Home

Ultra-high energy neutrino simulations

1. o oO 8 o a a ooo ao Im ooOaa ood a ro O OF ao O ae ugli Estimated total charge p e 0 O lo 810 o oo 0a oO aoa o ood ooo OOOO Oo oO oad Ooadcao o0 119 Estimated total charge p e 10 log l O IOoOoaos E m00 RHH j3ue yuz y JO ouISO N S ulilonalo Pd BFE 5 fe o sue yuz SY JO oUISO S gein Figure 5 6 Event distribution as a function of the total charge observed in the neutrino telescope and the cosine of the zenith angle for downward going UHE neutrino induced events were simulated with SIRENE The atmospheric muon background was simulated with MUPAGE muons top and atmospheric background muons bottom at the trigger level Signal High Energy Simulations Results 10 Event rates per year 10 10 1 0 8 0 6 0 4 0 2 0 Cosine of the Zenith angle Figure 5 7 Distribution of the cosine of the zenith angle rates per year for atmospheric muons dashed and muons induced by downward going UHE astrophysical neutrinos solid at the trigger level Event rates per year 1 2 3 4 5 6 7 log i gemd total charge p e Figure 5 8 Distribution of the total charge in p e of the digitised signal induced by downward going ultra high energy UHE neutrinos solid and atmo
2. 2 4 6 8 10 12 log4q Energy GeV Figure 3 11 Upward going muon neutrino rates simulated with GENHEN dashed and ANIS solid line assuming a diffuse AGN like spectrum for energies between 10 and 10 GeV Only events with a neutrino or secondary muons which could reach the telescope are displayed The produced muon secondaries were propagated towards the Can of the ANTARES telescope using the lepton propagator MMC and re weighted to a dif fuse AGN like neutrino flux E7 10 GeV cm sr s at the surface of Earth In parallel 10 upward going muon neutrinos were generated with GENHEN assuming an E 4 GeV m sr s power law interacting spectrum The pro duced muon secondaries were propagated towards the Can of ANTARES using the lepton propagator MUM and were re weighted to a neutrino flux E 10 GeV cm sr s at the surface of Earth In Figure 3 11 only charged current CC neutrino interactions of the deep inelastic DIS channel are shown with events occuring inside or at the Can and events with a secondary muon which could reach the Can This event selection is done automatically with GENHEN as the program provides an option which 70 3 4 Comparison between ANIS and GENHEN number of events per year EPE ef ne Bil ge a GN rah Hr 1 2 3 4 5 6 7 8 9 log4 o Energy GeV Figure 3 12 Atmospheric muon neutrino rates simulated with GENHEN dashed and ANIS solid like Only events with a neutrino
3. x 6 5 x 10 x Ermo 4 3 Nye 13 7132exp 5 2 cos 6 0 75 4 4 77 Detector simulation and Track reconstruction Hadronic showers Hadronic showers are created in case of a photo nuclear interaction i e hadrons produced in charged current inelastic neutrino nucleus scattering at the neutrino interaction vertex see Chapter 3 As for electromagnetic showers hadronic showers can be considered point like on the large length scale of a neutrino tele scope When an event occurs within the Can of the telescope the hadronic show ers should in principle be simulated as they may produce charged particles and yield Cherenkov photons 55 They create a significant number of photons espe cially at high energies However the high energy neutrino events are likely to be well reconstructed irrespective of whether the light from the hadronic showers is considered The contribution of Cherenkov light produced by charged hadronic showers at the neutrino interaction vertex is thus not simulated in SIRENE Nev ertheless for energies below 100 PeV the simulation of hadronic showers can be added with the help of the ANTARES GEANT based Monte Carlo program GEAMU 84 85 An interface between SIRENE and this hadronic simulation program is available 4 1 2 Photon field simulation All Cherenkov photons are propagated for each individual source towards every optical module of the telescope The Cherenkov photons produced by
4. cles gain energy The maximum energy they can possibly obtain depends on the charge of the particle the shock velocity the magnetic field at the site and the size of the accelerating source This is known as the Hillas criterion 5 which is illustrated in figure 2 where acceleration limits for possible high energy cosmic ray sources are shown Astrophysical objects which can accelerate protons above CONTENTS 10 eV are displayed on the top right corner above the dotted line The two solid lines represent the proton energies above 10 eV above and iron nuclei with energies above 10 eV below As can be seen in the figure only a few objects 15 h Neutron Protons 9 100 Eev Protons pn 1 Zev u u 5 0 D Ta g Ad H u id Fe 100 EeV 4 hot spots amp g lobes S 3 yr m Colliding o galaxies rd SNR disk halo Galactic 9 3 6 h9 12 f 15 18 21 1 au l1 pe 1 kpc 1 Mpe log size km Figure 2 A Hillas plot 5 The strength of the electromagnetic field is shown as a function of the size of the object in which protons and nuclei are accelerated to become high energy cosmic rays Objects above the dotted line are able to accelerate protons above 10 eV assuming an efficient diffuse shock acceleration mechanism are able to accelerate cosmic rays up to the highest energies This is discussed below Cosmic rays with energies below the knee are thought to originate inside the Galaxy most proba
5. cm 3 than the water below 1 0404 cm For ANTARES three spheres define the Earth which is enclosed inside air The first sphere with the radius of the Earth defines the sphere for the sea water at the ANTARES site above the telescope center This sphere extends down to z 2 196849 x 10 meters The second sphere contains for the sea water at the ANTARES site below the detector center Its radius is 6 3713 x 10 2 196849 x 10 6 369103151 x 10 meters The sphere passes through the center of the tele scope which defines the origin of the altitudes z 0 in the MMC geometry The third and last sphere is the seabed rock below the detector and consist of standard rock The radius of the sphere is 6 3713 x 10 2475 6 368825 x 10 meters The sphere goes through z 278 151 meters MMC uses a media definition file in which the rock and sea water surrounding ANTARES are detailed Such a file can be defined as follows media definition file include with mediadef this file use the following syntax all det vcut ecut conti rho medium sphere z r det vcut ecut conti rho medium all 0 0 05 5 e2 1 1 0 Air sphere 2196 840 6371300 0 0 05 5 e2 1 0 99456 Antares Water sphere 0 6369103 151 1 0 05 5 e2 1 1 0 Antares Water 141 Implementation of ANTARES in MMC sphere 278 151 6368825 0 0 05 5 e2 1 1 0 Standard Rock A new medium Antares water was implemented in MMC with the prop erties of the Mediterranean sea
6. rn 3 4 8 Z l fes Pstart 1 ri m t start eos mT F c c a The model is similar to the one used in the detector simulation programs KM3 90 and GEASIM 84 The hit time and wavelength of the Cherenkov photons which can reach the optical module OM are then randomly sampled from the generated distribu tions so that each event is unique Both the direct and scattered photons from muons are sampled from the same distributions The arrival times for the scattered photons are smeared using a truncated normal distribution This truncated distribution is the probability dis tribution of the normally distributed hit time whose values are bound by the standard deviation 460 7 2 54 0 0005 try The variable t is the amount of time a reference photon with a wavelength of 460 nm needs to travel from the muon track to the OM Only the photons which scatter with relatively small angles are taken into account in this parametrisation They are more likely to be observed with a time delay of less than typically 10 ns with respect to direct photons Photons which scatter with relatively wide angles and are delayed more have been neglected The high energy neutrino events are likely to be well reconstructed irrespective of whether these photons are considered In Figure 4 4 the simulated time residuals r tpi A are shown between the hit time tp obtained from the PMTs and the predicted arrival time ie of the photons
7. that the muon track will cause the measured hits 7 The likelihood function is defined as L Oly f 7I where f i denotes the probability density function PDF that specifies the probability of observing the hits 7 given the muon track According to the expression of the single hit PDFs in Equation 4 10 the likelihood can be expressed by N L tf ti Oinci Ailt h tilt Pudu li Purdy binei Purdy Ai Purdy i 1 with p the position and dy the direction of the muon tracks at a fixed time t Minimum Chi Squared Estimate This approach defines a Chi Square x merit function SNZ fe D iii i 1 f with the same convention as for the maximum likelihood The variable g is the central deviation of the chi square distribution If the single hit PDFs are ex pressed in terms of time residuals the chi square function becomes N hie FAN 2 ti ti Pudu X 3 et i The best fit parameters are determined by the minimisation of this expression over the muon parameter space A merit function measures the agreement between the observed data and the fitting model for a particular choice of the parameters 97 Detector simulation and Track reconstruction M estimator The performance of the likelihood maximisation or the chi square minimisation depends on the assumed distribution that has been chosen to describe the ob served data These approaches can thus become inefficient and biased when small de
8. 1 3 Detector layout Two syntactic foam buoys placed at the top maintain each line vertically The anchor is made of a dead weight and a recoverable structure the bottom string socket BSS The recovery of a line is achieved by releasing part of the BSS after issuing an acoustic command from a ship at the sea surface The BSS is connected with an electro optical cable the interconnecting link IL to the junction box JB which provides electrical power to the lines and allows for the transmission of data The backbone of each detector line consists of an electro optical cable It pro vides mechanical strength to the lines electrical power and an optical data link to the storeys A full line is approximately 450 m long with an active length of 350 m It is composed of twenty five storeys each carrying three optical mod ules OMs mechanically supported by a titanium frame A storey of the detector is shown in Figure 1 4 The vertical distance between two consecutive storeys is 14 5 m Together five storeys constitute one sector of a detection line which has its own communication channel to shore The storey closest to the seabed is placed at a height of 100 m so the bottom part of the line is uninstrumented to provide a larger region for observing Cherenkov cones while they are develop ing The three OMs of a storey point downwards at an angle of 45 with respect to the vertical in order to increase the sensitivity to up going muons The chosen
9. 94 J D Zornoza Effect of the ars saturation on the energy reconstruction ANTARES internal note ANTARES SOFT 2003 003 95 W J Metzger Statistical methods in data analysis Katholieke Universiteit Nijmegen Nijmegen The Netherlands 2001 HEN 343 96 M de Jong Partial linearisation of the track fit problem ANTARES internal note ANTARES Soft 2007 001 97 A Romeyer Etude de la sensibilit du d tecteur ANTARES un flux diffus de neutrinos cosmiques de haute energie PhD thesis Universit e Paris VII Paris France April 2003 164 BIBLIOGRAPHY 98 A Oppelt Etude de la r solution angulaire du t lescope a neutrinos ANTARES PhD thesis Universit de la M ditterran e Aix Marseille II France April 2001 99 J D Zornoza Sensitivity to diffuse fluxes and energy spectrum reconstruction in the ANTARES neutrino telescope PhD thesis Universidad de Valencia Valencia Spain January 2005 100 G Carminati et al Mupage A muon generator for parametric formulas ANTARES internal note ANTARES Phys 2006 003 101 G Battistoni et al Astropart Phys 3 1575 1995 102 J Ranft Phys Rev D 51 64 1995 103 J Horandel Astrop Phys 19 193 2003 104 S P Ahlen et al Nuclear Physics B 370 432 1992 105 G J Feldman and R D Cousins Phys Rev D 57 3873 3889 1998 106 G C Hill and K Rawlins Unbiased cut selection for optimal upper limits in neutrino detectors the model rejection
10. AGASA oe HiRes 1 2 monocular x x Auger 2007 gt vo Tp 1025 S 0E lt 1023 1018 1019 1020 102 Energy eV Figure 3 The measured cosmic ray energy spectrum for energies between 1018 eV and 107 eV Recent measurements from AGASA 12 HiRes 11 and Auger 13 are shown The plot was taken from reference 3 the existence of the GZK limit However the flux measurements by AGASA 12 seem in contradiction with the data of the other two observatories They are in favor of a non acceleration scenario for explaining the origins of ultra high energy cosmic rays This alternative model predicts that the excess of cosmic rays measured above the GZK limit results from the decay of very massive long lived particles such as topological defects 14 The discrepancy between the observed spectra is often attributed to a failing energy calibration of the AGASA detector 15 but this issue remains controver sial Further information on the origin of ultra high energy cosmic rays could be obtained by searching for high energy neutrinos High fluxes of neutrinos are expected regardless of the cosmic ray production scenario They are believed to be produced in the acceleration of cosmic rays or by the decay of topological de fects The detection of ultra high energy neutrinos could conclusively close the debate on the existence of the GZK limit and the origin of the very high energy cosmic rays It would also help understand the physics of extrem
11. of the photo multiplier tubes prevail Above 10 TeV an angular resolution of about 0 3 can be achieved with ANTARES Propagation through the Earth Also for the propagation process through the Earth only DIS interactions are sim ulated It is now well established 38 39 that the deep inelastic CC interaction 47 Neutrino event generators Vv V W W a p p Py Py N Hadronic shower Figure 3 1 Neutrino or anti neutrino deep inelastic scattering DIS event A charged current interaction between a muon neutrino or muon anti neutrino and a nucleon proton or neutron is depicted The incoming neutrino or anti neutrino creates a W_ or W boson and turns into a muon The W interacts with an individual quark within the nucleon resulting in its disintegration The struck quark and the remain ing two quarks shower into a variety of hadrons dissipating the large amount of energy transferred to the nucleon length for neutrinos in rock is comparable with the diameter of the Earth at ap proximately 40 TeV Above this energy the Earth becomes gradually opaque to neutrinos Below this energy the Earth is getting more and more transparent to neutrinos Deep inelastic scattering with nuclei is therefore also the major source of the attenuation of the neutrino flux while passing through the Earth Electron neutrinos and muon neutrinos are absorbed in CC interactions and are driven to lower energies in NC interact
12. one receiver being placed in every sector The receivers are composed of an hydrophone with dedicated electronics The method used for determining the relative position is to measure the travel time of acoustic pulses between the receiving hydrophones and the emitter recei vers Apart from the hydrophones and emitter receivers on the detector lines several transponders are located as independent units anchored at different places to the seabed in the vicinity of the telescope These transponder units contain their own electronics and are controlled remotely by the emitter receivers mod ules on the detection lines of the telescope While the acoustic system provides the relative positioning the global positioning of ANTARES is obtained by acous tic triangulation of the emitter receiver modules and a surface ship equipped with GPS The acoustic positioning precision strongly depends on the precision by which the sound velocity in water has been measured as it directly determines the relative distance between the hydrophones Environmental parameters such as the temperature the salinity of the sea wa ter and the evolution of the pressure at the telescope s site influence the sound velocity and need to be measured continuously Sound velocimeters are placed at various places of the telescope along the detection strings Two of them are also equipped with conductivity temperature and pressure probes while conductivity temperature depth CTD dev
13. only Instrumented Figure 3 4 Definition of the telescope geometry inside the Monte Carlo event generator GENHEN including a definition of the Can is commonly used to describe the production of high energy neutrinos The en ergy is sampled in the range Ey in Emax such that EY is uniformly distributed The event generation in E and cos is therefore made such that the events are uniformly flat distributed over phase space The energy spectrum from which the events are drawn corresponds either to the flux of neutrinos at the interaction vertex close to the telescope or to the flux of primary neutrinos entering the opposite side of the Earth A typical value of the spectral index a is 1 4 for the interacting neutrino flux as it gives reasonable statis tics in all energy ranges This value corresponds to a spectral index y a 1 2 4 for the neutrino flux at the surface of the Earth In the case that the drawing spectrum is the primary neutrino spectrum a full propagation through the Earth is required If no full propagation through the Earth is performed the shadow ing effect of the Earth is contained in event weights see next section in the form of a transmission probability taking only charged current CC scatterings into account 54 For a neutrino v with an energy E and a zenith direction the probability of survival through the Earth or the transmission probability Ptrans can be expressed by Pirans Ey exp Na0v
14. put format of ANIS and MMC has been modified such that the software pack ages could be integrated into the ANTARES Monte Carlo simulation chain This enabled a comparison with GENHEN 55 a neutrino event generator that in cludes lepton propagation When suitably tuned both programs were found to require roughly the same amount of CPU time and give comparable event rates Nevertheless only ANIS can be used in association with MMC above 1016 eV to 134 5 3 Discussion generate ultra high energy neutrino events in the telescope A fast and flexible detector simulation program for neutrino telescopes has been developed to simulate the hits from Cherenkov photons at ultra high en ergies The software package named SIRENE was primarily designed for the future cubic kilometer sized detector KM3NeT 79 but it has been applied to ANTARES as the program allows the use of alternative geometries belonging to other neutrino telescopes SIRENE is capable to model the detector response for neutrino events with energies as high as 10 eV The program does not only allow the simulation of variable telescope geometries but also of different configura tions and characteristics of the photo multiplier tubes inside the optical modules of the telescope SIRENE has been integrated into the ANTARES Monte Carlo simulation chain thus enabling a comparison with the detector simulation pack age KM3 which is commonly used for ANTARES simulations The ANTARES ev
15. ulate the leptons passing through the Cherenkov telescope to obtain the detector response The energy thresholds are artificial and depend upon characteristics of the muon propagation algorithm but also on the configuration of the detector The energy thresholds should therefore be chosen carefully in order to get the best accuracy within a reasonable calculation time since the stochastic treatment of the energy can lead to a very large number of separate energy loss events 50 Bremsstrahlung also known as decelerating radiation is electromagnetic ra diation which is produced by the acceleration of the muon when deflected by an atomic nucleus Muons can also radiate a virtual photon which can interact with a nucleus and create a positron electron pair This is the direct pair production interaction Bremsstrahlung and pair production are the dominant processes at high energy E gt 1 TeV as they contribute up to 40 and 50 of the average muon energy loss respectively The inelastic interaction of muons can be de scribed via the exchange of a quasi real photon between the muon and a nucleon It is often referred to as photo nuclear interactions of the muons Its contribu tion is rather small but it becomes more important at high energy This process contributes up to 10 of the total energy loss around 1 TeV The photo nuclear interaction is essentially a low Q process Q lt 1 GeV and its description is model dependent Simplifications can
16. which has been generated with GENHEN A summary of the main characteristics of the program is given in the following sections 3 2 1 Program summary GENHEN simulates neutrino and anti neutrino interactions with conventional matter for high energy neutrinos up to 10 GeV This upper limit on the neutrino energy is due to the lepton propagator programs which are used in GENHEN see Section 3 2 3 For the propagation of the neutrinos through the Earth the Preliminary Reference Earth Model 40 34 is used to calculate the density pro file Details on this Earth model as implemented in GENHEN can be found in 56 The detector reference system of ANTARES that is used by GENHEN is cen tered at the center of gravity of the telescope such that x y z North West Up with z 0 at the seabed level Straight upward going vertical events are charac terized conventionally by a zenith angle 0 Upward going events therefore have a zenith angle such that 0 lt cos lt 1 The code is written in the pro gramming language Fortran 57 The particle identification numbers follow the scheme used in the GEANT 58 detector simulation program Since different event topologies require different input conditions to be generated GENHEN 52 3 2 GENHEN number of events per year 2 3 4 J 6 7 8 g log40 Energy GeV Figure 3 3 Upward going muon neutrino energy spectrum simulated with GENHEN assuming a diffuse AGN like flux with
17. 1 the differential atmospheric flux for elec tron or muon neutrinos and Feen E 0 GeV lem sr s7 the generated dif ferential flux at the surface of the Earth The atmospheric neutrino flux contained in R and Ri has been derived from the Volkova flux using the Lipari para meters 78 ANIS uses pre calculated tables with the atmospheric neutrino flux data The procedure on how to use these weights is described in detail in 74 An example on how to use ANIS for simulating a neutrino flux in ANTARES is given in appendix A 64 3 3 ANIS All Neutrino Interaction Simulation 3 3 3 The Muon Propagator MMC The muon propagator MMC propagates secondary charged leptons and neutri nos from neutrino nucleon interactions towards the telescope The program takes into account ionization losses bremsstrahlung photo nuclear interactions and pair production including effects such as Landau Pomeranchuk Migdal LPM and dielectric suppression decay and Moli re scattering Several parametriza tions for bremsstrahlung and photo nuclear cross sections are available All parti cles which have been produced during the propagation are also propagated until they disapear or exit the detector volume MMC can propagate leptons with energies from their rest mass up to 1 eV using extrapolations of the known cross sections at high energies A full de scription of the program is given in 50 MMC has been originally developed by Dmitry Chir
18. 1 1 Ultra high energy neutrino simulation 5 1 2 Atmospheric muon background simulation Data selection and analysis anar abarth Pee SS 5 2 1 Track energy estimate os at ee eee Berk Ber Hee 5 2 2 Uncorrelated background and triggering CONTENTS 323 FPS event selection we ode ts wai een Pot Hem HP 117 524 Expected rates and neutrino flux limit 118 5 2 5 Model rejection potential 0 00040 122 5 2 6 Average flux upper limits aoe de we SE oh Sa 127 5 2 7 Comparison with existing bounds 127 5 2 8 Effective detector area ete onde el A Ge hw See 129 50 DISCUSSION A aria oren sae Gey ded de Dok a ea Sk a Ak Rt 131 Summary and conclusions 133 A Implementation of ANTARES in ANIS 139 B Implementation of ANTARES in MMC 141 C SIRENE software implementation 145 C 0 1 Classes to build the geometry of a telescope 145 C 0 2 Classes for the tracking of the particles 147 C 0 3 Classes to model the scattering of the particles 148 Acknowledgements 151 List of figures 153 List of tables 157 Bibliography 159 iii Introduction stronomy with several thousand years of history is one of the oldest of the natural sciences its roots dat ing back to antiquity For all time mankind has been gazing at the sky seeking answers to nature s mys teries While early astronomers watched the regular movements of visible celestia
19. A full description of the program is given in 74 In Figure 3 5 an example is shown of a diffuse AGN like flux of muon neutrinos 59 Neutrino event generators generated with ANIS assuming E2 10 GeV cm s sr Comparison be tween Figures 3 3 and 3 5 shows that the ANIS spectrum extends to higher ener gies A summary of the properties of ANIS is discussed in the following sections number of events per year 2 4 6 8 10 12 log4g Energy GeV Figure 3 5 Upward going muon neutrino energy spectrum generated using ANIS as suming a diffuse AGN like flux with energies between 10 GeV and 10 GeV Events with a neutrino or secondary muons which could reach the Can were selected for further processing with a detector simulation Few events can be seen at high energies due to the absorption of neutrinos in the Earth 3 3 1 Program summary ANIS can generate neutrinos and anti neutrinos up to energies of 101 GeV The program is written in the language C which makes it flexible and allows new physics processes to be implemented easily at a later stage ANIS is independent of any external software package Instead the program uses pre calculated tables to provide cross sections for various final states which consist of pairs of the vari ables x and y and rely on the type of interaction and the flavor of the neutrino that is generated This makes the program fast and already implemented physics processes can be adapted by o
20. ENT o o tT NC N oo x a en le le o t301 p AP o S0 p ap Figure 4 12 Distribution of the reconstructed error 4 on the direction of the muon obtained with the AartStrategy procedure The performance of the linear prefit top left the M estimator top right the maximum likelihood with the original PDFs bottom left and the final fit with the improved PDFs bottom right are shown separately 104 4 4 Reconstruction ficiency of the ML estimate depends therefore on the starting point but also on the PDFs themselves As seen in the previous section the PDFs used in Aart Strategy involve cross sections that describe the muon energy loss for energies below 10 GeV Since SIRENE uses a high energy approximation for the muon cross sections and neglects the processes involved at low energy the presently used PDFs are less suitable Moreover the PDFs strongly depend on the model that is used to determine the hit time as well as the parameters which define the relations between the muon track and the optical modules OMs in the tele scope The calculations of the photon path length but also the incident angle of the photons on the OMs and the amplitude of the hits differ in KM3 GEASIM and SIRENE A different tracking algorithm involving specific muon cross sec tions and medium properties is used as compared to the ones used for deter mining the PDF parametrisation Hence new PDFs need to be determined f
21. Equation 3 10 Neer is the total number of events to be generated As a next step the number of events N per bin 7 is scaled by the ratio of the corresponding Generation Volume V for that bin i and the entire Generation Volume V gen cor responding to the total generated spectrum The obtained value is smeared with a Poisson distribution P in order to account for the fluctuations in the generated spectrum NEEDE BIS Veo N vee 3 11 Veen At initialisation the GENHEN program computes the interaction cross sections of the neutrinos with the surrounding matter for the topology chosen and ini tialises the lepton propagation code see Section 3 2 3 Once the information is available the event loop starts The energy E of the neutrino is sampled from the chosen generation spectrum If the propagation through the Earth is not activated the energy is drawn from the neutrino flux at the interaction vertex close to the telescope accounting for the different densities of the media around ANTARES If a full propagation is required the energy is taken from the neu trino flux at the surface of the Earth The coordinates and direction cosines of the neutrino interaction vertex are generated uniformly in the Generation Volume V depending on the value of E If the interaction vertex is generated outside the Can volume cuts on the distance of closest approach to the Can and the neutrino direction are performed to exclude events with secondaries which wi
