The Antares Neutrino Telescope and Multi-Messenger Astronomy
aa r X i v : . [ a s t r o - ph . H E ] A p r The
Antares
Neutrino Telescope andMulti-Messenger Astronomy
Thierry PRADIER ∗ , on behalf of the Antares
Collaboration ‡ ∗ University of Strasbourg (France) &
Institut Pluridisciplinaire Hubert Curien
Abstract.
Antares is currently the largest neutrino telescope operating in theNorthern Hemisphere, aiming at the detection of high-energy neutrinos fromastrophysical sources. Such observations would provide important clues about theprocesses at work in those sources, and possibly help solve the puzzle of ultra-highenergy cosmic rays. In this context,
Antares is developing several programs toimprove its capabilities of revealing possible spatial and/or temporal correlations ofneutrinos with other cosmic messengers: photons, cosmic rays and gravitational waves.The neutrino telescope and its most recent results are presented, together with thesemulti-messenger programs.
1. Introduction
Astroparticle physics has entered an exciting period with the recent development ofexperimental techniques that have opened new windows of observation of the cosmicradiation in all its components: photons, cosmic rays, but also gravitational waves andhigh energy neutrinos, that could be detected both by IceCube [1] and
Antares .The advantage of using neutrinos as new messengers lies firstly on their weakinteraction cross-section ; unlike protons or γ , they provide a cosmological-rangeunaltered information from the very heart of their sources. Secondly, charged particlesare deflected by magnetic fields. Neutrinos on the other hand point directly to theirsources and exact production site.The neutrinos Antares is aiming at are typically TeV neutrinos fromAGNs or galactic sources (microquasars), 30 orders of magnitude lower in fluxthan solar neutrinos. The detection of those specific neutrinos requires underwater/ice instruments, or alternatively acoustic/radio techniques in the PeV-EeVrange and air showers arrays above 1 EeV. In spite of efforts in those variousenergy ranges, since the detection of the MeV neutrino burst from SN 1987A by
Kamiokande/Baksan/imb/Mont-Blanc [2] no astrophysical source for neutrinosabove a few GeV has ever been identified.Sources for TeV ν are typically compact objects (neutron stars/black holes), fromwhich often emerge relativistic plasma jets with a still unclear composition - purely ‡ http://antares.in2p3.fr he Antares
Neutrino Telescope and Multi-Messenger Astronomy γ -rays. These photons can be producedby e − via inverse compton effect (on ambient photon field)/synchrotron radiation, or byprotons/nuclei via photoproduction of π /π ± : p/A + p/γ −→ π π ± where π → γγ and π ± → ν µ µ, with µ → ν µ ν e e (1)In the former scenario, no neutrinos are produced, whereas in the latter, theneutrino flux is directly related to the gamma flux: a TeV neutrino detection fromgamma sources would then yield a unique way to probe the inner processes of the mostpowerful events in the universe. Several hints exist which indicates that hadrons couldbe accelerated up to very high energies. Firstly, the combined radio, X-rays and γ -rays observations of the shell-type supernova remnant RX J1713.7-3946 [3] favour theproduction of photons via π decay (figure 1, left). Secondly, the correlations betweenX and γ for the Blazar 1ES1959+650 [4] prove the existence of γ flares not visible in X(figure 1, right), which is difficult to account for in purely leptonic models. Figure 1.
