Astrophysical counterparts of IceCube events

# Are both BL Lacs and pulsar wind nebulae the astrophysical counterparts of IceCube neutrino events?

European Southern Observatory, Karl-Schwarzschild-Str. 2, D-85748 Garching bei München, Germany
Technische Universität München, James-Frank-Str. 1, D-85748 Garching bei München, Germany
###### Abstract

IceCube has recently reported the discovery of high-energy neutrinos of astrophysical origin, opening up the PeV (10 eV) sky. Because of their large positional uncertainties, these events have not yet been associated to any astrophysical source. We have found plausible astronomical counterparts in the GeV – TeV bands by looking for sources in the available large area high-energy -ray catalogues within the error circles of the IceCube events. We then built the spectral energy distribution of these sources and compared it with the energy and flux of the corresponding neutrino. Likely counterparts include mostly BL Lacs and two Galactic pulsar wind nebulae. On the one hand many objects, including the starburst galaxy NGC 253 and Centaurus A, despite being spatially coincident with neutrino events, are too weak to be reconciled with the neutrino flux. On the other hand, various GeV powerful objects cannot be assessed as possible counterparts due to their lack of TeV data. The definitive association between high-energy astrophysical neutrinos and our candidates will be significantly helped by new TeV observations but will be confirmed or disproved only by further IceCube data. Either way, this will have momentous implications for blazar jets, high-energy astrophysics, and cosmic-ray and neutrino astronomy.

###### keywords:
BL Lacertae objects: general — gamma-rays: galaxies — neutrinos — pulsars: general — radiation mechanisms: non-thermal
pagerange: Are both BL Lacs and pulsar wind nebulae the astrophysical counterparts of IceCube neutrino events?Are both BL Lacs and pulsar wind nebulae the astrophysical counterparts of IceCube neutrino events?pubyear: 2014

## 1 Introduction

The IceCube South Pole Neutrino Observatory has reported the first evidence of high-energy astrophysical neutrinos222In this paper neutrino means both neutrino and antineutrino. (Aartsen et al., 2013; IceCube Collaboration, 2013), and more recently has confirmed and strengthened these observations by publishing a sample of 35 events with a deposited energy from 30 TeV to 2 PeV (IceCube Collaboration, 2014). With this enlarged sample the null hypothesis that all events are associated with the atmospheric background can be rejected at the level. If the observation of ultra-high energy cosmic rays revealed the existence of extreme cosmic accelerators, the IceCube neutrinos show that hadronic particle physics is in action in astrophysical sites at an energy scale somewhat higher than any man-made accelerator. IceCube is therefore opening a new window at the high-energy frontier of particle- and astro-physics. Motivated by this discovery we investigate here plausible -ray counterparts of the IceCube events and discuss possible new scenarios. The detection of high-energy neutrinos up to the PeV (10 eV) scale implies the existence of a class of astrophysical objects accelerating protons up to at least 10 eV, which then collide with other protons ( collisions) or photons ( collisions). High-energy -rays with energy and flux about a factor two higher than the neutrinos at the source, and therefore reaching the TeV range for the IceCube events, are also expected as secondary products in both cases (Kelner, Aharonian, & Bugayov, 2006; Kelner & Aharonian, 2008). In the following we refer to these -rays as neutrino twins. The study of these twin photons would provide the most direct way to shed light on the origin of the IceCube neutrinos. The twin photons, however, cannot be at the moment investigated due to the fact that present -ray telescopes reach only TeV. Moreover, depending on the sources and their distance, absorption of the twin photons might dilute the direct photon-neutrino connection.

The topology of the IceCube detections are broadly classified in two types: 1. cascade-like, characterised by a compact spherical energy deposition; 2. track-like, defined by a dominant linear topology from the induced muon. A large majority of the 35 IceCube events are characterised by a cascade-like topology, which, unlike the track-like topology, can only be reconstructed with a resolution in the tens of degrees. The association of the IceCube astrophysical neutrinos with astronomical sources is therefore not only limited by the missing twin photons but also by the relatively poor resolution of the events. Moreover, about half of the 35 events are expected to be background produced in the atmosphere (muons and neutrinos) and so mostly concentrated at lower energies.

IceCube has performed a series of tests on the incoming direction of the events showing no significant deviation from an assumed isotropic distribution nor prominent point sources. The all-sky integrated flux of neutrinos in the 60 TeV – 3 PeV range is at the level of GeV cm s sr per flavour for an spectrum. We note here that nearly half of the events reported by IceCube are on the Galactic plane (best fit position ) with a few relatively close to the Galactic centre. The remaining half are at higher Galactic latitude. Hence both Galactic (e.g., Fox, Kashiyama, & Mészarós, 2013; Taylor, Gabici, & Aharonian, 2014) as well as extra-galactic (e.g., Murase, Inoue, & Dermer, 2014; Fang et al., 2014) scenarios provide valid explanations for the IceCube detections.

In this work we study -ray sources which fall within the median angular error of the IceCube events333The IceCube Collaboration has reported up to now only the median angular error of their events.. We refrain from modelling candidate sources and specific scenarios, addressing instead the question of the counterparts of the IceCube events from a purely phenomenological point of view by using the highest-energy all-sky catalogues. In the absence of photon observations above 60 TeV, in fact, we investigate possible counterparts in the GeV TeV band. Each neutrino is studied independently as if it were a single detection in the sky. Both Galactic as well as extra-galactic sources are treated equally.