22. Nap A the number of atoms per volume of the medium N4 is the Avogadro number p the density of the medium and A the mass per mole Xo the mean range of muons in the medium k the energy of the photons radiated and y k E the energy transferred to the photons The truncated total cross section for emitting photons with a minimum en 2 Ke i ergy Kinin is Kinax do HEE S Kye I dk Kinin dk with Kmax gt Ep the maximum energy radiated by the photons At high energies Ey gt gt Kinin the total cross section can then be approximated by 4 1 E xL en 4 2 MEn 3 Nat Xo A factor 2 5 is used to scale the expression in Equation 4 2 to account for electron positron pair production 50 and photo nuclear interactions 10 whose con tributions are relevant at high energies The mean free path of the muon is the inverse of the resulting total cross sec tion The total number of mean free paths nz the muon travels before interacting See Chapter 3 2 As can be seen in the next paragraph this is relevant for the implementation of electromag netic showers in SIRENE 75 Detector simulation and Track reconstruction is determined in a differential approach to particle transport which is used in many simulations 53 First the quantity nj is evaluated ny lny with 7 a random number uniformly distributed in the range 0 1 The corre sponding distance is then given by dx ny Ly with L the mean free
23. Omax the upper edge of the range of the zenith an gle In the upward direction the extension of the Can is either determined by the maximum lepton range in sea water or by the sea surface whichever is the smallest for downward going neutrinos If upward going events are simulated the Generation Volume stops at the top of the Can The Can and the Generation Volume are tools used in the generation method The final result of the simulation should not depend on the details of Veen or the Can of course The generation process is statistical and encompasses the entire generation volume Veen Algorithm The energy E and the direction of the neutrinos that need to be simulated are cho sen to reflect the ANTARES neutrino telescope properties or specific interests of the user The direction is defined by the zenith angle 0 and the azimuth angle x It is sampled uniformly in the cosine of the zenith angle range 6 nin max and in the azimuth angle range 0 271 For the energy spectrum only power law spec tra of the type p E x E 7 are supported in GENHEN Such an energy spectrum is motivated by the theory of Fermi shock acceleration see Introduction which The default Can which is commonly used in ANTARES simulations was also exploited in this work It is defined by the lower extend dZnin 278 15 m the upper extend dZmax 341 47 m and the radial extend dR 266 11 m 54 3 2 GENHEN nnn Can Particle e se tracking reet
24. Q the energy transfer in the target frame v the Bjorken scaling variable x the relative energy transfer or inelasticity y and the total energy of the outgoing hadrons in their center of mass frame W The basic relations between the kinematic variables are presented in the following equations s p pn 2ME M 2MEy 3 5a Q 0 p pl Ey Ep 4EyE sin Jo gt 0 3 5b y HN Z E Ep Ex M 3 50 A2 2 q Sai ae 3 5d Jpn 2Mv 49 Neutrino event generators apn gE PPN Ey E 2MEs 3 5e y W E pk Ey E M pp Q 2Mv M 356 where E and p are the energy and the four momentum of the incident neutrino respectively E and p are the energy and the four momentum of the outgo ing muon The variables q p p pn and px are the four momenta of the ex changed boson W or W the incoming nucleon N and the outgoing hadronic final states X while p is the four momentum of one single hadron The basic kinematic relations 3 5d and 3 5e lead to E Q _ Q Y DME s M 3 6 which shows that for a given energy E DIS neutrino nucleon scatterings can be characterized by two variables such as x y or x Q Cross section To calculate the rate of high energy events in a neutrino detector the differential DIS neutrino nucleon cross sections needs to be evaluated in the whole range of the kinematic variables 0 lt x lt 1 and 0 l
25. a source which can be either a muon track or a full electromagnetic shower are tracked using a scattering model of the water at the site of the neutrino telescope The position direction arrival time and wavelength of the photons hitting the optical modules of the telescope are recorded Medium properties The medium in which the photons travel from their emission point on the muon track to the optical modules OMs of the telescope influences their probability of detection It is thus necessary to take the main properties of that medium into account when tracking the Cherenkov photons to the detector The water properties and the scattering model chosen to represent the ANTARES site 86 are implemented in SIRENE Alternative scattering models representing different medium properties can be added Index of refraction and group velocity The index of refraction nmedium Of a medium is a measure for how much the speed of light is reduced in a transpar ent medium It is defined as the ratio of the speed of light in vacuum reference medium to the phase velocity vp of the photons in the medium of interest C N medium 77 78 4 1 SIRENE The phase velocity is the rate at which the phase of a light pulse propagates in the transparent medium and depends on the wavelength as A A Ww ETE with A the wavelength T the period w the angular frequency and k the wavenum ber of the photons in the medium As a result a dispersion of the light
26. anomalous behaviour of CTEQ5 at very small x lt 1076 has been overcome 74 ANIS uses two different extrapolations for small x and large Q in order to correct for this anomaly in the CTEQ5 distribution sets 68 3 4 Comparison between ANIS and GENHEN ANIS only simulates DIS neutrino scattering with nuclei as the program fo cus on high energy neutrinos Since GENHEN also generates the QE and RES channels only DIS events have been selected after they were produced for com parison with the results of ANIS When GENHEN propagates neutrinos through the Earth only the DIS channel is simulated as absorption is important only at high energies 34 2 CPU time Comparison The Linux time command is used to acquire timing information about the high energy neutrino event generators GENHEN and ANIS The speed of both pro grams depends on the range of energies chosen and the drawing spectrum On a single Linux machine Intel R Pentium R 4 CPU 2 80 GHz 10 muon neu trinos have been generated with GENHEN and ANIS assuming a flux F E E at the surface of the Earth Only Charged Current CC interactions at the neutrino nucleon interaction vertex close to the detector are taken into account All events are written to disk The system CPU time used by the two programs are shown in Table 3 1 Since GENHEN is often used without the full simula tion of the Earth for muon neutrinos timing statistics for a run performed with out the full propagat
27. arrangement of the OMs leads to the detection of light in the lower hemisphere with high efficiency and has also some acceptance for muon directions above the horizontal plane 1 3 2 Electronics containers Each detection line of ANTARES is instrumented with electronics containers and special containers housing acoustic devices and calibration equipment Every storey is associated with a local control module LCM located at the center of the titanium optical module frame OMF A string control module SCM is placed at the base of each line in a container at the bottom string socket BSS which also houses the string power module SPM The first LCM on the line is linked to the SCM and the last LCM is connected to the top buoy The LCM and SCM contain ers hold the electronic boards for all readout functionalities at the storey level and at the detection line level respectively The electronics in the containers host the data acquisition and slow control systems of the neutrino telescope and consti tute a single node of the data transmission network receiving and transmitting data slow control and clock commands from the telescope to the shore station and vice versa In each sector of a detection line there are four slave LCMs and one master LCM MLCM Each LCM slave controls the data transmission of the three optical modules OMs of a storey and sends these data to the MLCM The MLCM collects all data from the corresponding sector and sends t
28. average upper limit over the expected signal from the model ie the model rejection potential or model rejection factor Hoo No MRF Ns The model rejection factor is shown in figure 5 12 as a function of the muon zenith angle and the total charge of the hits for the simulated signal and background The model rejection factor reaches a minimum for cos 0 2 and a total charge of the hits NPE 10 photo electrons p e The best constraints on the simulated The table gives upper limits for background rates up to 15 A linear interpolation has been used to determine the 90 upper limit within the range of given background rates For values of the expected background larger than 15 an extrapolation has been used assuming the statistical fluctuations are of order VNp 55 122 5 2 Data selection and analysis Events per year above limiting value 0 6 0 5 0 4 0 3 0 2 0 1 0 Cosine of the zenith angle 90 CL average upper limit 0 6 0 5 0 4 0 3 0 2 0 1 0 Cosine of the zenith angle Figure 5 10 Integrated distribution of the cosine of the muon zenith angle for the flux shown in Figure 5 9 Signal solid events induced by the simulated AGN like neutrino flux and atmospheric background dashed events are shown in the top figure The average upper limit is shown in the bottom figure 123 High Energy Simulations Results Event rates per year 4 4 5 5 55 6 5 7 log
29. designed 93 Detector simulation and Track reconstruction for processing events of the highest energies the discrepencies at lower energies can be attributed to the parametrisation of the muon total cross section which neglects low energy effects see Section 4 1 The remaining differences between the spectra are assumed to be caused by the use of two different photon tracking algorithms Hit time residuals In Figure 4 9 the time residuals r tpi a between the hit time tp obtained from the PMTs and the predicted arrival time t of the photons on the OMs is shown Both distributions are peaked at r 0 The distribution produced using SIRENE is broader compared to the one computed with KM3 This is due to the fact that SIRENE takes the dimensions of the OM sphere into account while KM3 does not The photons may hit the OM earlier or later than they would have arrived at its center 10 0 20 40 60 80 100 Residual r ns Figure 4 9 Distribution of the time residuals between the hit time obtained from the PMTs and the theoretical arrival time of the emitted Cherenkov photons on the OMs Photons from muons and electromagnetic showers are taken into account The time resid uals calculated with SIRENE are shown by a solid line while the time residuals given by KM3 are represented by a dashed line It can be concluded that KM3 and SIRENE give similar results although dif 94 4 4 Reconstruction ferences remain
30. ee a a 1 3 2 Electronics containers 4 lt i 442 ale eae aa 1 3 3 Optical modules a iene gta a aces Bear inne Boa ee 1 3 4 Sea monitoring instrumentation 1 4 Development and Construction o saoo 1 4 1 Demonstrator Line aoaaa 0000004 1 4 2 Prototype Sector Line and Mini Instrumentation Line 1 4 3 Line Zero ne 14 MIO Sas ante ae ear et gre Neta tthe agate Ae Sees 14 5 The complete telescope ee The ANTARES data acquisition system 2L BELEDEN 2 1 1 Readout of the data zee Re OR 2 1 2 The ANTARES Trigger dta Bolk enk 3 3 amp dds US Dataformat sse ver akte donte ae ele Moy a y Pew ato 214 The Clock SV SUC INN te ca Gerad anes denied nen en ee ee a 2 2 Relative time calibration 0 0 002 0000084 2 2 1 The optical beacon system erm Peas BOGS HAEG 2 2 2 Relative time calibration of the MILOM 223 The ARS TVC calibration aaa aa aa 2 3 The Data Quality Monitoring System 2 3 1 Software Modules ms sac de woedden he oh hh he a 11 13 15 15 16 17 18 20 21 21 21 24 24 26 29 30 30 31 33 33 34 34 35 35 38 39 CONTENTS 23 2 The Kolmogorov Smirnov test 0 0004 3 Neutrino event generators ii 3 1 3 2 3 3 3 4 Event Generators and Propagators o oo aaa 3 1 1 Neutrino Generators ee 3 1 2 Lepton propagators gente dent ded ta Poy to H GENHEN Aad NA at a ot RE A de ON 3 2 1 Progra
31. energies between 10 GeV and 10 GeV Events with a neutrino or secondary muons which could reach the Can were selected for further processing with a detector simulation Few events can be seen at high energies due to the absorption of neutrinos in the Earth simulates one neutrino or anti neutrino flavor at a time assuming only one type of interaction essentially to improve the speed of the simulation A package based on the Monte Carlo programs LEPTO 59 and RSQ 60 has been implemented to simulate the neutrino interactions 54 The total cross sec tions and event kinematics of neutrino nucleon CC and NC interactions in the matter surrounding ANTARES are computed with the help of LEPTO for the DIS channel LEPTO integrates the differential cross section given by Equation 3 7 over the full range of neutrino interactions relevant to ANTARES As the LEPTO program is accurate up to neutrino energies of 10 TeV GENHEN uses an extrap olation of the model to calculate the neutrino nucleon cross sections and kine matics to generate DIS interactions up to an energy of 10 GeV 55 GENHEN uses various parton density parametrizations of the C TEQ group 42 provided by the library PDFLIB 61 for the DIS channel The recommended and used parametrization is the last tested version CTEQ6D The nucleon and A resonant RES and the low energy quasi elastic QE parts of the neutrino nucleon inter action are generated with the program RSQ GENHEN simula
32. energy Ey in Emax and directional ranges cos Onin COS Omax with 0 the zenith angle of the neutrinos are chosen together with the type of gener ation spectrum Only power law spectra F E x E are supported but other 62 3 3 ANIS All Neutrino Interaction Simulation detector hei ght Figure 3 7 Definition of the telescope geometry assumed by the Monte Carlo event gen erator ANIS The geometry is anchored to the detector center of gravity A cylindrical Final Volume is defined for each neutrino event The axis of the Final Volume is parallel to the direction of motion of the neutrino 74 types could be added at a later stage Neutrinos are generated uniformly on the surface of the Earth Their energy is sampled from the chosen power law spec trum Neutrinos are subsequently propagated through the Earth in small steps towards the telescope Neutrinos that survive the propagation through the Earth are simulated to interact within the Final Volume All generated events are recorded or sampled according to their interaction probability If all events are written to output they need to be properly weighted in order to obtain distributions of physics variables The weights used in ANIS for this purpose are defined below Neutrino event weights Four event weights are defined in the program to produce a physical spectrum of the events once weighted with their interaction probability These weights include a normaliz
33. from a PMT will cause a hit with the measured amplitude The second and third routines rely on the calculation of the time residuals between the hit time t obtained from the PMTs and the predicted arrival time t of the corresponding photons on the OMs The best estimates of the track parameters are the values that maximise the quantity where r t ph is the hit time residual for each hit i and g r 6 In f r with f r the PDF of finding a hit i with residual r as in Equation 4 10 Uncorre lated background hits are not taken into account The PDFs describe the signal only Due to the shape of the hit time residual distribution the PDFs are rela tively flat for large residuals It is therefore rather difficult to find the maximum of the corresponding likelihood function To overcome this problem the third routine makes use of the M estimator with g r 6 2 1 r 2 2 which is linear in r for large values of the residual The accuracy of the reconstruction algorithm is improved by iterating the second and third routines with gradu ally improving start positions The best track estimate at each step is the one for which the likelihood function has reached a maximum The final track es timate is based on the Maximum Likelihood ML approach using a PDF that takes the background hits into account A quality cut based on the value of the likelihood function at the fitted maximum has been implemented to better dis criminate be
34. i gE unated total charge p e Event rates per year 4 45 5 5 5 6 6 5 7 log i Estimated total charge p e Figure 5 11 Integrated distribution of the estimated total charge of the hits for the flux shown in Figure 5 9 In order to place a more accurate constraint on the simulated neu trino flux the atmospheric spectrum has been extrapolated to higher values assuming an E 7 spectrum Signal solid events induced by the simulated AGN like neutrino flux and atmospheric background dashed events are shown in the top figure The average upper limit is shown in the bottom figure 124 5 2 Data selection and analysis 10 Oo S 5 oO D 10 3 a gt 1 107 0 6 0 5 0 4 0 3 0 2 0 1 0 Cosine of the zenith angle iol o 2 Q 8 10 _ oO oO D hen B 10 1 4 45 5 5 5 6 6 5 7 log i o Estimated total charge p e Figure 5 12 Model rejection factor as defined in the text for the diffuse flux of downward going UHE neutrinos as a function of the muon zenith angle top and to tal charge bottom cut variables 125 High Energy Simulations Results neutrino flux are therefore found by accepting events with cos gt 0 2 and a total charge of the hits NPE gt 10 p e The expected event rates corresponding to the final event selection criteria defined above are shown in Figure 5 13 as a function of the cosine of the muon zenith angle direction and the total charge of th
35. interactions in matter in or around ANTARES have a range long enough to reach the telescope As a muon travels through matter it looses energy due to ionization including delta ray pro duction and excitation processes bremsstrahlung photo nuclear interaction and pair production The relative importance of these processes depends on both the target material and the muon energy The process of muon pair production in muon propagation is usually neglected due to the very small cross sections in volved 47 The energy loss of a muon in matter below a few hundred GeV is continuous and dominated by ionisation as described in the Bethe Bloch rela tion 48 The energy transferred to the free electrons during a collision is rather small but knock on electrons also called delta rays can be emitted At high muon energies E gt 1 TeV the radiative processes become prevalent In some rare cases the radiative energy loss can be very large As this happens only rarely the energy loss cannot be treated as a uniform and continuous process Instead a division between a continuous and a discrete energy loss regime is usually introduced in the lepton propagation programs via an energy cut Ect and a relative energy loss cut Vey see Sections 3 2 3 and 3 3 3 Below these cuts all losses are con sidered as continuous 49 While Ecut is preferably used in the evaluation of the transport of muons and taus down to the detector location vet is applied to sim
36. meson interactions The neutrino flux is taken to follow an E depen dence as expected from Fermi acceleration The Waxman and Bahcall limit is 127 High Energy Simulations Results 0 A coal ce Laren Neel alc ayva AAVV AAN Figure 5 14 Comparison of the calculated upper flux limits for diffuse muon neutrinos as derived from different experiments and model calculations The limits are given as a function of the logarithm of the neutrino energy The measured atmospheric muon flux is shown using AMANDA II data 108 with the central values for different calcula tions 109 The Waxman and Bahcall bound is from 107 the ANTARES limit from 99 the AMANDA B10 limit at ultra high energy from 110 The AMANDA II limit is from 108 and for ultra energy from 111 The RICE limit is from 112 and the BAIKAL limit from 113 The IceCube limit is from 114 and the KM3NeT limit from rarmrnins 5 6 7 8 9 10 11 log E GeV ANTARES UHE v limit 1 year ANTARES UHE v limit 3 years ANTARES v limit 1 year ANTARES v limit 3 years AMANDA B10 1997 UHE all flavors 3 AMANDA II 2000 2002 UHE all flavors 3 AMANDA II 2000 9 limit RICE 1999 2005 all flavors 3 Baikal 1998 2002 all flavors 3 Full Icecube 1 year KM3NeT 1 year prelim W amp B limit 2 transparent sources Bartol2004 Naumov et al ROPM HKKM2004 Martin et al GBW AMANDA atm V data 79 The assumption has
37. of muons with ener gies between 107 and 10 GeV See Chapter 4 110 5 1 Monte Carlo simulations 0 45 dP dO 0 4 0 35 0 3 0 25 0 2 0 15 0 1 1 08 0 6 04 02 0 02 04 06 08 1 Cosine of the Zenith angle O Figure 5 1 Distribution of the cosine of the zenith angle for neutrinos which can pro duce a muon able to reach the detector Neutrino interactions were simulated with ANIS assuming an AGN like spectrum Muon secondaries were propagated to the default Can of ANTARES with MMC The probability distribution is shown for neutrinos with an energy below 10 GeV dotted between 107 10 GeV dashed and 10 101 GeV solid line In the ANTARES reference frame a zenith angle of 0 corresponds to an upward going neutrino while an angle of 180 corresponds to a downward going neu trino In Figure 5 1 the distribution of the cosine of the zenith angle of the generated muon neutrinos is shown for various energy ranges Only events which can pro duce a muon able to reach the detector Can see Chapter 2 are shown As can be seen in the figure the generated neutrino events concentrate near the horizon as the energy increases This is expected as the Earth becomes opaque to UHE neu trinos and the path length through the atmosphere and the sea water is longest in the horizontal direction A major fraction of muons from UHE neutrinos will therefore reach the detector in the downward going horizontal direc
38. of the new detector software SIRENE 91 Detector simulation and Track reconstruction has been compared with KM3 and GEASIM using the ANTARES geometry The results of this comparison are presented in the next section 43 Comparison with the ANTARES detector simula tions 4 3 1 KM3 and GEASIM The ANTARES collaboration has developed two Monte Carlo programs for the simulation of the response of the detector to passing muons each simulating dif ferent aspects of the events The package KM3 simulates the detector response to the Cherenkov light produced by muons and secondary electromagnetic parti cles from neutrino interactions in the media surrounding the telescope including scattering in sea water Pre calculated data tables generated with a full GEANT simulation 58 are used to parametrise the amount of Cherenkov light reach ing the optical modules OMs of the detector The use of these photon tables makes the simulation faster The second simulation program is GEASIM which accounts for hadronic ef fects not included in KM3 GEASIM performs the tracking of all particles except for the Cherenkov photons through the entire telescope volume It uses an ana lytical function to estimate the number of photons detected assuming there is no light scattering in the sea water Both KM3 and GEASIM are written in the FORTRAN language and are run ning under UNIX A detailed description of the programs and the methods cho sen for t
39. of the pro gram can be envisaged An important issue is the addition of electron and tau leptons as Cherenkov light sources into the program Specific tracking methods for these particles also need to be developed Simulation of the muon energy loss at high energies should be extended with the implementation of electron positron pair production and photo nuclear inter actions with the associated generation of hadronic showers Continuous muon energy loss due to ionisation should be implemented for the use of SIRENE in low energy studies Concerning the photon tracking algorithm one single distribution with corre lated wavelengths and times should be determined as the arrival times of the photons on the optical modules OMs of the telescope depend on the wave length Different wavelength and hit time distributions should also be imple mented for both the direct and scattered photons A parametrisation for the ar rival time distribution of the photons which scatter with a wide angle and reach 107 Detector simulation and Track reconstruction the OMs with a relatively long time delay needs to be implemented For a proper reconstruction of UHE muon tracks generated new PDFs have to be determined for use in the reconstruction programs 108 Chapter 5 High Energy Simulations Results The sensitivity of the ANTARES telescope to a diffuse flux of cosmic neutrinos with en ergies above 10 GeV has been estimated using the newly devel
40. offsets between the responses of the front end chips of the line to a flash of an optical beacon All electronics and calibration systems were found to contribute less than 0 5 ns to the overall timing resolution This result enables us to predict an angular resolution of less than 0 3 for the ANTARES telescope The sensitivity of ANTARES to ultra high energy neutrinos has been evalu ated in terms of Monte Carlo simulations Three distinct stages of simulation can be recognised the neutrino event generation the propagation of the produced leptons towards the telescope and the detector response simulation The sim ulation programs used to simulate events in ANTARES accept neutrino events within a limited range of energies only i e below 10 eV and thus cannot be exploited for the study of ultra high energy neutrino events Improvements of already existing packages and the development of new programs have thus been required The neutrino event generator ANIS 74 was initially designed for use with the AMANDA neutrino telescope but it has been adapted to ANTARES to gener ate ultra high energy neutrino interactions in the surrounding water and seabed The produced secondary leptons have been propagated towards and through ANTARES using the lepton propagator MMC 50 which is interfaced with ANIS and can process ultra high energy events The telescope geometry and the sea water properties at the ANTARES site have been implemented in MMC The out
41. on the OMs The contribution from the muons and the EM showers are shown separately The broad tails at large positive values of r are due to light scattering All distributions are peaked around r 0 ns Direct photons from muons carry the most precise timing information while scattered photons from muons are delayed with respect to this reference time The distribution of the arrival time for direct photons from EM showers occurs in a broader time range This is due to the fact that the photons are not emitted at a fixed angle Oc but according to an angular distribution see Section 4 1 1 85 Detector simulation and Track reconstruction 20 0 20 40 60 80 100 Residual r ns Figure 4 4 Simulated distributions of the time residuals r in ns between the hit time obtained from the PMTs and the predicted arrival time of the photons on the OMs The contribution from direct photons emitted by muons is shown by the dashed line The con tribution from all muons is represented with a dotted line The time residual for photons from electromagnetic showers is displayed by the dotted dashed line The contribution for all photons is shown by the solid line Effective photon flux detected The photons inside the optical module OM need to be tracked towards the photomultipler tube PMT to determine the av erage number of photo electrons which are produced The effective photon field strength is computed taking into account the reduction of the l
42. optical module OM 0 19 The Prototype Sector Line PSL 2 s6 a05 oS a BS ee Oe 22 The Mini Instrumentation Line MIL 23 The Mini Instrumentation Line with one Optical Module MILOM 25 Downward going muon event reconstructed using Line 1 26 Sensitivity to point like neutrino sources with ANTARES 5 Line Ae er i a od ET a Se he MH ate Ee ca Serge 4 rde ws 8 27 Distribution of measured time differences from OMs of the MILOM and the Led Beacon used as reference 0000 36 Raw and Integrated TVC spectra sacar area ade de aa zn oel 37 Correlations between ARS time stamp TS and Time to Voltage Con verter IVC 22 ede eee Get aad ei ed A aa ae tre 38 Time stability of the dynamic ranges of the ARS TVC 39 Diagram showing the data flow between the different modules of PhAntOM 40 Sample picture of the ANTARES Histogram Presenter window 42 Neutrino or anti neutrino deep inelastic scattering DIS event 48 Density profile of the Earth as a function of the distance from the center Of the Earth oe ati SEERDE eS 49 Upward going muon neutrino energy spectrum simulated with GENIETEN it Steden Std et ede ane neden Ve thts Soe 53 Telescope geometry as defined in GENHEN 55 Upward going muon neutrino energy spectrum generated using ANIS a iets sce eta cert etd ese Ne crane eh eee area at cit ea at aa 60 153 LIST OF FIGURES 154 3 6 3 7 3 8 3
43. or secondary muons which could reach the telescope are displayed The parametrization from Lipari 78 was used allows to store events at or inside the Can For ANIS the program Atmflux which belongs to the MMC package was used to process the muons produced in neutrino nuclei interactions As can be seen in Figure 3 11 the energy spectra obtained with ANIS and GENHEN are comparable The remaining differences between the two spectra can possibly be attributed to the use of two different generation algorithms The low event statistics at energies above E gt 10 GeV are due to the opacity of the Earth for very high energy neutrinos see Introduction For a study of ultra high energy UHE cosmic neutrinos with ANTARES one therefore needs to focus on downward going events A method to select such events in order to arrive at an estimate of the sensitivity of ANTARES to these rare neutrino events is described in Chapter 5 In Figure 3 12 the neutrino rates for an atmospheric neutrino flux are com pared using either the Barthol model for GENHEN or the Lipari model for ANIS The rates are comparable The small differences are most likely due to the slightly different atmospheric flux model and the generation method used It is con cluded that when suitably tuned both GENHEN and ANIS can be used to simulate high energy neutrino spectra for ANTARES However for ultra high energy downward going neutrinos of E gt 10 GeV only ANIS can be
