Neutrino Telescope Array Letter of Intent: A Large Array of High Resolution Imaging Atmospheric Cherenkov and Fluorescence Detectors for Survey of Air-showers from Cosmic Tau Neutrinos in the PeV-EeV Energy Range
aa r X i v : . [ a s t r o - ph . I M ] J u l Neutrino Telescope Array Letter of Intent:
A Large Array of High Resolution Imaging Atmospheric Cherenkovand Fluorescence Detectors for Survey of Air-showers from CosmicTau Neutrinos in the PeV-EeV Energy Range
Makoto Sasaki ∗ and George Wei-Shu Hou † ICRR, The University of Tokyo, Kashiwa, Chiba 277-8582, Japan Department of Physics, National Taiwan University, Taipei 10617, TaiwanJuly 20, 2015
Abstract
This Letter of Intent (LoI) describes the outline and plan for the Neutrino TelescopeArray (NTA) project. High-energy neutrinos provide unique and indisputable evidencefor hadronic acceleration, as well as a most accurate probe into the hidden sector of tradi-tional astronomy or physics, such as dark matter. However, their extremely low flux andinteraction cross section make their detection extraordinarily difficult. Recently, IceCubehas reported astronomical neutrino candidates in excess of expectation from atmosphericsecondaries, but is limited by the water Cherenkov detection method. A next generationhigh-energy neutrino telescope should be capable of establishing indisputable evidencefor cosmic high-energy neutrinos. It should not only have orders-of-magnitude largersensitivity, but also enough pointing accuracy to probe known or unknown astronomicalobjects, without suffering from atmospheric secondaries. The proposed installation is alarge array of compound eye stations of imaging atmospheric Cherenkov and fluorescencedetectors, with wide field of view and refined observational ability of air showers from cos-mic tau neutrinos in the PeV-EeV energy range. This advanced optical complex systemis based substantially on the development of All-sky Survey High Resolution Air-showerdetector (Ashra) and applies the tau shower Earth-skimming method to survey PeV-EeV ν τ s. It allows wide (30 ◦ × ◦ ) and deep ( ∼
400 Mpc) survey observation for PeV-EeV ν τ s assuming the standard GRB neutrino fluence. In addition, it enjoys the pointing ac-curacy of better than 0.2 ◦ in essentially background-free conditions. With the advancedimaging of Earth-skimming tau showers in the wide field of view, we aim for clear discov-ery and identification of astronomical ν τ sources, providing inescapable evidence of theastrophysical hadronic model for acceleration and/or propagation of extremely high en-ergy protons in the precisely determined direction. In this LoI, we present main featuresof the NTA detector, scientific goal and observational objects, Earth-skimming detectionmethod, the NTA detector, the expected detector performance, and brief summaries oftime frame, organization, and funding. Keywards:
Astroparticle physics, Neutrino astronomy, Hadron acceleration, Cosmic rayorigin, Dark matter, Gamma ray burst, Active galactic nuclei, Neutrino telescope, PeV-EeVneutrinos, Earth-skimming detection, Tau neutrino, Ashra, Neutrino Telescope Array ∗ [email protected] † [email protected] Introduction
High-energy neutrinos uniquely provide indisputable evidence for hadronic acceleration inthe universe. High-energy charged cosmic rays have been observed for a long time, but theirorigin is still a mystery. The energy spectrum follows globally a broken E − α power law, where α = 2 . ∼ .
1, which indicates shock acceleration. Several astronomical object classes havebeen proposed as potential hadronic accelerators. The galactic and extragalactic magneticfields prevent us from using the arrival direction observed on Earth to reveal the actualsources. So far, standard astronomical observational data, spanning the electromagneticwavelengths from radio to γ -ray, have not succeeded in revealing direct evidence of thenon-thermal process. On the other hand, high-energy neutrinos should be produced at theaccelerators through charged pion production in collisions with radiation fields or the ambientmatter, in reactions such as: p + γ → ∆ + → π + p, π + + np + nucleus → π + X ( π = π , π ± ) . Subsequent decay gives the approximate neutrino flavour ratio ν e : ν µ : ν τ = 1 : 2 : 0 atthe sources, which is turned into the ratio of ν e : ν µ : ν τ = 1 : 1 : 1 by neutrino oscillationupon arrival at Earth. The photopion ( pγ ) reaction is typically the main neutrino generationprocess where extra galactic sources like jets and cores of active galactic nuclei (AGN) and γ -ray burst (GRB) jets have been widely studied. Some sources like starburst galaxies (SBGs)may emit the neutrino fluxes mainly through the hadronuclear ( pp ) reaction [1]. For manyastronomical objects, ambient photons are expected to be in the UV region. In that case,the kinetic threshold for photopion production through delta resonance is in the range ofseveral PeV.The IceCube Collaboration claims the first observation of two PeV-Energy neutrinos, withmoderate 2.8 σ excess over Monte Carlo expectation of background events [2]. They furtherextended their analysis to lower energy region [3]. Fitting to the observed photoelectron spec-trum, they estimate the diffuse neutrino flux to be E φ ν e + ν µ + ν τ = 3 . × − GeV sr − s − ,assuming an E − power law flux. The fact that no more events occur in the higher energyregion favours neutrinos from astronomical objects but not cosmogenic neutrinos [4], if theevents are true neutrino signals. Assuming astronomical objects where the observed neutri-nos were produced, the estimated fluences of the neutrino beams are rather high. It thereforebecomes plausible that a next generation high-energy neutrino telescope, with higher sensi-tivity for high-energy neutrinos and wide field of view, could make clear discovery of hadronaccelerators in the Universe.For cosmogenic neutrinos, produced by the photopion process of protons with the cosmicmicrowave background, the energy threshold is around 10 . eV. However, the Pierre AugerObservatory (Auger) claims that heavier components dominate the highest energy regionaround 10 . eV [5]. If true, the flux estimate of cosmogenic neutrinos is much suppressed.Both the cosmic ray flux spectrum around threshold and the density of cosmic microwavebackground are observed so well, the detection of cosmogenic neutrinos provide a good checkof the Auger results on cosmic ray composition. Besides high-energy neutrino detection,the observation of PeV γ -rays could also have provided a clear proof of hadron acceleration,from the subsequent π → γγ decay in the above process. However, in the PeV range,photons are absorbed by interaction with cosmic microwave photons into electron pairs.Therefore, especially the observation of PeV-EeV neutrinos with precise pointing accuracywould provide unique and particularly important identification of astronomical cosmic ray2rigins, as well as examination of cosmogenic neutrino production from extragalactic hadronpropagation.A final answer to the mystery of cosmic ray origins requires the observation of high-energy neutrinos. High-energy neutrinos can be the most accurate probe into hidden sectorof traditional astronomy or physics, such as dark matter. However, their extremely low fluxand interaction cross section make their detection extraordinarily difficult. To discover clear,indisputable evidence of cosmic high-energy neutrinos, a next generation detector shouldhave orders-of-magnitude larger sensitivity. To probe into the association with known orunknown astronomical objects as a telescope for very high-energy (VHE) neutrinos with theenergies above 1 PeV, it should also have enough pointing accuracy, without suffering frombackground events of atmospheric secondaries.The Earth-skimming tau neutrino technique enjoys a large target mass by detecting ex-tensive air-showers produced by tau lepton decays in the atmosphere. The tau leptons,produced by VHE tau neutrinos that interact with the Earth matter they traverse, emergeout of a mountain or the ground facing the detector. This method has detection sensitivityin the PeV-EeV region, and can be used to search for neutrinos originating from hadronacceleration in astronomical objects. Additional advantages are perfect shielding of cos-mic ray secondaries, precise arrival direction determination, and negligible background fromatmospheric neutrinos.The All-sky Survey High Resolution Air-shower detector Phase I (Ashra-1) is an optical-telescope based detector system [6] optimized to detect VHE particles aiming for “multi-particle astronomy” [7, 8]. It is distinguished by two features: (1) an ultra wide opticalsystem in which 42-degree FOV (field of view) is demagnified to 1-inch diameter phospherscreen on an output window by using photon and electron optics [9]; (2) high resolutionimaging system with a trigger. Ashra-1 combines these unique features, resulting in verycost-effective pixels compared to conventional photomultiplier arrays at the focal surface ofan optical telescope (Fig. 1). Ashra-1 can observe the whole sky with a few arc minutesresolution, with 12 detector units pointing at different directions, where a detector unitconsists of a few Light Collectors (LC) pointing at the same direction.The Ashra-1 detector system is designed so that the focal image is split into trigger/imagecapture devices after amplification. This feature enables one to simultaneously access 3 kindsof phenomena that have different time scales, i.e., Cherenkov emission (ns), fluorescence ( µ s),and starlight (s), without sacrificing the signal to noise ratio. By fully utilizing these distinctfeatures, Ashra aims to undertake full-fledged astronomical observation using VHE particles,commencing with the first detection of VHE neutrinos using Earth and mountain as target[10]. It can also be used to optically observe transient objects like GRBs, as it monitors thewhole sky simultaneously [11, 12]. The principal demonstration phase, Ashra-1, has beenrunning at the Mauna Loa site at 3300 m above sea level on Hawaii Island since 2008. Thedeployed main and sub stations at the Mauna Loa site are shown in Fig. 2 Ashra-1 succeededin the first search for PeV-EeV tau neutrinos originating from a GRB in the commissioningrun [10], demonstrating the great sensitivity around 100 PeV with the earth-skimming ν τ technique. Ashra-1 has achieved the best instantaneous sensitivity in the energy regionaround 100 PeV since January 2012 after trigger upgrade.The NuTel [13] project, conceived and started in 2001, was an effort concurrent with thedevelopment of Ashra-1. It purposed the fast construction of a limited neutrino telescopefor detecting ν τ -originated air-showers in the energy range of 1 to 1000 PeV, with possiblesources such as AGNs, GRBs, the Galactic Center (GC), etc. It took in the possibility ofa base up the smaller Mt. Hualalai, which provides a wide baseline view of Mauna Loa.A multi-anode PMT-based readout electronics, aimed for Cherenkov light in the UV, wasquickly built, but the project got cut since 2004. The group still built two 2m telescopes,3igure 1: An Ashra-1 lightcollector toward Mauna Kea. Figure 2: The Ashra-1 main and sub stations at theMauna Loa site.and conducted a mountain test up 2200 m in Taiwan. An issue was the estimated event rateof ∼ . The key technical feature of the Ashra-1 detector rests on the use of electrostatic lenses,rather than optical lens systems, to generate convergent beams. This enables us to realise ahigh resolution over a wide field of view. This electron optics requires: • wide angle high precision optics [14]; a Schmidt type optical system with modifiedBaker-Nunn optics allows a compromise between wide 42 ◦ field of view and 1 arc minresolution on the focal sphere of the light collector (Fig. 1), with pupil diameter of 1 m; • photoelectric lens imaging-tube [9]; in addition to the optical system, the world’s largestimaging-tube uses electrostatic lens to generate convergent beams from photo cathodeof 20 inch diameter to output phosphorus window of 1 inch diameter, enabling a verylow cost and high performance image sensor that provides high resolution over a wideFOV; and • image pipeline [15]; the image transportation from imaging-tube (image intensifier) toa trigger device and image sensor of fine pixels (CCD+CMOS) with high gain andresolution, enables very fine images with parallel self-trigger systems that trigger foroptical flash, atmospheric Cherenkov and fluorescence light separately.Based on these achievements from Ashra-1, we start to form a new collaboration forrealizing the next generation large Neutrino Telescope Array (NTA). The conceptual layoutfor the NTA observatory considers three site stations for a 25 km-side triangle, watchingthe total air mass surrounded by the mountains of Mauna Loa, Mauna Kea, and Hualalai(Fig. 3). A single site station at the center of the triangle has half-sky coverage. This config-uration allows for tremendous instantaneous sensitivity (equivalent to >
