Pis'ma v ZhETF, vol. 100, iss. 12, pp. 865-869
© 2014 December 25
IceCube astrophysical neutrinos without a spectral cutoff and 1015-1017 eV cosmic gamma radiation
O.Kalashev, S. Troitsky1') Institute for Nuclear Research of the RAS, 117312 Moscow, Russia
Submitted 11 October 2014
We present a range of unbroken power-law fits to the astrophysical-neutrino spectrum consistent with the most recent published IceCube data at the 68 % confidence level. Assuming that the neutrinos originate in decays of 7r-mesons, we estimate accompanying gamma-ray fluxes for various distributions of sources, taking propagation effects into account. We then briefly discuss existing experimental results constraining PeV to EeV diffuse gamma-ray flux and their systematic uncertainties. Several scenarios are marginally consistent both with the KASKADE and CASA-MIA upper limits at 101B-1016 eV and with the EAS-MSU tentative detection at -1017 eV, given large systematic errors of the measurements. Future searches for the diffuse gamma-ray background at sub-PeV to sub-EeV energies just below present upper limits will give a crucial diagnostic tool for distinguishing between the Galactic and extragalactic models of the origin of the IceCube events.
DOI: 10.7868/S0370274X14240011
The observation of an excess of high-energy neutrinos above the atmospheric background by the IceCube observatory [1-3] gave a strong boost to astroparticle physics (see, e.g., Ref. [4] for a review and references). While a firm conclusion about the astrophysical origin of these events would require future studies and a confirmation by an independent experiment, numerous scenarios have been put forward to explain the observation. Not surprisingly, present low statistics does not allow to single out a unique explanation of the origin of these events.
Absence of observed events with energies E > 3 PeV is often considered as an argument for the presence of a spectral cutoff at these energies. Indeed, the experimental exposure for electron antineutrinos peaks around ~ 6.3 PeV because of the Glashow resonance, and one would expect additional events while none is detected (see, e.g., Ref. [5] for a detailed discussion). This cutoff would add further uncertainty to astrophysical explanations because the maximal energy of 3 PeV is not singled out by any general argument. In this note, we assume that the neutrino spectrum continues beyond the highest observed energies and the absence of events at the Glashow resonance is a statistical fluctuation, not an indication of a cutoff. We will see that this assumption agrees with the data perfectly.
The general conventional model for production of energetic astrophysical neutrinos implies their creation in decays of charged tt-mesons, ir±, produced in turn
^e-mail: st@ms2.inr.ac.ru
in high-energy hadronic or photohadronic interactions. These 7r±'s are necessary accompanied by neutral 7r°'s which decay to photons. The energetic photons, therefore, have to accompany energetic neutrinos, see e.g. Refs. [6, 7] for discussions and estimates and Refs. [810] for more detailed model analyses in the context of the IceCube result. Since the neutrinos propagate freely through the Universe while the photons may be absorbed, a comparison of the two fluxes may give important information about the distribution of sources. In what follows, we will estimate the gamma-ray flux expected in various scenarios, starting from the IceCube data. Throughout the paper, we will assume that:
• the neutrino spectrum follows an unbroken power law,
where the diffuse flux F is measured in cm 2 x s_1 x sr_1;
• all neutrinos originate from 7r± decays, and the mechanism producing these 7r±'s provides for equal amounts of 7r+, 7t~, and ir°;
• the neutrino mixing is maximal, so, given previous assumptions, the neutrinos arrive to the observer with 1:1:1 flavor ratio.
Deviations from these assumptions are certainly present but their account is beyond the precision of the present data.
6 Письма в ЖЭТФ том 100 вып. 11-12 2014 865
To proceed, we first need to quantify the astrophysi-cal neutrino fiux consistent with observations. This part of the study may be useful for other phenomenological considerations, so we discuss it in some detail.
IceCube reported 37 high-energy astrophysical neutrino candidate events at the background expectation of ~ 15. This small statistics is expected to agree with various descriptions of the spectrum. The original IceCube paper [3] does not present a range of allowed spectral fits and quotes only two benchmark fits, one with fixed a = 2.0 and the best fit with a = 2.3. However, all necessary information to obtain the allowed region of parameters (N, a) at a given confidence level is published, and we will use it in this study. Namely, the energies of 36 of the observed events and the energy-dependent number of background events are given in Ref. [3] while the energy dependence of the exposure is presented in Ref. [7]. There are however two subtle points in the fitting procedure.
Firstly, the number of events in the sample, and especially in high-energy bins, is so small that the standard chi-square method would give a biased result, and the binned Poisson likelyhood does not follow the chi-square distribution. This is easy to cure, however, with the Monte-Carlo procedure described e.g. in Ref. [11], which is used here.