44. organised according to the ARS identifier and the data type Each SPE hit contains information such as the IP ad dress of the frame which depends on the identifiers of the ARS and the PMT that has been hit Using the Histogrammer package of PhAntOM the online monitor ing system of ANTARES one dimensional histograms can be made such as TVC distributions of SPE hits from physics events see Section 2 3 The data format is described in the ANTARES internal note 26 2 1 4 The Clock System The purpose of the clock system is to provide a common timing signal to all front end chips of the telescope in order to synchronise and control the data flow This is essential since the timing of the Cherenkov photons hits from muon tracks needs to be determined precisely The system consists of a main clock generator on shore a clock distribution system made of a bidirectional optical fibre network and a clock signal transceiver in each local control module LCM The on shore master clock system drives the slave clock system in the LCMs It defines a common frequency for all Analogue Ring Sampler ARS chips of the telescope All clocks are locked to the on shore reference clock frequency of 20 MHz that corresponds to a time period of 50 ns On shore the reference clock sends a sinusoidal electrical signal which is con verted to an optical signal and transmited to the Junction Box JB via the main electro optical cable MEOC that connects the shore to
45. path of the muon in the medium Electromagnetic showers Muons may give rise to a succession of electromagnetic showers EM showers in the electric field of nuclei on their way through the detector These EM showers are composed of charged particles electrons positrons and photons The elec trons and positrons suffer Bremsstrahlung and generate Cherenkov photons in the sea water The photons emitted may in their turn create an electron positron pair which may also lead to the emission of Cherenkov photons This mechanism continues until the energy of the secondary particles falls below the critical energy for which the energy losses via ionisation and the radiative processes become equal see Chapter 3 The energy at which electrons stop radiating Cherenkov photons has been estimated to be Ec 0 7 x 10 3 GeV 81 At this point no further EM showers are produced On the length scale of high energy neutrino telescopes EM showers can be considered as point like i e as localised events along the track Statistical fluctu ations between electromagnetic cascades are small if they contain a large number of secondary particles This is true for showers of energies much higher than their Cherenkov thresholds 82 As the energy of the EM showers in SIRENE has a minimum energy of 10 MeV the EM showers become independent of the initial muon For that reason all EM showers can be considered identical In SIRENE an EM shower characterized by it
46. potential technique 2002 astro ph 0209350 107 E Waxman and J Bahcall Phys Rev D 59 023002 1999 108 K Hoshina Diffuse high energy neutrino searches in amanda ii and ice cube Results and future prospects In Journal of Physics Conference Series 120 2008 062007 Proceedings of TAUP 2007 109 M Honda and T K Gaisser Ann Rev Nucl Part Sci 52 153 2002 110 M Ackermann et al Flux limits on ultra high energy neutrinos with amanda b10 Astropart Phys 22 339 353 2005 111 M Ackermann et al Search for ultra high energy neutrinos with amanda ii Astrophysical Journal 675 1014 1024 2008 112 I Kravchenko et al Rice limits on the diffuse ultrahigh energy neutrino flux Physics Review D 73 082002 2006 113 V Aynutdinov et al Status and new results from the baikal detector Pro ceedings of Ninth Int Conf on Topics in Astroparticle and Underground Physics TAUP 2005 114 J Arhens et al Sensitivity of the icecube detector to astrophysical sources of high energy muon neutrinos Astropart Phys 20 507 532 2004 165 BIBLIOGRAPHY 115 A Okada Astropart Phys 2 393 1994 116 B Stroustrup C Programming Language The 3rd Edition Addison Wesley Professional 1997 117 F Cassol Gendet 1 0 a program to generate detector files ANTARES internal note ANTARES SOFT 1999 007 166
47. prototype instrumentation line MILOM see Section 1 4 4 in preparation of using it for the entire telescope is detailed in the next chapter 1 4 Development and Construction 1 4 1 Demonstrator Line The ANTARES project was launched in 1996 After a design study period a first prototype line consisting of seven optical modules OMs and several instruments for position monitoring and calibration was deployed in 1999 and connected to the shore at 1100 m depth about 37 km from Marseille The line used an already existing undersea cable donated by France Telecom This made it possible to take data in a telephone exchange centre in Marseille with the prototype line The main purpose of this prototype line referred to as the Demonstrator Line was to investigate the feasability of the overall principle of the ANTARES telescope and various specific aspects of the ANTARES design in particular the acoustic positioning system The line was in operation until June 2000 and allowed the detection and reconstruction of the first muon signals In parallel a detailed programme of in situ measurements has been under taken since October 1996 in order to find a suitable location for the ANTARES telescope considering several determining parameters such as the depth the wa ter transparency the optical background and the strength of the deep sea cur rents This has led to the selection of a site near Toulon 42 50 N 6 10 E at a depth of 2475 m In Octob
48. scribed in 89 rescaled to coincide with the maximum absorption length which was measured in situ at the reference wavelength of 460 nm 86 A flux of pho tons travelling a distance through a transparent medium will then be attenuated by the absorption factor A A depending on Lnedium with A N exp nn where is the wavelength of a photon Tracking algorithm Since the number of events to be processed by the detector simulation is large the number of secondary particles produced is also large As the distances traversed by the muons are also large propagation of the total number of emitted photons is not feasible In SIRENE an alternative method has been implemented For each source Cherenkov photons are generated with a wavelength distri bution between 300 nm and 600 nm The corresponding arrival times on every optical module OM in the telescope are calculated using the medium properties described above Direct and scattered photons from muons and direct photons from electromagnetic EM showers are recorded As a first approximation the wavelength and time distributions of the generated photons are taken to be in dependent Both wavelength and time distributions are normalised to the total number of generated photons In SIRENE photons are propagated to a so called virtual module defined as a sphere encompassing one or more OMs of the telescope The concept of a virtual module allows efficient comparison betwe
49. seen in the figure both programs give comparable results The same settings were used in the frame work of this thesis for the comparison of GENHEN MUM and ANIS MMC for the ANTARES telescope described in the following section 3 4 Comparison between ANIS and GENHEN 3 4 1 General differences Both ANIS and GENHEN are Monte Carlo MC event generators for neutrino Cherenkov telescopes While GENHEN has been developed specifically for AN 66 3 4 Comparison between ANIS and GENHEN Nn o D k muons with that energy o a muons with that energy gt oa D eee priitiit verbaal tipi tii tii tits EEEN TN 6 98 100 0 10 20 30 40 50 60 70 80 90 100 energy TeV energy TeV EEN eT Pour tipi tii Pi i 80 82 84 86 88 90 92 94 9 Figure 3 9 Comparison of the final energy distributions of 500000 muons with initial energy 100 TeV which were propagated through 1 km of water calculated by MMC solid line and MUM dashed with the same parametrizations of all cross sections and value of energy cutoff Veut 1 10 Left a close up of the picture on the right 50 TARES ANIS can be used for any high energy neutrino Cherenkov detector Both algorithms can generate neutrinos of all flavors ve Vy Vz taking into account all relevant interactions with atomic nuclei and electrons GENHEN can simulate neutrino events up to 10 GeV which is the usual energy range of interest for neutrino telescopes of the
50. size of ANTARES ANIS can generate events up to 1012 GeV with special emphasis on the simulation of the highest energy neutrinos For this reason quasi elastic QE and resonant RES interactions are neglected in ANIS while they are properly accounted for in GENHEN Note that these processes are only relevant at low energies E lt 10 GeV GENHEN is written in the language Fortran whereas ANIS has been imple mented in C The use of the C programming language makes the ANIS code fast and more flexible GENHEN uses the ANTARES Monte Carlo event libraries while ANIS does not link to any library that is specific to a given detector ANIS uses the HepMC and Vector packages of the CLHEP library to record the neutrino events energies positions and directions The neutrino nucleon interaction cross sections are described in ANIS us ing a parametrization for deep inelastic DIS scattering based on the CTEQ5 parametrization whereas GENHEN can use various parametrizations of the CTEQ group 42 In the framework of this thesis CTEQ6 which is the latest parametriza tion available was used As can be seen in Figure 3 10 there is a small dif 67 Neutrino event generators ference between the cross sections calculated with the CTEQ5 and the CTEQ6 parametrizations for neutrinos with energies in excess of 10 GeV _ 2 ao N aC 9 10 ob O 5 10 ob Cc A 10 D 2 10 10 1 2 3 4 5 6 7 8 9 log4o Ene
51. that can be attributed to different methods and amount of detail employed by the two programs In contrast to KM3 SIRENE is suited for neu trino simulations at ultra high energy that is above 10 GeV The methods used to reconstruct the SIRENE data after their digitisation and processing through the trigger software are described in Section 4 4 Possible extensions of the program are presented in the last section of this chapter 4 4 Reconstruction Several dedicated programs are available within the ANTARES collaboration to reconstruct the muon position direction and energy considering the time and amplitude of the physics events recorded by the trigger software The task of the reconstruction routine is to estimate track parameters 8 that are compatible with the observed hits The relation between the position and orientation of the muons with the arrival time of the emitted photons is described by the model prediction given by Equation 4 7 Hence the track reconstruction is non lin ear Therefore approaches such as maximum likelihood minimum chi square or other least square methods 95 are usually applied in an iterative process The procedures can be used in a linearised form by keeping one of the parameters constant For instance it has been demonstrated 96 that the hit time is a lin ear function of the muon position once the track direction is fixed The iterative process allows then to find a unique solution for the muon positi
52. the JB at the ANTARES site A passive splitter inside the JB distributes the signal to the detection lines of the telescope In each string control module SCM the reference clock signal is regenerated to compensate for the optical power losses along the 40 km long MEOC cable The signal is then distributed to the individual line sectors and finally to each LCM The LCM internal clock register can be reset by an external reset time stamp RTS signal from the on shore master clock which defines the period for all ARS chips In the LCM the received optical signal is converted into an electrical serial data stream which is transmitted to the six ARS chips The clock signal is decoded so any associated run command is also made available to the LCM and to all associated front end chips The ARS counts the LCM clock periods since the last 33 The ANTARES data acquisition system reset and provides the arrival time of photons on the PMTs by means of a time stamp TS with a resolution of 50 ns that corresponds to the master clock period A Time to Voltage Converter TVC is used to interpolate between clock pulses It gives an analogue signal proportional to the precise time of arrival inside the LCM clock period which is the time between two clock stamps measured with a resolution of 8 bits corresponding to 50ns x 256 0 2 ns The decoding of the TS and TVC raw data is necessary to obtain calibrated time information Time offsets between the loca
53. used 6 Atmflux allows to propagate leptons produced in neutrino nuclei interactions from the in teraction vertex towards the Can of the telescope and gives an estimate of the energy lost along the path If a muon reaches the Can or a neutrino interaction occurs inside the Can the energy is positive 71 Neutrino event generators 72 Chapter 4 Detector simulation and Track reconstruction In this chapter a new detector simulation program for neutrino telescopes is presented The software package named SIRENE is primarily designed for use with the future cu bic kilometer size detector KM3NeT but has been applied to ANTARES as the program allows to be easily adapted for other neutrino telescopes The performance of SIRENE has been compared with that of KM3 another detector simulation package To validate the quality of the events generated using SIRENE they have been subjected to two different track reconstruction programs The results of the comparisons are presented wo different stages of Monte Carlo simulation can be discerned when modelling cosmic neutrino tele scopes the neutrino event generation and the detec tor response Monte Carlo event generators are de scribed in the previous chapter of this thesis The present chapter focuses on the simulations which model the Cherenkov light created by secondary leptons produced in neutrino interactions in the in strumented volume of the telescope of interest It also includ
54. was com pared with the currently most competitive upper bounds made on the diffuse flux of ultra high energy neutrinos The upper limit predicted after one year of observation with ANTARES is about on order of magnitude above the theoretical bound calculated by Waxman and Bahcall 107 and about a factor of four above the upper limit placed by the AMANDA II 111 telescope The expected sensi tivity after three years of operation is enhanced to about a factor of two above the Waxman and Bahcall limit Compared with the results reported by AMANDA II the sensitivity of ANTARES expected after three years of data taking shows an improvement of almost a factor two These results are encouraging However it should be noted that the present study is aimed at an order of magnitude esti mate No full energy nor track geometry reconstruction was included Nonethe less a fairly competitive upper limit for the neutrino event rate has been found 136 5 3 Discussion It is anticipated that the cubic kilometer sized KM3NeT 79 project will lead to an event rate increase of about a factor twenty and hence a substantical reduction in the upper limit allowing to probe the Waxman and Bahcall range 137 High Energy Simulations Results 138 Appendix A Implementation of ANTARES in ANIS On input the program ANIS uses a steering file which contains all run specific information Below is an example of such a file for the ANTARES telescope which is s
55. 005 003 160 BIBLIOGRAPHY 31 C Colnard Physics data of ANTARES online monitoring phantom instal lation and users manual ANTARES Internal note ANTARES Soft 2004 005 32 C Colnard The ANTARES histogram presenter installation and users manual ANTARES Internal note ANTARES Soft 2004 003 33 Eadie et al Statistical Methods in Experimental Physics Elsevier Science Ltd 1983 2nd Repr edition 34 R Gandhi et al Ultrahigh energy neutrino interactions 1995 hep ph 9512364 35 S L Glashow Phys Rev 118 316 317 1960 36 S Jadach et al Comput Phys Commun 64 275 1991 37 H Athar et al Phys Rev D 62 103007 2000 38 R Gandhi et al Ultrahigh energy neutrino interactions Astropart Phys 5 81 110 1996 39 R Gandhi Ultra high energy neutrinos A review of theoretical and phe nomenological issues 2001 hep ph 0011176 40 A M Dziewonski and D L Anderson Phys Earth Planet Interior 25 297 1981 41 G Sterman et al Handbook of Perturbative QCD Review of Modern Physics 67 157 248 1995 42 J Pumplin et al New generation of parton distribution functions with un certainties from qcd analysis Journal of High Energy Physics 012 0207 2002 43 A D Martin et al Parton distributions for the LHC 2009 arXiv 0901 0002 44 J Pumplin et al New Generation of Parton Distribution with Uncertainties from Global QCD Analysis 2002 hep ph 0201195 45 ZEUS coll
56. 1 SIRENE Energy below 1 TeV el ar a pg A 0 8 0 6 0 4 0 2 100 200 300 ce 500 Distance m NPE Energy below 10 PeV o 100 200 300 500 Distance m Figure 4 5 Simulated number of detected Cherenkov photons NPE per OM as a func tion of the distance of closest approach of the muon track to the hit OM The total number of detected photons i e from muons and EM showers is shown as a solid line The contri bution from muons only is shown by a dashed line In the top plot the number of photons for muons with an energy Eu lt 1 TeV is shown In the bottom plot the number of photons for muons with an energy Ey lt 10 PeV is shown 89 Detector simulation and Track reconstruction g 10 1 2 3 4 5 6 7 logo Muon energy GeV Figure 4 6 Distribution of the number of detected Cherenkov photons NPE from elec tromagnetic showers solid and muons dashed as a function of the muon energy the simulation The default value is 80 photons which is far beyond the satura tion level of the PMT cathode at 8 npe Thereafter the hits need to be processed by a simulation of the detector electronics to be able to study the performance of the trigger and the reconstruction software This final step is presented in the next section 4 2 Signal digitisation and triggering As seen in Chapter 2 the front end ARS chip integrates the analogue signal of the PMT by summing the total number of detected p
57. 9 Vadim Kuzmin and Georgiy Zatsepin 10 indepen dently predicted that high energy cosmic rays traveling through outer space can scatter with the omnipresent photons of the cosmic microwave background 1 to create charged and neutral pions via the A resonance p yrAtant n la pty At part 1b when the cosmic rays energy crosses the threshold of 5 x 101 eV The mean free path associated with the interaction drops considerably above this energy thresh old which is referred to as the GZK limit This results in the prediction that the observed cosmic ray flux at Earth will be strongly suppressed above 5 x 10 eV At these extreme energies only cosmic rays coming from within the local super cluster of galaxies that is from astrophysical sources within 100 Mpc from Earth can be detected With the emergence of large high sensitivity cosmic ray telescopes the GZK limit can now be measured The High Resolution Fly s Eye HiRes detector 11 the AGASA surface array detector 12 and the Pierre Auger Observatory PAO experiment 13 have studied the cosmic ray flux in this energy domain Figure 3 presents an expanded view of the cosmic ray spectrum shown in Figure 1 Only the more recent measurements are shown for energies between 1018 eV and 107 eV As can be seen in the figure both the HiRes and the PAO results show a steepening in the cosmic ray flux beyond 3 5 x 101 eV which is consistent with CONTENTS 1026
58. 9 3 10 3 11 3 12 4 1 4 2 4 3 4 4 4 5 4 6 4 7 4 8 4 9 4 10 4 11 4 12 4 13 4 14 5 1 Differential neutrino nucleon cross sections used by ANIS 62 Telescope geometry as defined in ANIS vn 63 Telescope geometry as defined in MMC 66 Comparison of the lepton propagators MMC and MUM 67 Comparison of the CTEQ6 and the parametrization used in ANIS for neutrino nucleon cross sections 2 0 0 00 0 0000 68 Comparison of upward going muon neutrino rates simulated with GENHEN and ANIS assuming a diffuse AGN like spectrum 70 Comparison of atmospheric muon neutrino rates simulated with GENHEN and ANIS esn achet begs bert beh ep aed ante aie 71 Model predictions and measurements of the group velocity of light in sea water at the ANTARES site 0 0000 80 Representative geometry for tracking the photons in SIRENE 82 Coordinates used to describe the detection of Cherenkov light in SIRENE inenen Agere ho de Deen gael sel de ae soo HA A 83 Distribution of the time residuals simulated with SIRENE 86 Distribution of the number of detected Cherenkov photons simu lated with SIRENE vaneen edad erde edi Belle we BS 89 Distribution of the number of detected Cherenkov photons from electromagnetic showers and muons as a function of the muon en OPN ae Ee A ee te oh ee Ps 90 Distribution of the time residuals between the digitized signal sim u
59. Chapter 3 This validates the choice of the Can parameters In Figure 4 6 the number of detected photons from EM showers is displayed as a function of energy The number increases roughly exponentially with the muon energy while the contribution from muons alone is constant This is in agreement with our expectations based on the known equation of the muon en ergy loss see Chapter 3 The amount of Cherenkov light detected becomes considerable for the high est energy events Unfortunatly the corresponding simulation becomes CPU intensive and yields output files of a very large size To avoid this problem when simulating high energy events and in order to make the simulation CPU effective SIRENE generates a maximum number of photons on the telescope s OMs and associates a weight to each of them The photon weight is equal to 1 if the number of photons reaching an OM is smaller than the maximum of photons allowed for the simulation If the number of photo electrons detected by an OM is larger than the maximum number of photo electrons allowed the weight is set equal to the ratio of the number of photons that should have been generated to the maximum number of photons allowed The maximum amount of Cherenkov photons allowed to be generated on each OM can be set manually at the start of The saturation level of the ARS of the ANTARES telescope is 8 photo electrons in the so called single photo electron SPE detection mode 94 88 4
60. DF which is denoted by f describes the probability of observing the hits 7 given a track Puda with p the position and d the direction of the muon at a certain time If the hits are statistically independent the PDF can be expressed as the mul tiplication of PDFs for individual hits 68 N FGO TTfilvil i 1 The probability of measuring a certain hit time depends on the muon track pa rameters and therefore on how the track and the optical modules OMs of the telescope relate to each other As was discussed in Section 4 1 the PDF of a sin gle hit can therefore be described in terms of the predicted hit time the photon path length and the angle of incidence finc of the photons on the OMs The PDF can also be written as a function of a more convenient quantity the time residual ri ti ph between the observed time t and the predicted time ge of hit i The resulting PDF of all hits can thus be expressed by N FOG J TAilrilli Pine A 4 10 1 The quantity A is the predicted amplitude of the hit 7 It has been added to the list of quantities defining the PDF since a hit is characterised by its time and amplitude Signal and Background The PDF of the hit time residuals has contributions from both signal and back ground hits The PDF of the signal is often parametrised with a continuous differentiable function of the time residual which favor the use of optimisation methods based on ma
61. DFs At high energies up to E lt 10 GeV lepton quark scattering occurs at small Bjorken x x lt 1075 where no data exist In the absence of any reliable PDF parametrization in this domain extrapolations become inevitable ANIS uses two approaches founded on different theoretical models that provide an extrapolation of the nucleon structure functions to small x and large Q one is based on the CTEQ5 parametrization and the other one on a non standard hard pomeron enhanced model 76 inspired by A Donnachie and P V Landshoff 77 In Figure 3 6 the DIS neutrino nuclear cross sections for CC and NC interactions for the two high energy extrapolations are compared Starting from the neutrino nucleon interaction vertex secondary muons and taus have to be propagated further towards the telescope using a dedicated lep ton propagation program ANIS does not include such a program but the muon propagator MMC 50 can be used to propagate the secondary muons towards the telescope A description of the program MMC and its interface with the ANTARES software is given in Section 3 3 3 The detector reference system in ANIS is centered at the center of gravity of the telescope The coordinate system is right handed with the x axis directed eastward and the z axis pointing away from the Earth center By convention straight downward going vertical events are characterized by a zenith angle 0 Upward going events therefore have a zenith angle su
62. E f eeoa 3 9 In the energy range covered by GENHEN the neutrino nucleon interaction total cross section is considered to rise linearly with the energy 63 64 55 Neutrino event generators with N4 the Avogadro number g the neutrino CC cross section and pg the Earth column depth in the neutrino direction defined by the zenith angle 0 This ap proximation is reasonable for muon and electron neutrinos 54 but NC inter actions need to be taken into account when simulating tau neutrinos and their regeneration chain Therefore a full simulation of the propagation though the Earth is necessary for tau neutrinos In order to limit the time required to simulate events within a large range of energies and obtain sufficient statistics at high energies the total energy range is divided into a number of equal divisions in logio E with E the energy of the neutrino This makes it possible to define a different Generation Volume V for each energy bin i and prevent large numbers of events from being generated at low energies in an interaction volume determined by the high energy range The number of events N generated in each energy bin i is calculated by weight ing the events with the chosen neutrino generation spectrum run E tAE N E ESS Neen X frr 3 10 Emin E dE where En and El are the extremes of the energy bin i and Emin and Emax are the extremes of the entire energy range of the neutrinos to be simulated In
63. ETRY defines the position and the orientation of each object in the telescope It is derived from the classes POSITION and DIREC TION ROTATION The class ROTATION is an utility class to define a three dimen sional rotation with spherical polar coordinates C 0 2 Classes for the tracking of the particles An interface has been developed within the ANTARES Monte Carlo generators producing ASCII format output of SIRENE such that the program can use input files created with GENHEN or ANIS and MMC The program MonteCarloEvent can then be used to translate the output of SIRENE writen in the ANTARES ASCII 147 SIRENE software implementation event format into the ANTARES ROOT event format for a more efficient storage and processing of the data with the trigger TriggerEfficiency and Reconstruc tion Aart Strategy ScanFit codes The entire ANTARES library for simulation codes ANTCC was changed in order to take the photon weights calculated by SIRENE into account For the tracking of the secondary particles created at the neutrino interaction vertices and the Cherenkov photons they generate the fol lowing classes are constructed PHOTON The class PHOTON defines the light emitted by a Cherenkov radia tion source The class PHOTON is derived from the class GEOMETRY SOURCE The class SOURCE represents the type of Cherenkov light source Currently muons and electromagnetic showers are implemented as possible sources but SIRENE is design
64. Heijboer Track Reconstruction and Point Source Searches with ANTARES PhD thesis Universiteit van Amsterdam Amsterdam the Netherlands June 2004 69 P Lipari and T Stanev Phys Rev D 44 3543 1991 70 P Antonioli et al A three dimensional code for muon propagation through the rock Music 1997 hep ph 9705408 71 I Sokalski Accuracy of the muon transportation Propmu music mum ANTARES internal note ANTARES Soft 2001 005 72 J Breitweg et al Eur Phys J C 7 609 1999 73 Yu M Andreev et al Phys Atom Nucl 57 2066 1994 74 A Gazizov and M Kowalski High energy neutrino generator for neutrino telescopes 2003 Proceedings of the 28th Internaltional Cosmic Ray Con ference Tsukuba Japan 75 M Dobbs and J B Hansen Comput Phys Commun 134 41 2001 76 A Z Gazizov and S I Yanush Phys Rev D 65 093003 2002 77 A Donnachie and P V Landshoff Phys Lett B 518 63 71 2001 78 P Lipari Astropart Phys 1 195 1993 79 U Katz KM3NeT Towards a km3 mediterranean neutrino telescope 2005 Proceedings of the VLVnT2 Workshop Catania Sicilie 80 Y S Tsai Rev Mod Phys 49 421 1977 81 B D Hartmann Reconstruction of Neutrino induced Hadronic and Electro magnetic Showers with the ANTARES Experiment PhD thesis Universitat Erlangen Nurnberg Erlangen Germany June 2006 163 BIBLIOGRAPHY 82 J Brunner Cherenkov light from he electromagnetic and had