Section 2 defines the list of IceCube events adopted in this paper, while Section 3 describes the investigated -ray counterparts, their selection, and their -ray and neutrino hybrid spectral energy distributions (SEDs). Section 4 discusses our results, while in Section 5 we summarise our conclusions. We adopt here the definitions used in Aharonian (2004) for -ray astronomy: “high energy” (HE) or GeV astronomy spans the 30 MeV to 30 GeV energy range while “very high energy” (VHE) or TeV astronomy refers to the 30 GeV to 30 TeV range. For a review of VHE astronomy we refer to Holder (2012).

## 2 The IceCube Astrophysical Neutrinos

The first high-energy astrophysical neutrinos ever detected have been reported by IceCube through the selection of very high-energy events with interaction vertex inside the detector (IceCube Collaboration, 2013). This containment strategy of very high-energy events is extremely efficient in rejecting the atmospheric background, including atmospheric neutrinos (Schönert et al., 2009). In doing so, the analysis strategy favours cascade-like events, which deposit most of their energy inside the detector, over track-like ones, which travel various kilometres outside the detector. While, on the one hand, through the containment strategy astrophysical neutrinos are successfully singled out from the entire sky, on the other hand, the poor resolution of the cascade-like events makes the association with possible counterparts extremely challenging. Moreover, the absorption of very high-energy neutrinos crossing the Earth favours the southern hemisphere over the northern one, contrary to other IceCube analysis strategies optimized at lower energies, which are more sensitive to northern hemisphere events (Aartsen et al., 2013).

In order to reduce the residual atmospheric background contamination, which might still be produced by mouns and atmospheric neutrinos and concentrate in the low-energy part of spectrum (see Fig. 2 in IceCube Collaboration, 2014), we consider here only IceCube events with energy TeV. Moreover, to somewhat limit the number of possible counterparts, we consider only events with median angular error . These two cuts reduce the sample from 35 to 18 events. These are listed in Tab. 1, which gives the deposited energy of the neutrino, the flux at the deposited energy in units, the coordinates, the median angular error in degrees, the Galactic latitude, and the time of detection in Modified Julian Days. By interpreting each single event as coming from one astrophysical counterpart, in fact, we have derived the flux per neutrino event assuming that the observed flux is spread over 1 dex in energy and that the spectrum is (equivalent to ). We use here an energy bin, which is somewhat larger than the IceCube average energy resolution. This is done to take into account the larger uncertainty due to the different topologies of the event and possible stochastic variations in the deposited energy of the single event. The resulting uncertainty in the flux estimate is almost fully absorbed in the Poissonian error and does not affect significantly the comparisons described in the following. Effective areas from the IceCube northern or southern hemisphere (depending on declination) and a live time of detection of 988 days (IceCube Collaboration, 2014) were also used. In order to take into account the fact that the deposited energy is lower than the neutrino energy, we have averaged the effective areas in two neighbouring bins (at the neutrino deposited energy and at the next highest one). The derived fluxes are in the range erg cm s (i.e., GeV cm s) and errors are Poissonian for one event (Gehrels, 1986).

The 18 “golden” events are assumed here below to be all of astrophysical origin. They span a range of deposited energies from 60 TeV to 2 PeV, with a mean of TeV. In or collision scenarios (Kelner, Aharonian, & Bugayov, 2006; Kelner & Aharonian, 2008), the energy range of primary protons would be above a few tens of PeV. In primary cosmic rays, this is the region around and above the knee, which can be interpreted as the cross road of Galactic and extra-galactic cosmic rays (Giacinti, Kachelriess, & Semikoz, 2014). Primary cosmic rays up to extremely high energies are significantly deviated by magnetic fields, hence the sources of these primary cosmic rays have not been identified yet. If the primary cosmic rays encounter enough target material, they interact producing secondary neutrinos (from charged mesons) and photons (from neutral mesons). As photons and neutrinos are neutral they do not feel the effect of magnetic fields and can therefore be used in principle to probe astronomical PeV accelerators, the so-called PeVatrons.

## 3 The counterparts of the IceCube astrophysical neutrinos

Neutrinos are intimately connected to their -ray counterparts. As a matter of fact, the twin photons coming from the same interactions that produced the IceCube neutrinos would have energies TeV and up to PeV, and therefore well above the energy range of VHE astronomy. We will assume in the following a simple direct connection between VHE astronomical data and the TeV – PeV band sampled by the IceCube data. Of course, the larger the gap between the highest photons detected and the IceCube neutrinos the more approximate such extrapolation becomes.

