinos, which corresponds to a detection rate of 36 atmospheric neutrinos per year. The detection rate of astrophysical neutrinos is kept fixed at $36{\mathrm{yr}}^{-1}$. We show results for different source assumptions (see the legend). For AGN, the spread corresponds to the range ${10}^{-5}–{10}^{-4}{\mathrm{Mpc}}^{-3}$ source density. The results are obtained assuming an angular uncertainty of 0.3°, and $T_{\mathrm{obs}}=10\mathrm{yr}$.