65. M which correspond to cos 1 In contrast cos 1 corresponds to the minimum efficiency of the OM i e when a photon flux hits the back of the OM which is painted black Arrival time of the photons The arrival time of the photons on the OMs is cal culated from the muon track parameters as th emit start _ fom Pemit fit start FO ter 4 6 which is the sum of the initial time of the muon track the time the muon travelled before reaching 7 and emitted light and the time it takes for the produced photons to reach the OM For ease of writing Equation 4 6 will be writen in the following as m tstart temit travel In SIRENE this expression is implemented as follows The time of emission of the photons on the muon track can be calculated as Fctosest Fstart Is Fclosest z Femit emit RE where Fojosest is defined in Figure 4 3 This expression can also be written as IFotosest Fstart _ Leos c c c temit The time the light travels in water is expressed as the ratio of the path length and the group velocity v of the photons in sea water group y Vo pP i l 1 r travel quae pater sin c 5 84 4 1 SIRENE The time of arrival of the photons on the OMs is thus calculated by 1 r 1 r thit tstart c rss TT Fstart Ee a owater x 4 7 If muons are relativistic particles and considering the Cherenkov approximation the hit time can be modelled by 1
66. PMT is protected by a magnetic shield from the Earth s magnetic field which could degrade the uniformity of its response The Earth s magnetic field would influence electron trajectories between the photo cathode and the first dynode of the PMT The magnetic shield has the form of a metal cage composed of two parts a hemispherical part which covers the photo cathode of the PMT and a flat part with a hole at its center for the neck of the PMT The hemispherical part of the magnetic shield is positioned in the gel while the flat part is mounted around the neck of the PMT The hemisphere which supports the PMT constitutes the major part of the apperture of the detector The other hemisphere is equipped with an electrical penetrator which provides the connection to the power and acquisition system outside the OM To reduce the sensitivity to photons hitting the back of the OM the inside of the back hemisphere has been covered with black paint The PMT selected for the ANTARES telescope is the hemispherical Hama matsu R7081 20 tube which has a diameter of 10 inches and a correspondingly large active area of about 500 cm This provides the OM with a large angular acceptance falling to half its maximum at about 70 from the axis The photo multiplier is sensitive to single photons in the wavelength range from 300 nm to 600 nm Its quantum efficiency is largest in the wavelength range from 350 to 450 nm and amounts to about 25 The timing resolution o
67. The B decay process of K in salty sea water can be expressed as follows d d g PREY gh a 1 3 The produced electron has a maximum energy of 1 3 MeV which is sufficient to give Cherenkov light detectable by ANTARES To study this environmental background the ANTARES telescope includes devices to monitor the deep sea parameters such as water salinity and water current in order to better understand the K and bio luminescent backgrounds 19 1 3 Detector layout The ANTARES telescope is located at a depth of 2475 m on the Mediterranean seabed The exact location of the telescope is 42 50 N 6 10 E which is approxi mately 40 km off the coast of Toulon in France The detector is monitored from an on shore control room known as the shore station in Institut Michel Pacha in La Seyne sur mer a French laboratory dedicated to research of Marine Biology and Physiology Electrical power to the telescope is transferred from the beach by a 15 The ANTARES telescope rad Shore station 3 Optical module Top buoy 60 m lydrophone Con pass O i pe S Mp o Oo luieter 2500 m i i dd 9 a electronics depth c Sd T r 1 EG od 9 N 300 m I 19 dd ji a 2 4 active d 39 79 d EE M R Oo 1 d 2 Electro optical T F f IT cable 40 km L 3e KET a I Xo 419 Electronics container T Aie P Link cables 100 m i Junction box l i Dan ja 5 SLi ij Anchor Ee Oe Acoustic beacon Figure 1 3 Sch
68. Ultra high energy neutrino simulations Front cover Inspired by Rabelais Henri Colnard Back cover Detail of a violin in faience Henri Colnard Martres Tolosane 1995 Ultra high energy neutrino simulations ACADEMISCH PROEFSCHRIFT TER VERKRIJGING VAN DE GRAAD VAN DOCTOR AAN DE UNIVERSITEIT VAN AMSTERDAM OP GEZAG VAN DE RECTOR MAGNIFICUS PROF DR D C VAN DEN BOOM TEN OVERSTAAN VAN EEN DOOR HET COLLEGE VOOR PROMOTIES INGESTELDE COMMISSIE IN HET OPENBAAR TE VERDEDIGEN IN DE AGNIETENKAPEL OP DONDERDAG 16 APRIL 2009 TE 10 00 UUR door Claudine Marie Marguerite Colnard geboren te Saint Gaudens Frankrijk Promotor Prof Dr G van der Steenhoven Co promotor Dr E de Wolf Faculteit der Natuurwetenschappen Wiskunde en Informatica A mes parents A philosopher once asked Are we human because we gaze at the stars or do we gaze at them because we are human Pointless really Do the stars gaze back Now that s a question Neil Gaiman Stardust 1999 Contents Introduction 1 2 0 1 Very high energy Cosmic Rays 0 0 2 The Greisen Zatsepin Kuz min limit 0 3 Active galactic nuclei 20 5 223 2009 418 SSO Be Ad 0 4 Discussion saei e a eare ede ee ee The ANTARES telescope 1 1 Neutrino detection principle tn od an ee Gear ge any 3 1 2 Background signal ss za oare geek anton a L3 Wetectorlay out saiad len eet ae Se ee ee 1 3 1 Detection lines an van
69. aboration A ZEUS next to leading order QCD analysis of data on deep inelastic scattering Phys Rev D 67 012007 2003 46 ZEUS collaboration An NLO QCD analysis of inclusive cross section and jet production data from the ZEUS experiment Eur Phys J C 42 1 16 2005 47 C Carloganu Muon Interactions at High Energies ANTARES internal note ANTARES PHYS 98 013 161 BIBLIOGRAPHY 48 W M Yao et al Journal of Physics G 33 1 2006 49 LA Sokalski et al Parametrization of atmospheric muon angular flux un derwater Phys Rev D 64 014016 2001 50 D Chirkin and W Rhode Propagating leptons through matter with muon monte carlo mmc arXiv hep ph 0407075v2 51 L B Bezrukov and E V Bugaev Sov J Nucl Phys 32 847 1980 52 H Olsen and L C Maximon Photon and electron polarization in high energy bremsstrahlung and pair production with screening Physical Re view 114 887 904 1959 53 J Allison et al Geant4 developments and applications IEEE Transactions on Nuclear Science 53 No 1 270 278 2006 54 A Labbate et al Genhen v6 ANTARES neutrino generator extension to all neutrino flavors and inclusion of propagation through the earth ANTARES internal note ANTARES SOFT 2004 010 55 David J L Bailey PhD thesis Wolfson College Oxford UK 2002 56 A Kouchner Possibilit d observation par le t lescope ANTARES de neutrinos de haute nergie associ s aux sursauts gamma et v
70. ace of Earth in the energy range 10 107 GeV The produced muon tracks were passed to SIRENE and processed by the 3D Trigger to search for time correlated hits in any direction requiring a minimum of 5L1 local coincidences The triggered hits were finally passed to the reconstruction program The performance of the reconstruction algorithm can be expressed by plotting the distribution of the reconstruction error a defined as the angle between the 101 Detector simulation and Track reconstruction true direction of the muon track and the direction of the reconstructed track such that ay COS l 4 ae with dy the direction of the muon and the direction of the reconstructed muon track In the next paragraphs results obtained with AartStrategy and Scan Fit are presented Results with AartStrategy In Figure 4 10 the distribution of the reconstruction errors using AartStrategy is shown As can be seen in the figure a first peak appears in the distribution around a 0 1 well reconstructed events and a second one around a 85 badly reconstructed events About 61 4 of the events are reconstructed with an error on the direction smaller than 10 and 56 4 are reconstructed within 1 from the true track When using KM3 for simulation of the detector response these numbers are smaller 68 A possible explanation is given below 0 45 0 4 0 35 dP dlog 100 0 3 0 2 0 15 0 1 0 05 al al log Reconstructi
71. active galac tic nuclei lying within 75 Mpc from Earth 13 Active galaxies are therefore also likely sources of high energy cosmic neutrinos The term active galactic nuclei or AGN refers to the existence of very energetic phenomena occuring in the core of some galaxies The amount of radiation emit ted by the central source is at least comparable to the energy radiated by all the stars in the galaxy Such galaxies are thought to possess a massive black hole at their center as high as 10 solar masses that powers their enormous energy out put Several types of AGN can be recognized based on the mass of their central engine The two main subclasses of AGN are Seyfert galaxies and quasars Both possess very luminous nuclei which appear almost starlike but with strong and broad emission lines from highly ionized gas These emission lines show evi dence for the presence of large amounts of very hot and fast moving gas accret ing around the galactic center Seyfert galaxies were the first active galactic nuclei to be identified They are named after Carl Keenan Seyfert who discovered them in 1943 Seyfert galaxies are thought to be powered by a moderate mass black hole Quasi stellar radio sources or quasars show emission lines in their spec tra which are even more prominent than for Seyfert galaxies They also present very large redshifts indicating by the Hubble law that they are at great distances CONTENTS Some of the quasars which hav
72. aded version of the MIL or Mini Instrumentation Line based on the final design of all electronics and mechanics of the telescope except for the improved connec tors described in Section 1 4 3 As can be seen in Figure 1 8 the MILOM consisted of a bottom string socket BSS and three storeys In addition to the instruments inherited from the MIL the MILOM contained four OMs a full triplet of three OMs on the second storey and an additional single OM at the top of the line The OMs were mounted on the MILOM to evaluate the telescope s response ahead of full production of the twelve detection lines which compose ANTARES With the MILOM also the modified electro magnetic cable EMC of the telescope could be tested in situ It was also used to monitor the deep sea environment and test the acoustic positioning and time calibration systems of the telescope The line was equipped with three sources of intense light a laser beacon and two LED beacons attached to the first and top storeys to illuminate the OMs and determine their relative time calibration In Figure 1 8 the layout of the MILOM is represented Following its deployment in March 2005 data taking started with the MILOM right after the connection of the line to the junction box JB on 12 April 2005 The raw data from the OMs could be analysed immediately thus demonstrating the success of all design changes The MILOM line was in operation until April 2007 24 1 4 Development and C
73. al storey This starts the digitization of all the SPE hits corresponding to a LO trigger that have been memorized in the pipeline cells of the ARS chips On shore the trigger software searches for position and time correlations be tween all hits that were selected by the L1 trigger A decoding algorithm converts the raw data of the ARS into calibrated time charge and geometry information During data taking calibration parameters of the detector are obtained from a ta ble in a database where the measured parameter values of the hardware devices are stored see Section 2 2 3 The geometry of the telescope is described in terms of positions and orientations of the PMTs The method used to correct the time information is explained in Section 2 2 3 The time and position of hits created by unscattered Cherenkov light from the same muon path are correlated This is called causality 24 A group of causally related L1 hits form a so called cluster When the trigger algorithm finds a sufficient number of correlated L1 hits within a cluster in any of the directions it is assumed that a muon signal is present in the data The data that corresponds to a group of correlated hits found by the trigger algorithm is referred to as an event Only events are stored on disk for off line analysis This reduces the amount of data by a factor of about 104 Events are written as simple ASCII files The basic concept of the used data format is explained in an ANTARES i
74. alidation des techniques de d tection a l aide d un prototype PhD thesis Universit Paris VII Paris France April 2001 57 L Nyhoff FORTRAN 90 for Engineers and Scientists Prentice Hall 1996 58 S Agostinelli et al Geant4 a simulation toolkit Nuclear Instruments and Methods A 506 250 303 2003 59 G Ingelman et al Lepto 6 5 A monte carlo generator for deep inelastic lepton nucleon scattering Computer Physics Communications 101 108 134 1997 60 G Barr PhD thesis Oxford University Oxford UK 1987 61 H Plothow Besch Pdflib Proton pion and photon parton density func tions of the nucleus user s manual 2000 CERN computer program library entry W5051 62 K Kuzmin et al Genhenv6r3 implementation of the glashow resonance and of the music transport code ANTARES internal note ANTARES SOFT 2004 012 162 BIBLIOGRAPHY 63 David J L Bailey Genhen v5r1 Software documentation ANTARES inter nal note ANTARES SOFT 2002 004 64 B R Martin and G Shaw Particle Physics Wiley 1997 Second Edition 65 J Brunner Updated tag list for the new ANTARES event format ANTARES internal note ANTARES SOFT 1999 003 66 T Montaruli and A Romeyer Conventional and prompt atmospheric neu trino fluxes ANTARES internal note ANTARES PHYS 2001 015 67 David J L Bailey Genhen v3rl Further optimisation for large energy ranges ANTARES internal note ANTARES SOFT 2000 6 68 A
75. an indirect measure of the muon energy As it depends on the position of the track relative to the telescope it contains position and direction information and can be used to estimate the muon energy without having to rely on a full geometrical reconstruction To include the simulation of the front ends electronics the total charge of the hits has been chosen as an estimate of the muon energy for the data analysis It was found that the atmospheric background can be rejected by excluding events with low charge values and vertical directions A model rejection potential method has been applied to optimise the experimental cuts and place the strongest constraints on the simulated AGN like diffuse neu trino flux The optimised selection criteria lower the effective area for neutrinos at energies below 1016 eV as compared with the standard effective area of the ANTARES telescope However the effective area is enhanced at higher energies up to 10 m at 10 eV The obtained effective area is comparable to the one of the neutrino telescope AMANDA II for the same flux model and energy range 111 An estimate of the average flux upper limit that can be obtained with AN TARES for the simulated AGN like diffuse muon neutrino flux has been placed at the 90 confidence level at Ep lt 12 x 10 GeVem sier after 1 year 5 2a Een v S41 x 10 GeVcm s ter after 3 years 5 2b This limit is valid for neutrino energies between 10 eV and 10 eV It
76. angle and total charge a EEN eo at Te 6 ada ae A 125 Distributions of the cosine of the zenith angle and the total charge of the hits for downward going ultra high energy muons and at mospheric muons using optimised cuts o ooo 126 Comparison of the ANTARES upper limits at ultra high energies with various experimental bounds on diffuse muon neutrino fluxes 128 Muon neutrino effective areas for UHE events in ANTARES 130 Final energy distribution of 100 TeV muons propagated through 300 meters of ANTARES Water simulated with MMC 143 Same as the right part of Figure B 1 but using the continuous ran domisation option tu sAnne od ee ee GY 144 155 LIST OF FIGURES C 1 Schematic overview of the classes of the detector simulation pro gran SIRENE eerdere ebr te Saye ac he Wh es wae a Soe ot et ae ee 156 List of Tables 3 1 CPU usage of ANIS and GENHEN 3 as nas 234983624 4 69 B 1 Chemical composition of the sea water at the ANTARES site 142 157 LIST OF TABLES 158 Bibliography 1 A A Penzias and R W Wilson Astrophys J 142 419 1965 2 J Learned and K Mannheim High energy neutrino astrophysics Annual Review Nuclear Particle Science 50 679 749 2000 3 C Amsler et al Physics Letters B 667 1 2008 4 E Fermi Phys Rev 75 1169 1949 5 A M Hillas Ann Rev Astron Astrophys 22 425 1984 6 D J Bird et al Astrophys J 424 491 1994 7 J Abraham et a
77. are compared one by one with the corresponding histograms from the reference file The Kolmogorov Smirnov test is one of the most useful and general nonpara metric methods for comparing two data samples as it is sensitive to differences in both the location and the shape of the empirical distribution functions of the two data sets The Kolmogorov Smirnov test is accurate but it is more sensitive at points near the median of the distribution than on its tails The Kolmogorov Smirnov test is explained in Section 2 3 2 The bin by bin comparison is based on the number of standard deviations by which the data from a histogram deviate from the data of a reference histogram The error per bin i of an histogram is computed as the square root of the bin s contents xj Of SAGs The standard deviation which is a measure of the spread or dispersion of a set of data is then defined as Pee i In the bin by bin comparison test the standard deviation is deduced from the reference histogram which is considered a perfect description of the true distri bution The difference between the bin contents of a histogram and that of the reference histogram is calculated for each bin and compared with the standard deviation 7 As soon as the difference is significant for one bin the two his tograms are considered incompatible 2 3 2 The Kolmogorov Smirnov test The Kolmogorov Smirnov test is based on the work of mathematicians Andrey Nikolaevich Kolmo
78. arger charge up to about 10 photo 117 High Energy Simulations Results electrons These observations suggest that the atmospheric background can be rejected by excluding events with vertical directions and relatively low charge values Cut on the zenith angle In figure 5 7 the distribution of the cosine of the muon zenith angle is shown for both signal and background events at the trigger level As can be seen in the figure atmospheric muons are peaking in the vertical direction and dominate over the astrophysical neutrino signal over almost the entire zenith angle range with the exception of horizontal events in the range 0 2 lt cos lt 0 A loose cut is placed on the zenith angle at cos gt 0 6 to reject the atmospheric vertical events while most signal events are kept Cut on the total charge of the hits In Figure 5 8 the distributions of the total charge of the hits induced by ultra high energy UHE signal events and atmospheric background events are shown at the trigger level As for the muon energy spectrum the distribution of atmo spheric events has been extrapolated beyond 104 photo electrons p e with a power law spectrum E following the primary cosmic ray spectrum The dis tribution has thus been estimated up to 10 p e As can be seen in the figure the distribution of UHE muons extends to a larger induced charge in the photo multiplier tubes PMTs of the telescope than that of atmosphe
79. ata Do not change unless you know what you are doing CC cteq5 cteq5_cc_nu data cteq5 final_ctegq5_cc_nu data 101010 110 CC cteq5 cteq5_cc_nubar data cteq5 final_cteq5_cc_nubar data 10101 110 NC cteg5 cteg5_nc_nu data cteq5 final_cteqb5_nc_nu data 101010 110 NC cteq5 cteq5_nc_nubar data cteq5 final_cteqb_nc_nubar data 10101 110 GR dummy dummy 10000 1 nannan tt AH The generated events can be re weighted to any relevent power law spectrum such as diffuse AGN like or other ultra high energy UHE neutrino and anti neutrino flux predictions The second value that follows the r run parameter in the steering file stands for the output format to be used by ANIS The AMANDA IceCube 2000 ASCII format is used for ANTARES as the lepton propagator MMC which is interfaced with ANIS is using this specific format 140 Appendix B Implementation of ANTARES in MMC The various media implemented in MMC are assumed to be concentric spheres centered at the center of the Earth each having different densities These spheres are surrounded by a main medium which fills all space The Earth s radius is assumed to be Rr 6 3713 x 10 meters the bedrock lies at b 2475 me ters below sea level at the ANTARES site No additional bed elevation zo 0 of the rock is assumed The center of gravity of the ANTARES telescope is at a depth D 2 196849 x 10 meters The sea water above the detector cen ter has a smaller density 1 0291
80. ata selection and analysis T ae a eoo00000 on 5 vt EEE Et yp we EL os EL EEE e S Ha mha aa a is oo o of QOQoOoaqaaQ a oo ooOoaoQhQadnhadgcadaa 0 o o as 2 o 0D oO 000000 0 0 oo o 2 o 2a OOo QOoQYaaodboaoo a o o o o D o O De O OQ o 0 O0 o 0 0 oO O o a 2 O o o 0 o0 ao 0 0 a a Q n a a tT 1 2 3 4 5 6 7 8 9 10 11 log rue muon energy GeV ze of Oo 2 o 6 5 oO oF 0 0 O O o D O 0 0 O O a ao a 5 OOODODOd 2 2 2 3 noOOoOgooaa z B OEE El elk lts 48e 2 e a gt e mua E E f eooo doos 000 JDOee Ld Een En 1 2 3 4 5 6 7 8 9 10 11 eg muon energy GeV Figure 5 4 Distribution of the total charge of the hits produced by one event as a function of the true muon track energy The energy distribution is assumed to follow an AGN like spectrum bottom The same distribution is also shown for a generic E spectrum top for illustrative purposes Downward going events simulated with SIRENE are shown at the trigger level 115 High Energy Simulations Results larger than about 10 photo electrons p e The performance of the PMT thus limits the readout of a large charge However the true muon energy is still a determining factor in the observed charge The total charge of the digitised hits is therefore chosen as an estimate of the muon energy without having to rely on track reconstruction 5 2 2 Uncorrelated background and triggering ANTARES does not only record hit
81. ation 150 Acknowledgements These years in Amsterdam have been a very unique experience both from a pro fessional and a personal points of view Iam very grateful I was given the chance to carry out my doctorate in the ANTARES group at Nikhef It has really been a pleasure to work in such a good and inspiring environment I would like first to thank my promotor Gerard van der Steenhoven and my co promotor Els de Wolf for their precious advices and guidance during my PhD You have pushed me to achieve more than I would have never done otherwise I might repeat myself x2 but thank you for your help and support I also wish to thank our team leader Maarten de Jong Thank you for your time whenever I droped by your office I learnt a lot by working with you es pecially on the time calibration and the monitoring of ANTARES and my C programming skills have dramatically improved Thank you very much to Paul Kooijman for helping me develop the software program SIRENE and for being there whenever I had a question on the related physics topics and my study of ultra high energy neutrinos I would like to give special thanks to Marek Kowalski and Dmitry Chirkin for allowing me to use the software programs ANIS and MMC Thank you also to Teresa Montaruli for helping me to learn the GENHEN software program Your many emails have guided me during my work on ultra high energy neutrino generators and lepton propagators I also had the opportunity to vi
82. ation constant Wyorm the interaction probability in the Final Volume P and two weights R and Ro that correspond to the atmospheric neutrino flux for electron and muon neutrinos A detailed description of these weights and a tutorial on how to use them can be found in 74 Here a brief de scription of these weights is given The normalization constant Wrorm expressed in GeV cm sr s year is computed as Io x Ig x F X Agen Worm N gen where 63 Neutrino event generators Neen is the total number of generated neutrinos Agen in cm is the surface of the cross section of the Final Volume with the detector instrumented volume F is the number of seconds per year Ig is the energy phase factor It is defined as 1 Ygen 1 ygen Emax Emin Ip In gt if Ygen 1 Ir is given in units of giga electron volt GeV Ig is the angular phase factor It is defined as Io 271 cOS Omax COS Omin Ig is given in units of steradian sr The normalization constant is chosen such that the event rate per year R is given by i Neen R Whorm y Pinti 2 i 1 The probability of interaction P dimensionless inside the Final Volume is de fined as Pint 1 exp Otot x T with to the neutrino total cross section for the event and t the column depth The atmospheric weights R77 and Rj dimensionless are defined as atm atm Rve vm Pe E atm Foon Ee with Fytmo E 0 GeV 1 em srm s
83. automatically compares the recorded histograms with a set of predefined reference histograms If the histograms are not compatible an error message is displayed on the screen and further data analysis will be necessary to identify the cause of this result The Histogram Presenter features a control panel that allows the analysis of histograms created by the Histogrammer and the Comparator The Histogram Presenter which can operate on any set of one dimensional histograms is described in detail in a separate internal note 32 38 2 3 The Data Quality Monitoring System TVC1 min B oR 8 5 TVC2 min Run number Run number x 220 220 g 216 E 216 5 210 N 210 2 206 206 200 200 196 196 190 190 186 186 100 10 16 20 26 180 9 5 10 15 20 Run number Run number Figure 2 4 Time stability of the dynamic ranges of the ARS TVC Top TV C min spectra of the first and second TVC slope respectively Bottom TV Cmax spectra of the first ramp and second TVC slope respectively Data were taken from runs 11236 through 11262 with the MILOM In the remaining of this section the other three modules are described in Sec tion 2 3 1 and the principles of the so called Kolmogorov Smirnov test used in the Comparator are given in Section 2 3 2 2 3 1 Software Modules Histogrammer The Histogrammer module consists of several programs dedicated to the physics quantity of interest The readout system of ANTARES consists of ARSs chips which digitiz
84. be made in the theoretical considerations of the process to obtain convenient and simple formulae for the cross section The most commonly used are the expressions given by Bezrukov and Bugaev 51 or 51 Neutrino event generators Borog and Petrukhin 52 which lead to results agreeing within 10 for the dif ferential cross section and within about 5 for the average energy loss provided that the same photo nuclear cross sections are used in the calculations 53 The total energy loss is determined by the summation of all individual contri butions The average rate of muon energy loss can be described by lt a E b BE 3 8 with a E the muon energy loss by ionization and b E the summation of brems strahlung photo nuclear interaction and pair production Below 1 TeV the first term dominates As can be seen from Equation 3 8 it is approximately indepen dent of the muon energy Above 1 TeV the second term of the equation becomes dominant where the energy loss is proportional to the energy of the muon 3 2 GENHEN The Monte Carlo event generator GENHEN generates high energy neutrino in teractions in the media surrounding the instrumented volume of the ANTARES neutrino telescope An extensive description of the lastest version of the pro gram can be found in an ANTARES internal note 54 and a Ph D thesis 55 In Figure 3 3 an example is shown of a diffuse AGN like flux of muon neutrinos as suming E 10 GeV cm s sr
85. bed in Section 3 3 Both ANIS and MMC 45 Neutrino event generators were primarily designed for use within the AMANDA collaboration but the pro grams have been adapted for ANTARES A comparison of the performance of the two neutrino generators is presented in Section 3 4 3 1 Event Generators and Propagators 3 1 1 Neutrino Generators Neutrino interactions of interest for ANTARES Neutrino generators are used to simulate neutrino interactions with nuclei and atomic electrons for each of the three neutrino flavors vy ve and vz Both chan nels of the neutrino nucleon weak interaction charged current CC and neutral current NC interactions are supported Over the energy range of interest for ANTARES the interactions of ve vy and T with electrons can generally be neglected in comparison to the neutrino in teractions with nuclei since the corresponding cross sections are much smaller than the neutrino nucleon cross sections at high energies me lt my The reac tions 7e gt WT p and ve W hadrons form important exeptions as an intermediate W boson is produced which decays in detectable secondary particles 34 This channel dominates over the 7 nucleon scatterings in a nar row region around the energy of the resonance E 6 3 10 GeV and needs to be taken into account by neutrino generators for high energy neutrinos and anti neutrinos The resonant scattering of anti neutrinos is named after physi cis
