Acta Polytechnica doi:10.14311/AP.2013.53.0756 Acta Polytechnica 53(Supplement):756–759, 2013 © Czech Technical University in Prague, 2013 available online at http://ojs.cvut.cz/ojs/index.php/ap LIMIT ON UHE NEUTRINO FLUXES FROM THE PIERRE AUGER OBSERVATORY S. Malderaa,∗, for the Pierre Auger Collaborationb,c a INAF-Osservatorio Astrofisico di Torino, c. Fiume 4, Torino, Italy b Av. San Matrín Norte 304 (5613) Malargüe, Prov. de Mendoza, Argentina c full author list: http: // www. auger. org/ archive/ authors_ 2012_ 06. html ∗ corresponding author: maldera@to.infn.it Abstract. The surface detector of the Pierre Auger Observatory is sensitive to Ultra High Energy (UHE) neutrinos. Neutrinos of all flavors can interact in the atmosphere producing inclined showers near the ground. Moreover, ultra high energy Earth-skimming tau neutrinos can be observed through the detection of showers induced by the decay of tau leptons created by interactions in the Earth’s crust. In both cases, neutrino showers can be identified through the time structure of the signals in the surface detector stations. Two sets of identification criteria have been designed to search for down-going and up-going neutrinos in the recorded data, with no candidates found. We will discuss the identification criteria used, and we will present the corresponding limits on the diffuse and point source neutrino fluxes. Keywords: UHE neutrinos, Pierre Auger Observatory. 1. Introduction All the proposed models for the origin of Ultra High Energy Cosmic Rays (UHECR, as are usually named cosmic rays with E > 1018 eV) predict a flux of high energy neutrinos, mainly via charged pion decay fol- lowing interactions on matter and radiation. Such interactions can occur at the acceleration site itself (“astrophysical neutrinos”) or during the propagation in the background radiation field (“cosmogenic neutri- nos”) [1, 2]. The detection of UHE neutrinos would open a new observation window to the universe. The Pierre Auger Observatory, is located in the province of Mendoza, Argentina at a mean altitude of 1400 m a.s.l. It uses two different techniques to detect the air showers: an array of about 1660 water Cherenkov detectors, placed at a distance of 1.5 km from each other, samples the particles at ground level over an area of about 3000 km2 (surface detector, SD [3]), while a fluorescence detector (FD [4]) ob- serves the ultra-violet light emitted by atmospheric nitrogen excited by the particles of the shower. The FD consists of 27 telescopes located at four sites at the edges of the ground array. As it operates only during moon-less nights it has a duty cycle of ∼ 15 %, while the SD has a duty cycle of almost 100 %. Although the main goal of the Auger Observatory is the detection of extensive air showers produced by UHECRs, it has a good detection and identifi- cation capability for neutrinos with energies above 1017 eV. At such energy, neutrinos of all flavors can interact in the atmosphere inducing a “down-going” (DG) shower that can be detected at ground. In addi- tion, Earth-skimming (ES) tau neutrinos can undergo charge current interaction inside the Earth, generat- ing a tau lepton that can emerge and decay in the atmosphere, giving an upward-going air shower. Even if tau neutrinos are not predicted to be produced at astrophysical sources, the flux at Earth will be equally distributed between all neutrino flavors due to neu- trino oscillations. The neutrino search is based on the data from the surface detector. Each SD station consists of a cylindrical polyethylene tank, 3.6 m in diameter and 1.2 m tall, containing 12 tons of puri- fied water. Three large photocatode photomultipliers detect the Cherenkov light emitted by relativistic par- ticles crossing the water volume and their signals are digitalized with samplig rate of 40 MHz. Two different trigger modes are implemented in the stations, a sim- ple threshold trigger and a time over threshold trigger (ToT) requiring that at least 13 samples are over a lower threshold within a sliding window of 3 µs (120 samples). 