100 giga ton water),with Cherenkov-fluorescence stereoscopic observation for PeV-EeV neutrinos in essentiallybackground-free conditions. With the demonstrated fine imaging of Earth-skimming taushowers and the significant improved detection solid angle (30 ◦ × ◦ ) for incoming tauneutrinos, we aim for clear discovery and identification of astronomical tau neutrino sources.Also interesting is the unique capability of cross observation between optical flashes, TeV-PeV γ rays, and PeV-EeV ν τ s, once one or more of these three kinds of self-triggers areobserved and associated with an astronomical object.4igure 3: Layout of the NTA Ob-servatory. Shaded region consists ofthree semicircles centered at Site1-3. Figure 4: NTA detector unit of four lightcollectors of the same type.From the current baseline design of the NTA detector system, each site has a group ofdetector units (DUs) for individual FOVs (Fig. 4). Each DU is composed of four LCs respon-sible for the same FOV. The LC has the design similar with Ashra-1 but with dimensionsscaled up by 1.5 to gather more light. Each LC is instrumented with a pupil lens, sevensegmented mirrors, a set of photoelectric imaging tube and image pipeline with a CMOSsensor. The DU has a unified trigger system which determine if image light gathered fromfour LCs through fiber-optic bundle transmission systems has enough confidence. Detaileddesign studies for the NTA detector are currently underway. The main scientific goal of the NTA project is: the clear discovery and identification of non-thermal hadronic process in the Universe.
This has not been directly confirmed by any observation so far and can be achieved byobserving PeV-EeV neutrino emission as direct evidence and sensitive probe for collectiveprocesses that accelerate particles to energies many orders of magnitude beyond thermalenergies. Fig. 5 shows measured and expected neutrino fluxes, and sensitive energy regionof NTA with the sensitive energy range of PeV-EeV.The multimessenger connection among Cosmic Rays, photons and neutrinos of differentparticles is crucial for comprehensive and deeper understandings of the fundamental non-thermal astrophysical processes. Multimessenger is a theme to much of the recent literature e.g. [16] [17] [18]. The measured fluxes of extremely-high energy cosmic rays (EHECR)with energies above 10 eV (EeV) inspire an associated flux of PeV-EeV cosmic neutrinos,although the production mechanisms of EHECR are still unknown. PeV-EeV neutrinos arepredicted as a result of the decay of charged pions generated in interactions of EHECRswithin the source objects (astrophysical neutrinos) and in their propagation through back-ground photon fields (cosmogenic neutrinos) [19]. Cosmic rays up to and even beyond thePeV (“knee”) are of Galactic origin. Around EeV between 10 and 10 . eV (“ankle”),at maximum, known Galactic source candidates are generically considered running out of5 Neutrino Energy [eV] ] - M e V - s r - s - (cid:9)(cid:9) F l u x [ c m NTA
Cosmological Solar Super Nova Burst (1987A)Super Nova Relic Geophysical ReactorAtmosphericGRB AGNCosmogenic -6 -4 -2
10 1 -30 -25 -20 -15 -10 -5 Figure 5: Measured and expected neutrino fluxes, and sensitive energy region of NTA (greenband).power and extragalactic sources start dominating the spectrum. On the other hand, fromrecent calculations, the maximum energy of accelerated particles may reach 5 × eV forFe ions in Type IIb Supernova Remnants (SNRs) [20] [21]. Adding that, both the detailedcomposition and galactic-extragalactic transitions in the PeV-EeV region is still unclear andto be understood [22] [23] [24]. Simultaneous searches for PeV gamma rays and neutrinoswould be useful to distinguish between galactic and extragalactic sources of cosmic rays [25].If EHECRs are produced from Galactic point sources, then those point sources are also emit-ting PeV gamma rays. We note that the detection of galactocentric PeV gamma rays in thefuture would be a signature of the presence of EeV cosmic accelerators in the Milky Way[26]. EHECR sources in our galaxy will plausibly be investigated by the multimessengerapproaches with very-high energy gamma rays and neutrinos in the PeV-EeV region, sincethe galactic size is within the observable distances of gamma rays around PeV even after thepropagations in the background photons (Fig. 6).IceCube recently reported the observation of three neutrinos with energies at 1-2 PeV [2]and 26 additional events at lower energies [3], which are significantly inconsistent with thebackground. Many authors have discussed the origins and physics scenarios of the IceCubesignals in the context of multimessenger of EHECRs, VHE gammas, and VHE neutrinosamong galactic and extragalactic sources, e.g. [17] [27].The following potential candidates are considered as search objectives with the NTAsurvey. 6igure 6: Particle energies and observable distances through the interactions with back-ground photons with distance regions for sources (colored boxes), and the highest observedgamma and proton energies (dashed lines). • Galactic Sources:The galactic supernova remnants (SNRs) are widely believed to be the dominant sourcefor the cosmic rays (CRs) at energies below the knee around PeV, most probablythrough the diffusive shock acceleration mechanism [28]. Recent calculations showthat Super Nova Remnant (SNR) acceleration in our galaxy can describe the wholeenergy spectrum of observed cosmic rays for the region from TeV up to the ankle,using different types of SNs and transition of composition in the galaxy [20]. GalacticGRBs, which are beamed away from Earth, can be the main source of Galactic cosmicrays at all energies [29]. From the observational point of view, Imaging Air CherenkovTelescopes (IACTs) have detected more than hundreds of TeV γ -ray sources, includingabout 30 SNRs [30]. There are three classes of such objects: shell-type supernovaremnants, pulsar wind nebulae, and binary systems. The expected neutrino fluxesfrom these sources and diffuse emission from cosmic ray interaction are calculated [31]. Shell-type Supernova remnants have long been considered as the likely accelerationsite for the bulk of the galactic cosmic rays. The morphology of γ -ray emission fromRXJ1713.7-3946 was studied [32]. A TeV γ -ray image of the SNR demonstrates thatVHE particles are accelerated at the spatially resolved remnant, which has a shellmorphology similar to that seen in X-rays. The energy spectrum indicates efficientacceleration of charged particles to energies beyond 100 TeV, consistent with currentideas of particle acceleration in young SNR shocks. Spatial correlations of the γ -rayemission with available target material seem to be present for the SNRs W28, IC443,RCW86 and RX J0852.0-4622 supernova in IACT data. The observations of γ -raysexceeding 10 TeV in the spectrum of the RX J0852.0-4622 supernova [33] has alsostrengthened the hypothesis that the hadronic acceleration is the process needed toexplain the hard and intense TeV γ -ray spectrum. Such a correlation is also seen7n the region of the Galactic Centre (GC), where the acceleration site of the cosmicrays is not clear [34]. However, the directional distribution of the 21 cascade eventssuggests weakly significant excess (“hotspot”) with a trial-corrected significance of 8 %[3]. Possible contributions from galactic neutrino sources like SNRs are consistentwith the present diffuse γ -ray limits [35]. If the neutrino spectrum is dominated bygalactic sources, the lack of observed CR anisotropy requires a soft neutrino spectrumwith index ∼ pp ) origin scenario [27]. The required index isconsistent with a spectral index 2.2 of a point-like γ -ray source at the Galactic center,which was reported by H.E.S.S. [36]. The possibility of discrimination between pp and pγ source models by combining the measured neutrino and γ -ray fluxes, will be oneexample for the multimessenger approach [17]. Pulsar Wind Nebulae (PWNe) are some of the brightest TeV γ -ray sources. The centralpulsar emits material into the nebulae such as the powerful Crab and Vela pulsars [37].A significant fraction of nuclei is suggested to exist in pulsar winds [38]. The decay ofpions produced in the interaction of these nuclei can dominate the TeV γ -ray emission,which suggests significant production of neutrinos should occur [39]. These nuclei andsignificant production of neutrinos may occur e.g. [39]. Pulsars could also be a strongsource of very-high energy neutrinos [40], although there is a pessimistic estimate ofthe fluxes [41]. Binary systems of a compact object and a massive star are well established galacticTeV γ -ray sources, which are classified into binary PWN or microquasars. In thebinary pulsar scenario, the spin-down of the neutron star is the energy source. In themicroquasar scenario, accretion is the power-source, and particle acceleration occursin relativistic jets produced close to the compact object (black hole or neutron star).The PSR 1259-63 system with 3.4-year period and the Be-star SS 2883 belong to theclass of binary PWN. LS 5039 and LSI +61 303, are the remaining well establishedsystems and expected as strong neutrino sources [42], of which acceleration site hasnot been revealed yet. Cyg X-1 is expected as the best γ -ray microquasar candidate,which hosts a black hole [43]. Undetected bright hard-spectrum sources beyond ∼ Our Galactic Center (GC) has also been proposed as neutrino sources. An intense dif-fuse emission of γ -rays with higher energies has been observed which likely implies thepresence of a source of cosmic ray protons and thus of neutrinos [34]. The GC regionis of particular interest because it is in the good sky view of NTA located on HawaiiIsland in the northern hemisphere. A general scenario of Galactic &
10 PeV cosmic-rayinteractions to produce PeV-EeV events [25], and plausible spectra of neutrino eventsas originating from Galactic cosmic rays [45], has been considered as well. IceCubehas announced detection of 26 neutrino events with energies in the ∼ ∼ • Extragalactic Sources:As the extragalactic candidates for PeV-EeV neutrino emission, Gamma Ray Bursts(GRB), Active Galactic Nuclei (AGN) and galaxy clusters are well motivated. PeV-EeV neutrinos are also directly linked with the physics of proton acceleration to ex-tremely high energy cosmic rays (EHECR) above EeV at cosmic ray origin objects.Recent measurements of the composition of EHECRs by the Pierre Auger Observatory(Auger) have suggested that the mean nuclear mass may increase with energy between2 EeV and 35 EeV [5].