86. been made that v Ve vr 1 1 1at Earth 128 5 2 Data selection and analysis interesting as it probes the sensitivity of the current and next generation of neu trino telescopes Upper limits have also been reported by several existing neutrino experi ments assuming the same benchmark E dependence In Figure 5 14 the 90 confidence level CL upper limits expected after one and three years of oper ation with ANTARES for the diffuse flux of ultra high energy muon neutrinos are shown as a function of the neutrino energy The most competitive experi mental upper limits at the 90 CL for diffuse muon neutrinos are shown as well for comparison The theoretical limit calculated by Waxman and Bahcall is also displayed The sensitivity of ANTARES expected after one year of operation for diffuse UHE neutrinos is about a factor eight above the Waxman and Bahcall bound After three years of data taking the sensitivity is enhanced to about a factor two above the Waxman and Bahcall limit As can be seen from the figure the presented analysis predicts a sensitivity after one year of observation which is about a factor four above the upper limit placed by the AMANDA II 111 tele scope on the diffuse flux of UHE muon neutrinos After three years of operation the sensitivity of ANTARES improves to about a factor of two below that same upper limit This upper limit has been determined by AMANDA II using data from three years 2000 to 2002 with a
87. bly from acceleration in the shock waves of supernovae rem lTron nuclei are the heaviest abundant nuclei observed in cosmic rays 0 2 The Greisen Zatsepin Kuz min limit nants 3 The knee could thus feature the highest energy that galactic accelerators can reach Cosmic rays with energies above the knee cannot be confined by the magnetic fields within the Galaxy They are thus believed to propagate through intergalactic space 6 what induces a steepening in the cosmic ray energy spec trum In the region around the ankle the galactic population of cosmic rays is overcome by the extragalactic population of cosmic rays 3 The ankle is conse quently considered to reflect the transition from galactic to extragalactic cosmic rays Above the ankle cosmic rays are difficult to detect since they hit the Earth at a rate of less than one per square kilometer per century Experimental evidence suggests that such high energy cosmic rays are mostly protons accelerated in ex tragalactic sources 7 Among the very few likely candidates 8 for these ultra high energy cosmic ray sources are active galactic nuclei AGN and gamma ray bursts GRBs in which also neutrino production may occur Moreover the asso ciated sources should be confined within a few 10 Mpc of the Earth to contribute to the observed cosmic ray spectrum due to the Greisen Zatsepin Kuz min or GZK limit 9 10 0 2 The Greisen Zatsepin Kuz min limit In 1965 Kenneth Greisen
88. bundles impinging on the ANTARES de tector surface has been simulated with the program MUPAGE 100 or MUon GEnerator from PArametric formulas MUPAGE is a parametrisation of the at mospheric muon flux at the depth of the detector simulating single and multiple underwater muons between 20 GeV and 500 TeV and up to 85 zenith angle All kinematic parameters of the muons are tuned with a Monte Carlo simulation of primary cosmic ray interactions and shower propagation in the atmosphere based on the program HEMAS 101 It allows the calculation of the high energy muon component E gt 500 GeV in extensive air showers EAS assuming a pri mary total energy between 10 and 107 eV Hadronic interactions are handled with DPMJET 102 which embodies a phenomenological model 103 using re sults from direct and indirect measurements of cosmic rays in the energy range between 10 GeV and 1 EeV Direct observations are used to extrapolate the energy spectra of each element to high energies The muons which could reach the sea 112 5 2 Data selection and analysis level are then transported through water towards the telescope with the lepton propagator MUSIC 70 see Chapter 4 A total number of 10 downward going atmospheric events consisting of bundles of at most 1000 muons have been generated with MUPAGE in the energy range 20 GeV 500 TeV at the default ANTARES Can see Chapter 4 for zenith angles with 1 lt cos lt 0 087 The
89. can mode Figure 2 6 Sample picture of the ANTARES Histogram Presenter window 2 3 The Data Quality Monitoring System distributions of data The Kolmogorov Smirnov test is a non parametric stas tistical test used to determine whether two independent distributions differ or whether an underlying distribution function differs from a theoritical distribu tion in either case based on finite samples Non parametric tests are often used instead of their parametric counterparts when certain assumptions about the un derlying population are questionable For example the Kolmogorov Smirnov test does not assume that the underlying population is normally distributed where as its parametric counterpart the x test does The two sample Kolmogorov Smirnov test is used to compare two continu ous distributions Sn and Sm with respectively n and m observations in order to determine whether the two independent data samples are compatible i e that they originate from the same distribution The computation of the test involves two empirical distribution functions Sy an Sm The empirical distribution func tion Sy x is a function of x which is equal to the fraction of data points that are less than or equal to x for each x R For N independent data points y1 y2 yn ordered from the smallest to the largest value Sy x is defined as Sn O lt x lt y 0 Sny Sx lt y Sn x gt yn el and hence Sy x can be calculated using wher
90. ce histograms for further usage by the Comparator The installation procedure of the soft ware package and the system requirements are described in detail in 32 which also contains a description of the main subpanels and buttons of the Histogram Presenter control window Via the graphical user interface histograms can be imported from a file The user can browse through system directories and file structures to select the histograms and display them on the Histogram Presen ter main window A sample picture of the Histogram Presenter graphical inter face is shown in figure 2 6 Once the histograms have been accumulated they can be edited via a graphical editor A file selection dialog permits to save the histograms as a picture in any format A print option is also available The dis 40 2 3 The Data Quality Monitoring System played histograms can also be scanned to provide numerical information about the selected histograms Comparator The main purpose of the Comparator is to compare histograms produced by the Histogrammer package with a set of reference histograms The program loops over all the input histograms and searches for the corresponding histograms in the reference file taking into account the name dimension binning and number of channels If the histograms are consistent two different methods can be ap plied for the actual comparison the Kolmogorov Smirnov test and a bin by bin comparison Histograms from the input file
91. ch that 1 lt cos lt 0 3 3 2 Generation method Geometry In ANIS neutrinos that survived the propagation through the Earth are simu lated to interact within a rotating cylinder whose z axis is parallel to the neutrino direction which is referred to as the Final Volume The Final Volume is centered 61 Neutrino event generators log 0 E cm 1 23 465 6 7 8 9 1011 12 log E GeV Figure 3 6 Differential neutrino nucleon cross sections in cm used by ANIS Both NC and CC v N interactions are represented At high energy ANIS uses an hard pomeron enhanced model HP or an extrapolation of the CTEQ5 paramatrization pQCD Both neutrino solid line and anti neutrino dashed cross sections are shown 74 The reso nant Vee reaction is also shown on the center of gravity of the telescope and two heights are defined a posi tive height in the target region of the original neutrino nucleon interaction and a negative height in the detector region The size of the interaction volume is op timized by extending the positive height in the direction of the neutrino nucleon interaction using the maximum muon range at the considered energy currently this option is only available for muon neutrinos The Final Volume is an auxil lary concept used in the generation method Since it is defined for each event it makes the generation process dynamic but the final results should not depend on it Algorithm Initially the
92. ctra to be not continuous for high Veut software see Chapter 3 The ANTARES event ASCII format makes use of tags defined to make the Input Output IO of the Monte Carlo data compatible with the ROOT framework which allows to analyse the data in a more efficient way A detailed description of the ANTARES event tags which exist for Monte Carlo simulated data can be found in the internal note 65 144 Appendix C SIRENE software implementation SIRENE runs on a UNIX platform and consists of a library of C classes 1 that enables the implementation of different sources of Cherenkov photons differ ent scattering models for the transparent medium surrounding the telescope and various detector geometries The program assumes that a neutrino tele scope is a collection of modules housing one or more photo multiplier tubes PMTs which allows any neutrino detector geometry and any configuration of the photo multiplier tubes inside these modules Sometimes for example in the case of ANTARES these modules are referred to as optical modules see Chapter 2 Figure C 1 shows a schematic overview of the detector simulation program SIRENE The most important C classes which compose the program are shown as well as their hierarchy These classes are described one by one in the next sub section C 0 1 Classes to build the geometry of a telescope The geometry of a Cherenkov neutrino telescope is determined by the number of modules housing p
93. detector simulation program SIRENE has been used to determine the resulting photon hits in the telescope In Figure 5 2 dP dO 107 10 1 08 06 04 02 0 02 04 06 08 1 Cosine of the Zenith angle O Figure 5 3 Distribution of the cosine of the zenith angle for atmospheric muons which could reach the detector The events have been simulated using MUPAGE In the ANTARES reference frame a zenith angle of 0 corresponds to an upward going muon while an angle of 180 corresponds to a downward going muon the event rates are shown with a dashed line as a function of the muon energy for those events which could produce a detectable signal As can be seen in the figure the distribution drops rapidly with increasing energy In Figure 5 3 the distribution of the cosine of the zenith angle is shown for at mospheric muons which could produce a detectable signal As can be seen in the figure most atmospheric muons reach the detector from the vertical downward going direction Such a spectrum is limited to an angular range of approximately cos lt 0 10 5 2 Data selection and analysis A dedicated data analysis method needs has been developed in order to reduce the amount of atmospheric muon background while preserving sensitivity to the 113 High Energy Simulations Results ultra high energy UHE signal Using this analysis method an estimate can be made of the upper limit that can be achieved with ANTARES on the as
94. downward going neutrinos of energy beyond 1016 eV can be observed This is of relevance not 0 4 Discussion only for the ANTARES telescope but also for the future cubic kilometer scale de tector KM3NeT It will provide information for the optimisation of the geometry of the telescope and the technologies required Rather than searching for neutrinos from a specific source a non localised flux of high energy neutrinos has been considered Such an approach gives an higher probability of detection in the case the neutrino flux from individual sources would be too small to be detected by a neutrino telescope of the size of ANTARES As muon neutrinos have an higher detectability the diffuse flux of muon neutri nos from AGN has been considered This neutrino flux can be modelled by a generic E spectrum with units of 10 GeV cm s sr 2 A description of the ANTARES neutrino telescope and the detection principle involved are given in Chapter 1 A chronological analysis of the development and construction of the telescope is presented in this chapter The expected performance of ANTARES in the high energy range relies on a good timing resolution between the signals recorded in the photo sensors of the telescope and a good functioning of the various mechanical and electronic components of the telescope A time calibration method and a diagnostic tool to determine whether components are functioning as expected during the opera tions have been dev
95. e astrophysical objects such as active galactic nuclei In the framework of this thesis only neutri nos from accelerated cosmic rays will be considered The predicted mechanism of production for such neutrinos is described below 0 3 Active galactic nuclei While confined by the high magnetic fields at the acceleration site some frac tion of the cosmic rays suffer resonant pion photoproduction with the ambient photons as in equation la and 1b This leads to electron and muon neutrinos through the decay of the produced charged pions mt ut Huy gt et e Vy Vy 2a T gt y Vy gt e e Vy t Vy 2b and the free neutron nopte h 2c According to the equations 2a 2b and 2c the flavour ratio of the high energy cos mic neutrino flux is typically Py Pv Pv 1 2 0 at the source where Q is the combined flux of neutrinos and anti neutrinos for the flavour l After prop agation over cosmological distances a flavour ratio of fy bv bv 1 1 1is observed at Earth due to the phenomenon of neutrino oscillations 16 All three flavors of neutrinos are therefore initially of the same importance for detection However the probability of detection of muon neutrinos is higher due to the long path length of the muons they produce in interaction with matter 0 3 Active galactic nuclei The Pierre Auger Observatory has recently reported a direct correlation between the highest energy cosmic rays observed and the presence of nearby
96. e attributed to the use of simple Gaussian PDFs except for the final fit by Scanfit However the ScanFit algorithm is less accurate than Aart Strategy since it has an angular resolution of almost 0 5 instead of 0 1 This is a known behaviour of the pro gram 21 However AartStrategy performs better in terms of angular accuracy The same remarks concerning the PDFs and the various steps used in the recon struction procedure which were made for AartStrategy also apply to ScanFit The algorithm should be further tuned for a better performance with the reconstruc tion of events generated with SIRENE Moreover ScanFit is not suitable for ultra high energy simulations An high energy reconstruction program needs to be 105 Detector simulation and Track reconstruction dP dlog 19 oO oO oO 2 o Pp ee wR 2 Ui N Ui w Ui MN w Tette Pa Eeeh eka E eE E E 2 2 O a Ke log Reconstruction error a 10 Figure 4 13 Distribution of the reconstruction error on the direction of the muon ob tained for events simulated with SIRENE and reconstructed using the ScanFit procedure Only upward going muons are shown developed for use with SIRENE above 107 GeV Energy Reconstruction results In Figure 4 14 the distribution of the reconstructed muon energy is shown as a function of the true muon energy for events simulated with SIRENE Muons from muon neutrinos were generated by ANIS in the energy range 10 10 GeV fol l
97. e been observed so far are about 10 billion of light years away The fact that they are visible at such distances implies that they emit enormous amounts of energy and can not be stars in our Galaxy Because of these observations quasars are believed to be active galactic nuclei with a high mass black hole The radiation cannot be produced by the supermassive black hole itself as it is invisible but by the surrounding interstellar gas accreting onto it Observations show that narrow beams of energetic particles are ejected in opposite directions from this accretion disk Even though uncertainties remain the mechanisms in volved in the production and acceleration of the jets are most likely due to ac celeration in the ambient magnetic fields The gas is attracted by the black hole and as it slowly spirals towards the center of the galaxy its gravitational poten tial energy is converted into thermal energy The thermal energy can accelerate jets of material from the accretion disk to relativistic speed Among the radiation products of the jets are protons which can interact with the ambient radiation in the AGN to give neutrinos according to the equations 2a and 2b With excep tional gravitational forces in the vicinity of their central massive black hole active galaxies are believed to possess the tremendous amounts of energy needed to ac celerate cosmic rays and produce neutrinos up to the highest energies 0 4 Discussion The main topic of t
98. e hits A total number of N 20 io oO gt 10 is E o H a pE wn 8 L Ka g 5 o gt L aa 1E EE nr 1 08 06 04 02 0 02 04 06 08 1 Cosine of the Zenith angle S 105 gt Iss E 5 C a L n 8 L Ss ken TE A E o ES gt i ical Ls 101 E 107 E E L E E er aa a m N w 4 5 6 7 log l gEstimated total charge p e Figure 5 13 Distributions of the cosine of the muon zenith angle top and the total charge in p e of the digitised signal bottom induced by downward going ultra high energy UHE neutrinos using optimised cut selections as defined in the text The astrophysical neutrino flux corresponds to an AGN like spectrum signal events remain and no background N 0 This corresponds to an average upper limit of 2 44 events for one year and 0 81 events per year for three years of observation with the ANTARES telescope 126 5 2 Data selection and analysis 5 2 6 Average flux upper limits The average flux upper limit or sensitivity of ANTARES to the diffuse flux of UHE muon neutrinos has been determined by scaling the assumed AGN like neutrino flux by the ratio of the obtained average upper limit flg9 N to the expected sig nal Ns P Ev oo P Ev x foe The predicted average flux upper limit regardless of systematic uncertainties is therefore given by E lt 1 2x10 GeVem s sr aft
99. e n x lt x is the number of points for which x lt x The empirical dis tribution is a step function that increases by at the value of each data point Kolmogorov suggested in 1933 to estimate the largest distance D ym between the two empirical distribution functions Sy and Sm by measuring the difference in the y direction in order to compare the two continuous distributions Sn and Sm If the two samples have been drawn from the same original distribution the em pirical distribution functions of both samples may be expected to be close to each other This is the null hypothesis Ho If the two empirical distributions differ the samples come from different populations This is the alternative hypothesis H4 Ho Sn x Sm x for all x ER Hi SN X SMm X for at least one value of x The distance or test statistic Dym between Sy and Sy for each x is the maximum deviation between the two integrated distribution functions It is defined as DNM pmax Sn xi Sm xi l 43 The ANTARES data acquisition system Normalized to the total number of entries the test statistic becomes NM TNM NEMOM l The null hypothesis Hp is rejected if the test statistic Tym is smaller than a chosen critical value a also called the significance level which defines the sensitivity of the test Assuming the null hypothesis Ho the probability that the Kolmogorov Smirnov test statistic exceeds a value z at t lt 0 converges for large values o
100. e the charge and the time of the analogue signals of the photo multiplier tubes PMTs The Histogrammer can process the different data types produced by the ARS Examples include the values of the time stamp the time to voltage converter TVC and the amplitude to voltage converter AVC of the physics events The Histogrammer builds one dimensional histograms for each of the data types produced by ANTARES For further analysis the histograms are stored in separate files The main directory in which the histograms are stored is divided into subdirectories that correspond to the different data types of 39 The ANTARES data acquisition system Physics data Quality criteria Physics data Histogrammer Define reference histograms Make histograms BS Rerarence histograms histograms g Reference compe histograms histograms g histograms Level of comparison Histogram Presenter Display histograms User criteria Figure 2 5 Diagram showing the data flow between the different modules of PhAntOM ANTARES These directories are subdivided further into subdirectories one for each variable of interest Histogram Presenter The Histogram Presenter is a stand alone program which can be used to manipu late one dimensional histograms using a graphical user interface GUI The user is free to use the program on any file containing one dimensional histograms The Histogram Presenter can also be used to identify a set of referen
101. eared with a Poisson Gaussian the Poisson distribution becomes normal for large npe distribution to take into ac count the statistical fluctuations between the different wavelength domains The Gaussian is used when the total number of npe is larger than the saturation level of the ARS of the photo multiplier tube PMT inside the OM The result rep resents the total number of photo electrons that actually need to be generated by the simulation Wavelengths and arrival times of each photon are pulled randomly from the generated wavelength and arrival time distributions of the emitted Cherenkov photons The positions and directions of the photons on the virtual module are also randomly sampled The output of the program is a set of photo electrons with their position direction energy and arrival time at the different OMs in the telescope Results The simulated number of detected Cherenkov photons per OM is shown in Figure 4 5 as a function of the distance of closest approach of the muon track to the hit OM As can be seen from the figure at low energies E lt 1 TeV the contribution from muons is still relatively important as compared to that of the EM showers Above this energy the number of photons from showers starts to dominate This is in agreement with the muon energy loss behaviour see Chap ter 3 Roughly 90 of the hits originate from within 300 m of the track which is inside the limits of the default Can volume of ANTARES see
102. ed such that new sources can be easily added EMSHOWER The class EMSHOWER defines a Bremsstrahlung initiated elec tromagnetic shower All its energy is deposited over a short distance and it is considered point like The class EMSHOWER is derived from the classes GEOM ETRY and SOURCE TRACK The class TRACK is a particle track with a starting time and energy The class TRACK is derived from the class GEOMETRY MUON The class MUON defines a muon A muon is characterised by a straight track losing energy and emitting Cherenkov light Stochastic losses which pro duce independent electromagnetic showers are implemented The class MUON is derived from the classes TRACK and SOURCE MODULE HIT The class MODULE HIT corresponds to an ABSTRACT MOD ULE with the list of all the photons that hit the corresponding virtual sphere The class MODULE HIT is derived from the class GEOMETRY C 0 3 Classes to model the scattering of the particles For the scattering of the Cherenkov photons inside the medium surrounding the instrumented volume of the neutrino telescope the following classes are imple mented MEDIUM The class MEDIUM defines the properties of the detection medium to include the effect of the medium on light propagation 148 ANTARES WATER The class ANTARES WATER represents the scattering model and water properties at the ANTARES telescope site The class ANTARES WA TER is derived from the class MEDIUM 149 SIRENE software implement
103. eloped in the framework of this thesis They are presented in Chapter 2 The readout system of the telescope is also described in this chapter The sensitivity of ANTARES to ultra high energy cosmic neutrinos has been estimated using Monte Carlo simulations Three distinct stages of simulation can be distinguished The first one consists in the generation of neutrino interactions in the water and seabed around the telescope The propagation of the resulting secondary leptons towards the telescope needs then to be modeled Finally the simulation of the detector response to the Cherenkov light induced in the photo sensors of the telescope by the relativistic moving secondary leptons is computed by a specific detector simulation Since the available simulations accept neutrino events within a limited range of energies only improvement of existing packages and the development of new programs were required The implementation of the neutrino generator ANIS and the lepton propagator MMC in the ANTARES Monte Carlo event simulation chain is described in Chapter 3 Both programs were initially designed for use with the AMANDA neutrino telescope but they have been adapted to ANTARES in order to generate ultra high energy neutrino events A comparison between ANIS and the program GENHEN which is gener ally used with ANTARES is also presented A new simulation program named SIRENE which is capable of modeling the detector response for neutrino events up to the hig
104. ematic view of the ANTARES telescope showing the main components of the neutrino detector 40 km long deep sea cable referred to as the main electro optical cable MEOC which also provides data transmission from and to the shore using optical fibers At the ANTARES site on the bottom of the sea the cable is connected to a junction box JB distributing the power to the detection lines which constitute the neu trino telescope The junction box JB was the first object that has been deployed in December 2002 The ANTARES telescope itself consists of a three dimensional array organised in twelve vertical independent detection lines carrying a total of 900 photo multiplier tubes operating at depths from 2375 to 2000 m A schematic overview of the ANTARES telescope is shown in Figure 1 3 As can be seen in the figure the detection lines are separated from each other by a distance of 60 to 75 m Each line also supports hydrophones for acoustic positioning and a LED optical beacon for time calibration cf Section 1 3 2 A thirteenth line called the instrumentation line is equipped with specific instruments to monitor the sea environmental parameters at the ANTARES site A detailed description of the instrumentation used on this line can be found in Section 1 4 4 1 3 1 Detection lines The detection lines of ANTARES are anchored to the Mediterranean seabed and can be released by means of an accoustic transmitter receiver system if necessary 16
105. en different configurations of optical modules OMs It reduces the total CPU time when comparing differ ent OM configurations inside the virtual module as only the propagation of the photons inside the virtual module has to be repeated The emitted Cherenkov light is propagated towards a plane intersection of the virtual sphere The orientation of the disk defined by the intersection is perpen dicular to the direction of the photon of 460 nm In the case of the ANTARES telescope the virtual sphere was chosen to enclose a single OM and the disk passes through the center of it see Figure 4 2 The radius of the virtual sphere gt The wavelength taken as a reference is 460 nm This corresponds to the maximum sensitivity of the optical modules OMs of the ANTARES telescope To gain running time a virtual sphere could also envelop an entire storey see Chapter 1 which is the basic component of the ANTARES telescope consisting of three optical modules OMs SIRENE has also been tested with this configuration To ease the implementation of the existing software generating the effect of the front end electronics it was decided to define for ANTARES the virtual module to be the optical module see Section 4 2 81 Detector simulation and Track reconstruction Figure 4 2 Tracking of a photon y from the muon track u towards the module centered at O The virtual sphere which is used in the simulation is shown together with the optimal
106. ent library has been modified to take the characteristics of SIRENE into ac count Both programs were found to give comparable results in the energy range below 1016 eV although differences remain that can be attributed to the track ing algorithms employed and the parametrisation of the muon total cross sec tion Moreover in contrast to KM3 SIRENE can be used to process ultra high energy neutrino events An interface with the trigger software of ANTARES has been developed to simulate the effect of the detector electronics and trigger the events produced by SIRENE In order to validate the quality of the events sim ulated by SIRENE they have been subjected after digitization to two distinct track reconstruction programs known as Aart Strategy and Scanfit which are available in the ANTARES collaboration It was found that SIRENE cannot eas ily be used with either reconstruction algorithm This can be partly attributed to the probability density functions adopted by both reconstruction programs that correspond to muon energies less than 104 eV These probability functions are therefore not suitable for simulating ultra high energy neutrino events with SIRENE Moreover the reconstruction programs were primarily designed for use with KM3 which involves a specific tracking algorithm that is not especially ad equate for SIRENE Therefore new probability density functions for describing the ultra high energy range need to be developed and further t
