THE FIRST COMBINED SEARCH FOR NEUTRINO POINT-SOURCES IN THE SOUTHERN HEMISPHERE WITH THE ANTARES AND ICECUBE NEUTRINO TELESCOPES

Neutrinos offer a unique insight into the universe because they interact only weakly and through gravity. Unlike charged particles, they can travel straight from the source to the Earth without being deflected by magnetic fields or being absorbed. Neutrinos are expected to originate from the same locations where the acceleration of cosmic rays take place (Gaisser et al. 1995; Learned & Mannheim 2000; Halzen & Hooper 2002; Becker 2008; Kelner & Aharonian 2008; Murase 2015). A large variety of classes of astrophysical objects are predicted to be sources of high-energy neutrinos, where galactic candidates include microquasars (Levinson & Waxman 2001; Romero et al. 2003; Bednarek 2005; Torres et al. 2005), supernova remnants (Alvarez-Muñiz & Halzen 2002; Guetta & Amato 2003; Cavasinni et al. 2006; Halzen et al. 2008; Vissani et al. 2011; González-García et al. 2014; Mandelartz & Becker Tjus 2015), or various objects close to the Galactic Center (Kistler & Beacom 2006; Fujita et al. 2015). Extragalactic sources comprise active galactic nuclei (Stecker et al. 1991; Nellen et al. 1993; Mannheim 1995; Rachen & Mészáros 1998; Atoyan & Dermer 2001; Mücke et al. 2003; Becker et al. 2005; Stecker 2005; Eichmann et al. 2012) and gamma-ray bursts (Waxman & Bahcall 1997, 2000; Mészáros & Waxman 2001; Razzaque et al. 2003; Becker et al. 2006; Murase & Nagataki 2006; Hümmer et al. 2012).

The low neutrino cross section also implies that their detection is challenging. The first attempts to make a skymap of neutrinos were performed by the KGF experiment (Krishnaswam et al. 1971). Until the development of neutrino telescopes, the MACRO experiment had the most stringent limits for astrophysical neutrino sources (Ambrosio et al. 2001). After the pioneering efforts to build large volume neutrino telescopes by the DUMAND (Roberts & Wilkins 1978; Bosetti et al. 1988), Baikal (Aynutdinov et al. 2008), and AMANDA (Ahrens et al. 2002) collaborations, the field is presently led by the IceCube (Achterberg et al. 2006) and ANTARES (Ageron et al. 2011) experiments. For a historical review of the development of neutrino telescopes, refer to Spiering (2012). IceCube, which is placed in the deep Antarctic ice, is the first detector to reach the cubic-kilometer size predicted to be necessary to detect cosmic neutrino fluxes (Berezinsky & Zatsepin 1977; Gaisser et al. 1995). Recently, IceCube reported the crucial discovery of a flux of neutrinos up to ∼ PeV energies, which cannot be explained by the background of atmospheric muons and neutrinos only (Aartsen et al. 2013c, 2013d). The specific origin of these events is currently unknown. Some authors propose that at least part of the flux may have a galactic origin (Fox et al. 2013; Razzaque 2013; Anchordoqui et al. 2014a, 2014b; Padovani & Resconi 2014; Ahlers et al. 2015; Bai et al. 2015), whereas others have focused on the extragalactic component (Cholis & Hooper 2013; Kalashev et al. 2013; Roulet et al. 2013; Stecker 2013). Meanwhile the ANTARES experiment has proven the feasibility of the Cherenkov telescope technique in sea water (Adrian-Martinez et al. 2012a, 2013b). While its instrument volume is significantly smaller than that of IceCube, its geographical location provides a view of the southern sky with significantly reduced background for neutrino energies below 100 TeV, and hence better sensitivity to many predicted Galactic sources of neutrinos in this part of the sky. The complementarity of the detectors with respect to southern sky sources, due to their different geographical location, size, and atmospheric muon background, allows for a gain in sensitivity by combining the analyses of data from both experiments in a joint search for point-like sources. The level of improvement depends on the details of the assumed astrophysical flux, in particular on its energy spectrum and the existence of a possible high-energy cut-off. The energy spectra are not yet known and predictions vary widely depending on the source model.