Gamma Ray Bursts (GRBs) eject the most energetic outflows in the observed Uni-verse, with jets of material expanding relativistically into the surrounding interstellarmatter with a Lorentz factor Γ of 100 or more. Energy dissipation processes involvingnonthermal interactions between particles are thought to play an important role inGRBs, but remain observationally unresolved. The detection of PeV–EeV neutrinos( ν s) from a GRB would provide direct evidence for the acceleration of hadrons into theEeV range, and of photopion interactions in the GRB. The GRB standard model [48],which is based on internal/external shock acceleration, has been used to describe thegeneral features of a GRB and the observed multi-wavelength afterglow. However, thestandard model cannot reproduce well the recent observational results [49]. The earlyX-ray afterglows detected by Swift exhibit a canonical behavior of steep-flat-steep intheir light curve [50]. In some of GRBs, precursor activities were observed [51]. Insome cases, the precursor preceded the main burst by several hundred seconds withsignificant energy emission. To better understand the ambiguous mechanisms of GRBs,observational probes of the optically thick region of the electromagnetic components,as well as hadron acceleration processes throughout the precursor, prompt, and after-glow phases are required. VHE ν s can be used as direct observational probes, whichare effective even in optically thick regions. The discovery of nearby low-luminosity(LL) GRB060218 suggests a much higher local event rate of LL-GRBs [52], which NTAcan easily search for. NTA can check the ratio between the observed neutrino eventrates from the Earth and the sky in the field of view of the detector, which means themeasurement of the diffuse neutrino background from all GRBs with less systematicerror. Active Galactic Nuclei consist of super-massive black holes with 10 ∼ solar massesin their centre. The black hole radiates huge amount of energy typically of the order of10 erg/s, which is transfered from gravitational energy after it accretes matter. Theenergy is expected to induce acceleration of particles. A special class of AGN, Blazers,has jet aligned closely to the line of sight, which can be strong gamma-ray sources.Many sources are reported at GeV and TeV energies by Fermi LAT [53]. Gammaray emission from blazars is often highly variable, e.g. PKS 2155-304, with the mostextreme variation observed an increase by two orders of magnitude within one hour[54]. We should observe the neutrino fluxes from such a source significantly within ashort time. An observation of outburst from the blazar 1ES 1959+650 [55] suggestsanother type of neutrino sources, which is TeV emission without being accompaniedby X-ray emission as synchrotron self-Compton (SSC) models typically predict. Ahadronic model does not require TeV emission accompanied by X-ray emission. Theobserved flares are encouraging sites to search for high energy neutrino emission. Starburst galaxies have unusually high rates of large-scale star formation processes. A9alactic-scale wind is driven by the collective effect of supernova explosions and massivestars at the central regions of the starburst galaxies. IACTs have detected the gammaray flux at several hundred GeV from the starburst galaxies NGC253 and M82 [56] [57].They suggests cosmic ray densities much higher than typical case expected in our ownGalaxy by two to three orders of magnitude. The diffuse neutrino flux from all starburstgalaxies is expected detectable with current detectors [1].
Cosmogenic Neutrinos are the secondary particles of the Greisen-Zatsepin-Kuzmin(GZK) process from the interaction of the highest energy cosmic rays with the cos-mic microwave background [19, 58, 59]. Various cosmogenic neutrino models (for ex-ample [60]) which assume primary cosmic ray protons predict neutrino fluxes. Theyrequire 4 π solid angle averaged neutrino effective area A ν to be more than 10 − km at 100 PeV to detect several cosmogenic neutrinos every year in case of full duty cycle.NTA satisfies this requirement well even assuming the duty of 10%.The predicted flux has large uncertainties due to dependence on source spectrum andon spatial distribution and cosmological evolution of the sources [22]. If EHECRs areheavy nuclei like irons, the yield of the cosmogenic neutrino is strongly suppressed [61].Therefore NTA has sufficient sensitivity for cosmogenic neutrinos to directly test thehypothesis of the observed highest energy cut-off of the cosmic ray spectrum due to asuppression induced by the GZK propagation of pure protons taking into account ofthe uncertainty of flux prediction even in the case of null result. • Dark Matter and New Particles
Weakly Interacting Massive Particles (WIMPs) are favoured dark matter candidates,which are preferentially discussed in the minimal supersymmetric standard model(MSSM) framework [62]. Indirect WIMP detection use secondary particles such as γ s, ν s, weak bosons, tau pairs and so on from annihilations. Direct WIMP detectionuses recoil nuclei from elastic WIMP-nucleus scattering. There is some complemen-tarity between direct and indirect searches for dark matter, given the astrophysicalassumptions inherent to the calculations. Both methods are sensitive to opposite ex-tremes of the velocity distribution of dark matter particles in the Galaxy (low-velocityparticles are captured more efficiently in the Sun, high-velocity particles leave clearersignals in direct detection experiments), as well as presenting different sensitivity tothe structure of the dark matter halo (a local void or clump can deplete or enhancethe possibilities for direct detection). IceCube has evaluated these data for evidence ofdark matter annihilations in the Sun, in the Galactic Center, and in the Galactic Halo,searching for an excess neutrino flux over the expected atmospheric neutrino back-ground, which provides the results of dark matter searches for WIMPs, Kaluza-Kleinmodes and super heavy candidates (Simpzillas), using the 79-string configurations ofIceCube [63]. Given that the Sun is essentially a proton target and that the muon fluxat the detector can be related to the capture rate of neutralinos in the Sun, the IceCubelimits on the spin-dependent neutralino-proton cross section are currently well belowthe reach of direct search experiments, proving that neutrino telescopes are competitivein this respect. The simple assumption that dark matter is a thermal relic limits themaximum mass of the dark matter particle, which turns out to be a few hundred TeVfor a thermal WIMP, the so called unitarity constraint. However, dark particles mighthave never experienced local chemical equilibrium during the evolution of the Universe,and their mass may be in the range much larger than the mass of thermal WIMPs,which have been called WIMPZILLAs [64, 65]. NTA can perform the most sensitiveindirect search for WIMPZILLAs, with much better effective detection area for tauneutrinos from annihilation in the Sun, especially above ∼
10 PeV, the complementary10ensitive energy region for IceCube.
Super-heavy particles (M & GeV) produced during inflation may be the dark mat-ter, independent of their interaction strength. Most popular ones are SIMPZILLAs,magnetic monopoles, supersymmetric Q-balls and nuclearites.
Strongly interacting super-heavy particles (SIMPZILLAs) will be captured by the Sun,and their annihilation in the center of the Sun will produce a flux of energetic tauneutrinos that should be detectable by neutrino telescopes [66].
Magnetic monopoles turn out to be consequence of most variants of Grand UnifiedTheories [67]. The electromagnetic energy losses of monopoles in the atmosphere,as well as neutrinos produced from monopole-antimonopole annihilations in the Sunand Earth, induce clear signatures in optical (Cherenkov and fluorescence) air-showerdetectors like NTA [68].
Nuclearites (strange quark matter or strangelets) are hypothetical aggregates of u, dand s quarks, combined with electrons to adjust electric neutrality. Nuclearites, likemeteors, produce visible light as they traverse the atmosphere. Their luminosity as afunction of their mass is L = 1 . × − ( M/ µ g) watt [69]. For example, the apparentvisual magnitude of a 20 g nuclearite at a height of 10 km is − .
4, equal to that of thebrightest star, Sirius. Atmospheric nuclearites at galactic velocities ( v ∼
250 km/s) caneasily be distinguished from ordinary meteors bounded to the Solar System, movingno faster than 72 km/s. It could be identified with clear evidence with the wide FOVhigh resolution optical detector of NTA.
Q-balls are hypothetical coherent states of quarks, sleptons and Higgs fields [70]. Neu-tral Q-ball (Supersymmetric Electrically Neutral Solitons, SENS) could catalyse protondecay along their path, similar to GUT monopoles. Electrically charged Q-ball (Su-persymmetric Electrically Charged Solitons, SECS) would produce light in a similarway as nuclearites.