### 3.1 γ-ray catalogues

In order to alleviate the problem of the missing spectral coverage of the -rays twins of the IceCube neutrinos, our approach is to look for counterparts of the IceCube neutrinos in currently available all-sky catalogues that cover the highest possible energies. Namely, in decreasing order of energy and priority:

1. TeVCat, an online catalogue for VHE astronomy, which includes at the time of writing 147 sources detected typically above GeV and reaching the TeV regime. Of these, 53 are blazars (50 BL Lacs and 3 flat-spectrum radio quasars [FSRQs]; see Sect. 3.2 for a description of the two blazar sub-classes) and 59 are Galactic sources, with the rest being mostly unclassified. TeVCat is a list of TeV sources, as there are no all-sky flux-limited TeV catalogues at the moment, given the very sparse sky coverage available at these energies. The High Energy Stereoscopic System (H.E.S.S.), however, has undertaken a TeV Galactic Plane survey covering a range of 250 to 65 in longitude at (Carrigan et al., 2013). There must be, therefore, many more TeV sources with fluxes comparable to the detected ones, which are still undetected, particularly at ;

2. the Wise High Synchrotron Peaked (WHSP) catalogue (Arsioli et al., 2014), which provides a large area () catalogue of blazars and blazar candidates selected to have the peak of the synchrotron emission at Hz and therefore expected to radiate strongly in the HE and VHE bands. For our purposes we selected the sub-sample of 76 sources with a “figure of merit” (FoM) on their potential detectability in the TeV band . The FoM is defined as the ratio between the synchrotron peak flux of a source and that of the faintest blazar in the WHSP sample already detected in the TeV band. The WHSP sources are all BL Lacs, with of them being known TeV sources and the remaining ones thought to be within reach of detection by current VHE instrumentation. Although technically not a -ray catalogue, this WHSP sub-sample represents at present the best way to compensate for the lack of full sky coverage in the TeV band for blazars. Moreover, of these sources have a Fermi  2FGL or 3FGL -ray counterpart (Arsioli et al., 2014);

3. the first Fermi-Large Area Telescope (LAT) catalogue of sources detected above 10 GeV (1FHL), which includes 514 sources, of which are blazars (277 BL Lacs and 53 FSRQs) or blazar candidates (58) (Ackermann et al., 2013). The remaining objects are unclassified () and Galactic ().

Our logic can be thus summarized: in the absence of VHE data reaching the TeV – PeV band, we use TeVCat as our starting point. To compensate for its incompleteness and limited sky coverage we then add the WHSP catalogue, which however covers the high Galactic latitude sky and includes only blazars. Furthermore, as WHSP provides for most of its sources only a FoM for the TeV detectability, we complement it by using the highest energy Fermi catalogue, that is 1FHL, which gives an all-sky view above 10 GeV for all astronomical sources. As it turns out, the gap between 10 GeV and the neutrino energies ( TeV) appears to be quite large for a sensible extrapolation. Nevertheless, we wanted to be as thourough as possible in our search without penalizing sources, which have not been observed yet in the TeV band but might still one day prove to be plausible IceCube counterparts.

### 3.2 Blazar counterparts

Blazars are those Active Galactic Nuclei (AGN) whose emission is dominated by a relativistic jet viewed at a relatively small angle with respect to the line of sight (Urry & Padovani, 1995). The two main blazar sub-classes, namely BL Lacertae objects (BL Lacs) and FSRQs, differ mainly in their optical spectra, with the former displaying strong, broad emission lines and the latter instead being characterized by optical spectra showing at most weak emission lines, sometimes exhibiting absorption features, and in many cases being completely featureless (see Giommi et al., 2012; Giommi, Padovani, & Polenta, 2013, for a recent re-evaluation of the two blazar classes). The strong non-thermal blazar radiation, which spans the entire electromagnetic spectrum, is composed of two broad humps, the low-energy one attributed to synchrotron radiation, and the high-energy one, usually thought to be due to inverse Compton radiation (see e.g. Abdo et al., 2010) or, alternatively, to hadronic processes (e.g. Böttcher et al., 2013, and reference therein). The peak of the synchrotron hump, , ranges from about  Hz to over  Hz reflecting the maximum energy at which particles can be accelerated (e.g. Giommi et al., 2012). Blazars with  Hz in their rest frame are called Low Synchrotron Peaked (LSP) sources, while those with  Hz  Hz, and  Hz are called Intermediate and High Synchrotron Peaked (ISP and HSP) sources respectively (Abdo et al., 2010). This definition extends the original division of BL Lacs into LBL and HBL sources first introduced by Padovani & Giommi (1995). Basically all HSP are BL Lacs (see Giommi et al., 2012; Padovani, Giommi, & Rau, 2012, for a possible explanation). Objects with large are obviously favoured to be VHE sources. Indeed, only one out of the fifty currently TeV-detected BL Lacs is an LSP.

Due to their very large luminosities and SEDs routinely reaching the HE and VHE bands, blazars are thought to be amongst the most powerful accelerators in the Universe and as a result have been considered prime candidate sources of ultra-high energy cosmic rays and neutrinos (Halzen & Vazquez, 1993; Protheroe, 1997; Mannheim, 1999; Dermer & Atoyan, 2001).

Tab. 2 shows the results of our search for blazars and gives the IceCube ID, the catalogues where the counterparts were found, the counterparts’ names and coordinates, ranked by energy (TeVCat first) and flux, the offset between the reconstructed position of the IceCube event and the blazar one, the redshift of the source (if available), the GeV flux (from the 1FHL catalogue), the (observed) flux above 200 GeV (from papers referenced in TeVCat) for the TeV-detected sources, and the blazar type, namely BL Lac, FSRQ, or unknown555This refers to the so-called “active galaxies of uncertain type” (AGU), most of which are expected to be blazars (e.g. Ackermann et al., 2013).. Note that given the very strong variability of blazars flux values should be taken only as approximate. Nevertheless, fluxes are important as, on average, a stronger neutrino source should also be a stronger -ray source, unless significant absorption is present (Sect. 1).