Left: Multiwavelength observations of the SNR RXJ 1713.7-39; the solidcurve at energies above 10 eV corresponds to π -decay γ -ray emission, whereas thedashed and dash-dotted curves indicate the inverse Compton (IC) and NonthermalBremsstrahlung (NB) emissions, respectively. Right: Whipple vs RXTE flux, for theBlazar 1ES1959+650, which shows the existence of orphan γ flares (crosses). The connection between high energy neutrino astronomy and both gamma-rayastronomy, charged cosmic rays and gravitational waves thus emphasizes that not onlya multi-wavelength but also a multi-messenger approach, combining data from differentobservatories, is suited for the study of the most powerful astrophysical sources in theUniverse.Sections 2 and 3 first describe the
Antares neutrino telescope and its latest physicsresults. Section 4 presents the different strategies imagined to detect neutrinos fromgamma-ray bursts, while section 5 details the correlations that can be performed withother observatories, such as
Auger , Hess , or
Virgo / LIGO . he Antares
Neutrino Telescope and Multi-Messenger Astronomy
2. The
Antares
Neutrino Telescope: principles & description
Antares is a three-dimensional grid of photomultiplier tubes, arranged in strings, ableto detect the ˇCerenkov photons induced by the passage in sea water of relativisticcharged particles, produced by the interaction of a cosmic neutrino in the Earth [5].The measurements of the time of the hits (with a time resolution of the order of ns)and the amplitude of the hits (with a resolution of about 30%), together with theposition of the hits (by measuring the position of each PMT, to reach a resolution ofabout 10 cm) are needed to achieve the reconstruction of those signals with the desiredresolution. Muon tracks initiated by the charged current interaction of a ν µ in theEarth are detected via their directional ˇCerenkov light and can be reconstructed withan angular resolution below 0.3 ◦ above 10 TeV. The resolution below this energy isdominated by the kinematics of the interaction. The energy resolution is quite poor,a factor 2-3 on average, restricted by the granularity/density of the light sensors andthe fact that the muon traverses the detector. Showers produced by ν e on the otherhand emit quasi-isotropic light, and can be reconstructed with a better energy resolution(roughly 30 %) but with a poorer angular resolution, typically 3-5 ◦ .The main physical backgrounds are twofold. Atmospheric muons produced in theupper atmosphere by the interaction of cosmic rays can be strongly suppressed becauseof their downward direction. Upward-going atmospheric neutrinos on the other handare more delicate to identify: they have exactly the same signature as the expectedcosmic signal Antares awaits for.The
Antares neutrino telescope, deployed at 2500 m below sea surface, 40 kmoff the coast of Toulon (Southern France) is composed of 12 strings, with 25 storeyseach containing a triplet of 10 “ photomultipliers oriented at 45 degrees downward to beoptimally sensitive to upward going muons [6]. Since May 2008, 12 lines are continuouslytaking data. A schematic description of the detector, together with the layout of thelines, and of one of the storey, can be found in figure 2. Figure 2.
The layout of the completed
Antares detector. Bottom left: optical storeywith its three photomultipliers. he Antares
Neutrino Telescope and Multi-Messenger Astronomy
3. Selected results from
Antares
The main results obtained recently, related to atmospheric muons and neutrino-inducedupward-going muons, are presented in this section.
The attenuation of the muon flux as a function of depth can be observed in figure 3 (left),as computed using the method described in [7]. Alternatively, the reconstructed zenithangle can be converted to an equivalent slant depth through the sea water, and a depthintensity relation can be extracted, also shown in figure 3 (right). The results are inagreement with previous measurements.
Figure 3.
Left: Attenuation of the flux of muons as a function of depth, asextracted using the method described in [7]. Right: Vertical depth intensity relationof atmospheric muons with E µ > GeV (black points).
The muons produced by the interaction of neutrinos can be isolated from theatmospheric muons by requiring that the muon trajectory is reconstructed as up-going.In figure 4 the zenith angular distribution of muons in the 2007+2008 data (5 lines and9-12 lines) sample is shown. A total of 1062 up-going neutrino candidates are found, ingood agreement with expectations from the atmospheric neutrino background.An all sky search, independent of assumption on the source location, has beenperformed on the 5-line data. No significant cluster was found. A search amongst apre-defined list of 24 of the most promising galactic and extra-galactic neutrino sources( e.g. supernova remnants, BL Lac objects) was performed. The corresponding fluxexpected sensitivities, assuming an E flux, are also plotted in figure 4 and compared he Antares
Neutrino Telescope and Multi-Messenger Astronomy Figure 4.