In this paper, a combined analysis using the point-source data samples of IceCube from 2008 to 2011 and of ANTARES from 2007 to 2012 is presented. This paper is structured as follows: in Section 2, the IceCube and ANTARES detectors are introduced. In Section 3 the samples from each experiment are described, while in Section 4 the search method is explained. Finally, the results are presented in Section 5 and the conclusions are discussed in Section 6.

IceCube is a cubic-kilometer neutrino telescope located at the geographic south Pole. It consists of a total of 5160 digital optical modules (DOMs) deployed in the Antarctic ice at depths from 1450 to 2450 m below the surface (Abbasi et al. 2010). Each DOM consists of a pressure-resistant sphere that houses electronics, calibration LEDs, and a 10'' PMT facing downward. The DOMs are configured in a hexagonal array of 86 vertical cables descending from the surface, called "strings," with 60 DOMs per string. The average horizontal distance is 125 m between strings, and the average vertical spacing is 17 m between DOMs on a string. A sub-array of eight strings (Deep Core) is also present in the core of the detector (Abbasi et al. 2012). These strings have a smaller separation to improve the sensitivity for lower energies. Construction of the detector began in 2005 and was completed six years later. The analysis presented in this paper is based on data from three years of the partially completed detector, when 40, 59, and 79 strings were deployed. Future joint analyses are envisioned that will be based on data from the full 86-string detector, including recent data samples that use outer-detector modules as vetoes to achieve sensitivity to lower-energy neutrinos.

ANTARES is the first neutrino telescope that operates in the sea (Ageron et al. 2011). It was completed in 2008, with the first lines operating from 2006. It is located in the Mediterranean Sea at a depth of 2475m at coordinates (42° 48' N, 6° 10' E), which is 40 km south of Toulon (France). It consists of an array of 885 Optical Modules (OMs) distributed along 12 lines of 350 m height and an inter-line separation of 60–75 m. An OM consists of a 10'' photomultiplier tube (PMT) contained inside a 17'' glass sphere. The OMs are grouped into triplets and face downward at an angle of 45° to optimize the detection of up-going muon neutrinos. There are 25 triplets (storeys) on each line, with a distance of about 15 m between each. Lines are kept vertical with a buoy at their top.

One of the main focuses of the ANTARES and IceCube neutrino telescopes is the observation of cosmic point-like sources of neutrino emission. At present, corresponding searches are mainly focused on the detection of muon neutrinos, which can be reconstructed with sub-degree angular resolution. Muon neutrinos are indirectly detected through the muon produced in their charged current interaction (CC) with a nucleus (N) inside or near the detector volume:

$$\begin{eqnarray}&&{\nu }_{\mu }+N\to {\mu }^{-}+X.\end{eqnarray} \tag{ 1 }$$

In this reaction, a muon and a hadronic shower, X, are produced.⁹¹ The ultra-relativistic muon can travel long distances (up to several kilometers) and, when crossing a suitable medium such as ice or water, induce Cherenkov radiation that can be detected by the photomultipliers (PMTs) of neutrino telescopes. The corresponding charge and time information of the detected photons is used to reconstruct the direction of the muon, which is almost collinear with the original neutrino for energies above the TeV range. The main backgrounds for cosmic neutrino searches are atmospheric muons and neutrinos produced in the decay of the secondary particles created in the interactions of cosmic rays with the nuclei of the atmosphere.

The data sample employed for this analysis corresponds to all events from the southern sky that were included in the three-year IceCube point-source analysis (Aartsen et al. 2013b), combined with the events in the latest ANTARES point-source analysis (Adrián-Martínez et al. 2014). The ANTARES sample contains data recorded from 2007 January 29 to 2012 December 31; for IceCube, the data was recorded from 2008 April 5 to 2011 May 13 with the partially completed detector, and without the use of the Deep Core strings.