2. Identification of ν candidates Neutrino induced showers have to be extracted from a large background of cosmic ray showers (protons or heavier nuclei). At large zenith angles (θ > 75°), nucleonic showers will be dominated at the ground by the muonic component, as the electromagnetic ones are almost completely absorbed in the atmosphere (old showers). On the other hand neutrino induced showers (DG or ES) initiated near the detector will be electromagnetic rich (young showers). The key point for the neutrino discrimination is therefore the selection of deeply interacting (young) inclined show- ers. At very large zenith angles (above 90° in the ES case) the standard event reconstruction algorithms are not reliable and, therefore, in order to select very 756 http://dx.doi.org/10.14311/AP.2013.53.0756 http://ojs.cvut.cz/ojs/index.php/ap http://www.auger.org/archive/authors_2012_06.html vol. 53 supplement/2013 Limit on UHE Neutrino Fluxes from the Pierre Auger Observatory inclined events, some more general shower characteris- tics are exploited: the shape of the shower footprint at ground and the apparent ground speed. The pattern of the triggered SD stations for an inclined shower is an elongated ellipse, with the major axis (L) along the shower arrival direction and the minor axis (W) perpendicular to that direction (see [5] for the details of the definitions of L and W). A very inclined event has a large L/W ratio. The ground speed is defined as Vij = dij/Δtij, where dij is the distance between two stations participating in the event, and Δtij is the difference between the start time of the signal in the two stations. The average ground speed (ob- tained by averaging over all the pairs of stations in the event) is close to the speed of light for a horizontal shower, while it is much larger in the case of a vertical event. A set of cuts (listed in Tab. 1) on the L/W, the average signal speed (V ) and its dispersion (σV ) have been implemented for the DG and ES channels to select showers in the zenith angle ranges between 75° < θ < 90° and 90° < θ < 96° respectively. In addition we require at least 3 and 4 triggered stations for the ES and DG channels and in the case of the DG analysis a further cut on the reconstructed zenith angle is used. The selection of events with a large electromag- netic component is based on the time structure of the signals recorded by the ground detectors. In fact, a shower with a significant electromagnetic compo- nent at the ground produces a signal in the triggered SD stations which is broader in time (hundreds of nanoseconds) with respect to the one given by “old”, muon-dominated showers (tens of nanoseconds). The main observable is the ratio of the integrated signal charge collected by the photomultipliers over its peak height (Area over Peak, AoP) normalized to that of isolated muons (periodically collected for calibration and monitoring purposes), which is sensitive to the time spread of the signal. In addition, stations with a broad signal usually satisfy the time over threshold (ToT) local trigger condition. The selection criteria have been optimized using Monte-Carlo simulations to estimate the expectations from neutrino induced showers while the background is estimated using a subsample of the SD data (train- ing data), to take into account the actual primary cosmic rays composition and all possible detector ef- fects that may not be reproduced by simulations. In the Earth-skimming analysis the selection of young showers is done applying a cut on the fraction of sta- tions satisfying the ToT trigger condition and with AoP > 1.4. In the down-going case, the discrim- ination is performed using the Fisher discriminant method [6] to improve the separation of neutrinos and background. The variables used are the AoP of the first (time ordered) 4 triggered stations and some combinations of them (their squares and their prod- uct), and the difference in AoP between “early” and “late” stations. The cut on the Fisher discriminant is fixed requiring 1 background event in 20 years. Down-going (DG) Earth skimming (ES) n. stations ≥ 4 n. stations ≥ 3 θrec > 75° L/W > 3 L/W > 5 V < 0.313 m ns−1 0.29 m ns−1 < V < 0.31 m ns−1 σV /V < 0.08 σV < 0.08 m ns−1 Fisher discriminant fraction of stations with based on AoP ToT trigger AND AoP > 1.4 greater than 60 % Table 1. Criteria used for the selection of inclined events (upper part) and for the neutrino discrimination (lower part). In the Earth-skimming case the training data sub- sample goes from 1 Nov 2004 to 31 Dec 2004, while it is from 1 Jan 2004 to 31 Oct 2007 for DG events1. The remaining events (search sample) were not used until all the selection algorithms were defined, and only at that point they were “unblinded”. 