Blazar counterparts were found for 16/18 neutrino events, in one case (ID 13) with an offset slightly larger ( times) than the median angular error. Namely, six (with two sources correlated with the same event) were found in TeVCat (all of them in 1FHL), eleven are WHSP sources (nine of which are in 1FHL as well), while all others were found in 1FHL.

We stress that 8/9 of the IceCube events with have WHSP counterparts. Even more strikingly, and as is the case for TeVCat, in all these cases the WHSP source(s) is (are) always the strongest one(s). This vindicates the use of a selection on synchrotron peak and flux to identify present or potential TeV emitters and is particularly important for events without TeVCat counterparts, namely ID 11, 19, 20, and 26. We also note that the strongest 1FHL counterpart of ID 22, 1H 1914194, which has and therefore is not included by definition in WHSP, fullfills the other criteria ( and FoM) and therefore is a very promising TeV candidate. Same story for PMNJ08161311, the strongest 1FHL counterpart of ID 27666We thank Paolo Giommi for pointing this out to us.. We have therefore marked these two sources as WHSP in Tab. 2.

Tab. 2 includes some well-known, bright BL Lacs, i.e., MKN 421, PKS 2005489, PG 1553+113. Only three sources, namely B2 0912+29, RX J0908.9+2311, and 1RXS J091211.9+275955, are associated with more than one neutrino event (ID 9 and 26). Out of the 61 unique objects in Tab. 2 only six are FSRQs, with none of them being the strongest source within the error circle. Most () neutrino events have more than one blazar counterpart.

For the two neutrino events for which no counterpart was found in the three catalogues used, we checked for completeness the Fermi-2FGL catalogue (Nolan et al., 2012), which includes 1,873 sources detected above 100 MeV777The 2FGL includes blazars or blazar candidates (), Galactic () and unclassified () sources.. In both cases we found a counterpart: 2FGL J0825.93216 (PKS 0823321, an AGU), with an offset of 1.6 (slightly above the median angular error) for ID 3 and 2FGL J0726.00053 (PKS 0723008, an other AGU), with an offset of 1 for ID 5. For both objects the HE emission reaches GeV, i.e. not too far from the 1FHL cutoff.

#### 3.2.1 Hybrid SEDs

To see how the neutrino and photon energetics compare, we have put together the -ray SEDs of all sources using the SED builder of the ASI Science Data Centre (ASDC) adding, if needed, VHE data taken from the literature. We have also included the flux per neutrino event at the specific energy, thereby building a hybrid photon – neutrino SED. We then performed an “energetic” diagnostic by checking if a simple extrapolation succeeded in connecting the most energetic -rays to the IceCube neutrino in the hybrid SED, taking into account the rather large uncertainty in the flux of the latter. If this was the case we considered the source to be a probable counterpart. Otherwise, we discarded the object. Anything more sophisticated would require detailed modelling, which goes beyond the scope of this paper. We show in Figs. 14 and Fig. 8 the SEDs of the TeV-detected blazars, which we now turn to comment:

• MKN 421999Fang et al. (2014) have compared the TeV flux of MKN 421 and the neutrino fluxes related to ID 9. (Fig. 1); this is the strongest -ray source in our sample and a simple extrapolation of its VHE spectrum has no problems in explaining the energy and flux of the corresponding IceCube event (ID 9), even taking into account the factor increase expected for the energy and flux of the twin photons associated with the neutrinos. This is a clear case that passes the “energetic” diagnostic. Fig. 1 shows also the SED, corrected for absorption by the extragalactic background light (EBL), of 1ES 1011+496, the other TeVCat BL Lac in the same error circle, which appears also to be a plausible counterpart. Finally, the SED of S4 0917+44, which is the weakest source amongst the other nine blazars associated with ID 9, is also displayed. While the -ray spectra of MKN 421 and 1ES 1011+496 are raising with energy, that of S4 0917+44 is falling, which makes it a very unlikely counterpart for the IceCube event. This shows how important the “energetic” diagnostic is, especially because of the relatively loose spatial one, given the large error circles. We note that, for our purposes, the extrapolation to -ray energies larger than the observed ones is not influenced by the EBL as neutrinos reflect the photon densities at the source and not at the observer after the interaction with the EBL photons. However, the comparison between observed photon and neutrino fluxes is affected, as the former will be absorbed by the EBL, above GeV, while the latter will not;

• PG 1553+113 (Fig. 2); in this case the EBL has a very strong effect and, after correcting for its absorption (assuming : see discussion in Aleksić et al., 2012), its VHE spectrum appears also consistent with the IceCube event (ID 17);

• PKS 2005489 (Fig. 3); despite being relatively local, the EBL has an effect also on the VHE spectrum of this BL Lac, which reaches relatively large energies ( TeV). However, even taking this into account, its VHE spectrum appears to be inconsistent with a neutrino detection (ID 12);

• H 2356309 (Fig. 4); the EBL has quite a strong effect also on the VHE spectrum of this BL Lac and, after correcting for it, its VHE spectrum appears consistent with the IceCube event (ID 10).