Left: Zenith distribution of reconstructed muons in the 2007+2008 data.Right: expected sensitivity for
Antares in the 5 line configuration. The expectedsensitivity of
Antares for one year with twelve lines is also shown. to published upper limits from other experiments (see [8] for
Amanda ). Also shown isthe predicted upper limit for
Antares after one full year of twelve line operation.
4. Searches for neutrinos from GRBs
Among all the possible astrophysical sources, transient sources offer one of the mostpromising perspectives for the detection of cosmic neutrinos thanks to the almostbackground free search. For instance, several models predict the production of high-energy neutrinos by gamma-ray bursts (GRBs) [9]: the detection of these neutrinoswould provide evidence for hadron acceleration by GRBs.Two different methods to detect transient sources can be used. The triggered search is based on the search for neutrino candidates in conjunction with an accurate timingand positional information provided by an external source. The rolling search is basedon the search for high energy or multiplet of neutrino events coming from the sameposition within a given time window.
Classically, GRBs or flare of AGNs are detected by gamma-ray satellites (Swift, Fermi)which deliver in real time an alert to the Gamma-ray bursts Coordinates Network (GCN,figure 5). The characteristics (direction and time of the detection) of this alert are thendistributed to the other observatories. The small difference in arrival time and positionexpected between photons and neutrinos allows a very efficient detection by reducingthe associated background. This method has been implemented in
Antares mainly forthe GRB detection since the end of 2006. Data triggered by more than 500 alerts havebeen stored up to now. he Antares
Neutrino Telescope and Multi-Messenger Astronomy Figure 5.
Left: the Gamma-ray bursts Coordinates Network (GCN). Right:
Antares response time to an alert.
Based on the time of the external alert, in complement to the standard acquisitionstrategy, an on-line running program stores the data coming from the whole detectorduring 2 minutes without any filtering. This allows to lower the energy threshold of theevent selection during the offline analysis with respect to the standard filtered data. Dueto a continuous buffering of data (covering 60s) and thanks to the very fast responsetime of the GCN network (see figure 5),
Antares is able to record data before thedetection of the GRB by the satellite [10]. The analysis of the data relying on thoseexternal alerts is on-going.Due to the very low background rate, even the detection of a small number ofneutrinos correlated with GRBs could set a discovery. But, due to the relatively smallfield of view of the gamma-ray satellites ( e.g. , Swift has a 1.4 sr field of view), onlya small fraction of the existing bursts are triggered. Moreover, choked GRBs withoutphotons counterpart can not be detected by this method. This justifies the use of acomplementary rolling search strategy.
TAToO project
This second method relies on the detection of a burst of neutrinos in temporal anddirectional coincidence. Applied to
Antares , the detection of 2 neutrinos within ashort time is almost statistically significant: the number of doublets due to atmosphericneutrino background events is of the order of 0.004 per year when a temporal window of900 s and a directional cone of 2 ◦ × ◦ are defined. It is also possible to search for singlecosmic neutrino events by requiring that the reconstructed muon energy is higher thana given energy threshold (typically above a few tens of TeV), for which the atmosphericneutrino background is negligible (see figure 6). When the neutrino telescope is running,this method is almost 100% efficient and applies whenever the neutrinos are emittedwith respect to the gamma flash. The main drawback is that a detection is notautomatically associated to an astronomical source. It is thus fundamental to organizea complementary follow-up program to confirm the detection. he Antares
Neutrino Telescope and Multi-Messenger Astronomy
Antares is organizing a follow-up program in collaborationwith TAROT (T´elescope `a Action Rapide pour les Objets Transitoires, RapidAction Telescope for Transient Objects), called
TAToO (TAROT-
Antares
Targetof Opportunity, figure 6). The TAROT network is composed of two 25 cm opticalrobotic telescopes located at Calern (South of France) and La Silla (Chile). The mainadvantages of the TAROT instruments are the large field of view of 1 . ◦ × . ◦ andtheir very fast positioning time (less than 10 s). A GRB afterglow requires a veryfast observation strategy in contrast to a core-collapse supernovae for which the opticalsignal will appear several days after the neutrino signal. The observational strategy iscomposed of a real time observation followed by few observations during the followingmonth. Depending on the neutrino trigger settings, an alert sent to TAROT is issuedat a rate of about one or two per month. The total latency is quite impressive: lessthan 10 ms/event for reconstruction, with a pointing accuracy of the order of 0 . ◦ above 10 TeV, less than 1s for alert sending, with a positioning of the telescopes in lessthan 10 seconds. A confirmation by an optical telescope of a neutrino alert will not Figure 6.