Detector performance differs not only between ANTARES and IceCube, but also between the three IceCube configurations as the detector grew—from 40 to 59, and then to 79 strings. The effective area is defined as the equivalent surface with a perfect efficiency that detects the same number of events as the detector for the final event selections used. For a source position of δ = −30°, the effective area for each IceCube configuration and for ANTARES is shown in Figure 1 (left). Due to its larger size, the effective area for the IceCube samples is larger for neutrino energies above ∼100 TeV. However, to view sources in the southern Sky, IceCube must contend with the down-going background of atmospheric muons, which becomes overwhelming at lower energies. To minimize these, the IceCube point-source analysis introduced a declination-dependent energy cut that strongly suppresses low energy events in the final data sample. ANTARES, which can use the Earth as a filter against atmospheric muons in the southern sky, thus maintains a larger effective area in this energy and declination range.

Zoom In Zoom Out Reset image size

A comparison of the median angular resolution of each sample can be seen in Figure 1 (right). The better resolution of the ANTARES sample is due to the longer photon scattering length in water compared to ice. The sensitivities reported by both experiments for the whole sky using the classical frequentist method (Neyman 1937) are shown in Figure 2. The sensitivities are calculated as the median 90% CL upper limits that would be obtained for background only.

Zoom In Zoom Out Reset image size

Different selection criteria are applied to each sample. A summary of these selections, which in all cases were developed with a data-blinding policy and were optimized to minimize the neutrino flux needed for a 5σ discovery in 50% of the experiments, is given below.

3.1. ANTARES

3.2. IceCube

3.3. Relative Fraction of Source Events for Different Source Assumptions

The relative fraction of expected source events from each sample needs to be calculated to estimate its respective weight in the likelihood that will be used to search for an excess of events from a particular direction (see Section 4). This fraction is defined as the ratio of the expected number of signal events for the given sample to that for all samples,

$$\begin{eqnarray}&&{C}^{j}(\delta ,d{\rm{\Phi }}/{{dE}}_{\nu })=\displaystyle \frac{{N}^{j}(\delta ,d{\rm{\Phi }}/{{dE}}_{\nu })}{{\displaystyle \sum }_{i}\;{N}^{i}(\delta ,d{\rm{\Phi }}/{{dE}}_{\nu })},\end{eqnarray} \tag{ 2 }$$

where the total number of expected events for the jth sample, N^j, with a given source declination, δ, and a given source spectrum, $\tfrac{d{\rm{\Phi }}}{{{dE}}_{\nu }}$, can be calculated as

$$\begin{eqnarray}&&{N}^{j}\left(\delta ,\displaystyle \frac{d{\rm{\Phi }}}{{{dE}}_{\nu }}\right)=\int {dt}\int {{dE}}_{\nu }{A}_{{\rm{eff}}}^{j}({E}_{\nu },\delta )\displaystyle \frac{d{\rm{\Phi }}}{{{dE}}_{\nu }}.\end{eqnarray} \tag{ 3 }$$

The time integration extends over the live time of each sample and ${A}_{\mathrm{eff}}^{j}({E}_{\nu },\delta )$ indicates the effective area of the corresponding detector layout j as a function of the neutrino energy, E_ν, and the declination of the source, δ.

Because each detector layout has a different response depending on the neutrino energy and declination, this relative fraction of source events needs to be calculated for different source spectra and source declinations. Figure 3 shows the relative fraction of signal events for an unbroken E⁻² spectrum, which corresponds to vanilla first order Fermi acceleration (Krymskii 1977; Blandford & Ostriker 1978). In this case, there is a significant contribution from all samples over most of the southern Sky, with the ANTARES contribution being more significant for declinations closer to δ = −90° and IceCube for declinations closer to 0°. The reason for this variability is mostly due to the declination-dependent energy cut applied in the IceCube samples to reduce the background of atmospheric muons.

Zoom In Zoom Out Reset image size

Other source assumptions are also considered in this analysis. The relative fraction of source events is calculated for an unbroken E^−2.5 power-law spectrum, as suggested in recent IceCube diffuse-flux searches (Aartsen et al. 2015), and for an E⁻² spectrum with exponential square-root cut-offs $\left(\tfrac{d{\rm{\Phi }}}{{dE}}\propto {E}^{-2}\mathrm{exp}\left[-\sqrt{\tfrac{E}{{E}_{\mathrm{cut} \mbox{-} \mathrm{off}}}}\right]\right)$ of 100 TeV, 300 TeV, and 1 PeV, because a square-root dependence may be expected from Galactic sources (Kappes et al. 2007). Figure 4 shows the relative fraction of source events for these cases. Compared with an unbroken E⁻² spectrum, the contribution of high-energy neutrinos in all of these cases is lower, and therefore the relative contribution of the ANTARES sample increases.