107. er 1 year E p lt 4 x 10 GeVem s sr after 3 years It is valid for neutrino energies between 10 and 10 GeV It applies to a pure flux of muon neutrinos Considering oscillations with a flavor composition at Earth such that v ve vz 1 1 1 the equivalent average upper limit for an equally mixed neutrino flavors flux would be about three times the obtained value that is Ebens v lt 3 6 x 10 GeVem siert after 1 year BO lt 1 2 x 10 GeVem s ier after 3 years 5 27 Comparison with existing bounds Theoretical upper bounds have been placed on the diffuse flux of neutrinos based on observations of gamma rays and cosmic rays The level of sensitivity obtained for the ANTARES telescope can be compared to the limit predicted by Waxman and Bahcall 107 for diffuse high energy muon neutrinos at Ep 0 9 4 5 x 10 8GeVem s tsr for energies above 10 GeV This bound was calculated with the assumption of a Vy Ve Vr 1 2 1 ratio The corresponding 90 confidence level CL upper limit on an equally mixed flavor neutrino flux is thus given by sat ane 4v 1 35 6 75 x 10 8GeVem s sr7 The Waxman and Bahcall limit constrains the neutrino flux using measurements of the cosmic ray spectrum above 10 eV The model assumes that neutrinos are produced in interactions of protons with the ambient photons or matter in sources which are optically thin for high energy protons to photo meson and nucleon
108. er 2001 a new 40 km long electro optical cable for the power and data transmission was succesfully deployed between the telescope site and the shore station in La Seyne sur mer One year later in December 2002 the Junction Box was installed on the seabed and connected to the main electro optical cable 1 4 2 Prototype Sector Line and Mini Instrumentation Line Between November 2002 and March 2003 the Prototype Sector Line PSL and the Mini Instrumentation Line MIL were deployed and connected to the Junc 21 The ANTARES telescope tion Box JB The two prototype lines allowed to test in deep sea conditions the almost final design for the ANTARES telescope using prototype electronics boards The PSL was a short prototype of a full optical detector line with only 15 instead of 75 OMs It consisted of a single sector of five storeys equipped with OMs A schematic view of the PSL is shown in Figure 1 6 The sector line also contained equipment to test the acoustic positioning system the standard acoustic receivers and transmitters RtRx at the base of the line and the receivers Rx on the first and fifth storey A LED optical beacon was placed at the second storey for calibration purposes LCM acoustic Rx2 LCM LCM 12m MLCM LED beacon LCM acoustic Rx1 100m SCM SPM acoustic Rx Tx P sensor 8 Link cable Figure 1 6 Schematic view of the Prototype Sector Line or PSL of ANTARES deployed in Decemb
109. er 2002 The PSL allowed to measure the level and variation of single photon counting rates in the OMs which is largely caused by bioluminescence created by deep sea 22 1 4 Development and Construction life forms Unfortunately a failure in the vertical line prevented the electronics to be synchronised and muons couldn t be reconstructed Mechanical Cable Sound Velocimeter ADCP Electro Mechanical Cable 3 fibres for DAQ Acoustic Positioning Modules receivers Optical Beacon CSTAR Electro Mechanical Cable 2 fibres for DAQ 1 for clock 100m LASER Acoustic JB 2 fibres for DAQ Beacon NE 1 for clock Positioning Modules Figure 1 7 Schematic view of the Mini Instrumentation Line or MIL of ANTARES deployed in February 2003 The MIL was a prototype of the instrumentation line of the ANTARES tele scope There was no OM integrated on this line but several oceanographic probes to monitor the marine environment as well as a seismograph to study the deep sea seismic activity in the region around the telescope As can be seen in Fig ure 1 7 the MIL was also equipped with acoustic positioning modules sound ve locimeters and instruments to probe the salinity conductivity and temperature of the deep sea environment The MIL also incorporated a laser beacon placed on the Bottom String Socket BSS and a LED Optical Beacon on the lower storey used for the relative time calibration of the sector line This l
110. erenkov photons generated in muon neutrino interac tion and the passage of the resulting muon through the detector are described in this section together with their implementation in SIRENE 74 4 1 SIRENE Muons Input to SIRENE are the starting position direction initial energy and time of the muons propagating from the neutrino interaction vertices towards the detector volume Only particles passing through the Can volume are taken into account The high energy muons propagate with the speed of light in vacuum c since they are relativistic particles produced by very high energy neutrinos The muon tracks are assumed to be straight lines The muons suffer stochastic energy losses which produce independent electromagnetic EM showers with lower energy The muon energy loss dE dx is proportional to the muon energy E deposit ing an energy dE in the form of an EM shower for every distance dx travelled The distance dx is calculated using the mean free path or interaction length Lj in the medium the muon traverses The mean free path depends on the muon interaction processes involved Bremsstrahlung dominates over the other muon interaction processes at moderate energies E gt 1 and contributes about 40 of the total energy loss at ultra high energies see Chapter 3 According to refer ence 80 at high energies the Bremsstrahlung cross section differential in energy is described by do 11 4 4 2 n aara 5 vet with nat
111. es a discussion of the reconstruction al gorithms used to extract observables from the sim ulated events The design of new neutrino telescopes such as KM3NeT 79 re quires the simulation of many different detector geometries and configurations of the photo sensors in the instrumented volume However most programs are not flexible enough to allow different neutrino telescope geometries and only ac cept neutrino events in a limited energy range Hence a fast detector simulation program capable of modeling the detector response for neutrino events up to the highest energies and allowing any configuration of the photo sensors in a neutrino telescope had to be developed The new detector simulation program 73 Detector simulation and Track reconstruction SIRENE does have these features In the framework of the research project de scribed in this thesis SIRENE has been integrated in the chain of simulation and reconstruction programs of ANTARES to make it possible to study neutrinos in the ultra high energy range A description of the algorithms implemented in SIRENE and the usage of SIRENE to simulate the response of ANTARES are presented in Section 4 1 The tracking method of the secondary particles essential for neutrino detection and the algorithm used to track the produced Cherenkov photons towards the photo multiplier tubes PMTs of the telescope are also discussed In Section 4 2 the simulation of the signal digitisation and tr
112. et te As a result the Universe is opaque to gamma rays from extragalactic sources and photons with energies in excess of 10 GeV cannot even survive the journey from the Galactic Center to the Earth 2 In addition gamma rays cannot penetrate nor escape the hot and dense regions which form the core of stars and other high energy astrophysical sources It is therefore impossible to investigate the proper CONTENTS ties of these objects or regions of the Universe by direct observations In the last decades a new field of research has emerged to overcome these limitations At a crossroad between astronomy particle physics and cosmology astropar ticle physics studies elementary particles of cosmic origin Through the design and construction of new types of infrastructure such as underground laboratories or especially designed telescopes antennas and satellites astroparticle physics opens a new window on the Universe searching not only for photons from astro physical objects but also for high energy cosmic rays neutrinos and gravitational waves Neutrinos are of particular interest as they are stable particles which interact very weakly with matter They can cross long distances and penetrate regions which are opaque to photons Neutrinos are electrically neutral and their trajec tory cannot be deflected by the ambiant magnetic fields in the Universe Con sequently the direction of the observed neutrinos at Earth points back directl
113. f ANTARES which is essential for the precise determination of the direction of a muon relies on the transit time spread TTS of the PMTs To measure the TTS of a photo multiplier tube a LED system consisting of a blue LED and a pulser are used The LED is glued onto the rear of the PMT bulb and illuminates the photo cathode The Glass sphere 15 mm Photomultiplier Optical gel Shielding Black paint Connector Figure 1 5 Schematic cross section of an optical module OM of ANTARES 19 The ANTARES telescope pulser card is also glued onto the photo multiplier tube and is triggered by a clock signal The measured single photon TTS of the PMT is about 2 6 ns FWHM 1 3 4 Sea monitoring instrumentation The detection lines of ANTARES also contain instruments for the acoustic posi tioning of the telescope and the monitoring of the deep sea properties As the detection lines are anchored to the seabed and maintained vertically by buoys they undergo drifts of a few tens of meters as measured at the top of the lines due to the deep sea underwater currents The relative positioning of the lines is based on underwater acoustic techniques monitoring these motions The sys tem consists of a three dimensional array of transponders and acoustic modules exchanging precisely timed acoustic signals between each other The acoustic modules are emitter receivers located on the BSS of each line and receivers dis tributed along the lines
114. f N and M to the confidence level of the test NM lim Pi NEMOM gt z gt a z Sj eeen k 1 The confidence level of the test a z depends on the value z see 33 The rel evant critical values z can be obtained from software programs that perform a Kolmogorov Smirnov test In practice the value z 1 36 is commonly used This value corresponds to a confidence level a z 0 05 for which the null hypoth esis is inadvertently rejected in 5 of the cases when it is in fact true In the Comparator module the Kolmogorov Smirmov test algorithm as implemented in ROOT 28 is used to test the compatibility in shape between two one dimensional histograms with unbinned data The returned value is calculated such that it will be uniformly distributed between zero and one for compatible histograms pro vided the data are not binned The result is 1 if the null hypothesis Ho can be rejected or 0 if it cannot be rejected If the returned probability is less than the chosen confidence level the two histograms are not compatible 44 Chapter 3 Neutrino event generators In this chapter two different neutrino Monte Carlo event generators ANIS and GEN HEN are described These generators are used to simulate neutrinos of all flavors and their interactions in the seabed rock and the water surrounding the neutrino detector ANTARES in the energy range covered by the telescope A comparison in performance of both neutrino event generators is pre
115. for muon energy between 20 GeV and 10 TeV MUM is the only lepton propagator available in GENHEN that can also propagate tau leptons which loose energy differently than muons The lat est version of MUSIC also contains the propagation of taus with the package TAUSIC but this version has not yet been included in GENHEN An ANTARES internal note details the comparison of the three lepton propagators 62 As can be seen in that paper MUM and MUSIC give comparable results for the muon propagation through matter but the MUM algorithm is faster Because of this MUM has been chosen as the lepton propagator in this work In this lepton propagator the photo nuclear cross sections according to Bezrukov Bugaev 51 or the parametrization of the Zeus group 72 can be used Bremsstrahlung is described according to Bezrukov Bugaev Andreev 73 or Koboulin Kelner and Petrukhin GEANT 4 58 The energy thresholds that determine the continu ous and stochastic regimes see Section 3 1 2 for their definition are parame ters for the initiation procedure of MUM and can be set to any value between 1074 lt veut lt 0 2 and 10 MeV lt Ecut lt 500 MeV A study of these energy cuts 49 leads to the choice of veut 0 01 and Ecut 10 MeV 3 3 ANIS All Neutrino Interaction Simulation ANIS is a Monte Carlo neutrino event generator for high energy neutrino tele scopes It has been developed by Marek Kowalski and Askhat Gazizov for the AMANDA collaboration
116. for ultra high energy neutrinos Ultra high energy neutrinos originate from the interactions of cosmic rays with ambient photons at the sites where they are produced 2 Since neutrinos are stable neutral particles which interact very weakly with mat ter they can travel across extragalactic distances to reach the Earth without being deflected by the electromagnetic fields of the Universe Neutrinos thus point back to their sources and may provide essential information on the inner workings of ultra high energy cosmic ray accelerators The small interaction probability of neutrinos make them difficult to detect Neutrino telescopes overcome this issue by observing the products of their inter actions through the Cherenkov effect in a transparent medium such as sea water or ice The main topic of this thesis is to study whether neutrino detectors can be used to search for ultra high energy neutrinos A correlation between ultra high energy cosmic rays and the several nearby active galactic nuclei AGN has recently been reported 13 which suggests that AGN are likely sources of ultra high energy neutrinos as well The sensitivity of the ANTARES neutrino telescope to ultra high energy neutrino events above 1016 eV and up to 10 eV has been estimated in the framework of this thesis At these ultra high energies the Earth is opaque to neutrinos and only downward going and horizontal neutrinos can arrive at the detector Since neutrino tele scopes s
117. g the last months of my PhD You really are the dearest of friends Je ne saurai terminer sans remercier mes parents et mon fr re Anthony Aussi loin que mes souvenirs remontent j ai toujours voulu devenir astrophysicienne et c est gr ce vous si aujourd hui mon r ve se r alise Vous m avez donn le courage et la force d arriver jusqu au bout de croire en moi quoi qu il arrive Je vous aime My last thought goes to you Petter You have supported me encouraged me and comforted me in the most difficult moments even though I seemed more like a troll than a swedish alva sometimes I would have never made it without you You mean everything to me Je t aime plus que tout Even though my doctorate has come to an end I will always keep the Nether lands in my heart I could thus not end these acknowledgements without this final word bedankt 152 List of Figures N 1 1 1 2 1 3 1 4 1 5 1 6 1 7 1 8 1 9 1 10 2 1 22 2 3 2 4 2 5 2 6 3 1 3 2 3 3 3 4 3 5 The measured cosmic ray spectrum between 10 and 10 eV 3 A Hilas plotr Ae aeta die a te Maton ine ir E ae Matas oh Ge wah A 4 The measured cosmic ray energy spectrum between 1018 eV and D a sha a ah a deer Dd 6 The ANTARES collaboration oes dain atin a at oe od 12 The Cherenkov ettech 5 3 4 4 oi tiga Aine ge ene Ae 14 Sketch of the ANTARES telescope eee 16 A storey of the ANTARES telescope ee 18 The ANTARES
118. gorov and Nikolai Vasilyevich Smirnov for the comparison of two one dimensional data samples While the x test is designed for discrete distributions and consequently in the continuous case only approximates the test statistic the Kolmogorov Smirnow test is designed to compare only continuous 4 The ANTARES data acquisition system 42 Control Buttons Draw Select All Clear Save Pic Edit Pic Print I Multiple selection Clear List Previous Import scan Next Display Layout m Bi Jano f ae Filter ROOT fles root Owner I Modified _ El 9 1K 8 8K 510 7 2K 7 2K 7 2K 8 4K colnard conard colnard conard calnard colnard conard colnard a The lower panel is switched to the import mode Draw Select All Clear Save Pic Edit Pic Print Buttons Multiple selection Clear List Previous g Display Layout Import Scan pes ns a SE le a _ Scan Panel General Histo Info Name Title Channels Range Entries HTVC_01593_5_ 256 0 5 255 5 2403 j Non empty bins Control Buttons 25 i oa Previous Import 27 Pa Next Clear EER Bin content 1 000000e 00 Bin range 22 50 23 50 b The lower panel is switched to the s
119. grams of figure 2 4 correspond to the runs 11248 11254 and 11246 for which there were no SPE hits recorded 37 The ANTARES data acquisition system e O Oe i 50ns gt timestamp 0x0 0x1 0x2 0x3 TVC tve max 0 tvc_min 0 Figure 2 3 Correlations between ARS time stamp T S and Time to Voltage Converter TVC 2 3 The Data Quality Monitoring System PhAntOM or Physics data of ANTARES Online Monitoring is the data qual ity monitoring system of ANTARES The main objective of PhAntOM is to com pare ANTARES data with a set of reference histograms in order to monitor the change in performance of the main components of the telescope during opera tions In an internal note 31 it is explained how to use and install the data monitoring package of ANTARES PhAntOM consists of three complementary software modules Histogrammer Comparator and Histogram Presenter that are described separately in the next sections A fourth module Watchdog monitors the operational conditions of PhAntOM The data flow between the various mod ules of PhAntOM is represented in figure 2 5 The PhAntOM system has been implemented to operate automatically during data taking with the ANTARES telescope The Histogrammer module makes one dimensional histograms of the data The Comparator module compares one dimensional histograms from two separate sets using output files of the Histogrammer for instance During data acquisition the program
120. gure 5 11 top The simulated background rates are relatively small and their fluctuations can be de scribed by Poisson statistics For this specific situation the method proposed by Feldman and Cousins 105 can be used to calculate the average upper limit that would be observed after hypothetical repetition of the experiment Table XII of reference 105 is used to determine the average upper limit of the signal which Feldman and Cousins refer to as sensitivity at the 90 confidence level CL that would be obtained over an ensemble of experiments with no true signal N 0 for each value N of the simulated background The resulting average upper limit at the 90 CL is shown as a function of the muon zenith angle in Figure 5 10 bot tom and as a function of the total charge of the hits in Figure 5 11 bottom As can be seen in the figure the average upper limit is enhanced towards 2 44 events N 0 at the highest zenith angle and charge values 5 2 5 Model rejection potential The method proposed by G Hill and K Rawlins 106 has been used to optimise the experimental cuts and develop a stronger upper limit on the AGN like neu trino flux model studied The method is based on signal and background expec tations from Monte Carlo simulations It is used to place the strongest constraints on theoretical signal models It is therefore suitable for this analysis The final selection values are the ones which minimise the ratio of the
121. he four consecutive steps of AartStrategy This can be seen in Figure 4 12 which shows the effect of the subse quent application of the linear pre fit the M estimator and the ML method with the original PDFs Although the results improve at each step the number of well reconstructed events does not increase much Still the final fit using the full PDF functions with background clearly gives the best result The last step would need a bet ter starting position for the ML estimate to perform better This behaviour was already observed in previous work on track reconstruction 68 In order to enhance the reconstruction efficiency the method used to choose the starting points needs improvement by using a different function g r in the M estimator In addition if the likelihood function yields multiple local maxima the algorithm has difficulties to determine the global maximum In most cases it will converge to a local maximum which is close to the starting point The ef 103 Detector simulation and Track reconstruction log 9 Reconstruction error a logo Reconstruction error a or cl T lo o EN D mM on no t HoH Mm HN WT Wm CO st N gt oo Re s N o NO Mp S LN Sr GO OS fe o o o o oO OT 0 0301 p dp 0 01301 p aP mMm mM AN A 8 8 EN EN le le H a H H N v Ai o g g E 2 D ar Oo oO gt gt H H Ae gt n sS g Q Q Q Q Q Q aa 5 af yo 9S o 2 i 20 0 Q e A A N N mM on
122. he integration of the signal by the second ARS of the PMT while the first chip is inactive As can be seen in the figure the peak at r 0 is shifted to the left com pared to the distribution without the simulation of the electronics This is due to the combination of the hits which are separated by less than 25 ns especially if the number of photons that was merged is large The trigger can also add a constant rate of uncorrelated background to the hits from B decay of K in salty sea water and bioluminescence see Chapter 1 In Figure 4 7 a rate of 70 kHz per PMT which is the uncorrelated background usually observed in situ was added This is reflected in the plateau at negative time residuals in Figure 4 7 L1 triggered hits are built from two or more LO hits at the same storey over a small time window of typically 20 ns Any of the available trigger algorithms can be applied to the digitized SIRENE data In this work the standard three dimensional trigger or 3D trigger see Chapter 2 was used to select the hits with minimum five local coincidences 5L1 hits The detector geometry as well as the AVC the TVC and the overall time calibrations are used by the trigger algorithm to convert the raw data into calibrated data for further processing by the recon struction and analysis programs The decoding is done by retrieving information from files obtained with the calibration methods described in Chapter 2 In order to validate the performance
123. hem to shore It receives orders from the shore station and distributes information to the LCMs through an Ethernet channel 17 The ANTARES telescope Figure 1 4 Schematic view of a complete storey or floor of the ANTARES telescope with three optical modules a cylindrical container housing the electronics in the center of the three optical modules an optical beacon above this cylinder and a hydrophone below the optical modules 13 3 Optical modules The optical module OM is the basic detector component of the ANTARES tele scope A high pressure resistant glass sphere houses a 10 inch Hamamatsu photo multiplier tube PMT and electronics to filter and amplify the signal The sphere consists of two hemispherical parts with an outer diameter of 17 inches and a wall thickness of 15 mm The light transmission through the glass is more than 95 for the blue wavelengths corresponding to those of the Cherenkov light emission The PMT is fixed to the front part of the sphere with a transparent silicon rubber gel with somewhat different optical properties as compared to the glass envelope and the photo multiplier window The refractive index of the gel is 1 404 which results in an additional optical transition between the water with a refractive in dex of 1 35 and the glass with an index of 1 47 Apart from an optical link the gel is also a mechanical link between the sphere and the PMT magnetic shield 18 1 3 Detector layout The
124. her make up a timeslice which contains all data from the 13 ms period All data contained in a particular timeslice are pro cessed by one PC in the farm Subsequent timeslices will go to different com puters in order to allow each PC sufficient time for processing The amount of data produced in the entire detector is too large to be stored for off line analysis Hence the data needs to be filtered before storage on disk This is the role of the trigger When a timeslice is completed it is passed to a fast software algorithm that searches for hits caused by a muon traversing the telescope 24 25 2 1 2 The ANTARES Trigger Background light from radioactive K decay and bioluminescence create uncor related hits in the telescope whereas atmospheric muons and signal muon in duced by charged current cosmic neutrino interactions in the media surrounding the ANTARES telescope will produce hits related in position and time as a conse quence of the properties of Cherenkov light The purpose of the ANTARES trig ger is to select these correlated hits in the Analogue Ring Sampler ARS pipeline data streams for digitisation readout and transfer to the online data processing system on shore The system can adjust to the variable background conditions in such a way as to retain the maximum of useful data and reject as much back ground as possible without loosing signal All PMT signals exceeding the low threshold values are transferred to shore The
125. her models can be used as well see 63 For details on how to use the weights of GENHEN we refer to 67 GENHEN directly links to the program NUFLUX to simulate the various atmospheric flux models NUFLUX is described in detail in 66 For more details on weights we refer to 68 and 55 58 3 3 ANIS All Neutrino Interaction Simulation 3 2 3 Lepton propagator GENHEN contains lepton propagation subroutines to calculate the maximum ranges which define the Generation Volume i e the neutrino interaction vol ume and the effective ranges that allow to compute effective areas see Chapter 5 The main role of the subroutines is to propagate the produced leptons from the interaction vertex to the Can of the telescope simulating their energy losses and survival probabilities including angular and lateral deflections due to multiple scattering The lepton propagator MUM Three dedicated lepton propagators MUM 49 PROPMU 69 and MUSIC 70 are included in GENHEN to calculate the cross sections free paths and energy losses for lepton interactions and simulate the propagation of the muons over large distances in the media around ANTARES While MUSIC and PROPMU in clude the angular and lateral deflections due to multiple scattering MUM ne glects muon scattering It has been noted however that PROPMU is perform ing badly 71 substantial numerical differences appear between the simulated and the predicted energy losses up to 20
126. hest energies is presented in Chapter 4 The program allows the im plementation of any neutrino telescope geometry SIRENE has been developed for the KM3NeT detector but it has been tested with the ANTARES telescope for the analysis presented in this thesis Results of the Monte Carlo simulation study are given in Chapter 5 A dedicated method has been developed to separate the ultra high energy neutrino signal from the large background of atmospheric CONTENTS muons using the characteristics of the simulated ultra high energy events in the ANTARES telescope Optimized event selection criteria based on the arrival di rection of the Cherenkov photons and the total charge deposited in the telescope have been applied Predictions on the diffuse AGN like muon neutrino flux limit and the associated neutrino effective area are presented 10 Chapter 1 The ANTARES telescope In this chapter the detection principle and the design of the ANTARES high energy cos mic neutrino telescope are described The technical development and construction of the detector are also presented he European scientific programme ANTARES or As tronomy with a Neutrino Telescope and Abyss en vironmental RESearch is dedicated to Neutrino As tronomy About 150 particle physicists astrophysi cists and engineers from 22 European research insti tutes and universities contribute to the project The ANTARES collaboration is assisted by marine envi ronment and technolog
127. his thesis is to study whether neutrino telescopes can be used to search for ultra high energy neutrinos These neutrinos are speculated to be produced in the Fermi acceleration of cosmic ray protons in extragalactic sources The recently reported results on ultra high energy cosmic ray production 7 13 indicate that active galactic nuclei are likely sources of high energy cosmic neu trinos Since cosmic rays with energies close to 10 eV have been observed neu trino beams of similar energy are expected as well 2 The small interaction probability of neutrinos make them difficult to detect However the products of their interactions are observable through the Cherenkov effect in a transparent medium such as water or ice The predicted fluxes of ultra high energy neutrinos are most likely within reach of the first generation of neu trino telescopes and certainly detectable by future kilometer scale neutrino obser vatories The neutrino telescope ANTARES is constructed to search for high en ergy neutrinos Its sensitivity to neutrino events above 1016 eV and up to the GZK cut off has been evaluated in the framework of this thesis At these ultra high en ergies the Earth is opaque to neutrinos and only downward going and horizontal neutrinos can arrive at the detector Since neutrino telescopes such as ANTARES are optimized for the detection of upward going neutrinos in the energy range from 10 eV to 10 eV it needs to be investigated whether