To detect VHE neutrinos, a large target volume is required in order to compensate for thevery small neutrino-nucleus cross section. On that basis, the secondary particles producedby the first neutrino interaction must be detected in one way or another. The detectionmethod using water and ice as a target detects Cherenkov light from secondary muons,taking advantage of the fact that ice and water are optically transmissive to some extent.This method can be categorized as the method in which the target and detection volumesare water and near-by rock surrounded by the water tank. On the other hand, the detectionmethod using air-showers aims at the detection of higher-energy neutrinos. This methodenables us to achieve a huge detection volume as the atmosphere has very high transmittance.However, it is difficult to obtain a larger target mass due to low atmospheric density. Thedetection method called Earth-skimming ν τ technique [71, 72, 73, 74, 75] can realize a hugetarget mass and detection volume at the same time, by dividing the target and detectionvolume utilizing the interaction process of ν τ . The detection method is described as follows(see Fig. 7). The VHE ν τ interacts in the Earth or mountain and produces tau lepton ( τ ). τ penetrates the Earth and/or mountain and appears in the atmosphere. Subsequently, itdecays and produces an air-shower. Cherenkov photons from the air-shower are detected.Owing to the separation of the first interaction where ν τ produces τ and the τ decay thatgenerates the air-shower, air-shower observation becomes possible while preserving the huge11igure 7: Schematic view of Cherenkov τ shower ES method. Mauna Kea is used as thetarget mass for neutrino charged current interaction. The produced air-shower is observedfrom Mauna Loa. In addition to the fact that the mountain can be viewed with large solidangle from the observatory, the distance of about 30 km from the observatory to the MaunaKea surface is appropriate for the air-shower development, resulting in the huge advantageof the Ashra-1 observatory.target mass required for the first interaction. “Cherenkov τ shower ES method” is definedas the detection method which detects Cherenkov photons from tau shower appearing fromthe Earth or the mountain fully utilizing this feature. We note, for example, that MaunaKea is over 3,200 km in volume and 9.3 tera tons in mass [76]. This section describes deflection of Cherenkov τ shower compared to the arrival directionof parent ν τ , in order to estimate the ability to trace back to the accelerator based on thedirection of the detected air-shower. We evaluate the deflection of the propagating particlein each step of neutrino charged current interaction, τ propagation in the Earth, tau decay,and production of extensive air-shower. We use PYTHIA [77] to evaluate neutrino chargedcurrent interaction. Since P τ < M W where P τ denotes the transverse momentum of aproduced τ and M W denotes the mass of the W boson, the deflection angle τ (∆ θ τ ) withrespect to the parent ν τ should be less than 0.3 arcmin for E τ > τ due to propagation inthe Earth. To estimate the energy loss of high energy leptons, the following parametrizationis generally adopted [79]: − (cid:28) dEdX (cid:29) = α + βE, where α denotes the nearly constant ionization loss, and β denotes the radiative energyloss due to Bremsstrahlung, pair production and photonuclear interaction. Since radiativeenergy loss is dominant for high energy τ s, these high energy processes must be included inthe “Physics List” of GEANT4. Thus, we apply the following processes originally definedfor muons to τ s, and estimated the deflection after propagating through 10 km of rock. • G4MuBremsstrahlung: Bremsstrahlung • G4MuPairProduction: e + e − pair production • G4MuNuclearInteraction: Photonuclear Interaction We modified the original G4MuPairProduction so that momentum conservation includes the producedparticles, resulting in the inclusion of deflection. au Lepton Deflection Angle [deg] -7 -6 -5 -4 -3 -2 -1
10 1 10 N u m b e r o f E ve n t s From right to left= 1 PeV τ E =10 PeV τ E PeV =10 τ E PeV =10 τ E GEANT4 (All processes except photonuclear)
Tau Lepton Deflection Angle [deg] -7 -6 -5 -4 -3 -2 -1
10 1 10 N u m b e r o f E ve n t s From right to left= 1 PeV τ E =10 PeV τ E PeV =10 τ E PeV =10 τ E ALLM reproduced (Photonuclear only)
Figure 8: The simulation results of deflection angle of τ s after propagating through 10 km ofrock: ( Left ) the GEANT4 result including all high energy processes except for photonuclearinteraction; (
Right ) the result of photonuclear interaction from custom simulation. Note thatthe decay of τ was switched off for the above simulations. The hatched histograms indicatethat the τ range is less than 10 km.To validate our GEANT4 simulation, we compare the energy dependence of β for Bremsstrahlung,pair production, and photonuclear interaction to Ref. [79]. The β energy dependence agreeswell for the former two processes, but we find that GEANT4 produces smaller values forphotonuclear interaction at higher energy, and that the difference is a factor of 3 at 10 GeV.We write accordingly a toy Monte Carlo simulation for photonuclear interaction using theformalism of Refs. [80, 81], reproducing the energy dependence of β within ±
30 % accuracy.Fig. 8 shows the simulation results for τ deflection after propagating through 10 km ofrock. The left panel shows the GEANT4 result including all high energy processes exceptphotonuclear interaction, while right panel shows our “homemade” simulation result for thelatter. These results indicate that photonuclear interaction becomes dominant for deflectionat 1 PeV and higher. Note that the decay of τ was switched off for the above simulations,and the hatched histograms indicate that the τ range is less than 10 km. For example, the τ range is 4.9 km at 100 PeV. We conclude that the deflection angle of τ s with energy greaterthan 1 PeV is much less than 1 arcmin.Next, the deflection due to τ decay is estimated by using the output of TAUOLA [82],taking into account τ polarization. From the mass m τ , the deflection angle must be lessthan 1 arcmin if the energy of the secondary particle is higher than 13 TeV. Using TAUOLAoutput, it was shown that the probability to have deflection greater than 1 arcmin is quitesmall from the decay of PeV τ s. We conclude that the deflection angle between decayparticles which produce the air-shower and parent ν τ is less than 1 arcmin.Finally, we evaluate the direction of the hadron air-shower using CORSIKA. At theshower maximum, we compare the direction of the parent particle (charged pion) to that ofelectrons and positrons, the dominant producers of Cherenkov photons. We find that theangle between the average direction of electrons and positrons and parent particle of theair-shower is coincident within 0 . ◦ at 1 PeV.In conclusion, we find that the arrival direction of PeV ν τ s is preserved within 0 . ◦ ,including the hadron air-shower generation. The accurate reconstruction of arrival directionby means of fine imaging will be a very powerful technique in the determination of the pointsources of PeV ν τ s. 13 The NTA Detector
The NTA observatory will consist of four sites, Site0, Site1, Site2, Site3, as shown in Fig. 3.The conceptual layout for the NTA observatory considers three site stations (Site1, Site2, andSite3) forming a 25 km-side triangle watching the total air mass surrounded by the mountainsof Mauna Loa, Mauna Kea, and Hualalai. A single site station, Site0, at the center of thetriangle has half-sky (extendable to full-sky) coverage. Each site has a centralized group ofDetector Units (DU). One detector unit (Fig. 4) has a few Light Collecting systems (LC)with segmented mirrors. The features of the system were studied with the Ashra-1 stationsite constructed on Mauna Loa (3300 m a.s.l.).In order to investigate the performance of the NTA detector, we shall use the followingsetup conditions of the assumed locations of observational sites of the NTA system on HawaiiIsland.1. The left side of Fig. 9 shows the layout of the NTA site locations. The local Cartesiancoordinate system is defined with the origin at Site0, as denoted by “0” in Fig. 9,the positive z-axis points to the zenith, and the positive y-axis points north. The x-ycoordinates of the site locations are from the projected x-y plane, with the z-coordinatesfrom the corresponding height of the topography data of Hawaii Island.2. The three observational sites, each located at the vertices of a triangle of equal 25 kmside length, are: Site1 at Mauna Loa Ashra-1 location, Site2 on the slope of MaunaKea, and Site3 on the slope of Hualalai.3. The central observational site, Site0, is at the center of gravity of the above three sitelocations. Site1–Site3 are equidistant 14.4 km from Site0.4. The location of Site1 and Site2 are set at the Ashra-1 Mauna Loa Observation Site(ML-OS), and at 25 km distant from ML-OS in the direction of Kilohana Girl ScoutCamp, respectively.5. After above settings, the locations of remaining two sites are automatically fixed.The right part of Fig. 9 shows the layout of the four sites, superimposed on the topographymap image of Hawaii, to be used as settings in the simulation program. Table 1 shows thex-y-z coordinates of the site locations and the detection FOV coverage, as determined fromthe above description. For the simulation study, given in the next section, we assume thateach LC has the total FOV of 32 ◦ × ◦ , trigger pixel FOV of 0.5 ◦ × ◦ , and image sensorpixel FOV of 0.125 ◦ × ◦ . The Site0 system consists of 12 LCs in the lower elevationangle regions, which together cover the half-sky solid angle that is π sr. The other sites haveonly 6 LCs in the lower elevation angle region, to cover the FOV of the half-sky solid anglewhich is π /2 sr. The bottom edge of the lower elevation angle region is defined to be − ◦ inelevation angle ( 9 ◦ below the horizon ).Fig. 10 shows the simulated panoramic views in altitude and azimuthal directions fromthe four NTA sites, with colours indicating the distance from the corresponding site. We investigate NTA performance with site location setup of previous Section.14igure 9: (left) The x-y coordinates of the four NTA site locations, with Site0 defined asthe origin; (right) the Hawaii Island topography map superimposed with the four NTA sitelocations. Site ID Location X [km] Y [km] Z [km] FOV [sr]Site0 Center 0.000 0.00 2.03 π Site1 Mauna Loa 9.91 − π /2Site2 Mauna Kea 4.12 13.82 1.70 π /2Site3 Hualalai − − π /2Table 1: The x-y-z coordinates and detection FOV coverage of the four NTA sites, whichare used in the simulation program. The location of Site0 is defined as the origin of thecoordinate system. For simulating the propagation of ν τ s and τ s in the Earth, we performed the followingprocedure and treatment.1. The density profile of the Earth is chosen according to the Preliminary Earth Model[83, 