Zoom In Zoom Out Reset image size

An unbinned ML ratio estimation has been performed to search for excesses of events that would indicate cosmic neutrinos coming from a common source. In order to estimate the significance of a cluster of events, this likelihood takes into account the energy and directional information of each event. Due to the different detector response, the data sample that an event belongs to is also taken into account. The likelihood, as a function of the total number of fitted signal events, n_s, can be expressed as:

$$\begin{eqnarray}&&L({n}_{s})=\displaystyle \prod _{j=1}^{4}\displaystyle \prod _{i=1}^{{N}^{j}}\left[\displaystyle \frac{{n}_{s}^{j}}{{N}^{j}}{S}_{i}^{j}+\left(1-\displaystyle \frac{{n}_{s}^{j}}{{N}^{j}}\right){B}_{i}^{j}\right]\end{eqnarray} \tag{ 4 }$$

where j marks one of the four data samples, i indicates an event belonging to the jth sample, ${S}_{i}^{j}$ is the value of the signal probability distribution function (PDF) for the ith event, ${B}_{i}^{j}$ indicates the value of the background PDF, N^j is the total number of events in the jth sample, and n^j_s is the number of signal events fitted for in the jth sample. Because a given evaluation of the likelihood refers to a single source hypothesis at a fixed sky location, the number of signal events n^j_s that are fitted for in each sample is related to the total number of signal events n_s by the relative contribution of each sample, ${n}_{s}^{j}\;=\;{n}_{s}\cdot {C}^{j}\left(\delta ,\tfrac{d{\rm{\Phi }}}{{dE}}\right)$.

The signal and background PDFs for the IceCube and ANTARES samples have slightly different definitions. The signal PDF for ANTARES is defined as

$$\begin{eqnarray}&&{S}^{\mathrm{ANT}}=\displaystyle \frac{1}{2\pi {\sigma }^{2}}\mathrm{exp}\left(-\displaystyle \frac{{\rm{\Delta }}{\rm{\Psi }}{({{\boldsymbol{x}}}_{s})}^{2}}{2{\sigma }^{2}}\right){P}_{s}^{\mathrm{ANT}}({{ \mathcal N }}^{\mathrm{hits}},\sigma ),\end{eqnarray} \tag{ 5 }$$

where ${{\boldsymbol{x}}}_{s}$ = (${\alpha }_{s}$, δ_s) indicates the source direction in equatorial coordinates, ${\rm{\Delta }}{\rm{\Psi }}({{\boldsymbol{x}}}_{s})$ is the angular distance of a given event to the source, and ${P}_{s}^{\mathrm{ANT}}({{ \mathcal N }}^{\mathrm{hits}},\sigma )$ is the probability of a signal event to be reconstructed with an angular error estimate of σ and a number of hits ${{ \mathcal N }}^{\mathrm{hits}}$. The number of hits is a proxy for the energy of the event. In this sense, an event with a higher number of hits (higher deposited energy) would be less likely to be produced by the expected background.

The definition of the signal PDFs for the IceCube samples is similar,

$$\begin{eqnarray}&&{S}^{\mathrm{IC}}=\displaystyle \frac{1}{2\pi {\sigma }^{2}}\mathrm{exp}\left(-\displaystyle \frac{{\rm{\Delta }}{\rm{\Psi }}{({{\boldsymbol{x}}}_{s})}^{2}}{2{\sigma }^{2}}\right){P}_{s}^{\mathrm{IC}}({ \mathcal E },\sigma | \delta )\end{eqnarray} \tag{ 6 }$$

where the main difference lies in the use of the reconstructed energy, ${ \mathcal E }$, and the declination dependence of the probability for a signal event to be reconstructed with a given σ and ${ \mathcal E }$. The declination dependence is needed mainly because of the event selection cut on reconstructed energy, which is designed to reduce the atmospheric muon background.