128. hits having a large charge are also included in the aforementioned L1 events The 3D trigger is suitable for UHE analysis Since UHE muons induce large photo electron hits in the PMTs the trigger will select them irrespective of the number of local coincidences involved The minimum of five correlated hits in the telescope has been shown to be sufficient to separate the hits due to atmo spheric muons from those from uncorrelated background 21 This selection will also improve the distinction between hits induced by UHE muons and the uncor related background The resulting event rates are shown in Figure 5 5 Compared with the event rates shown in Figure 5 2 the background has been reduced as hits due to K decay and bioluminescence are now removed However at trig ger level no distinction can be made between events from atmospheric muons or cosmic neutrinos Atmospheric muon bundles can spread over a large area and hit many optical modules OMs within a time window of a few nanosec onds 100 104 The trigger can therefore wrongly identify them as correlated The average uncorrelated background rate observed in situ 116 5 2 Data selection and analysis Event rates per year 10 Keser tivities lived eaaada er log ee energy GeV Figure 5 5 Muon energy distribution at the trigger level induced by downward going ultra high energy UHE astrophysical neutrinos solid and cosmic rays interaction in the atmos
129. hoto multiplier tubes as well as their positions and their characteristics The type of photo multiplier tubes their position and direction within the modules of the telescope and their characteristics have also to be spec ified An interface has been developed such that SIRENE may use files produced with GENDET 117 to describe the geometry of the ANTARES telescope With this convention the reference system is such that the x axis is directed towards the North the y axis towards the West and the z axis is pointing upward The origin of the coordinate system is located at the center of gravity of the telescope To enable the software reconstruction of a neutrino telescope the following classes are implemented 1The C programming language allows to define program specific datatypes through the use of classes Examples of these data types are known as objects and can contain member variables constants member functions and overloaded operators defined by the programmer 116 145 SIRENE software implementation X Y Z 0 9 ROTATION GEOMETRY PHOTON e Wavelength Time Type Time Surface Energy Quantum efficiency ST Water properties Transition Time thin Transition Time Spread Amplitude Spread ABSTRACT MODULE IDENTIFIER initial time Water properties e Initial energy MODULE Initial speed e Length ANTARES WATER medium TELESCOPE Scattering Length Scattering angle Index of refrac
130. hotons over a typical time window 25 ns This is simulated by the trigger software see Chapter 2 which translates the amplitude and time information of each hit into AVC timestamp and TVC values respectively A so called LO trigger occurs when the PMT signal exceeds a certain threshold In SIRENE hits with an amplitude larger than 0 3 of a single photo electron amplitude are selected Since consecutive photons arriv ing ona PMT within less than 25 ns cannot be observed separately due to the ARS integration they are combined into one single signal The trigger adds the ampli tudes of the photons and the hit time is considered to be the time of the earliest photon This is illustrated in Figure 4 7 where the time residuals r tpit fe be tween the hit time t obtained from the digitized PMT signal and the predicted arrival time fi of the photons on the OMs are shown The tail of the peak at r 0 90 4 2 Signal digitisation and triggering dN dr ns 40 20 0 20 40 60 80 100 120 140 Residual r ns Figure 4 7 Time residuals r thi ie between the hit time tpi obtained from the digitized PMT signal and the predicted arrival time tl of the photons on the OMs solid The distribution without simulation of the electronics is shown dashed for comparison is depleted compared to the distribution before digitisation see Figure 4 4 and a smaller second peak can be seen above 25 ns The additional peak corresponds to t
131. ia Genova Villefranche sur Mer Figure 1 1 The ANTARES collaboration consists of 150 scientists from 23 European institutes in France Germany Italy the Netherlands Russia and Spain energetic astrophysical phenomena of the Universe such as supernovae remnants SNRs active galactic nuclei AGN and gamma ray bursts GRBs or the central parts of stars which are opaque to photons The ANTARES telescope will open a new window on our Universe and may also shed light on the mysterious nature and origin of dark matter and the cosmic radiation which bombards the Earth s atmosphere at high energies Catching the weakly interacting neutrinos requires a telescope with a very large detection volume to increase the chance of detec tion With an instrumented volume of about 0 02 km and an aperture of 0 1 km ANTARES constitutes a first step towards the future European cubic kilometer size neutrino telescope KM3NeT This giant photo detector array will be large enough to capture and study the neutrinos associated with the rarest and highest energy cosmic rays up to the Greisen Zatsepin and Kuz min GZK cutoff 9 10 ANTARES is sensitive to the light produced by the secondary particles cre ated in neutrino interactions in the seabed rock and water in and surrounding the detector In Section 1 1 the detection principle of high energy cosmic neu trinos employed by the ANTARES telescope is described The various optical backgrounds which have
132. ices also provide an estimate of the salinity and temperature of the deep sea water These instruments provide information on en vironmental properties that are relevant for both the associated sciences oceanog raphy and the neutrino telescope data analysis Tiltmeters and compasses are placed in each BSS and optical module frame OMF of the telescope to give an additional measure of the position of the dif ferent OMs as well as their orientation All these oceanographic intruments are also present on the so called intrumentation line of ANTARES whose purpose is to monitor the deep sea parameters and to provide calibration parameters of the telescope This additional line is dedicated entirely to marine property mea 20 1 4 Development and Construction surements and runs independently of the rest of the telescope It also contains devices to measure the attenuation of the light absorption in sea water in order to determine the optical water properties at the telescope s site which vary in time The instrumentation line is equipped with an acoustic Doppler current profiler ADCP to monitor the water current flow The position of the optical modules is measured with the optical beacon de vice of ANTARES which consists of a set of LED beacons at four different places along the detection lines and a laser beacon fixed to the BSS of the instrumen tation line A procedure to determine the relative timing calibration that was developed for the
133. iffuse muon neu trinos from downward going and horizontal events has been simulated Using new dedicated software packages and an optimised set of selection criteria a fairly competitive upper limit for the neutrino event rate has been found The event rates however remain low when using the neutrino telescope ANTARES It is anticipated that the cubic kilometer sized KM3NeT 79 telescope will lead to an event rate increase with a factor twenty or more and hence a reduction in the upper limit that probes the Waxman and Bahcall range cf Figure 5 14 131 High Energy Simulations Results 132 Summary and conclusions Extremelly energetic cosmic particles with energies as high as 10 eV have been observed in extensive air shower arrays 11 12 7 The origin of these ultra high energy cosmic rays is one of the intriguing puzzles of astroparticle physics In order to solve this problem the astrophysical objects which are the source of cosmic rays and the processes involved in the production of particles with such extreme energies need to be unraveled In the last several years considerable advances have been achieved in the field of cosmic ray research Even though uncertainties remain and alternative models can be envisaged 12 evidence may be emerging for a diffuse shock acceleration scenario involving extragalactic sources 11 7 Further information on the ori gin of ultra high energy cosmic rays can be obtained by searching
134. igger algorithms of ANTARES are pre sented A comparison of SIRENE with the existing detector simulation program KM3 is described in Section 4 3 Finally Section 4 4 focuses on the reconstruc tion of the energy and track geometry of the muons which caused the hits in the detector 4 1 SIRENE The primary objective of SIRENE is to simulate the response to neutrino events of any type of high energy neutrino telescope considering all scattering processes in the optically transparent medium water or ice surrounding the telescope in which Cherenkov radiation occurs Any type of secondary particle capable of generating detectable Cherenkov photons can be implemented and processed by the detector simulation program In the context of this thesis which focuses on the detection of ultra high en ergy neutrinos Cherenkov photons from muon neutrino interactions are primar ily considered since the detection probability of such particles is higher than that of the other neutrino flavors due to the longer path length of the produced muons The tracking of the secondary particles and the evaluation of the Cherenkov pho ton field they generate for the sea water conditions at the ANTARES site are de scribed in this section The corresponding software implementation with the most important C classes of the SIRENE libraries and a short summary of their functions are presented in appendix C 4 1 1 Tracking secondary particles The various sources of Ch
135. ight flux due to absorption by the OM glass and optical gel as well as the quantum efficiency of the PMT and the angle of incidence inc of the photons This is calculated for each OM in the telescope by weighting the photon flux arriving on the glass sphere by the effective area A f of the OM which depends on the wavelength of the Cherenkov photon and the OM orientation The quantity Aff is the product of the OM angular acceptance Ag Oi the light transmission through the OM lass pau A and the optical gel ps 8 pP 8 transmission transmission efficiency QE A of the PMTs A as well as the quantum l 1 Aeff Pine A Ag inc x ps a A x EE x QE A transmission with finc the incident angle with respect to the OM orientation axis 86 4 1 SIRENE The angular acceptance Ag 6inc of the OMs accounts for the differences in detection probability of photons that arrive with different angles with respect to the photo cathode The OM of the ANTARES telescope has been described in Chapter 1 The angular acceptance function Ag 6in has been determined by the Gamelle measurements 91 It is described by a fit to these results 92 which is normalised such that the angular acceptance A gives the maximum OM effi ciency for cos inc 1 Ag cos binc 0 26 0 67 cos0 0 31 cos One 0 23 C08 binc The transmission of the photons through the glass of the OM psiass A and transmission through the optical gel ps s
136. in this case the high energy photon from muon Bremsstrahlung Considering that the energy is equally divided among the electrons or positrons and the photons in the shower the final number of radiating electrons is N 0 5 Eemo Ec with Ec the critical energy for the electron energy loss This number is then smeared with a Gaussian distribution G N to take into account statistical fluc tuations in the number of produced electrons Finally the number of emitted Cherenkov photons is computed with G Ne N The value of the constant 6 5 x 104 was obtained from empirical fits to GEANT4 simulations It is the number of photons radiated per centimeter times the inte grated path length of electrons in an EM shower of 1 GeV It is assumed that all the radiated Cherenkov photons are emitted isotropically in azimuth p from the shower axis The angular distribution of the EM shower depends on the zenith angle 9 according to a parametrisation of the angular dis tribution in cos given by dN dcos The angular distribution is normalised such as to integrate to 1 It is considered to be the same for all particles and shower energies since the amount of electrons positrons and gamma photons produced in the shower reaches an equilibrium The form of Expression 4 4 and the values of the constants used result from fits to GEANT4 SGEANT4 or GEOmetry ANd Tracking is a platform to simulate the passage of charged parti cles through matter 58
137. ine was operated until July 2003 proving the functionality of the various deep sea instruments con tained in the ANTARES telescope design 23 The ANTARES telescope 1 4 3 Line Zero Because of the problems encountered with the PSL the design of the electro magnetic cable EMC had to be improved In order to validate the new design in September 2005 a dedicated complete 25 storey line known as Line Zero 20 was deployed Line Zero was identical to a full scale ANTARES detector line ex cept for the abscence of photo multiplier tubes PMTs in the glass spheres of the optical modules OMs It was equipped with an autonomous monitoring system to check the water tightness of the electronic containers such as the Local Control Modules LCM and the String Control Modules SCM The Line Zero also allowed to monitor the attenuation of the optical fibres in side the vertical EMC linking the LCMs of adjacent storeys together The data collected by Line Zero showed no evidence for water leaks inside the electronic containers but did reveal a sharp increase in optical attenuation in some of the fibres after some time This could be attributed to the effect of the pressure push ing the fibres into the LCMs By modifying the design of the connectors between the EMC segments and the LCM this problem was solved in the final design of the detector lines 1 4 4 MILOM The MILOM or Mini Instrumentation Line with an Optical Module was an up gr
138. inside the medium is induced An empirical equation for the index of refraction of sea water based on a parameter fit by Quan and Fry 87 corrected for the pressure at the ANTARES site was determined as a function of the wavelength A of the photons 55 86 As can be seen in Figure 4 1 the model describes well the refractive index of the sea water at the ANTARES site for the range of wavelengths of interest 300 nm lt A lt 600 nm The empirical equation is given by 16 256 4382 1 1455 x 10 A2 A3 l n A 1 3201 336 x 107 4 5 In SIRENE the photons are assumed to propagate in the transparent medium around the telescope with the group velocity v4 which is defined as 8 medium iz c Ug A z pmedium 8 with c the speed of light in vacuum and n d the effective refractive index group velocity in the medium or group index of refraction which is defined be low The group velocity is the speed with which the variations in the shape of the photon waves amplitude i e is the modulation or envelope of the waves propagate through a transparent medium It can be written as dw ginedium mn 8 dk with w the angular frequency and k the wave number of the light waves Since w and k depend on the refractive index n medium Of the medium and the wavelength A of the photons such that ck N medium and k 27 age 4The index of refraction was used in earlier detector simulations but it was found to in trod
139. ion through the Earth are also shown The amount of time Number of events at the Can ANIS 44589 43 48 75 GENHEN 31095 49 27 53 Table 3 1 CPU usage of ANIS and GENHEN for the generation of 10 up going events The number of events with a neutrino or secondary muons which could reach the Can of the telescope are also given required for the simulation using ANIS is somewhat smaller compared to that required using GENHEN while it generates substantially more events at the de tector This is caused by the fact that GENHEN includes the propagation of the produced leptons towards the telescope while ANIS does not include this pro cess However the lepton propagator MMC which is used in conjugation with ANIS takes only about 5 minutes to propagate the secondary muons produced in DIS neutrino nucleon interactions with ANIS ANIS in association with MMC is therefore slightly faster than GENHEN This can be attributed to the use of pre calculated tables which implies that ANIS does not need to make a link to any other package for tau decay simulation cross section and atmospheric flux calcu lations We conclude that ANIS and GENHEN require roughly an equal amount of CPU time per event generated 69 Neutrino event generators 34 3 Event rate comparison To compare the output of both codes 10 upward going muon neutrinos were generated with ANIS assuming an E power law spectrum at the surface of Earth number of events per year
140. ions Given the energies involved QE and RES neutrino nucleon and interactions can also in this case be safely neglected for the calculation of the neutrino attenuation Tau neutrinos are a particular case They are never absorbed in CC interac tions as they regenerate v T Vr as was already mentioned above The Earth is thus transparent to tau neutrinos at all energies To quantify the phenomenon of absorption of the neutrinos in the Earth with more precision neutrinos are propagated through the Earth using various den 48 3 1 Event Generators and Propagators sity profiles which model the interior of our planet The most widely used den sity profile is taken from the Preliminary Reference Earth Model 40 34 which is shown in Figure 3 2 In this model the Earth consists of concentric spheres with different densities depending on the radial distance r to the center of the planet 14 12 density g cm3 10 3 L L 1 1 1 1 1 1 1 L L 1 L L L 1 L 1 1 1 1 1 L L L L L 1 x 0 1000 2000 3000 4000 5000 6000 7000 distance from the center of the Earth m Figure 3 2 Density profile of the Earth as a function of the distance from the center of the Earth Kinematics The kinematic variables describing the v u deep inelastic scattering DIS are the total center of mass energy s the square of the negative invariant momentum transfer between the incoming neutrino and the outgoing muon
141. iscriminate between the penetrating UHE neutrino events and the atmospheric background is presented in Section 5 2 which also includes a calculation of the upper limit for UHE neutrino detection with ANTARES and a comparison to existing upper limits The chapter is concluded with a discussion on these results 5 1 Monte Carlo simulations 5 1 1 Ultra high energy neutrino simulation The Pierre Auger Observatory has found evidence suggesting that Active Galac tic Nuclei AGN are likely sources of the highest energy cosmic rays 13 The present study therefore focuses on diffuse neutrino fluxes from AGN and the simulations are based on assumed AGN like neutrino spectra At the energies of interest for this analysis the values of the interaction cross section for neutrinos and anti neutrinos are nearly identical see Chapter 4 Hen ce only neutrinos need to be considered The Monte Carlo program ANIS 74 has been used to generate a flux of 10 muon neutrinos and their interactions in the media surrounding the ANTARES telescope The simulation assumes a generic AGN like spectrum of the form E 10 GeV cm s sr7 for ener gies ranging between 10 GeV and 10 GeV Muons produced in charged current interactions are propagated towards the telescope using the program MMC 50 The Probability Density Functions PDFs describing the signal distribution accounting for uncorrelated background hits are empirical fits to Monte Carlo simulations
142. ised by empirical fits to Monte Carlo simulations of muons traversing the telescope with an energy ranging from 10 to 10 GeV and assuming a flat E power law spectrum The cross sections involved take into account all muon energy loss processes Photons from electromagnetic EM showers are simulated The peak of the distribution is fitted with a Gaussian function while the tail is approximated by an exponential These two functions are joined together by a third order polynomial function However background hits can degrade the performance of the track recon struction For this reason PDFs in which background hits are taken into account have been developed and included in the final step of the fitting procedure 68 98 4 4 Reconstruction In these improved PDFs the signal is represented by the functions described above Description of the procedure The reconstruction procedure consists of four consecutive fitting algorithms each of them improving the results of the previous step The first routine is a straight line fit of the spatial coordinates of the muon track as a function of the hit time called Linear Prefit It is used as first step of the reconstruction algorithm since it does not require any starting position The position of the hits are approximated by the most likely points of closest approach of the track to the optical modules OMs The points are estimated from the likelihood that a track at a certain distance
143. ituated at depth D 2196 849 meters In this example N 10000 upward going muon neutrinos and anti neutrinos 50 each are generated be tween 10 GeV and 10 GeV assuming an energy spectrum F E E GeV em sr s71 cross sections are evaluated in the framework of pQCD using the CTEQ5 parton distribution functions PDFs The chosen size of the neu trino interaction volume Final Volume has been set equal to a cylinder with a radius of R 600 meters an upper detector region and a lower target region heights both equal to HAheight HBheight 600 meters This Final Volume en compasses the ANTARES Can R 266 11 meters HAheight 278 15 meters and HBheight 341 47 meters for every events in the energy range of interest It therefore allows to generate all the events which can produce detectable secon daries in the ANTARES telescope without missing any To optimize the geome try of the Final Volume its size is adjusted for each event to the maximum muon range at the considered energy A detailed description of the various parameters can be found in 74 HERRERA EREA REAR Raa RHE Steering file for running ANIS with the ANTARES telescope HERRERA EREA RARE AEEA The flux information p 1 10 tei2 1 0 14 The geometry information c 600 600 600 2196 849 1 The run information 10000 2 1 1234 H RH OQ bh H H 139 Implementation of ANTARES in ANIS the data directory anis v1 8 2 data The cross section processes and d
144. k 113 Distribution of the total charge of the hits as a function of the true muon energy for downward going ultra high energy events simu lated with SIRENE 2e 2G 4 2 deed ol Wed wld Ae le dh th sd 115 Energy distribution at the trigger level of ultra high energy muons simulated with SIRENE and atmospheric muons simulated with WELT VAG By ganosi setae eat aca aoe a ee nae aoe ee 117 Distribution of the total charge of the hits as a function of the muon zenith angle for downward going ultra high energy muons simu lated with SIRENE and atmospheric muons simulated with MU Distribution of the cosine of the zenith angle for downward going ultra high energy muons and atmospheric muons at the trigger level 120 Distribution of the total charge of the hits induced by downward going high energy muons and atmospheric muons at the trigger ENE 120 Distributions of the cosine of the zenith angle and the total charge of the hits for downward going ultra high energy muons and at mospheric muons after applying cuts on the muon direction and the estimated total charge of the hits ee 121 Integrated distribution of the cosine of the zenith angle for the flux shown in Figure 5 9 desen vint stig Bn hg Re Ge ee 123 Integrated distribution of the estimated charge of the hits for the flux shown in Figure SN wen pet Sh eat ce doe dede el GO 124 Model rejection factor for the diffuse flux of downward going UHE neutrinos as a function of the muon zenith
145. kin for the AMANDA telescope The program is written in the object oriented language Java with a C interface to improve the flexibility and readability of the code The current version of MMC at the time of the writing of this thesis i e version 1 4 6 contains the water properties at the ANTARES site so that ANIS and MMC can be used in the full ANTARES Monte Carlo chain An example on how to use MMC for simulating a neutrino flux in ANTARES is given in appendix B 930 Use of MMC for the ANTARES telescope MMC can be used to propagate the secondary muon and tau leptons created at the neutrino verteces by ANIS towards the Can of the ANTARES telescope As can be seen in Figure 3 8 the sea water at the ANTARES site is divided in three regions one from the surface of the sea to the point where the track of the muon enters the cylindrical Can around the telescope The second region starts here and ranges to the point where the muon trajectory exits the Can while the rest of the track lies in the third region We assume that the relative energy loss cut Veut and the energy cut Ect are interchangable and related by Ecut Vey Ey where Ey is the energy of the propagated muon The muons are propagated for a fixed Vey or Ecut through the medium until the particle reaches a point where it looses an energy AE that is more than the cutoff energy AE gt Ecut or v AE Ey gt Veut By default in the first region the muon relative energy loss cut
146. l Observation of the suppression of the flux of cosmic rays above 4 x 10 ev Phys Rev Lett 101 061101 2008 8 D E Torres and L A Anchordoqui Rept Prog Phys 67 16631730 2004 astro ph 0402371 9 K Greisen Phys Rev Lett 16 748 1966 10 G T Zatsepin and V A Kuzmin Pisma Zh Eksp Teor Fiz 4 114 1966 11 R U Abbasi et al Phys Rev Lett 100 101101 2008 12 M Takeda et al Astropart Phys 19 447 2003 13 J Abraham et al Correlation of the highest energy cosmic rays with the positions of nearby active galactic nuclei Astropart Phys 29 188 204 2008 14 M Drees The top down interpretation of ultra high energy cosmic rays J Phys Soc Jpn 77 16 18 2008 Suppl B 15 Shinozaki 2006 talk presented at GZK40 Moscow 16 Q R Ahmad et al Direct evidence for neutrino flavor transformation from neutral current interactions in the Sudbury neutrino observatory Phys Rev Lett 89 011301 2002 159 BIBLIOGRAPHY 17 M A Markov In Proc Annual International Conference on High Energy Physics 1960 Proceedings of Annual International Conference on High Energy Physics Manchester 18 P Amram et al Sedimentation and fouling of optical surface at the ANTARES site Astropart Phys 19 253 267 2003 19 J A Aguilar et al First results of the instrumentation line for the deep sea ANTARES neutrino telescope Astropart Phys 26 314 2006 20 M Ageron et al Studies
147. l objects with just the naked eye Galileo could observe the craters on the Moon and even the explosion of a supernova using one of the first built telescopes at the beginning of the 16 century Since this period astronomy has continued its evolution over time and remarkable progress has been made thanks to a continuous dialogue between theory and observations Nowadays most of our knowledge of the Universe comes from the observation of photons Beside their stability and electrical neutrality photons are easy to detect and their spectrum contains detailed information about the chemical and physical properties of their source Moreover they are abundantly produced in the Universe as electromagnetic radiation is released in a variety of physical processes and thus constitute a crucial cosmic information carrier The recent advent of multi wavelengths telescopes allows the observation of the en tire electromagnetic spectrum with wavelengths well beyond the limited range of visible light used in the past However the ultra high energy domain with energies far beyond the limits set by present day particle accelerators remains largely unexplored by conventional astronomical methods Even though gamma ray astronomy has been successful observations are lim ited by attenuation of high energy photons from distant sources due to electron positron pair production in interactions with the omnipresent 2 7 K cosmic mi crowave radiation 1 Ye yemB
148. l slave clocks are induced by differences in the optical path lengths involved in the distribution of the reference clock signal A precise measurement of the arrival times of Cherenkov photons on the PMTs requires the knowledge of these relative time offsets which are discussed in the next section 2 2 Relative time calibration The angular resolution of ANTARES depends on the relative timing resolution of the photo multiplier tubes PMTs of the telescope A procedure that can be used for the relative time calibration of a single line of ANTARES is presented in this section This method was applied to time calibrate the prototype instrumentation line MILOM thus providing preliminary information on the relative time calibra tion of the entire telescope An internal ANTARES note 29 describes in detail the results of this relative time calibration The method consists of the measurements of the time offsets between the re sponses of the ARS chips of the line to a flash of the optical beacon These time offsets are required to correct the time information obtained during normal data acquisition The optical beacon system is described in Section 2 2 1 The dedi cated time calibration procedure using this trigger is described in Section 2 2 2 A calibration procedure for the TVC of the ARS chips is presented in Section 2 2 3 2 2 1 The optical beacon system The system used for the relative time calibration of the photo multiplier tubes PMTs con