84]. It depends strongly on the depth in the Earth as shown in Fig. 11. Wemodified the profile just for the density of the ground surface in the radius range of r > into 2.9 g/cm , which is suitable for Basalt rock as themost common type of rock in the Earth’s crust and most of the ocean floor around theIsland of Hawaii.2. We took into account both charged current interaction (CC) and neutral current in-teraction (NC) of ν τ s and τ s in the Earth. The energy dependence of the inelasticityparameter y for CC and NC based on the CTEQ4 parametrization are shown in [83],and no difference is seen between those for CC and NC.3. We implement ν τ → τ → ν τ regeneration in the simulation. Because of the shortlifetime of the tau, regeneration can be an important effect as the ν τ passes through asignificant column depth through the Earth [79].4. In simulating τ propagation, the current position of τ is evaluated at every step of theenergy loss rate of 10%, unless the τ comes out of the Earth or decays.15igure 10: Panoramic views simulating the topographical image from NTA Site0 (top left),Site1(top right), Site2 (bottom left), Site3 (bottom right). Nearby obstacles with distanceless than 3 km are neglected.Figure 11: PREM (Preliminary Earth Model) density distribution of the Earth [83]5. In the lab frame, ν τ from τ decay on average carries a fraction 0.4 of the τ energy [85].We used this constant average value of 0.4 as the energy of ν τ from τ decay, withouttaking into account the energy distribution. The error from this approximation canbe neglected for the moment, because of the good agreement between results withour simulation and with ANIS (All Neutrino Interaction Simulation) [86], as shown inFig. 12.Fig. 12 shows the distribution in the plane of E τ and dip angle (minus elevation angle; − θ elev ) in the case of primary ν τ energy of 10 eV.The left side shows the case of neglecting any effect from ν τ → τ → ν τ regeneration orNC interaction of ν τ in the Earth, while the right side shows the case of taking into accountboth effects from ν τ → τ → ν τ regeneration and NC interaction of ν τ in the Earth. Eachbin content in these figures is given by: d N τ dE τ d Ω ( E τ , θ ) × d log E τ · πdθ. τ and dip angle (minus elevation angle; − θ elev ) plane in thecase of primary ν τ energy of 10 eV.Fig. 12 (bottom) shows the result using ANIS [86], which is approved for use for AMANDAand IceCube, and acknowledged well for detailed interactions, decays, and propagation of ν τ s and τ s. In general, the results with our simulation and ANIS agrees reasonably well.From detailed comparison, we should consider systematic errors of ∼
12% on produced τ fluxin the Earth in using our simulation program. Before evaluation of the performance of NTA using our simulation program, we summarizeour settings at the following three steps.1. Simulation for the Earth-skimming ν τ → τ • ν τ (CTEQ4) [87] • Inelasticity parameter [83] • Energy loss in the Earth [88, 79]2. Air-shower simulation: τ → Cherenkov and fluorescence light • τ Decay (approximated; [89]) • Air-shower generation (Gaisser-Hillas + NKG) [89]17 τ Energy CTEQ4 σ CC L CC θ celev eV 6.342 × − cm × g/cm − ◦ eV 1.749 × − cm × g/cm − ◦ eV 4.436 × − cm × g/cm − ◦ eV 1.049 × − cm × g/cm − ◦ eV 2.379 × − cm × g/cm − ◦ Table 2: Based on CTEQ4 [87], differential ν τ CC cross section ( σ CC ), corresponding inter-action length ( L CC ), and the critical angle ( θ celev ) such that the chord thickness at the criticalangle corresponds to L CC . For the Earth density profile, we refer to the parametrization ofPREM [86, 83].3. Detector simulation: • light collection and throughput of light • simplified triggering logic • Event reconstruction is not implemented yet.We assume the following input parameters.Light Collection Area: A = 7 .
07 m (equivalent with the effective pupil diameter of φ ǫ filt = 90%Quantum efficiency of photoelectric tube: ǫ QE = 24%LC FOV: 32 ◦ × ◦ Trigger pixel FOV: 0 . ◦ × . ◦ / trigger pixelExposure time in trigger pixel: t trigpix = 50 nsRequired trigger condition:To estimate the detection sensitivity of NTA the event candidates must satisfy thefollowing requirements: • total number of photoelectrons detected in one LC must satisfy: N LCpe > , • S/N estimated in the track-associated box of the width of 4 pixels and the lengthof 64 pixels, which includes the candidate event air-shower track, must satisfy:
S/N > , where the standard deviation of night sky background in the track-associated boxwith the exposure of 50 ns is estimated as σ ( N BGpe ) = 15 . . × photons / m / sr /µ s.A simulated Earth-skimming τ shower event with primary ν τ energy of E ν = 10 eVusing the above settings is shown in Fig. 13. The reconstructed air-shower axis with simplefits to the Cherenkov and fluorescence hit map images taken by the two sites reproduces theprimary ν τ arrival direction with an error of 0.08 ◦ .18igure 13: A simulated Earth-skimming τ shower event with primary E ν τ = 10 eV, whichhas both fluorescence image taken by Site0 and Cherenkov by Site1. (top) Global hit mapview in the NTA system; (bottom left) air-shower fluorescence image taken by Site0, and(bottom right) Cherenkov image from the same event taken by Site1. The trigger pixel andfine image FOVs are 0 . ◦ × . ◦ and 0 . ◦ × . ◦ , respectively. We estimate the effective detection area for ν τ fluence from a point source with our simulationprogram for Earth-skimming τ showers incorporating the appropriate Earth model [84], thetopography around the NTA observatory, the interaction and propagation process of ν τ and τ in the Earth [83, 86], the decay of τ and generation of air-shower, with parameter choicesas described before.We define the critical dip angle (minus critical elevation angle; − θ celev ) as the chordthickness at the dip angle − θ c elev that corresponds to the CC interaction length L CC ( E ν ),determined by the interaction cross-section σ CC ( E CC ) for a ν τ traveling with energy E ν .Table 2 shows differential cross sections of ν τ CC interaction σ CC based on CTEQ4 [87], thecorresponding interaction length L CC , and θ c elev for each E ν τ between 1 PeV and 10 EeV.Taking into account the critical dip angles for the energies of ν τ in Table 2, we estimatedthe effective detection areas for ν τ from a point source with azimuthal arrival directioncorresponding to that of the Mauna Loa summit ( φ ) and that of the Hualalai summit ( φ ),with respect to the central site of Site0, and dip angles of 2 . ◦ , − . ◦ , − . ◦ , − . ◦ , − . ◦ ,and − . ◦ , as shown in Fig. 15.Fig. 16 shows the differential sensitivities, as a function of E ν τ for a point source of ν τ ,19igure 14: Similar with 13 but in the case of the detection of stereoscopic fluorescencesignals.Figure 15: Estimated effective detection area simulated for ν τ from a point source withazimuthal arrival direction corresponding to (left) Mauna Loa ( φ ) and (right) Hualalai ( φ )with respect to the central site of Site0, and dip angles of 2.0 ◦ (black open circle), − ◦ (green star), − ◦ (blue filled box), − ◦ (red filled circle), − ◦ (yellow filled triangle),and − ◦ (black filled triangle). 20igure 16: Comparison of differential sensitivities as function of E ν τ for a pointsource of ν τ , calculated as the Feldman-Cousins 90% CL limit event number fornull expected events, using a light collector (LC) from Ashra-1 commissioning [10]and the NTA layout of LCs, in the cases of (left) θ elev = − ◦ , (right) θ elev =+2 ◦ (opencircle) , − ◦ (green) , − ◦ (blue) , − ◦ (yellow) , and − ◦ (black). The sensitivitiespublished from IceCube [91] and Pierre Auger Observatory [92] are shown, as well as theo-retical estimates used for the former (solid lines) and recalculated by H¨ummer et al. (dashedlines) [93] assuming the distance of z ∼ . E ν . The 2.3events is the Feldman-Cousins 90% CL limit event number for null expected events.Fig. 17 (top) shows the diffuse sensitivities for ν τ fluxes with NTA for 3 year observationtime. Both differential and integral sensitivities are given. The sensitivity limit is definedas 2 . E ν / ( S Ω eff · ∆ T ). Also shown is the comparison between NTA, Pierre Auger Obser-vatory [92] and IceCube [95], various model predictions for cosmogenic ν s, as well as otherexperiments of RICE, AMANDA, and ANITA are superimposed [96]. For the diffuse sen-sitivities of NTA, we assume the duty of 10% for 3 years observation ( ∼ . × s) andtrigger conditions as described before. From Fig. 15, NTA can survey ν τ point sources with best sensitivity in detection solid anglefor ν τ defined as − ◦ < θ elev < ◦ × ◦ < φ azi < ◦ in the primary ν τ energy region of10 PeV < E ν τ < z < . ∼
400 Mpc.With the observational conditions assumed as follows: • Solar elevation angle: < − ◦ • Lunar bright surface ratio: < . • Ideal weather efficiency: 100%the total duty is estimated to be 20.5%, which corresponds to maximum observation time of1800 hours per year.Fig. 18 shows the exposure map for the observation with NTA on Hawaii Island (the Site1position: 19 ◦ ′ ′′ N, 155 ◦ ′ ′′ W, 3294 m a.s.l.), with Mollweide projection in Galactic21
Neutrino Energy [GeV] ] - s r - s - F l u x [ G e V c m (cid:9) E PAO2012IceCube2012ANITA2010RICE2012WB LimitNTA 3yr
Ahlers2010 (p,best)Kotera2010 (p,FRII)Kotera2010 (Fe,SFR1) -9 -8 -7 -6 -5 -4 Figure 17: Diffuse sensitivities for ν τ fluxes with NTA for 3 years observation time. Bothdifferential sensitivity (curve) and integral sensitivity assuming the E − flux spectrum (hor-izontal line) are shown. The sensitivity limit is defined as 2 . E ν / ( S Ω eff · ∆ T ). Comparisonamong NTA, Pierre Auger Observatory [92] and IceCube [95], same as the top one but vari-ous model predictions for cosmogenic ν s, as well as other experiments of RICE, AMANDA,and ANITA are superimposed [96]. For NTA, the duty of 10% for 3 year observation ( ∼ × s) is assumed.Figure 18: Exposure map for observation with NTA on Hawaii Island (the Site1 position:19 ◦ ′ ′′ N, 155 ◦ ′ ′′ W, 3294 m a.s.l.), with Mollweide projection in Galactic (left) andEquatorial (right) coordinates. Maximum observation time is normalized to 1000 hours peryear (red), where NTA can detect with maximum efficiency (total duty of 11.4%).(left) and Equatorial (right) coordinates on the celestial sphere. The maximum observationtime is normalized to 1000 hours per year, as shown in red in the figure where NTA candetect with maximum efficiency, which means total duty of 11.4%, corresponding to abouthalf the above ideal case. The location of NTA on Hawaii Island allows us to enjoy a surveyof our galactic center for more than several hundred hours each year.22igure 19: (left) Modified layout of NTA sites, and (right) ratio of the two sets of effectivedetection area for ν τ s as a function of E ν τ , obtained with modified and regular layouts(Fig. 9). The location of Site0 is changed into the midpoint between Site2 and Site3. To check the effect of changing the site layout on the detection sensitivity of NTA, wechanged only the location of Site0 into the midpoint between Site2 and Site3, as shown inFig. 