Background events are expected to be distributed uniformly in right ascension. The background PDFs are obtained from the experimental data itself. The definition of the PDFs are:

$$\begin{eqnarray}&&{B}^{\mathrm{ANT}}=\displaystyle \frac{{B}^{\mathrm{ANT}}(\delta )}{2\pi }{P}_{b}^{\mathrm{ANT}}({{ \mathcal N }}^{\mathrm{hits}},\sigma ),\end{eqnarray} \tag{ 7 }$$

$$\begin{eqnarray}&&{B}^{\mathrm{IC}}=\displaystyle \frac{{B}^{\mathrm{IC}}(\delta )}{2\pi }{P}_{b}^{\mathrm{IC}}({ \mathcal E },\sigma | \delta ),\end{eqnarray} \tag{ 8 }$$

where B^j(δ) is the per-solid-angle rate of observed events as a function of the declination in the corresponding sample. ${P}_{b}^{\mathrm{ANT}}({{ \mathcal N }}^{\mathrm{hits}},\sigma )$ and ${P}_{b}^{\mathrm{IC}}({ \mathcal E },\sigma | \delta )$ characterize the distributions for background event properties, in analogy with the definitions of P^ANT_s and P^IC_s for the signal events given above.

The test statistic, TS, is determined from the likelihood (Equation (4)) as TS = $\mathrm{log}L({\hat{n}}_{s})-\mathrm{log}L({n}_{s}\;=\;0)$, where ${\hat{n}}_{s}$ is the value that maximizes the likelihood. The larger the TS, the lower the probability (p-value) of the observation to be produced by the expected background. Simulations are performed to obtain the distributions of the TS. The significance (specifically, the p-value) of an observation is determined by the fraction of TS values that are larger than the observed TS.

The TS is calculated as a preliminary step to obtain the post-trial p-values of a search. TS distributions for the fixed-source, background-only hypothesis have been calculated in steps of 1° in declination from pseudo–data sets of randomized data. Because these distributions vary with declination, the preliminary TS is turned into a "pre-trial p-value" by comparing the TS obtained at the source location from the data to the background TS distribution for the corresponding declination. The post-trial significance is then estimated with pseudo–data sets and according to the type of search, as explained with the results in Section 5.

Two different searches for point-like neutrino sources are performed. In the candidate list search, a possible excess of neutrino events is looked for at the location of 40 preselected neutrino source candidates. Because the location of these sources is fixed (at known locations with an uncertainty below the angular resolution of all samples), only the number of signal events n_s is a free parameter in the likelihood maximization. These candidates correspond to all sources in the southern sky considered in the previous candidate source list searches performed in the ANTARES and IceCube point-source analyses (Aartsen et al. 2013b; Adrián-Martínez et al. 2014).

The second search is a "full sky" search, looking for a significant point-like excess anywhere in the southern sky. For this purpose, the likelihood is evaluated in steps of 1° × 1° over the whole scanned region. In this case, the source position is an additional free parameter of the likelihood to fit the best position within the 1° × 1° boundaries.

Both the full southern sky and candidate list searches are performed using an E⁻² source spectrum in the signal PDFs. The main virtue of the energy term in the PDFs is to add power to distinguish signal neutrinos from the softer spectra of atmospheric neutrinos and atmospheric muons. Limits for the sources in the candidate list have also been calculated for the source spectra mentioned in Section 3.3.

Results from the full southern sky and candidate list searches are detailed below.

5.1. Full Southern-sky Search

No significant event clusters are found over the expected background. The most significant cluster is located at equatorial coordinates α = 3328, δ = −461, with best-fit n_s = 7.9 and a pre-trial p-value of 6.0 × 10⁻⁷. Figure 5 shows this pre-trial p-value compared with the distribution of the smallest p-values found when performing the same analysis on many pseudo–data sets (constructed by randomizing the right ascension coordinates of the real data). It is found that 24% of pseudo–data sets have a smaller p-value somewhere in the sky than is found in the real data; the post-trial significance is thus 24% (0.7σ in the one-sided sigma convention). The direction of this cluster is consistent with, but also less significant than the second most significant cluster in the previous ANTARES point-source analysis. Figure 6 shows the position of this cluster and the pre-trial p-values for all directions in the southern sky (a smaller step of 02 × 02 is used to plot this map).