We point out that also some of the 1FHL blazars in Tab. 2 display SEDs, which appear at a first glance not inconsistent with the corresponding neutrino event. However, the gap between the highest observable energy and that of the neutrino is much ( times) larger than that typical of TeV-detected sources, making any possible neutrino – photon association harder to pin down. We make an exception for WHSP sources, for which we have very strong hints of a possible TeV emission. As an example, we show in Fig. 5 the SED of SUMSS J014347584550, a WHSP source and the strongest counterpart of ID 20 (one of the three PeV events). A simple extrapolation of the HE spectrum appears not inconsistent with the energy and flux of the corresponding IceCube event, given its rising SED. The same argument applies to 1RXS J054357.3553206 (ID 19), 1H 1914194 (ID 22) and PMN J08161311 (ID 27).

We present our list of the most plausible counterparts in Sect. 4.

### 3.3 Non-blazar counterparts

Alternative scenarios for PeV neutrino sources beyond the blazar one include -ray bursts (GRB) (Asano & Mészáros, 2014, and references therein), supernovae remnants (Villante & Vissani, 2008, and references therein), pulsar wind nebulae (PWN; Bednarek, 2003), and star-forming galaxies (Tamborra, Ando, & Murase, 2014, and references therein). At present, the IceCube Collaboration reports no evidence of a GRB – neutrino connection (Abbasi et al., 2012), although this not based on the same events considered here.

The non-blazar counterparts present in the catalogues studied in this work are listed in Tab. 3, where we give also the GeV flux (from the 1FHL catalogue) and the HE flux in Crab units (as provided by TeVCat), if available. Sources were found for six neutrino events, four of them on the Galactic plane () and one of which actually consistent with the position of the Galactic centre. The error circles of the two events off the Galactic plane include one and two non-blazar counterparts respectively: NGC 253, a starburst galaxy (ID 10), and the Large Magellanic Cloud (LMC) plus an unidentified source (ID 19). While NGC 253 and the unidentified source have fluxes times smaller than those of the brightest blazars in the same error circle, this is not the case for the LMC, which is times brighter than 1RXS J054357.3553206, which however is a WHSP source. ID 27 has a single, weak non-blazar counterpart which, being an unidentified 1FHL source, might be a blazar as well. For the other three events we found a very large number () of counterparts, many with fluxes larger than those of the blazar counterparts. For example, Kookaburra J1420607, a PWN associated with ID 35 has a flux above 200 GeV times larger than that of 1ES 1312423, the BL Lac in the same error circle in TeVCat. And two supernova remnants associated with ID 14 and 33 have a GeV flux times larger than the respective brightest associated BL Lacs. We address in more detail the comparison between blazar and non-blazar counterparts in Sect. 3.3.1.

We note that of the thirty-seven classified sources in Tab. 3 all but three are Galactic, the exceptions being (highlighted in bold face in the table) the LMC, NGC 253, mentioned above, and Centaurus A, a radio galaxy associated with ID 35. This latter source is particularly interesting in this respect as the region around it is populated by a number of ultra high-energy cosmic ray events larger than the rest of the sky, although recent results show a chance probability for this to occur at a level of (Kampert et al., 2012). However, we point out that, amongst the many -ray sources in the error circle of ID 35, Centaurus A is one of the weakest.

#### 3.3.1 Hybrid SEDs

As done for blazars, we have put together the hybrid SEDs of all sources. It turned out that only two Galactic sources passed the “energetic” diagnostic being actually better at that than the blazars in the same error circle. We show here some of the most promising Galactic counterparts of the IceCube events.

Fig. 6 shows the -ray SEDs of three sources in the error circle of ID 14, namely VERA J18233454, the strongest blazar, HESS J1809193, the second strongest TeV source and a PWN, and the Galactic Centre. The latter two reach TeV but only HESS J1809193 seems to be a plausible astronomical counterpart, since the Galactic Centre appears to be too soft. The BL Lac has a strongly rising SED but reaches only GeV, which makes any extrapolation to the PeV range very uncertain.

Fig. 7 shows the -ray SEDs of three sources in the error circle of ID 33, namely 1RXS J193109.5+093714, the strongest BL Lac, MGRO J1908+06, the strongest TeV source and a PWN, and HESS J1857+026, an extended Galactic source. The latter two both reach TeV. However, only MGRO J1908+06 seems to be a plausible astronomical counterpart, since HESS J1857+026 appears to be too soft. The BL Lac reaches only GeV with an upper limit at TeV a factor of 3 below the fluxes of the two Galactic sources.