Left: a typical GRB neutrino spectrum. Right: Concept of
TAToO . only give the nature of the source but also allow to increase the precision of the sourcedirection determination in order to trigger other observatories. The TAToO program isoperational since February 2009, and several neutrino alerts have already been sent [11].A torough analysis of these alerts is underway.
5. Correlations with other observatories
As already pointed out, high energy cosmic neutrinos could be produced in astrophysicalsources that are also potential emitters of charged cosmic rays, gamma-rays orgravitational waves. Correlations with dedicated instruments could then bring a harvestof unique information about these sources.
Auger and ultra-high energy cosmic rays
The Pierre Auger Observatory reported an anisotropy in the arrival directions of ultra-high energy cosmic rays (UHECR) [12]. Correlation with Active Galactic Nuclei (AGN) he Antares
Neutrino Telescope and Multi-Messenger Astronomy
Antares .Instead of searching for such a localized excess, neutrino arrival direction can becorrelated with the arrival directions of ultra high energy cosmic rays, as describedin [13]. By obtaining the probability density function of the number of neutrino eventswithin specific angular distance from observed UHECRs, the number of neutrino eventsin the vicinity of observed ultra-high energy cosmic rays, necessary to claim a discoverywith a chosen significance, can be calculated. For example, for 27 UHECR, a correlationsignificance of 5 σ is reached with 2%- 25% of neutrino events falling in 1-10 ◦ bins aroundthe original UHECR direction. Possible observed correlation of the arrival directions ofthose two messengers would provide a strong indication of hadronic acceleration theory. γ sources observed by Hess
The exact composition of the emission of the sources observed by
Hess and othergamma-ray observatories is still an open issue. Some of these sources could have anon-negligible hadronic component, and could then be neutrino emitters. In the caseof continuous sources (as opposed to transients), a stacking analysis can be performed,which consists in summing up all the events detected in a given angular region aroundthe direction of the γ source, when below the horizon. Extended sources can be dividedin different regions of interest to increase the signal-to-noise ratio. Obviously, addingmore and more sources to the analysis also increases the number of background eventsdetected. It is found that an optimum of 25 sources is needed to reach a significanceclose to 3 σ after 5 years of Antares data taking, assuming that all
Hess sources arehadronic, after which the sensitivity decreases.There have also been many discussions about the possibility to detect muonsproduced by high energy gamma-rays in underground, underice or underwaterneutrino telescopes. In contrast to upward-going muons from neutrinos, downward-going muons from gamma rays suffer from a high atmospheric muon background.Therefore the sensitivity of a neutrino telescope to gamma ray induced muons is quitelower than atmospheric ˇCerenkov telescopes. However it monitors continuously alldirections. There are at least three processes by which a photon can produce muons :photoproduction, muon pair production and charm decay. It is then possible to lookfor a global excess of the muon flux in a direction correlated with the position of thesource (and within a given time window, if the source is transient). If the statistics issufficient, which depends on the source spectral index and flux, accurately measured fora number of galactic sources, calibration of both the absolute pointing and the angularresolution of the neutrino telescope could be performed [14]. he Antares
Neutrino Telescope and Multi-Messenger Astronomy gwhen project Coincident searches of high-energy neutrinos (HEN) and gravitational waves (GW) arealso of great interest and are detailed in [15]. Both GW and HEN are alternativecosmic messengers that carry information from the innermost regions of the astrophysicalengines. Such messengers could also reveal new, hidden sources that were not observedby conventional photon-based astronomy. The observation of such coincidence wouldbe a landmark event and can also provide observational evidence that GW and HENoriginate from a common astrophysical source.The network of GW detectors formed by the
LIGO and
Virgo interferometers(figure 7) can determine the direction/time of GW bursts in connection with neutrinoevents observed in
Antares . The
Virgo / LIGO network started a data-taking phasemid-2009. It monitors roughly 30% of the sky in common with
Antares : figure 7 showsthe daily averaged visibility of some sources by the network. Joint searches are ongoingwithin a dedicated gwhen working group: investigations [16] led for
Antares indicatethat a joint optimized
Virgo + LIGO / Antares analysis is needed, in order to increasethe coincident detection efficiency, while keeping the coincident false-alarm rate as lowas possible, to enhance the significance of such a detection.