Zoom In Zoom Out Reset image size

Zoom In Zoom Out Reset image size

5.2. Candidate List Search

The results of the candidate source list search are presented in Table 2. No statistically significant excess is found. The most significant excess for any object on the list corresponds to HESS J1741-302, with a pre-trial p-value of 0.003. To account for trial factors, the search is performed on the same list of sources using pseudo–data sets; the distribution of smallest p-values for these searches is shown in Figure 7. We find that 11% of randomized data sets have a smaller p-value for any source than that found for the real data; the post-trial significance of the source list search is thus 11% (1.2σ in the one-sided sigma convention).

Zoom In Zoom Out Reset image size

Table 2.  Fitted Number of Source Events, n_s, Pre-trial p-values, p, and 90% C.L. Flux Limits, ${{\rm{\Phi }}}_{\nu }^{90\mathrm{CL}}$ for the Different Source Spectra for the 40 Candidate Sources

Name	δ (°)	α (°)	ns	p	${\phi }_{{E}^{-2}}^{90\mathrm{CL}}$	${\phi }_{{E}^{-2.5}}^{90\mathrm{CL}}$	${\phi }_{{E}_{c}\;=\;1\;\mathrm{PeV}}^{90\%\mathrm{CL}}$	${\phi }_{{E}_{c}\;=\;300\;\mathrm{TeV}}^{90\mathrm{CL}}$	${\phi }_{{E}_{c}\;=\;100\;\mathrm{TeV}}^{90\mathrm{CL}}$
3C279	−5.8	−166.0	1.1	0.05	3.1E-09	1.0E-06	6.5E-09	9.2E-09	1.4E-08
HESSJ1837-069	−7.0	−80.6	-	-	1.6E-09	9.3E-07	3.6E-09	5.1E-09	8.9E-09
QSO2022-077	−7.6	−53.6	-	-	1.9E-09	9.9E-07	4.0E-09	6.1E-09	1.2E-08
PKS1406-076	−7.9	−147.8	-	-	2.2E-09	7.7E-07	4.3E-09	6.7E-09	1.0E-08
HESSJ1834-087	−8.8	−81.3	-	-	2.2E-09	1.1E-06	5.0E-09	8.0E-09	1.4E-08
PKS0727-11	−11.7	112.6	-	-	3.0E-09	1.6E-06	7.2E-09	1.2E-08	2.1E-08
1ES0347-121	−12.0	57.4	-	-	3.7E-09	1.7E-06	8.6E-09	1.3E-08	2.2E-08
QSO1730-130	−13.1	−96.7	-	-	3.3E-09	1.5E-06	7.5E-09	1.2E-08	2.0E-08
LS5039	−14.8	−83.4	-	-	4.2E-09	2.1E-06	1.1E-08	1.7E-08	2.9E-08
W28	−23.3	−89.6	-	-	6.3E-09	2.8E-06	1.7E-08	2.5E-08	4.0E-08
PKS0454-234	−23.4	74.3	-	-	7.4E-09	4.4E-06	2.6E-08	4.0E-08	5.7E-08