Fig. 8 shows the -ray SEDs of four sources in the error circle of ID 35, which is the IceCube event with the largest energy ( PeV). These include: 1ES 1312432, the only blazar in TeVCat, Centaurus A, and two PWNe, HESS J1356645 and Kookaburra J1418609. The PWNe are more than one order of magnitude brighter than the two extragalactic ones at TeV but still appear to fail the “energetic” test. We therefore cannot match the most energetic IceCube event with a plausible astronomical counterpart. However, as discussed in Sec. 3.2.1, we cannot exclude that one of the 1FHL sources in the same error box is a TeV emitter and responsible for the neutrino emission. Note that in the case of the radio galaxy and the BL Lac the EBL does not make much of a difference: Centaurus A is at while the correction for 1ES 1312432, which has a redshift in between those of PKS 2005489 and H 2356309 (), will therefore be between a factor and (see Figs. 3 and 4). Interestingly, the predicted neutrino flux from Centaurus A at the energy of ID 35 is times smaller than the flux connected with that event (Saba, Tjus, & Halzen, 2013).

In Fig. 4 we plot the -ray SED of NGC 253 as well, a starburst galaxy also in the error circle of ID 10. As could be anticipated by its relatively low VHE flux, the SED drops steeply and clearly fails the “energetic” diagnostic. Given its very low redshift () the effect of the EBL is negligible and therefore NGC 253 is an extremely unlikely counterpart of the IceCube event. As was the case for Centaurus A, a very recent paper by Yoast-Hull et al. (2014) predicts for NGC 253 a neutrino flux at the energy of ID 10 times smaller than the flux connected with that event. We present our list of most plausible counterparts in Sect. 4.

## 4 Discussion

Our goal was to investigate the origin of the first high-energy astrophysical neutrinos ever detected through a model independent approach. The most natural sources to begin with are -ray emitters. Starting from Tab. 2 and 3, we have studied the photon – neutrino hybrid SEDs of all possible counterparts. This turned out to be a very powerful discriminant for the likelihood of the association, especially given the relatively large error circles of the neutrino detections.

In the comparisons presented in Fig. 1 to Fig. 8, a gap is present typically from TeV to (neutrino) energies TeV (apart from Fig. 5 where the gap is much larger). Any possible extrapolation to close such a gap is highly non trivial. Moreover, strong variability and different emission peak energies displayed by blazars complicate even further the matter. We have nevertheless singled out -ray sources, which we believe are plausible counterparts.

Only 10% (6/47) of the TeVCat counterparts passed the “energetic” test, the others being too -ray faint. In order to make up for the rather scanty sky coverage of TeVCat, particularly in the high-latitude sky, we have also taken into consideration the WHSP blazars still not TeV-detected, applying the “energetic” test to them as well but considering in this case the strong likelihood that they might be TeV emitters. This way we singled out four WHSP-only objects and ranked them as probable counterparts. The 1FHL-only sources, although in some cases intriguing from the energetic side, turned out to have too large of a gap between their highest electromagnetic energy and that of the IceCube event to make a sensible link between photons and neutrinos. Moreover, without even a hint of a possible emission in the VHE band, we decided not to rank in this work the 1FHL-only sources as probable candidates.

Our findings are summarized in Tab. 4, which gives the most probable neutrino counterparts for 9/18 of the IceCube events. We stress that this does not mean that the remaining nine events have no astronomical counterparts but rather that there is less information available for the sources in their error circles for us to be able to judge on their association likelihood.

The counterparts turned out to be mostly HSP BL Lacs and two PWNe, pushing forward a mixed scenario of Galactic and extra-galactic neutrino sources, so far not yet considered in the literature. We note that one of the three PeV events (ID 14) appears to have only one plausible counterpart, a Galactic one. The second PeV IceCube event (ID 20) has one WHSP BL Lac counterpart but no Galactic ones, while the third and most energetic PeV event (ID 35) is at present without likely counterparts. NGC 253 (ID 10), which is the only starburst galaxy in our sample, is too weak when compared to other candidates. Same story for Centaurus A (ID 35), which is the only radio galaxy in our sample and an all-time favoured as a potential source by the cosmic ray community.

For completeness, we have also checked that our neutrino fluxes are consistent with the IceCube upper limits provided by Aartsen et al. (2013) (see their Tab. 2 and 3) assuming an () spectrum. This is especially important for the northern sources, where the neutrino non-detections are particularly stringent and at the level of the fluxes here derived (within the rather larger error bars). If the proposed counterparts in Tab. 4 are indeed neutrino sources, this suggests that a direct detection by IceCube is within reach.

To check the significance of our cross-correlations, we have re-done them by shifting the positions of the -ray catalogues by varying amounts always larger than the largest error radius (). Based on this, the chance of having seven or more neutrino events associated with TeVCat sources is , which reduces to for five neutrino events and blazar only counterparts. This simply reflects the incompleteness of TeVCat101010As shown by taking a random half of the WHSP sample and obtaining a value of for four neutrino events, which is much larger than the probability derived for the whole sample, as detailed here.. As regards WHSP, the chance of having eight neutrino events associated with that sample is 4, while that of having 16 neutrino events associated with 1FHL sources is , with the same probability for 1FHL blazars only becoming . Even if some of these values are intriguing, we are fully aware that, having used the median angular errors for the neutrino events as given by the IceCube team provides only lower limits to these numbers.