19 (left), and repeated the sequence of simulation for diffuse sources as before. Theright side of Fig. 19 shows the ratio of the two sets of effective detection area for ν τ s as afunction of E ν τ , which are obtained with modified layout and regular one. We do not seeany significant change over all energies in the PeV-EeV region. The layout can therefore beadapted to practical concerns. As discussed in Section 2.2, a Cherenkov τ shower with E > ν τ to within 0.1 ◦ accuracy. This means that the ability of the de-tector to reconstruct the arrival direction results in the precise identification of the VHEneutrino sources and leads to the realization of “multi-particle astronomy”. Owing to itshigh-resolution imaging capability, the NTA detector has a huge potential to improve thereconstruction of the arrival direction of ν τ -induced air-showers.NTA will observe quasi-horizontal air-showers with the primary energies between PeVand EeV. Note that the location of the shower maximum in the atmosphere, i.e., the depthof maximum development X max is expected to be roughly in the range of 500-1000 g/cm from the first interaction in the atmosphere for tau decays of different energies [97][98]. Thedepth range corresponds to the length of 6-12 km along the air-shower axis assuming theaveraged atmospheric pressure of 0.7 atm. The shower axis is defined as the extension ofthe initial momentum vector of the incident tau decay particle in the direction of cascadepropagation. The particle density in the shower core, i.e., in the central region, is veryhigh and drops rapidly with increasing the core distance from the shower axis. Electronlateral distribution functions (LDFs) of air-showers are well described by Nichimura-Kamata-Greisen (NKG) functions [99][100]. The LDFs measured by KASCADE in the energy range23rom 5 × eV up to 10 eV can be reproduced accurately for the fit parameter of electronlateral distributions r e ∼ s ∼ r M and age parameter s in the NKG functions [101]; the latter are used inthe Monte Carlo simulation here. In addition, Monte Carlo simulations of air-showers findsteeper LDFs than the NKG distribution in higher energies [98].Let us first we try a back-of-envelope estimate, for example, an image of 1000 photo-electrons detected by a light collector, which originated from an air-shower trajectory withthe Gaussian LDF of σ = 30 m the track length of 6 km as a typical event. From the image,we can determine the shower detector plane (SDP) with accuracy of ∼ ∼ r e , at the distance of 10 km as a typical case.The simulated air-shower event shown in Fig. 14 provides a more concrete and realisticexample. The ν τ is generated at the energy of 100 PeV with the elevation angle of − . ◦ ,which is a quasi-horizontally upward event. The converted τ emerges from the earth withenergy 73 PeV and decays into particles which induce an air-shower of 46 PeV. The fluores-cence light generated from the air-shower is triggered by light collectors deployed at Site0and Site2. The closest approaches or the impact parameters ( R P ’s) to the air-shower axisare 9.8 km and 12 km from the Site0 and Site2 respectively, and the distances to the X max from Site0 and Site2 are 16 km and 12 km respectively. Due to the design of the layout ofthe four sites, i.e. the 25 km-side triangle (Site1,2,3) with the centered site (Site0) as shownin Fig. 3, almost all air-shower axes which pass the air volume above the inner triangle areaof NTA have the closest approach of less than 12.5 km to one of the four sites. Therefore, thesimulated event shown in Fig. 14 is an example with relatively poor signal of all generatedevents of the same primary energies.Fig. 20 shows photo-electron images (red cross marks) on the plane of the altitude andazimuthal angles, which are triggered and taken from the same event shown in Fig. 14 bylight collectors at Site0 and Site2. The SDPs that correspond to the true air-shower axis(light green dotted line) and tilted by ± ◦ (light green solid line) are indicated on thesame figure. Note that the LDFs used in this analysis is the traditional NKG functions withthe fixed Moli`ere radius r M = 79 m and a variable age parameter s in the NKG form, notthe steeper LDF recently measured by KASCADE. Even with the traditional NKG formthe dominant components of photo-electrons of the image is between the boundaries of theSDPs tilted by ± ◦ with respect to that of the true air-shower axis. For the simplefit adopted here, to eliminate the statistical fluctuation in the altitude angle coordinate,profile histograms are used to display and fit the mean values of altitude angles in rebinnedazimuthal bins, which are also shown in Fig. 20. The mean values based on the profilehistograms reproduce the locations of true air-shower axis well. The image and trigger pixelresolutions assumed here are 0.125 ◦ and 0.5 ◦ respectively. With the combined images takenat Site0 and Site2, the SDP can be reproduced with fit error of 0.02 ◦ . Note that we canfurther improve the reconstruction accuracy here especiallyly in the higher energies by usingmore sophisticated likelihood fit analysis and expected shower developments for each showerenergy with the advantage of high resolution images, beyond our present simple treatmentof profiling the lateral distribution on the alt-azimuth coordinate plane. In the literature[102], we have described the detailed Monte Carlo study of the angular resolution only24igure 20: Simulated photo-electron images of air-shower development (red cross marks)in the alt-azimuth coordinates of light collectors installed at Site0 and Site2. The showerdetector plane (SDP) with the true air-shower axis (light green dotted line) and those tiltedby ± ◦ with respect to the SDP based on the true air-shower axis of which primary energyis 10 eV and altitude angle is − . ◦ .for monostatic observation of Cherenkov images of τ showers generated by earth-skimmingPeV-EeV ν τ events with the Ashra-1 detector system. In that work, we confirm that alikelihood analysis with fine images of shower core structures compared with Monte Carloexpectation of air-shower development improves significantly the arrival direction resolution.The sophisticated and completed likelihood analysis using Monte Carlo simulated air-showerdevelopments is beyond the scope of this LoI, before determining detailed detector designbut with only assumed baseline concepts.For pointing back to ν τ sources, we simultaneously fit observed data with the four param-eters of ( φ SDP , θ SDP ) of the normal unit vector of SDP and the impact parameter R P andthe arrival direction angle χ of the air-shower axis constrained on the SDP of the event. Thedetailed definitions of χ and R P on SDP can be seen in the Fly’s Eye detector paper [103].In the case of reconstruction of quasi-horizontal air-showers, however, purely geometricalbistatic method is not useful, since the opening angles between two of SDPs observed withlight collectors at different sites are nearly flat, i.e. 180 ◦ and strongly correlated with eachother. For the determination of ν τ source positions, particularly for χ and R P on SDP, weshould fully utilize the timing information recorded by the trigger pixel sensor in the NTAsystem as well as the image data. Our realistic baseline design of the trigger pixel sensorsystem has its pixel FOV of 0.5 ◦ × ◦ and the least time stamp resolution of each triggerpixel of 10 ns. An advantage of the baseline trigger design based on the developments ofAshra-1 is that we can optimize pixel and timing resolutions of imaging system and triggersystems independently. For each event, after determining these four parameters, they aretransformed into another set of parameters, i.e. the arrival direction of the air-shower axis( φ AS , θ AS ) and the position (X τ ,Y τ ,Z τ (X τ , Y τ )) where the τ emerges from the mountain.The Z coordinate of the emerging point Z τ (X τ , Y τ ) is obtained from a topographic map asa function of X and Y coordinates. For the aim of partially demonstrating the performanceof simultaneous fit of parameters needed to point back to sources of observed ν τ candidates,25e have prepared for Monte Carlo data simulating ν τ events with the energies of every halfdecade between 10 . eV and 10 eV with fixed altitude angle of − . ◦ and fixed azimuthalangle toward the peak of Mauna Kea.Fig. 21 shows the results of fitting Monte Carlo data. The blue filled square marks witherror bars show the total RMS resolution of the reconstructed τ arrival direction as functionof logarithmic energies of generated ν τ s. The red filled triangle marks with error bars showthe total RMS resolution of the polar angle component of the reconstructed τ arrival directionas function of logarithmic energies of generated ν τ s, which is important to eliminate cosmicray background events due to misreconstruction of the arrival directions as described in thenext subsection. Fig. 21 also shows the event rate of multistatic observation, that is ratio ofevents observed with DUs at two or more sites, which is another result from this Monte Carlosimulation study. Although the accuracy of the angular resolution increases, the multistaticrate is seen saturated above 10 eV, which indicates the limitation of this simple fit methodusing profiling LDS at each bin of longitudinal development along the air-shower axis.This simple method works well when the core structure is negligible. Although theNTA detectors with high resolution imagers resolve out the shower core structures for highenergy events, we treated the shower axis as a line without any lateral structure. At higherenergies, the effect of the shower core structure become significant. Again likelihood analysiswith fitting functions made of enough number of Monte Carlo events taking into accountthe shower lateral and longitudinal development more consistently, the resolution should berecovered particularly in the higher energy region. Even using the simple fit method forquasi-horizontal τ shower events with Monte Carlo events with the altitude angle of − . ◦ ,we have confirmed the pointing resolution of 0.1 ◦ -0.06 ◦ . For the altitude angle or inclinationof SDP, they can be determined within an error of 0.06-0.02 ◦ . Detailed and precise MonteCarlo studies of the NTA detector system will be performed elsewhere, at step of the detectordesign report, after the publication of this LoI.For the moment, we quote the estimate of the angular resolution to be less than 0 . ◦ asa fairly conservative estimate from simulated events of monostatic Cherenkov images in themost pessimistic case of reconstruction of images taken with the NTA detector system. In this subsection, we evaluate the background events due to air-showers. Background eventsdue to the detector itself are discussed in Ref. [10]. Air-shower