1ES1101-232	−23.5	165.9	-	-	6.4E-09	2.8E-06	1.7E-08	2.6E-08	4.1E-08
Galactic Center	−29.0	−93.6	-	-	7.6E-09	2.6E-06	1.8E-08	2.6E-08	3.8E-08
PKS1622-297	−29.9	−113.5	-	-	8.9E-09	2.6E-06	1.8E-08	2.6E-08	3.8E-08
HESSJ1741-302	−30.2	−94.8	1.6	0.003	2.5E-08	7.5E-06	5.5E-08	7.2E-08	1.0E-07
PKS2155.304	−30.2	−30.3	-	-	7.8E-09	2.6E-06	1.8E-08	2.6E-08	3.8E-08
H2356-309	−30.6	−0.2	-	-	7.9E-09	2.6E-06	1.8E-08	2.7E-08	3.8E-08
PKS0548-322	−32.3	87.7	0.9	0.07	1.6E-08	5.0E-06	3.8E-08	4.9E-08	6.7E-08
PKS1454-354	−35.7	−135.6	-	-	8.9E-09	3.5E-06	2.0E-08	2.8E-08	4.5E-08
PKS0426-380	−37.9	67.2	-	-	8.6E-09	2.5E-06	1.7E-08	2.5E-08	3.5E-08
RXJ1713.7-3946	−39.8	−101.8	-	-	8.7E-09	2.4E-06	1.7E-08	2.4E-08	3.3E-08
CenA	−43.0	−158.6	-	-	8.1E-09	2.2E-06	1.5E-08	2.1E-08	3.0E-08
PKS0537-441	−44.1	84.7	-	-	8.2E-09	2.2E-06	1.5E-08	2.1E-08	3.0E-08
VelaX	−45.6	128.8	-	-	8.3E-09	2.2E-06	1.5E-08	2.1E-08	2.9E-08
RXJ0852.0-4622	−46.4	133.0	-	-	9.5E-09	2.1E-06	1.5E-08	2.1E-08	2.9E-08
HESSJ1632-478	−47.8	−112.0	-	-	8.6E-09	2.1E-06	1.6E-08	2.1E-08	2.9E-08
PKS2005-489	−48.8	−57.6	-	-	1.0E-08	2.3E-06	1.7E-08	2.2E-08	2.9E-08
GX339-4	−48.8	−104.3	-	-	8.7E-09	2.2E-06	1.6E-08	2.1E-08	2.9E-08
HESSJ1616-508	−51.0	−116.0	-	-	1.1E-08	2.3E-06	1.8E-08	2.2E-08	3.0E-08
HESSJ1614-518	−51.8	−116.4	-	-	9.3E-09	2.9E-06	1.8E-08	2.7E-08	3.8E-08
CirX-1	−57.2	−129.8	-	-	9.1E-09	2.1E-06	1.6E-08	2.1E-08	2.8E-08
HESSJ1023-575	−57.8	155.8	0.8	0.08	1.7E-08	3.5E-06	2.8E-08	3.5E-08	4.7E-08
HESSJ1503-582	−58.7	−133.6	-	-	9.1E-09	2.1E-06	1.6E-08	2.0E-08	2.7E-08
MSH15-52	−59.2	−131.5	-	-	9.1E-09	2.1E-06	1.5E-08	2.0E-08	2.7E-08
ESO139-G12	−59.9	−95.6	0.8	0.07	1.8E-08	3.9E-06	2.9E-08	3.7E-08	5.1E-08
HESSJ1507-622	−62.3	−133.3	-	-	9.1E-09	2.0E-06	1.5E-08	2.0E-08	2.6E-08
RCW86	−62.5	−139.3	0.2	0.11	1.4E-08	2.0E-06	1.5E-08	2.0E-08	2.6E-08
HESSJ1303-631	−63.2	−164.2	-	-	9.1E-09	2.0E-06	1.5E-08	1.9E-08	2.6E-08
PSRB1259-63	−63.5	−164.3	-	-	9.1E-09	2.0E-06	1.5E-08	1.9E-08	2.6E-08
HESSJ1356-645	−64.5	−151.0	-	-	9.1E-09	2.0E-06	1.5E-08	1.9E-08	2.6E-08