We stress that, while better statistics are undoubtedly needed on the neutrino side, we have also shown that more TeV observations reaching also higher energies than currently available are badly needed to bridge the gap between photon and neutrino energies. Moreover, based on the energies and fluxes listed in Tab. 1, the merging of “classical” and neutrino astronomy will require sensitivities erg cm s at TeV and erg cm s at PeV. We note that, while the former values are within reach of the Cherenkov Telescope Array (CTA) even for relatively short exposure times (see Fig. 7 of Barnacka et al., 2013), the latter are not, as the maximum energy that CTA is expected to reach is only TeV. Needless to say, all high FoM WHSP sources, but in particular those in Tab. 2, are obvious targets for current TeV facilities.

Further upcoming IceCube observations will be able to confirm or disprove the associations suggested here. We can think of two possible future scenarios:

• if our conclusions are proven to be wrong, this will mean the following: 1. the existence of a (new?) population of Fermi(1FHL)-undetected TeV sources either still to be detected, given the poor sky coverage of TeV telescopes, or totally absorbed by the extragalactic background light; in the latter case, the high-energy neutrino background will not be resolved into individual sources and will remain diffuse (in IceCube lingo); 2. a complete decoupling between -ray photons and neutrino production in the Universe; 3. a mixture of the two scenarios described above. In all these cases, the identification of neutrino sources will be extremely challenging;

• if our conclusions are at least partially confirmed, we will have shown for the first time the unambiguous presence of hadronic processes in blazars and pulsar wind nebulae, with very important consequences for our understanding of jet and high-energy astrophysics.

## 5 Conclusions

We have taken a very simple approach to tackle the issue of the astronomical counterparts of the IceCube neutrinos by looking for high-energy ( GeV) sources within the error circles of the events. We found counterparts for sixteen out of the eighteen events considered in this work. By studying their hybrid photon – neutrino SEDs we have narrowed down our search and come up with a list of most probable counterparts for nine IceCube neutrinos. Interestingly, there is no single class of sources, which we can connect the IceCube events with. Instead, the available data suggest a mixed scenario of Galactic and extra-galactic neutrino sources. These include BL Lacs of the HSP type (i.e. with peak of the synchrotron emission at Hz) and pulsar wind nebulae. The still TeV-undetected sources, which we have singled out as probable neutrino counterparts, are obvious candidates for detection by current TeV facilities. CTA should reach the sensitivities required for a merging of “classical” and neutrino astronomy at TeV. Further upcoming IceCube observations will be able to confirm or disprove the associations suggested here. Whatever the outcome of these tests, this will have crucial implications for blazar jets, high-energy astrophysics, and cosmic-ray and neutrino astronomy.

## Acknowledgments

We thank Paolo Giommi, Stefan Coenders, Luigi Costamante, Andreas Gross, and Sirin Odrowski for useful discussions and suggestions and the many teams, which have produced the data and catalogues used in this paper for making this work possible. E. R. is supported by a Heisenberg Professorship of the Deutsche Forschungsgemeinschaft (DFG RE 2262/4-1). We acknowledge the use of data and software facilities from the ASDC, managed by the Italian Space Agency (ASI). This research has made use of the VizieR catalogue access tool, CDS, Strasbourg, France and of the NASA/IPAC Extragalactic Database (NED), which is operated by the Jet Propulsion Laboratory, California Institute of Technology, under contract with the National Aeronautics and Space Administration.