3. Exposure and limits to diffuse flux The neutrino selection criteria are applied to the search samples (from 1 Jan 2004 to 31 May 2010 for ES and from 1 Nov 2007 to 31 May 2010 for DG) for both the down-going and Earth-skimming searches, resulting in zero candidates being found. To be able to give an upper limit to the neutrino flux, we need to compute the neutrino exposure of the surface detector. This is obtained by folding the array aperture, the efficiency of the neutrino identification (given by the combination of the trigger probability of the array and the selection cuts) and the ν interaction probability, and integrating over time. For down-going neutrinos the identification efficiency depends on the neutrino flavor and interaction channel, its energy, incident zenith angle θ, and depth of the interaction. In the case of Earth-skimming ντ it depends mainly on the neutrino energy and on the height above ground of the shower induced by tau decay. The exposure is computed by means of Monte-Carlo simulations, taking into account the time evolution of the SD array. The first interaction between a neu- trino and nucleon is simulated with the HERWIG package [7], the development of the shower in the at- mosphere with AIRES [8], and the response of the SD detectors using the Auger simulation framework. In the case of ντ CC interactions, the decay of the τ lepton is simulated with TAUOLA [9]. For the DG ντ chan- nel a detailed description of the topography around the detector was used to estimate the contribution from ντ interacting in the Andes. The accumulated exposure corresponding to the search periods for the 1As DG ν-induced showers are more similar to cosmic ray showers, more statistics is required for the evaluation of the background. 757 S. Maldera, for the Pierre Auger Collaboration Acta Polytechnica Figure 1. Exposure of the surface detector for ES and DG neutrino induced showers, corresponding re- spectively to 3.5 and 2 years of the complete and fully efficient surface detector (the deployment of the SD was concluded by the end of 2008). Earth-skimming down-going k < 3.2 × 10−8 GeVcm2 s sr k < 1.7 × 10 −7 GeV cm2 s sr ∼ 0.16 < Eν[EeV] < 20.0 ∼ 0.1 < Eν[EeV] < 100.0 Table 2. 90 % C.L. upper limits to the single flavor neutrino flux and corresponding energy range, for Earth-skimming and down-going searches. Earth-skimming and down-going analysis are shown in Fig. 1. The main sources of uncertainties for DG neutrinos came from the ν induced shower simulation and the hadronic models used (+9 %, −33 %), and from neutrino cross-sections (±7 %). For ES neutrinos the dominant uncertainties comes from the tau en- ergy losses (+25 %, −10 %), shower simulation (+20 %, −5 %), and topography (+18 %, 0 %). Assuming a neutrino flux f(Eν) = k E−2, and a complete neu- trino mixing (flavor ratio 1:1:1), the upper limits to the single flavor neutrino flux are derived using a semi-Bayesian extension [10] of the Feldman–Cousins approach [11], that allows one to include in the limits the systematic uncertainties on the exposure cited above. The 90 % C.L. limits are shown in Tab. 2, together with the energy range in which they are derived. The resulting flux upper limits are also shown in Fig. 2, together with the results from IceCube Neu- trino Observatory [12] and the ANITA experiment [13], and flux predictions for cosmogenic neutrinos under different assumptions [14, 15]. The limits on the dif- ferential flux are also shown, assuming a spectrum of the form E−2 in each energy bin. 4. Limits to point-like sources As the neutrino search is limited to large zenith angles (75° to 96°), a point-like source can be seen only for a fraction of the sidereal day, depending on its decli- nation. In fact, a point in the sky with declination [eV]νE 17 10 18 10 19 10 20 10 21 10 ] -1 s r -1 s -2 ) [ G e V c m ν (E Φ ν2 k = E -11 10 -10 10 -9 10 -8 10 -7 10 -6 10 -5 10 Single flavour neutrino limits (90% CL) limitsν Auger downward-going Auger Earth-skimming IceCube 2011a Anita 2009 Cosmogenic models Ahlers 2010 Kotera 2010 Figure 2. Integrated (solid lines) and differential (dashed lines) 90 % C.L. upper limits to the single flavor diffuse flux of down-going ν and Earth-skimming ντ, as a function of neutrino energy. Limits from the IceCube [12] and ANITA [13] experiments and predictions for cosmogenic neutrinos under different assumptions [14, 15] are also shown. δ and right ascension α at a given sidereal time t is seen at the observatory latitude (λ = −35.2°) with a zenith angle θ given by cos(θ) = sin