Note. Units for the flux limits for the E^−2.5 spectra, ${\phi }_{{E}^{-2.5}}^{90\mathrm{CL}}$, are given in GeV^1.5 cm⁻² s⁻¹, whereas the rest are in GeV cm⁻² s⁻¹. The sources are sorted by their declination. Dashes (-) in the fitted number of source events and pre-trial p-values indicate sources with n_s ≤ 0.001.

Download table as:  ASCIITypeset image

Table 2 provides the pre-trial p-values, best-fit signal events n_s, and flux upper limits (under different assumptions of the energy spectrum) for all the candidate source objects. Figure 8 shows the Neyman sensitivities and limits for this search (assuming an E⁻² spectrum) in comparison with the previously published ANTARES and IceCube analyses of the same data. The point-source sensitivity in a substantial region of the sky, centered approximately at the declination of the Galactic center (δ = −29°), can be seen to have improved by up to a factor of 2. A maximum gain of at most $\sqrt{2}$ would be expected in a background-dominated sample; however, the low number of effective background events (with reconstructed energy and direction mimicking an astrophysical neutrino) is very low, so gains of more than $\sqrt{2}$ are possible. Similar gains in other regions of the sky can be seen for different energy spectra in Figure 9.

Zoom In Zoom Out Reset image size

Zoom In Zoom Out Reset image size

We present the first combined point-source analysis of data from the ANTARES and IceCube detectors. Their different characteristics, in particular IceCube's larger size and ANTARES' location in the Northern hemisphere, complement each other for southern sky searches. We calculated the sensitivity to point sources and, with respect to an analysis of either data set alone, found that up to a factor of 2 improvement is achieved in different regions of the southern sky, depending on the energy spectrum of the source. Two joint analyses of the data sets have been performed: a search over the whole southern sky for a point-like excess of neutrino events, and a targeted analysis of 40 preselected candidate source objects. The largest excess in the southern sky search has a post-trial probability of 24% (significance of 0.7σ), located at α = 3328, δ = −461 in equatorial coordinates. In the source list search, the candidate with the highest significance corresponds to HESS J1741-302, with a post-trial probability of 11% (significance of 1.2σ). Both results are compatible with the background-only hypothesis. Flux upper limits for each source candidate have been calculated for E⁻2 and E^−2.5 power-law energy spectra, as well as for E⁻² spectra with cut-offs at energies of 1 PeV, 300 TeV, and 100 TeV. Because of their complementary nature, with IceCube providing more sensitivity at higher energies and ANTARES at lower energies, a joint analysis of future data sets will continue to provide the best point-source sensitivity in critical overlap regions in the southern sky, where neutrino emission from Galactic sources in particular may be found.

The authors of the ANTARES collaboration acknowledge the financial support of the funding agencies: Centre National de la Recherche Scientifique (CNRS), Commissariat à l'énergie atomique et aux énergies alternatives (CEA), Commission Européenne (FEDER fund and Marie Curie Program), Région Île-de-France (DIM-ACAV) Région Alsace (contrat CPER), Région Provence-Alpes-Côte d'Azur, Département du Var and Ville de La Seyne-sur-Mer, France; Bundesministerium für Bildung und Forschung (BMBF), Germany; Istituto Nazionale di Fisica Nucleare (INFN), Italy; Stichting voor Fundamenteel Onderzoek der Materie (FOM), Nederlandse organisatie voor Wetenschappelijk Onderzoek (NWO), the Netherlands; Council of the President of the Russian Federation for young scientists and leading scientific schools supporting grants, Russia; National Authority for Scientific Research (ANCS), Romania; Ministerio de Economía y Competitividad (MINECO), Prometeo and Grisolía programs of Generalitat Valenciana and MultiDark, Spain; Agence de l'Oriental and CNRST, Morocco. We also acknowledge the technical support of Ifremer, AIM and Foselev Marine for the sea operation and the CC-IN2P3 for the computing facilities.

The authors of the IceCube collaboration acknowledge the support from the following agencies: U.S. National Science Foundation-Office of Polar Programs, U.S. National Science Foundation-Physics Division, University of Wisconsin Alumni Research Foundation, the Grid Laboratory Of Wisconsin (GLOW) grid infrastructure at the University of Wisconsin—Madison, the Open Science Grid (OSG) grid infrastructure; U.S. Department of Energy, and National Energy Research Scientific Computing Center, the Louisiana Optical Network Initiative (LONI) grid computing resources; Natural Sciences and Engineering Research Council of Canada, WestGrid and Compute/Calcul Canada; Swedish Research Council, Swedish Polar Research Secretariat, Swedish National Infrastructure for Computing (SNIC), and Knut and Alice Wallenberg Foundation, Sweden; German Ministry for Education and Research (BMBF), Deutsche Forschungsgemeinschaft (DFG), Helmholtz Alliance for Astroparticle Physics (HAP), Research Department of Plasmas with Complex Interactions (Bochum), Germany; Fund for Scientific Research (FNRS-FWO), FWO Odysseus programme, Flanders Institute to encourage scientific and technological research in industry (IWT), Belgian Federal Science Policy Office (Belspo); University of Oxford, United Kingdom; Marsden Fund, New Zealand; Australian Research Council; Japan Society for Promotion of Science (JSPS); the Swiss National Science Foundation (SNSF), Switzerland; National Research Foundation of Korea (NRF); Danish National Research Foundation, Denmark (DNRF).