## References

• Aartsen et al. (2013) Aartsen M. G., et al., 2013, PhRvL, 111, 021103
• Aartsen et al. (2013) Aartsen M. G., et al., 2013, ApJ, 779, 132
• Abbasi et al. (2012) Abbasi R., et al., 2012, Nature, 484, 351
• Abdo et al. (2010) Abdo A. A., et al., 2010, ApJ, 716, 30
• Abramowski et al. (2012) Abramowski A., et al., 2012, ApJ, 757, 158
• Acero et al. (2013) Acero F., et al., 2013, ApJ, 773, 77
• Ackermann et al. (2012) Ackermann M., et al., 2012, ApJ, 753, 83
• Ackermann et al. (2013) Ackermann M., et al., 2013, ApJS, 209, 34
• Aharonian et al. (2002) Aharonian F., et al., 2002, A&A, 393, 89
• Aharonian (2004) Aharonian F., 2004, Very high energy cosmic gamma radiation: a crucial window on the extreme Universe (River Edge, NJ: World Scientific Publishing)
• Aharonian et al. (2004) Aharonian F., et al., 2004, A&A, 425, L13
• Aharonian et al. (2006) Aharonian F., et al., 2006, A&A, 456, 245
• Aharonian et al. (2008) Aharonian F., et al., 2008, A&A, 477, 353
• Aleksić et al. (2012) Aleksić J., et al., 2012, ApJ, 748, 46
• Arsioli et al. (2014) Arsioli B., Fraga B., Giommi P., Padovani P., Marrese M., 2014, A&A, submitted
• Asano & Mészáros (2014) Asano K., Mészáros P., 2014, ApJ, 785, 54
• Barnacka et al. (2013) Barnacka A., Bogacz L., Grudzińska M., Frankowski A., Janiak M., Lubiński P., Moderski R., for the CTA Consortium, 2013, Proceedings ICRC2013, Rio de Janeiro (Brazil) (arXiv:1307.3409)
• Bednarek (2003) Bednarek W., 2003, A&A, 407, 1
• Böttcher et al. (2013) Böttcher M., Reimer A., Sweeney K., Prakash A., 2013, ApJ, 768, 54
• Carrigan et al. (2013) Carrigan S., Brun F., Chaves R. C. G., Deil C., Gast H., Marandon V., for the H. E. S. S. collaboration, 2013, Proceedings of the 48th Rencontres de Moriond (2013), La Thuile (Italy) (arXiv:1307.4868)
• Costamante (2012) Costamante L., 2012, 2011 Fermi Symposium proceedings, arXiv:1208.0808
• Dermer & Atoyan (2001) Dermer C. D., Atoyan A., 2001, Proc. 27th ICRC, Hamburg, Germany (arXiv:astro-ph/0107200)
• Fang et al. (2014) Fang K., Fujii T., Linden T., Olinto A. V., 2014, ApJ, submitted (arXiv:1404.6237)
• Fox, Kashiyama, & Mészarós (2013) Fox D. B., Kashiyama K., Mészarós P., 2013, ApJ, 774, 74
• Franceschini, Rodighiero, & Vaccari (2008) Franceschini A., Rodighiero G., Vaccari M., 2008, A&A, 487, 837
• Gehrels (1986) Gehrels N., 1986, ApJ, 303, 336
• Giacinti, Kachelriess, & Semikoz (2014) Giacinti G., Kachelriess M., Semikoz D. V., 2014, ApJ, submitted (arXiv:1403.3380)
• Giommi et al. (2012) Giommi P., et al., 2012, A&A, 541, A160
• Giommi et al. (2012) Giommi P., Padovani P., Polenta G., Turriziani S., D’Elia V., Piranomonte S., 2012, MNRAS, 420, 2899
• Giommi, Padovani, & Polenta (2013) Giommi P., Padovani P., Polenta G., 2013, MNRAS, 431, 1914
• Halzen & Vazquez (1993) Halzen F., Vazquez R. A., 1993, ICRC, 1, 447
• HESS Collaboration et al. (2013) HESS Collaboration, et al., 2013, MNRAS, 434, 1889
• Holder (2012) Holder J., 2012, APh, 39, 61
• IceCube Collaboration (2013) IceCube Collaboration, 2013, Sci, 342, 1242856
• IceCube Collaboration (2014) IceCube Collaboration, 2014, Physical Review Letters, submitted (arXiv:1405.5303)
• Kampert et al. (2012) Kampert K.-H., for the Pierre Auger Collaboration, 2012, Proceedings ICRC 2011, Beijing, China (arXiv:1207.4823)
• Kelner, Aharonian, & Bugayov (2006) Kelner S. R., Aharonian F. A., Bugayov V. V., 2006, PhRvD, 74, 034018
• Kelner & Aharonian (2008) Kelner S. R., Aharonian F. A., 2008, PhRvD, 78, 034013
• Mannheim (1999) Mannheim K., 1999, APh, 11, 49
• Massaro et al. (2013) Massaro F., D’Abrusco R., Paggi A., Masetti N., Giroletti M., Tosti G., Smith H. A., Funk S., 2013, ApJS, 209, 10
• Murase, Inoue, & Dermer (2014) Murase K., Inoue Y., Dermer C. D., 2014, arXiv, arXiv:1403.4089
• Nolan et al. (2012) Nolan P. L., et al., 2012, ApJS, 199, 31
• Padovani & Giommi (1995) Padovani P., Giommi P., 1995, ApJ, 444, 567
• Padovani, Giommi, & Rau (2012) Padovani P., Giommi P., Rau A., 2012, MNRAS, 422, L48
• Protheroe (1997) Protheroe R. J., 1997, ASPC, 121, 585
• Renaud et al. (2008) Renaud M., Hoppe S., Komin N., Moulin E., Marandon V., Clapson A.-C., 2008, AIPC, 1085, 285
• Saba, Tjus, & Halzen (2013) Saba I., Tjus J. B., Halzen F., 2013, APh, 48, 30
• Schönert et al. (2009) Schönert S., Gaisser T. K., Resconi E., Schulz O., 2009, PhRvD, 79, 043009
• Şentürk et al. (2013) Şentürk G. D., Errando M., Böttcher M., Mukherjee R., 2013, ApJ, 764, 119
• Tamborra, Ando, & Murase (2014) Tamborra I., Ando S., Murase K., 2014, JCAP, submitted (arXiv:1404.1189)
• Taylor, Gabici, & Aharonian (2014) Taylor A. M., Gabici S., Aharonian F., 2014, PhRvD, 89, 103003
• Urry & Padovani (1995) Urry C. M., Padovani P., 1995, PASP, 107, 803
• Villante & Vissani (2008) Villante F. L., Vissani F., 2008, PhRvD, 78, 103007
• Yoast-Hull et al. (2014) Yoast-Hull T. M., Gallagher J. S., III, Zweibel E. G., Everett J. E., 2014, ApJ, 780, 137
You are adding the first comment!
How to quickly get a good reply:
• Give credit where it’s due by listing out the positive aspects of a paper before getting into which changes should be made.
• Be specific in your critique, and provide supporting evidence with appropriate references to substantiate general statements.
• Your comment should inspire ideas to flow and help the author improves the paper.

The better we are at sharing our knowledge with each other, the faster we move forward.
The feedback must be of minimum 40 characters and the title a minimum of 5 characters