Figure 7.
Left: the network of groundbased GW interferometers. Thelocation of
Antares is also shown. Right: averaged visibility skymap for
Antares / Virgo + LIGO , where the location of known microquasars/soft-gammarepeaters, potential emitters of GW bursts and HEN, is also shown. This daily(normalized) averaged visibility takes into account the interferometers’ beam patterns,and the fact that a neutrino source can be detected only when below the horizon.
6. Conclusions
Antares is now taking data since 2008 and has demonstrated the possibility to operateand get competitive physics results from a neutrino telescope under the sea in theNorthern Hemisphere, in spite of its reduced size with respect to IceCube. It shouldalso be noted that, because of its location,
Antares can observe the Galactic Centreand most of
Hess sources on the galactic plane. he Antares
Neutrino Telescope and Multi-Messenger Astronomy
Antares detector not only enhances its own capabilities as a neutrino telescope, butalso contributes to the global effort of understanding the most violent phenomena inour Universe. In addition to offline searches for spatio-temporal correlations with othercosmic messengers (photons, cosmic rays and gravitational waves),
Antares has thecapability to handle external alerts in real time and to trigger follow-up observationswith the small latency time required for the study of transient sources. The possibilityto store a few minutes of raw data in coincidence with a GCN alert also brings newopportunities for offline analysis. This could be extended in the future to handle alertsinvolving other messengers, such as gravitational waves, to complete the gwhen project.Finally, the extension of the follow-up programs to other instruments in differentranges of wavelengths (X, radio) would undoubtely contribute to the development ofthe astrophysical potential of
Antares . Aknowledgements:
Great thanks to the Organizing Committee, especially FulvioRicci, for trusting me for this talk. [1] F. Halzen, for
IceCube , Proceedings of
GWDAW14 (2010)[2] K. S. Hirata et al. , Phys. Rev. Lett. (1987) 1490; R. M. Bionta et al , Phys. Rev.Lett. (1987) 1494 ; E. N. Alexeyev et al , PZETF (1987) 461; Mont-Blanc (controversial) :V. L. Dadykin et al. , PZETF (1987) 464[3] E. G. Berezhko & H. J. V¨olk, arXiv:0707.4647v1 , 30th icrc Veritas
Collaboration), arXiv:astro-ph/0305577v1 , 28th icrc
International Conference on High-Energy Physics
578 (1960)[6]
Antares
Collaboration,
Nuclear Instruments and Methods in Physics Research A
555 (2005) 132[7]
Antares
Collaboration,
Astropart. Phys. (2010) 86-90[8] Amanda
Collaboration, Phys. Rev. D (2007) 042008[9] e.g. S. Razzaque et al. , Phys. Rev. D (2004) 023001[10] M. Bouwhuis, for Antares , Proceedings of 31st ICRC conference, Lodz (2009) arXiv:astro-ph/0908.0818 [11] D. Dornic, for
Antares , Proceedings of 31st ICRC conference, Lodz (2009) arXiv:astro-ph/0908.0804 [12]
Pierre Auger Collaboration , Science (2007) 938[13] J. Petrovic, arXiv:astro-ph/0908.1235 [14] G. Guillard for
Antares , Proceedings of 31st ICRC conference, Lodz (2009) arXiv:astro-ph/0908.0855 [15] E. Chassande-Mottin, for
Antares / LIGO / Virgo , Proceedings of
GWDAW14 (2010)[16] Th. Pradier,