background candidates arenormal cosmic rays, muons, muon neutrinos, τ s, and ν τ s. From simple flux calculations, it isshown that the neutrino components through mountain, prompt τ components and muonsare negligible [104, 105, 106, 107, 108]. Thus, we consider normal cosmic rays with largezenith angles, of which arrival directions are misreconstructed, as the dominant backgroundcontamination in this study. To evaluate the rate of the background contamination dueto the directional misreconstruction, we count the air-shower events which pass throughthe sky region above mountain edges or horizon within assumed gap angles. To simulatenormal cosmic ray air-showers we use CORSIKA in the same way as with the ν τ simulationwith a thinning parameter of 10 − as a result of confirming it to be acceptable. Assumingthe maximum weather efficiency of 100%, 1750 hr of observation time on Hawaii Island isexpected in one year. We use the trigger pixel threshold of 20 photoelectrons as a realisticsensitivity which is same as the Ashra-1 DU one. The results are shown in Table 3.From this Monte Carlo study, we estimate the background contamination rate for theobservation of NTA. We require that the arrival direction of the candidate event must emergefrom earth or mountain. Background events may pass through the requirement of the direc-tion of air-shower axis direction due to the misreconstruction of air-shower events. However,the background rate should be very low and almost free. From the geography of Hawaii Is-26 og E (eV)15.5 16 16.5 17 17.5 18 A ngu l a r R e s o l u t i on ( deg ) M u l t i s t a t i c R a t i o Figure 21: Angular resolutions of recon-structed arrival directions of τ from MaunaKea. The RMS resolutions of the recon-structed τ arrival direction (blue filled squaremarks with errors) and the polar angle com-ponents (red filled square marks with er-rors) improves with energy. Multistatic rate(shaded histogram), i.e. ratio of events ob-served from two or more sites saturates above10 eV. Gap Angle (deg)1 − − E v en t R a t e / B i n ( / y r / deg ) − − Cosmic Rays
Figure 22: Estimate for cosmic ray back-ground contaminations due to misreconstruc-tion of their arrival directions. True anglewith respect of the mountain edge (Hatchedhistogram) and observed angles (red filled his-togram) assuming the angular resolution of0.2 ◦ after smering out the true angular dis-tribution. Two probability density functionsfor signal τ s with the arrival directions of 0.0 ◦ and -0.3 ◦ . The integrated areas of these his-tograms are normalized to be the expectedevent rate/year/DU.Gap angle δθ ( ◦ ) 0.1 0.3 1.0 3.0CR rate (showers/yr/DU) 0.082 0.55 4.3 39Table 3: The annual rate of cosmic ray showers which pass through the sky regions withinassumed gap angles and detected with NTA with the trigger pixel threshold of 20 photoelec-trons.land and the layout of the NTA sites, this directional misreconstruction is potentially causedonly near the edge between earth/mountain and atmosphere. We make Monte Carlo esti-mates for ordinary cosmic ray air-showers of which axis pass through the gap spaces of thesolid angle just above mountain edge as described above.For this aim, we make a histogram of differential event rate evaluated from the resultlisted in Table 3 and smear out the distribution by the angular resolution to quantitativelyestimated the leakage into our τ shower candidate sample from the outside of the fiducialvolume of NTA. Fig. 22 shows the differential distribution of true air-shower axis directionsfor ordinary cosmic rays (black shaded histogram) and the expected observed air-showerdirections after smearing them by very conservative angular resolution of 0.2 ◦ (red pearskinfinish), which is derived after detailed Monte Carlo studies on the directional reconstruc-tion for Cherenkov air-showers in comparison of the reconstruction with fluorescence light27etection as shown in the previous section.As a result, the rates of cosmic rays which pass over the mountain edge (0 ◦ in Fig. 22)and constrained boundary of solid angle below the mountain edge by 0.3 ◦ are estimated tobe 0.08 events/year and 0.006 events/year for one DU with the FOV of 32 ◦ × ◦ . Fig. 22also shows the probability density function for one event just on the edge (pink curve) andthat on the fiducial limit 0.3 ◦ below the mountain edge, where all histograms are normalizedfor the integrated area to be annual event rate for each DU respectively.In the case of the fiducial solid angle restricted after cutting < . ◦ with respect to themountain edge, total CR contamination rate in the NTA detector system with 30 DUs isestimated to be ∼ ν τ sto be ∼
90 %. We can realize the almost background-free condition without sacrificing thesuperior sensitivity of NTA for detecting Earth-skimming ν τ s using good advantages of highresolution images and the pointing accuracy. At the present time, we are investigating various options for the site, organization, and thedesign of instruments. Also, we intend to invite other groups to either contribute directly tothis project, or to join us on the site with their complementary instruments. The resultingsynergy effects would benefit all parties, avoid unnecessary parallel technical developments,and lead to cost savings for the different projects. It is clear that collaboration forming iskey to success of the NTA scientific goal. The time frame for the proposed project is thusgiven both by considerations of budgetary and scientific aspects. In March 19–20, 2014,a preliminary workshop (VHEPA2014) was held at Kashiwa campus of the University ofTokyo to discuss the design of the project and plans with interested colleagues. In April 8-9,2015 a small workshop, successive to VHEPA2014, was held at National Taiwan University todiscuss the scientific goals and promotion of the project. We plan to have an informal meetingto discuss post-IceCube project new detector project at the 34th International Cosmic RayConference (ICRC) held from July 30 to August 6, 2015, in The Hague.In January, 2016 we plan to hold a workshop as VHEPA2016 at University of HawaiiManoa to discuss more detailed physics and NTA potential performance, funding processes,and make ready for a white paper of the project as basic documents to use for the fundingrequests in each country.We have already set up the International Executive Board (IEB) of NTA for decisionmaking and steering the collaboration since October 12, 2012. Some IEB members havealready submitted funding requests to exchange information, detector design, meetings, andconstruction of the detector. IEB selects the representative who becomes the spokespersonof the collaboration. Each country has a Local Institutional Board (LIB), which is composedof representatives from institutes in the country. LIB selects the representative who becomesa member of IEB. We will set up various Working Groups (WGs) as real working bodies.WG leaders are nominated by LIB and decided by IEB.Major decisions concerning the hardware implementation should be undertaken in 2015,toward the publication of Project Proposal. In the subsequent two years, components shouldbe developed and tested, and we should continue to request the Japanese government for con-struction funding. We can eliminate critical developments by using experience from Ashra-1and NuTel projects. We aim at installing the first detector site and starting commissioningoperation in early 2017, if we succeed in the primary Japanese funding in time. We plan tostart the operation using at least a part of Site0 and Site1 of proposed four sites by 2018,with the primary construction budget covering at least 1/4 of full operation cost. Once wesucceed in the primary funding request and start construction of the first sites, we start the28equest for matching funds to the governments of various collaborating countries. We aim atstarting construction of Site2 and Site3 with the matching funds from countries other thanJapan.Since major choices concerning implementation details are still open at this time, it doesnot seem appropriate to discuss a detailed cost breakdown. To provide a guideline, however,we have estimated in some detail the cost of one design variant, with major componentsof the detector either covered by offers from potential manufacturers (light collector mount,mirrors, trigger-readout sensors and electronics, and so on), or extrapolating from well-knowncosts of the Ashra-1 instruments. On this basis, we estimate the production cost per lightcollector in the 20M yen range, plus total R&D costs of about 100M to 150M yen. Onedetector unit (DU), as shown in Fig. 1, requires four light collectors and one trigger andreadout unit, so the rough cost estimate is 100M yen per DU. We plan to build at least 12,6, 6, and 6 DUs at Site0, Site1, Site2, and Site3, respectively, for the coverage of FOV asshown in Table 1, assuming the FOV for DU to be 32 ◦ × ◦ . 30 DUs are needed for coveringthe total FOV. We do not include infrastructure costs such as site access, site preparation,light collector shelters, networking, and so on. Roughly speaking, at least 100M yen per DUis needed from the experience of construction of Ashra-1 at Mauna Loa. The current crudeestimate is 5000M yen for the construction of NTA.The Ashra-1 collaboration has agreed to continue the observation at the Mauna Loa siteas well as explore the NTA system, at least by the time NTA starts the construction. In orderto explore the hardware and software components, the Ashra-1 experience is recognized asan important and useful demonstration of the challenging new detection techniques. Acknowledgment
We thank P. Binder and J. Goldman from University of Hawaii Hilo for their useful commentsand help. Many thanks to the Ashra-1 and NuTel collaborations for their cooperation.The Ashra Experiment is supported by the Coordination Fund for Promoting Science andTechnology and by a Grant-in-Aid for Scientific Research from the Ministry of Education,Culture, Sports, Science and Technology of Japan.
References [1] A. Loeb and E. Waxman.
Journal of Cosmology and Astroparticle Physics ,2006(05):003, 2006.[2] M. G. Aartsen et al.
Phys. Rev. Lett. , 111:021103, 2013.[3] IceCube Collaboration et al.
Science , 342(6161):1242856, 2013.[4] M G Aartsen, R Abbasi, M Ackermann, and J Adams. Probing the origin of cosmicrays with extremely high energy neutrinos using the IceCube Observatory.
PhysicalReview D , 2013.[5] J. Abraham et al.
Phys. Rev. Lett. , 104:091101, 2010.[6] M. Sasaki.
J. Phys. Soc. Jpn. , 77SB:83, 2008.[7] M. Sasaki.
Proc. of ICRR2000 Satellite Symposium: Workshop of ComprehensiveStudy of the High Energy Universe , pages 109–124, 2000.[8] S. Barwick.
Physica Scripta. , T85:106, 2000.299] Y. Asaoka and M. Sasaki.
Nucl. Instrum. Methods Phys. Res. A , 647:34, 2011.[10] Y. Aita et al.
The Astrophysical Journal Letters , 736(1):L12, 2011.[11] Y. Aita et al.
GCN Circular, 8632 , 2008.[12] Y. Asaoka et al.