λ sin δ + cos λ cos δ sin(2πt/T − α) (1) being T the duration of sidereal day. The declination range accessible with this analysis is between −65° and 55°, while the regions near the terrestrial poles are not observable. The point source exposure is evaluated in a similar way as the diffuse exposure but avoiding to integrate over the solid angle [16]. Assuming a differential flux f(Eν) = kPS E−2, 1:1:1 neutrino flavor ratio, the upper limits to kPS are derived as function of source declination, in the same way as in the diffuse case. The 90 % C.L. upper limits for the DG and ES analyses are shown in Fig. 3 as a function of declination. In both analyses we have a region of about 100° in declination were the sensitivity is almost constant, and the limits on kPS are at the levels of ≈ 5 × 10−7 for the ES analysis and ≈ 2.5 × 10−6 GeV cm−2 s−1 for the DG analysis. The shapes of the declination dependent limits are determined mainly by the fraction of time in which a source is within the zenith angle range of DG and ES searches. The upper limits are derived for neutrino energies from 1.6 × 1017 to 2.0 × 1019 eV and from 1 × 1017 to 1 × 1020 eV for the Earth-skimming and the down-going analysis respectively. In Fig. 4 we show the limits on kPS for the particular case of the active galaxy Centaurus A (δ = −43°), a potential acceleration site for UHECR. Neutrino flux predictions for three different models of UHE ν production in the jets and in the core of CenA are also shown [17–19]. The expected number of events for a flux like in [17] is about 0.1 for the ES and 0.02 for the DG searches respectively. They are about one order of magnitude smaller for the flux predicted in [18]. 758 vol. 53 supplement/2013 Limit on UHE Neutrino Fluxes from the Pierre Auger Observatory [deg]δSource declination -80 -60 -40 -20 0 20 40 60 80 ] -1 s -2 ) [G e V c m ν F (E ν2 = E P S k -710 -6 10 -5 10 Single flavour neutrino limits (90% CL) Auger downward-going Auger Earth-skimming Figure 3. Upper limits (90 % C.L.) for the integral flux of single flavour neutrinos from a point-like source as a function of source declination. [eV] ν E 15 10 16 10 17 10 18 10 19 10 20 10 21 10 22 10 23 10 ] -1 s -2 ) [G e V c m ν F (E ν2 = E P S k -10 10 -9 10 -8 10 -710 -6 10 -5 10 -410 Centaurus A - Single flavour neutrino limits (90% CL) Kachelriess 2009 Cuoco 2008 Auger Earth-skimming Auger Downward-going LUNASKA 2008 IceCube 2011b Figure 4. Upper limits (90 % C.L.) on the integral flux of neutrinos from Centaurus A, together with the limits from other experiments (IceCube [20], LU- NASKA [21]) in different energy ranges and theoretical predictions. References [1] F. Halzen et al., Reports on Progress in Physics 65, (2002) 1025 [2] L. Anchordoqui, T.Montaruli, Annual Review of Nuclear and Particle Science 60, (2010) 129–162 [3] I. Allekotte et al., Nuclear Instruments and Methods in Physics Research Section A 586, (2008) 409 [4] J. Abraham et al., The Pierre Auger Collaboration, Nuclear Instruments and Methods in Physics Research Section A 620, (2010) 227 [5] J. Abraham et al., The Pierre Auger Collaboration, Physical Review D 79, (2009) 102001 [6] R. Fisher, Annals of Eugenics 7, (1936) 179 [7] G. Corcella, et al., JHEP 101, (2001) 10 [8] S. Sciutto, AIRES: A System for Air Shower Simulations, arXiv:astro-ph/9911331. [9] S. Jadach et al., Computer Physics Communications 76(3), (1993) 361–380 [10] J. Conrad et al., Physical Review D 67, (2003) 12002 [11] G.J. Feldman, R.D. Cousins, Physical Review D 57, (1998) 3873 [12] R. Abbasi et al., IceCube Collaboration, Physical Review D 83, (2011) 092003 [13] P. Gorham et al., ANITA Collaboration, Physical Review D 82, (2010) 022004 [14] M. Ahlers et al., Astroparticle Physics 34, (2010) 106 [15] K. Kotera et al., Journal of Cosmology and Astroparticle Physics 10, (2010) 013 [16] P. Abreu et al., Pierre Auger Collaboration, Astrophysical Journal Letters 755, (2012) L4 [17] A. Cuoco et al., Physical Review D 78 (2008), 023007 [18] M. Kachelriess et al., New Journal of Physics 11, (2009) 065017 [19] L.A. Anchordoqui arXiv:1104.0509 [hep-ph]. [20] R. Abbasi et al., The IceCube Collaboration, Astrophysical Journal 732, (2011), 18 [21] C.W. James et al., Monthly Notices of the Royal Astronomical Society 410, (2011) 885–889 759 Acta Polytechnica 53(Supplement):756–759, 2013 1 Introduction 2 Identification of nu candidates 3 Exposure and limits to diffuse flux 4 Limits to point-like sources References