Secondly, individual neutrino energies cannot be measured and it is only the deposited energy in the ice which is reported. For cascade events (to which three highest-energy ones belong), the difference between the two energies may be neglected, while for track events the deposited energy gives only a lower limit on the true neutrino energy, which may be several times higher [12]. However, the present IceCube data allow for a nice trick to overcome this problem, suggested in Ref. [7]. It exploits the fact that, occasionally or not, there are no observed events with 400 TeV < E < 1 PeV, while all three events above 1 PeV are showers. This means that, no matter shower or track, the energies of all neutrinos except these three do not exceed 1 PeV. We therefore adopt the approach of Ref. [7] and use only three bins in the analysis: E < 1 PeV, 1 PeV < E < 2 PeV, and E > 2 PeV. We use events with E > 40 TeV in the fits.
The allowed ranges of the spectral fits are presented in Figs. 1 and 2. The best-fit values are:
a = 2.4, N = 8.7 • 10~8 TeV-1 • cm"2
sr
-l
(2)
Fig. 1. (Color online) The parameter space (normalization N versus spectral index a) for unbroken power-law fits, Eq. (1), of the astrophysical neutrino spectrum. The thick and thin contours bound the regions of parameters consistent with the IceCube data [3] at the 68 and 95 % C.L., respectively. The star denotes our best-fit value, Eq. (2). The triangle and the circle denote benchmark fits of Ref. [3] with q = 2.3 and 2.0, respectively
£ o
O
6 7
log(£/GeV)
only slightly different from the best fit of Ref. [3] (the difference may originate in the data included in the fit, E > 40 TeV vs. E > 60 TeV, and in some details of the fitting procedure which is not described in Ref. [3]).
Fig. 2. (Color online) The IceCube astrophysical neutrino spectrum [3] (data points) together with the range of 68 % C.L. allowed power-law fluxes determined in Fig. 1 (shadow)
The data are well described by a power law without any cutoff for a wide range of spectral indices.
Within our assumptions, it is easy to estimate the accompanying flux of photons from 7r° decays. A simple estimate of the gamma-ray flux injected by an optically thin source is given by [4, 13]
dF:'
«
dEv
= 2
dF-,
Ev=E1j2
dE
nucbMa B >K3T® TOM 100 Bun. 11-12 2014
where F^ and F7 are the fluxes of neutrino (per flavor, that is 1/3 of the total neutrino flux within our assumptions) and photons at energies Ev and E7, respectively.
On their way from the source to the observer, energetic gamma rays participate in electromagnetic cascades, driven by the electron-positron pair production on cosmic background radiations (the relevant contributions here come from the cosmic microwave, infrared and ultraviolet backgrounds, depending on the gamma-ray energies) and by the inverse Compton scattering of electrons and positrons, which produces secondary energetic photons. Therefore, besides gamma rays from 7r° decays, also electrons and positrons from 7r± decays contribute to the cascade and should be taken into account (their contribution is more important for lower-energy, <PeV, part of the observed spectrum and for hard spectral indices). The injected flux Fe of electrons and positrons at the energy Ee is equal to neutrino (per flavour) flux, in the same approximation:
dFe dEK
dF„
dE„
The photon attenuation length [14] is as short as ~ 10 kpc for PeV photons, increasing rapidly at both lower and higher energies. Therefore, mostly Galactic sources may contribute to the gamma-ray flux at PeV-EeV energies, and the observed spectrum of these gamma rays is very sensitive to the distribution of sources in the Galaxy and in its immediate neighbourhood. In what follows, we use publicly available transport equation based code written by one of us [15, 16] to simulate electron-photon cascade propagation. In the case of Galactic source distribution for simplicity we neglect interactions of photons with Galactic infrared and optical backgrounds. This may lead to at most 5 % error in the 7-ray flux predictions in the energy range 30 TeV < E1 < 300 TeV only. We also take into account synchrotron losses of electrons in the ~ 10~6 G galactic magnetic field. For the extragalactic infrared and optical background we use the estimate of Ref. [17].
We use several benchmark source distributions n(r) in the Universe. We measure r = (x,y,z) from the Galactic Center and account for a non-central position of the Sun in the Galaxy (assuming that the Sun is 8.5 kpc from the center).
1. Stellar distribution: a
Для дальнейшего прочтения статьи необходимо приобрести полный текст. Статьи высылаются в формате PDF на указанную при оплате почту. Время доставки составляет менее 10 минут. Стоимость одной статьи — 150 рублей.