GCN Circular, 11291 , 2010.[13] George WS Hou and MA Huang. arXiv preprint astro-ph/0204145 , 2002.[14] M. Sasaki, A. Kusaka, and Y. Asaoka.
Nuclear Instruments and Methods in PhysicsResearch A , 492:49–56, 2002.[15] M. Sasaki, Y. Asaoka, and M. Jobashi.
Nuclear Instruments and Methods in PhysicsResearch A , 501:359–366, 2003.[16] J. K. Becker.
Physics Reports , 2008.[17] K. Murase, M. Ahlers, and B. C. Lacki.
Phys. Rev. D , 88:121301, 2013.[18] P. Baerwald, M. Bustamante, and W. Winter.
The Astrophysical Journal , 768(2):186,2013.[19] V. S. Berezinsky and G. T. Zatsepin.
Physics Letter B , 28:423, 1969.[20] V. Ptuskin, V. Zirakashvili, and E.-S. Seo.
The Astrophysical Journal , 718(1):31, 2010.[21] A. Calvez, A. Kusenko, and S. Nagataki.
Phys. Rev. Lett. , 105:091101, 2010.[22] K. Kotera and A. V. Olinto.
Astronomy and Astrophysics , 2011.[23] R. Aloisio, V. Berezinsky, and A. Gazizov.
Journal of Physics: Conference Series ,2012.[24] G. Giacinti, M. Kachelrieß, D. V. Semikoz, and G. Sigl.
Journal of Cosmology andAstroparticle Physics , 2012(07), 2012.[25] N. Gupta.
Astroparticle Physics , 48(0):75 – 77, 2013.[26] N. Gupta.
Astroparticle Physics , 2011.[27] L. A. Anchordoqui et al.
Journal of High Energy Astrophysics , 1:1–30, 2014.[28] A. M. Hillas.
Journal of Physics G: Nuclear and Particle Physics , 31(5):R95, 2005.[29] A. Dar and R. Plaga. A&A, 349:259–266, 1999.[30] J. Hinton.
New Journal of Physics , 11(5):055005, 2009.[31] A. Kappes et al.
The Astrophysical Journal , 656(2):870, 2007.[32] F. A. Aharonian et al. Nature, 432:75–77, 2004.[33] F. Aharonian et al. A&A, 437:L7–L10, 2005.[34] F Aharonian et al.
Nature , 439(7077):695–698, 2006.[35] M. Ahlers and K. Murase.
Physical Review D , 2014.[36] F. Aharonian et al.
Astronomy & Astrophysics , 425(1):L13–L17, 2004.3037] F. Aharonian et al. A&A, 448:L43–L47, 2006.[38] M. Hoshino, J. Arons, Y. A. Gallant, and A. B. Langdon. ApJ, 390:454–479, 1992.[39] D. Horns et al. A&A, 451:L51–L54, 2006.[40] B. Link and F. Burgio.
Monthly Notices of the Royal Astronomical Society , 371(1):375–379, 2006.[41] A. Bhadra and R. K. Dey.
Monthly Notices of the Royal Astronomical Society ,395(3):1371–1375, 2009.[42] F.A. Aharonian et al.
J. Phys. Conf. Ser. , 39:408, 2007.[43] I. F. Mirabel.
Science , 312:1759, 2006.[44] A. A. Abdo et al.
The Astrophysical Journal , 688(2):1078, 2008.[45] L. A. Anchordoqui et al.
ArXiv e-prints , 2013.[46] M. Su, T. R. Slatyer, and D. P. Finkbeiner.
The Astrophysical Journal , 724(2):1044,2010.[47] S. Razzaque.
Phys. Rev. D , 88:081302, 2013.[48] P. M´es´zaros.
Rep. Prog. Phys. , 69:2259, 2006. and references therein.[49] M. Ackermann et al. ApJ, 716(2):1178, 2010.[50] J. A. Nousek et al. ApJ, 642(1):389, 2006.[51] D. Burlon et al.
The Astrophysical Journal Letters , 685:L19, 2008.[52] K. Murase et al.
The Astrophysical Journal Letters , 651:L5, 2006.[53] T. Kifune G. Sinnis F.A. Aharonian, J. Buckley.
Rep. Progr. Phys. , 71:096901, 2008.[54] W. Benbow et al.
Proc. 30th Int. Cosmic Ray Conf., Merida, Mexico , 3:1081, 2007.[55] H. Krawczynski et al.
Astrophys. J. , 601:151, 2004.[56] F. Acero and others (H.E.E.S. Collaboration).
Science , 326:1080, 2009.[57] N. Karlsson et al. arXiv preprint arXiv:0912.3807 , 2009.[58] K. Greisen.
Phys. Rev. Lett. , 16:748–750, 1966.[59] G.T. Zatsepin and V.A. Kuz’min.
JETP Lett.(USSR)(Engl. Transl.) , 4, 1966.[60] M. Ahlers et al.
Astroparticle Physics , 34(2):106–115, 2010.[61] M. Ave et al.
Astroparticle Physics , 23(1):19–29, 2005.[62] R.C. Cotta, J.S. Gainer, J.L. Hewett, and T.G. Rizzo.
New Journal of Physics ,11(10):105026, 2009.[63] M. G. Aartsen et al.
Phys. Rev. Lett. , 110:131302, 2013.[64] D.J. Chung, E.W. Kolb, and A. Riotto.
Physical Review D , 59(2):023501, 1998.3165] P. Blasi, R. Dick, and E.W. Kolb.
Astroparticle Physics , 18(1):57 – 66, 2002.[66] I.F.M. Albuquerque, L. Hui, and E.W. Kolb.
Phys. Rev. D , 64:083504, 2001.[67] G. Giacomelli, S. Manzoor, E. Medinaceli, and L. Patrizii.
Journal of Physics: Con-ference Series , 116(1):012005, 2008.[68] S.D. Wick, T.W. Kephart, T.J. Weiler, and P.L. Biermann.
Astroparticle Physics ,18(6):663 – 687, 2003.[69] A. De Ra´ujula and S.L. Glashow.
Nature , 312(5996):734–737, 1984.[70] A. Kusenko.
Physics Letters B , 405(1-2):108 – 113, 1997.[71] G. Domokos and S. Kovesi-Domokos. 433:390, 1998.[72] A. Letessier-Selvon.
AIP Conf. Proc. , 566:157, 2000.[73] E. Zas H. Athar, G. Parente.
Physical Review D , 62:093010, 2000.[74] D. Fargion. ApJ, 570:909, 2002.[75] F. Wilczek T. Yu J. Feng, P. Fisher. Phys. Rev. Lett., 88:161102, 2002.[76] E. Wolfe, W. Wise, and G. Dalrymple.
U.S. Geological Survey Professional Paper ,1557:129, 1997.[77] T. Sj¨ostrand et al.
Comput. Phys. Commun. , 135:238, 2001.[78] S. Agostinelli et al.
Nucl. Instrum. Methods A , 506:250, 2003.[79] S. Iyer Dutta et al.
Physical Review D , 63:094020, 2001.[80] H. Abramowicz and A. Levy. arXiv:hep-ph/9712415v2 , 1997.[81] B. Badelek and J. Kwiecinski.
Rev. Mod. Phys. , 68:445, 1996.[82] S. Jadach et al.
Comput. Phys. Commun. , 76:361, 1993.[83] R. Gandhi et al.
Astropart. Phys. , 5:81, 1996.[84] A. M. Dziewonski and D. L. Anderson.
Physics of the Earth and Planetary Interiors ,25:297, 1981.[85] J. F. Beacom, P. Crotty, and E. W. Kolb.
Phys. Rev. D , 66:021302, 2002.[86] A. Gazizov and M. Kowalski.
Computer physics communications , 172(3):203–213, 2005.[87] R. Gandhi et al.
Physical Review D , 58:093009, 1998.[88] J. Tseng et al.
Physical Review D , 68:063003, 2003.[89] M. Sasaki, Y. Asaoka, and M. Jobashi.
Astropart. Phys. , 19:37, 2003.[90] M. Sasaki.
J. Phys. Soc. Jpn. , 70(Suppl. B):129, 2001.[91] R. Abbasi et al.
Nature , 484(Supplemental information):351, 2012.[92] J. Abraham et al.
Physical Review D , 79:102001, 2009.3293] S. H¨ummer, P. Baerwald, and W. Winter.
Phys. Rev. Lett. , 108:231101, 2012.[94] G. J. Feldman and R. D. Cousins.
Physical Review D , 57(7):3873, 1998.[95] R. Abbasi et al.
Phys. Rev. D , 83:092003, 2011.[96] T.K. Gaisser. High energy neutrinos from space. arXiv preprint arXiv:1201.6651 , 2012.[97] PKF Grieder. Average Development and Properties of the Hadronic Moonic and Elec-tromagnetic Components in Showers of 10**4 to 10**7 GEV Derived from AN All-. . . .
International Cosmic Ray Conference , 1979.[98] A M Hillas and J Lapikens. Electron-photon cascades in the atmosphere and in detec-tors.
International Cosmic Ray Conference , 1977.[99] K Greisen. Cosmic Ray Showers.
Annual Review of Nuclear and Particle Science ,10:63–108, 1960.[100] K Kamata and J Nishimura. The Lateral and the Angular Structure Functions ofElectron Showers.
Progress of Theoretical Physics Supplement , 6:93–155, 1958.[101] T Antoni et al. Electron, muon, and hadron lateral distributions measured in airshowers by the KASCADE experiment.
Astroparticle Physics , 14(4), 2001.[102] Y. Asaoka and M. Sasaki.
Astroparticle Physics , 41:7–16, 2013.[103] R. M. Baltrusaitis et al.
Nucl. Instrum. Methods A , 240:410, 1985.[104] Y. Aita et al. , 2009.[105] E. Zas et al.
Astropart.Phys. , 1:297, 1993.[106] R. Enberg et al.
Physical Review D , 78:043005, 2008.[107] A. D. Martin et al.
Acta Physica B , 34:3273, 2003.[108] H. Athar et al.