149. lated with SIRENE and the predicted hit time 91 Comparison of the muon energy distributions simulated with SIRENE and ON in eae en ees eg lettin ating tne ae 93 Comparison of the time residual distributions simulated with SIRENE and KM3 cod a Oe os nn ae pe oa ene ee Ge 94 Distribution of the reconstruction error on the direction of the muon determined using AartStrategy 0020000008 102 Distribution of the reconstructed error on the position using Aart PALE Ve Nanotech hae Oca ep ale oe ty Oe Ge Gs 103 Distribution of the reconstructed error on the muon direction using PATTO ALC BY ede oee oet ee ke Gad Rp A oD Aad DH oe ed ned 104 Distribution of the reconstructed error on the muon direction using Scanfit for events simulated with SIRENE 106 Distribution of the reconstructed muon energy as a function of the true energy for events simulated with SIRENE 107 Distribution of the cosine of the zenith angle for high energy events Simulated Wath ANIS terende aod ae A antenne a elated atl de ee ats 111 5 2 5 3 5 4 5 5 5 6 5 7 5 8 5 9 5 10 5 11 5 12 5 13 5 14 5 15 B 1 B 2 LIST OF FIGURES Energy distribution of downward going ultra high energy muons simulated using ANIS and atmospheric muons simulated with MU EEE EEEN 112 Distribution of the cosine of the zenith angle for atmospheric muons simulated with MUPAGE 3 ent de meh Goad ee oetans he on a
150. lg es 7 if Ygen 1 Emax 1 min The energy phase factor results from the integration over the simulated en ergy range E iin Emax in which neutrinos were generated Ig in sr is the angular phase factor Ig 27 cos Omax COS Omin It results from the integration over the simulated angular range 0min 9max assuming that the event generation is isotropic As the solid angle Q is com monly specified by the zenith angle 0 and the azimuth angle x the angular distribution is flat in cos and x o E in m is the neutrino nucleus interaction cross section pN in m is the number of target nuclei per m3 that is the target density Prrans is the probability of transmission through the Earth Fis the number of seconds per year The weights w to be assigned to each event t are therefore De EE eee X gmorel E 6 l gen In the GENHEN output tags three types of weight can be distinguished for each event The weight w1 in water equivalent m is the Can volume Veen The so called generation weight w2 in units of GeV m sr s year does not depends on the flux model and is expressed as w2 Ip IgE o E 0NAVgenPtransE The global weight w3 in year is defined as w3 w2datm which corresponds to the generation weight w2 folded with a neutrino flux model atm to give a rate of events per year As a default the atmospheric neutrino flux due to Barthol is added to the event weight w3 but ot
151. livetime of 456 8 days This is currently the most restrictive experimental bound placed by a neutrino telescope on a diffuse neutrino flux at UHE energies The upper limit calculated for ANTARES in this work is also about a factor two below the Radio Ice Cherenkov Experiment RICE bound 112 after one year of observation After three years of data taking the sensitivity of ANTARES increases to a factor of five below the RICE limit This bound has been determined by RICE using data from seven years 1999 to 2005 with a livetime of 20500 hours that is around 854 1 days This is currently the best limit at ultra high energies It should be noted that the present study is aimed at an order of magnitude estimate and does not account for systematic uncertainties No full energy nor track geometry reconstruction was included Nonetheless a competitive upper limit has been found 5 2 8 Effective detector area The effective area for neutrinos represents the area of an ideal detector capa ble of detecting neutrinos with full efficiency It describes the performance of a telescope including its efficiency to observe neutrinos The effective area Agr is defined as Nevents T dE de DA EB where Nevents is the number of events T the detector live time and the incom ing neutrino flux The effective area depends on the energy E and the zenith 129 High Energy Simulations Results Effective area m 1 2 3 4 5 6 7 log o Neutri
152. ll not reach the Can and therefore cannot create detectable light and events with secondaries 56 3 2 GENHEN that can reach the telescope but with insufficient energy to pass the detection threshold If the event survives the cuts or if the vertex is generated inside the Can the interaction is taking place close to the detector choosing a specific reac tion channel based on the relative cross sections at the sampled energy If a full propagation through the Earth is required it is checked whether the event has an energy vertex and direction that allow the neutrino to make at least one interac tion in the propagation through the Earth otherwise that event is discarded If the neutrino undergoes a charged current CC interaction with nuclei deep in the Earth the event is discarded as no emerging muon will survive the propagation If it undergoes a neutral current NC interaction with nuclei the final energy when leaving the Earth and entering the sea water is calculated This becomes the energy at the interaction vertex close to the telescope As was mentioned before the interaction channel is chosen based on the rel ative cross sections at the sampled energy If leptons are produced in the inter action they are propagated to determine whether they reach the Can using one of the available lepton propagator programs If an event contains a neutrino or a lepton which reaches the Can volume a weight is associated to it which de pends
153. ly oriented virtual module which is the grey plane perpendicular to the photon direction In this example the virtual sphere was chosen to encompass one single optical module OM The sphere and the module have the same radius is identical to that of the OM so the OM and the abstract module have the same geometric characteristics and the number of detected photons is closest to reality The tracking of the photons to the optical modules is illustrated in Figure 4 3 The quantities relevant 68 to determine the amount of light reaching the OM are the expected photon field path length The predicted angle of incidence of the photons on the OM i e the angle between the direction of the photon field and the pointing direction of the OM the predicted arrival time t of the photons on the OM These three properties characterize the position and the orientation of a given OM relative to the muon track To determine the path length of the emitted photons the angle of incidence and the true arrival time at the OM the location of the OM Fom is required as well as the start and end position of the muon track start and 82 4 1 SIRENE r closest Figure 4 3 Coordinates used to describe the detection of Cherenkov light The muon track u is characterized by its start position Fstart at time tstart and end position Tong at time teng Cherenkov photons are emitted in Po at time temit under the angle Oc The shockf
154. m summary 4 2 ara 44 444 ee nie ae Ev 32 2 Generation method ee 3 2 3 Lepton propagator eten rd red Re ra ee ANIS All Neutrino Interaction Simulation 3 3 1 Program summary ae Ul bemest gaa eh a a 3 3 2 Generation method 24 case ea Gele ele ESS 3 3 3 The Muon Propagator MMC 0 Comparison between ANIS and GENHEN 3 4 1 General differences oc 6 sa ho ae see en 3 42 CPU tme Comparison sose e eh a tet e e ote 3 4 3 Event rate comparison alsa aie a ales Detector simulation and Track reconstruction 4 1 4 2 4 3 4 4 4 5 SIRENE aara ee one din ee a a A a aa nt el a G a AA 41 1 Tracking secondary particles o oo aaa 41 2 Photon field simulation ooa ee Signal digitisation and triggering 0004 Comparison with the ANTARES detector simulations 4 3 1 KM3 and GEASIM 224 3 sos Aa ts 4 44 4a ei 28 4 3 2 Simulation results 0 2 0000 Reconstruci n avn Bod ete ath a a a ete hte ee Aa 441 Probability Density Functions 0 442 Linear least square methods 0 4 443 Track reconstruction with AartStrategy 4 4 4 Track reconstruction with ScanFit 445 Energy Reconstruction a 4 4 6 Performance of the Reconstruction algorithms Future extentions 0 0 0 a High Energy Simulations Results 5 1 5 2 Monte Carlo simulations o oo a a 5
155. mined Once the time stamp is corrected the TVC value needs to be calibrated to obtain the hit time Ideally the dynamic range of the TVC is 0 to 255 As can be seen in Figure 2 2 in reality the dynamic range is restricted between a lower and an upper bound TVC nin and TV Cmax respectively A function is needed to translate the TVC bi nary value into a measure of the time t in nanoseconds For TVC values between TVC in and TV Cmax this function is given as t ns Tsiope 0 1 ns bit x tvc TVCminl0 1 2 1 with Clock Period TV Cmax 0 1 TVC ain 0 1 where tvc is the value of the TVC The label 0 or 1 stands for the two TVCs that are part of the ARS The clock period is set to 50 ns The two TVC slopes are assumed linear as is the time transfer function Hence only the parameter Tstopel0 1 needs to be evaluated for the calibration The slope can be determined once we know the dynamic range of the TVC Between TV Cmin and TV Cmax there Tslopel0 1 2 2 36 2 2 Relative time calibration 60000 60000 50000 50000 40000 Counts Counts 40000 30000 30000 20000 20000 10000 10000 1 eee GAE A ON EEE a EE E A FOO 1 1 po PE AAE AE ES TEES A A ES E E AE E EESTE ARAD i 0 50 100 150 200 250 0 5 100 150 200 250 TVC value TVC value Counts Fie SEE OE E E S 0 y ES lll 200 250 0 50 100 150 200 250 TVC val
156. mission A as well as the quantum efficiency QE A of the PMTs are wavelength dependent Their parametrisations as a function of the wavelength are needed to estimate the amount of light which can reach the PMT before digitisation of the signal see Section 4 2 Measurements of these wavelength dependences by Hamamatsu for the PMT and laboratory measure ments for the OM main characteristics 93 91 are used in SIRENE Computation The Cherenkov wavelength range is divided into bins of width dX Aj Aj 10 nm from 300 nm to 610 nm According to Equation 1 in Chapter 1 the amount of Cherenkov light produced by a muon per unit of track length r in the wavelength domain A A 1 is given by dr E A2 pen Apia The average number of photons emitted by the muon track for a given wave length bin A Ai 1 is then calculated by En 2 dN N A Fend Fstari 7 The average number of photons detected by the OM in each wavelength domain can then be computed as the number of photons emitted by the muon track in this wavelength bin multiplied by the OM effective area N etectea A Ny A Aeff Finer A 4 9 The total number of photo electrons npe detected by the OM can then be com puted by taking the integral of equation 4 9 over the full Cherenkov wavelength range A 610nm Nape i Ny A Aeff Orne A AA A 300nm 87 Detector simulation and Track reconstruction The total number of photo electrons is sm
157. nd chips ARS of the PMT It also translates the recorded amplitude AVC time TVC and position of the digitised hits into calibrated information which can be used for further analysis The calibrated amplitude is determined using a linear dependency of the recorded AVC value on the num ber of photo electrons 25 The calibrated signal amplitude in the PMT is thus proportional to the number of detected photo electrons In Figure 5 4 the distribution of the estimated total charge of the hits induced in the telescope by the simulated AGN like muon neutrino spectrum is shown as a function of the true initial muon track energy The distribution is also shown for a generic E spectrum for comparison The muon energy considered in the figure is the energy of the incoming muon at the surface of the default Can volume of ANTARES see Chapter 2 The total charge of the hits in an event has been estimated by taking the sum of the calibrated hit amplitudes in every PMTs of the telescope As expected the distribution shows a correlation between the true muon energy and the digitised charge of the signal Despite the width of the distribution the total charge is seen to increase proportionally with the energy up to about 10 GeV Above this energy the distribution becomes flatter due to the signal integration of the front end chips of the PMT The ensemble of PMTs which constitute the telescope is seen to saturate when receiving a charge 114 5 2 D
158. nly modifying the cross section data in the tables without changing the code The HepMC 75 Monte Carlo event record and the vector package of the CLHEP library which is a set of high energy physics utility classes are used ANIS only simulates deep inelastic DIS neutrino interactions with matter as this channel is dominant at the energy range of interest for modern neutrino 60 3 3 ANIS All Neutrino Interaction Simulation telescopes Both charged current CC and neutral current NC scattering are implemented as well as neutrino interactions with atomic electrons In the latter case as was explained in Section 3 2 only the Glashow resonant Zee W anything interactions are relevant in the energy range of interest All produced hadrons are considered to be pions since Cherenkov telescopes cannot distin guish between different types of hadrons ANIS also simulates the propagation of neutrinos through the Earth using a density profile which follows the Preliminary Earth Model 40 Neutrino ab sorption in CC scatterings with nuclei and energy loss in NC interactions with nuclei or Glashow scattering with electrons are simulated The vr T vr re generation chain through the Earth is also included with the help of the program TAUOLA 36 which provides data tables with the final products of the decay chain Deep inelastic scattering is described using the CTEQ5 parametrization 42 for the parton distribution functions P
159. no energy GeV 104 Effective area m 3 aoe TM 4 5 6 7 8 9 10 11 log Neutrino energy GeV Figure 5 15 Muon neutrino effective area in square meters for ANTARES as a function of the neutrino energy in GeV The effective area for events which survive the selection cuts for UHE neutrino detection as described in Section 5 3 5 is shown bottom The standard effective area for upward going neutrino energies below 10 GeV is also shown for comparison top 130 5 3 Discussion angle 0 of the neutrinos In Figure 5 15 the effective area of the ANTARES tele scope is shown as a function of the true incoming neutrino energy for those events which could survive the optimised selection cuts described in the previous subsections The optimised selection criteria lower the effective area at energies below 107 GeV as compared with the standard effective area of the ANTARES telescope However the effective area is enhanced at higher energies As can be seen in the figure the detection area increases with the neutrino energy up to 104 m at 10 GeV In the nearly horizontal angular range a large area is covered by the muon tracks what yields an high detection efficiency Moreover high energy events produce a large amount of light which induce hits in the telescope even when the neutrino interaction occurs far from the telescope 5 3 Discussion In the presented study the flux of ultra high energy UHE d
160. nsparent medium through which it passes The minimum velocity at which this process can occur is defined as the Cherenkov threshold 1 gt Pe with n the index of refraction of the transparent medium and B v c the ve locity of the particle v expressed as a fraction of the speed of light in vacuum c The energy is released in the form of a coherent electromagnetic wavefront The number of Cherenkov photons emitted by a particle with unit charge per unit wavelength and per unit track length is dN 2ra 1 anda a2 Ben 1 1 with the fine structure constant and A the wavelength of the emitted photon Cherenkov radiation can be detected by the three dimensional array of photo sensors constituting the ANTARES telescope The detection of Cherenkov pho tons will make it possible to reconstruct the trajectory of the secondary particle which had emitted them and subsequently the direction of the neutrino which points back straight to its source Cherenkov photons are emitted only at a cer tain fixed angle with respect to the direction of motion of the charged particles resulting in a cone of blue light The angle of emission is related to the charged particle velocity and the refractive index of the local medium This angle is called 13 The ANTARES telescope ue Figure 1 2 Charged particles moving at relativistic speed through a transparent medium are described by a track of length dx emitting photons under the Cherenk
161. nternal note 26 The trigger software is composed of several algorithms each with a large number of parameters Amoung them are the trigger1D the trigger3D and the triggerOB The trigger1D is a one dimen sional trigger which searches for correlated hits from a muon in a single direc tion The trigger3D is a three dimensional trigger which searches for correlated hits from a muon in many directions by applying the trigger1D in each one of them The triggerOB is the optical beacon trigger 27 This algorithm searches for time and position correlations between hits generated by the optical beacon device and the rest of the hits In this thesis we are mostly interested in the optical beacon trigger for time calibration see 2 2 of the telescope and the trigger3D for the detection of high energy cosmic neutrinos 32 2 1 Data Acquisition 2 1 3 Data format The ANTARES DAQ system uses the ROOT 28 framework to write and read data All event files are stored into a CVS software repository the antares daq The data frames are organised in structures implemented in C classes inher ited from ROOT TObject The data is accessible by mean of the standard STL methods The timeslices which contain all SPE hits are structured in the class SPE_TimeSlice The physics events are structured in the PhysicsEvent class A physics event contains a list of all SPE hits as well as a list of all hits that made up the cluster that caused the trigger The data are
162. of a full scale mechanical prototype line for the ANTARES neutrino telescope and test of a prototype instrument for deep sea acoustic noise measurements Nuclear Instruments and Methods in Physics Research A 581 695 708 2007 21 R Bruijn The ANTARES Neutrino Telescope Performance Studies and Analysis of First Data PhD thesis Universiteit van Amsterdam Amsterdam the Netherlands March 2007 22 M Ageron et al Performance of the first ANTARES detector line 2008 arXiv 0812 2095 23 J A Aguilar et al Sensitivity to point like neutrino sources with ANTARES 5 line data 2008 ANTARES public plot ANTPLOT PHYS 2008 019 24 B van Rens Detection of magnetic monopoles below the Cerenkov limit PhD thesis Universiteit van Amsterdam Amsterdam the Netherlands July 2006 25 M de Jong The ANTARES trigger software ANTARES internal note ANTARES Soft 2004 001 26 M de Jong The ANTARES data format ANTARES internal note ANTARES Soft 2004 006 27 J A Aguilar Analysis of the optical beacon system and search for point like sources in the ANTARES neutrino telescope PhD thesis Universidad de Va lencia Valencia Spain December 2007 28 http root cern ch The ROOT Users Guide is available from the ROOT homepage 29 C Colnard Time calibration using the optical beacon trigger ANTARES internal note ANTARES Cali 2005 004 30 C Colnard The ars tvc calibration ANTARES internal note ANTARES Cali 2
163. off is taken to be Veut 0 05 dimensionless In the second region the absolute energy cutoff is fixed at Ecut 500 MeV and finally in the third region the muon is propagated in one step with a relative cutoff of Veut 1 all loss is continuous to the point where it is lost Only secondaries and interactions in the second region are recorded into the output For more details on the implementation of the ANTARES main characteristics in MMC see appendix B 4 As defined in Section 3 2 65 Neutrino event generators Figure 3 8 Geometry used by the muon Monte Carlo propagator MMC 50 to propagate leptons through the media towards the ANTARES telescope There are three propagation regions before the detector the propagation is done with a fixed Veut inside the detector the propagation is done with fixed Eet and after the detector fast propagation with Veut 1 Comparison with the ANTARES lepton propagator MUM The performance of MMC has been compared with that of the lepton propaga tor MUM for the same settings Veut 10 3 ZEUS 72 parametrization of the photo nuclear cross sections and Bezrukov Bugaev Andreev 73 paramatriza tion of bremsstrahlung for both standard rock and water Figure 3 9 shows the final energy distributions of 5 x 10 muons with initial energy 100 TeV which were propagated through 1 km of water calculated by MMC and MUM The de tails of this comparison can be found in 50 As can be
164. on and time for each direction in each step This concept was exploited by the reconstruction algorithms that are used below The events produced by SIRENE were reconstructed with an algorithm known as AartStrategy 68 and a more recent algorithm called ScanFit 21 The algo rithms are based on different approaches to overcome the non linearity of the problem mentioned above The reconstruction routines also provide information on the precision and quality of the proposed fits and the associated uncertain ties of the results For the energy reconstruction an independent program 97 is used The probability of observing a certain set of hits considering a given set of muon paramaters is described in the next section Probability density functions PDFs the maximum likelihood principle a derived method called M estimator and the minimum chi square evaluation which are both used by AartStrategy and ScanFit are also presented Details of the tracking and energy reconstruction procedures as well as results of their application to SIRENE events are discussed in subsequent subsections 8A muon track is described by five independent parameters Two parameters are necessary to define the position of the muon at a given time Two angles the zenith and azimuth angles describe its direction The last parameter is the time 95 Detector simulation and Track reconstruction 441 Probability Density Functions The probability density function P
165. on error q 10 3 2 1 Figure 4 10 Distribution of the reconstructed error x on the direction of the muon obtained with the AartStrategy procedure Only upward going events are shown In Figure 4 11 the error on the reconstructed position of the muon which is de fined as the distance of closest approach between the true and the reconstructed muon tracks is shown Even though the position of the muon track does not play a leading role in track reconstruction this result is shown for completeness As can be seen in the figure most events are reconstructed within 1m from the true 102 4 4 Reconstruction muon track A peak appears in the distribution around 0 2 0 3 m with a large tail The tail disapears and the distribution becomes approximately Gaussian for events which were reconstructed within 1 of the true direction demonstrating a correlation between the directional and the position errors The same behaviour can be observed when using KM3 for the simulation of the detector response 68 600 500 400 Number of events 300 200 100 m E ee oe Were l 7 OE Error on the reconstructed position m Figure 4 11 Distribution of the reconstructed error 4 on the position of the muon ob tained with AartStrategy The distribution is shown for all events solid line and for the events which were reconstructed within 1 of the true direction dotted line The accuracy of the reconstruction depends on t
166. on the energy bin The event weights are described in the next paragraph and calculated using the prescriptions of References 65 and 66 The weights are stored along with the events This can be used to adapt the generated sam ple and obtain physical spectra The final energy of the neutrino that propagates through the Earth is also recorded Neutrino event distribution To obtain a distribution corresponding to a model specified by a given neutrino flux pi E at the surface of Earth the sample of generated events needs to be weighted by the ratio of the model neutrino flux to the generated flux The differential flux of neutrinos in units of GEV m 2s sr that is simulated at the surface of Earth 3 can be expressed by Neen Ey ee du v v Veen Ie loo Ev oNAE Prrans Ev Ov F where Veen in m is the volume in which the neutrino interactions were gener ated The flux corresponding to the events that are generated is the flux of neutrinos arriving at the telescope The transmission probability through the Earth given in Equation 3 9 needs to be introduced in order to get the flux corresponding to the incoming neutrinos at Earth If a full propagation through the Earth is simulated there is no need to introduce the probability of transmission through the Earth in the weights In this case Ptrans is set to 1 57 Neutrino event generators Ir in GeV is the energy phase factor 1 1 E Hp Ygen
167. on to current experiments is also shown The sensitivity is given as a differential neutrino flux multi plied with E 23 27 The ANTARES telescope 28 Chapter 2 The ANTARES data acquisition system In this chapter the data acquisition system of the ANTARES telescope is presented At tention is given to a relative timing calibration procedure using the optical beacon device of the detector and the data quality monitoring system which were both developed in the framework of this thesis uon tracks produced by a high energy cosmic muon neutrino interaction in the rock and sea water surrounding ANTARES can be reconstructed by the three dimensional array of optical modules OMs y which measures the arrival times of Cherenkov pho tons hitting the photo multiplier tubes PMTs com posing the telescope The data acquisition DAQ y system collects the time and amplitude information of the PMTs and converts the data such that they are suitable for the analysis and reconstruction soft ware A description of the DAQ system is given in Section 2 1 The quality of the reconstruction is expressed in terms of the three dimensional pointing accuracy or more precisely the angular resolution of the telescope is determined by the relative timing of each PMT signal with respect to the others An in situ calibration system is used to monitor the varying environ mental working conditions of the apparatus and the long term variations of the
168. onstruction a Buoys i 15m Electro Mechanical Cable Z LED Beacon Storey 3 Acoustic Doppler Current Profiler TOP LCM 41 Optical Modules 50m in Spy Hydrophone Sound Velocimeter Storey 2 MLCM 3 Optical Modules 5m LED Beacon St orey 4 CT Seabird conductivity Temperature probe ed ea CSTAR Light Transmission meter Electro Mechanical Cable aen xN Acoustic Positioning Hydrophone Rx 100m e Acoustic Positioning Transducer RxTx Bottom a Laser Beacon String Socket Acoustic Releasable a SCM Transponders Figure 1 8 Schematic view of the Mini Instrumentation Line with an Optical Module MILOM of ANTARES which was deployed in March 2005 and recovered again in April 2007 25 The ANTARES telescope 1 4 5 The complete telescope Between March 2006 and May 2008 the twelve detection lines of ANTARES have been deployed and put into operation successfully Several hours after the con nection of Line One the first muon signals could be detected In Figure 1 9 the 250 Z m 200 150 100 50 150 200 250 3000 2000 1000 0 1000 2000 3000 Time ns Figure 1 9 Example of a downward going muon event reconstructed with hits recorded by the first detection line of ANTARES Line 1 The abscissa represents the relative time of the hit on the optical module OM and the ordinate represents the altitude above the seabed altitude zero being the centre of the instrumented area of
169. oped chain of simulation programs An average upper limit of the event rate is determined and the corresponding effective areas for neutrinos are presented he study of the diffuse neutrino flux that originates from discrete sources which cannot be individually resolved or from interactions of cosmic rays with in tergalactic matter or radiation may yield important cosmological information Such measurements are of particular interest for neutrino energies in excess of 1016 eV in the so called ultra high energy UHE range Indeed the origin of UHE neutrinos and the associated high energy cosmic rays has remained a mystery for many years see Chapter 1 Recent re sults appear to correlate extragalactic supermassive black holes at the center of nearby active galaxies with the observation of UHE cosmic rays on Earth The detection of UHE neutrinos will provide crucial in formation on the sources of UHE cosmic rays and the processes involved in the production of such extremely energetic particles Above 107 GeV the Earth is opaque to muon neutrinos At the same time the neutrino interaction probability is still insufficient for the limited mass of sea wa ter above the telescope to yield a reasonable event rate Hence only horizontally traveling neutrinos have a chance of being detected in an underwater neutrino telescope When searching for neutrinos in the UHE domain the background consists of atmospheric muons which can reach the detect
170. or They are produced in large air showers by interactions of cosmic ray primaries with the Earth s atmosphere Es pecially multiple atmospheric muons in a short time slot can be mis reconstructed 109 High Energy Simulations Results as secondary leptons from charged current deep inelastic neutrino scattering in teractions The distribution and intensity of the atmospheric muons therefore need to be known in order to discriminate this large background from the UHE neutrino signal The techniques usually employed for neutrino searches see Chapter 2 need to be reconsidered when looking for UHE neutrinos The Earth cannot be used as a shield against downward going atmospheric muons Moreover the reconstruc tion algorithms typically used by ANTARES are optimised for muon energies lower than 10 GeV and not necessarily suited for the reconstruction of UHE muon tracks However trigger selections and cuts on the energy and zenith an gle can be applied based on the expectation that the signal is dominated by the downward going almost horizontal direction while the background is mostly vertically downward going The generation of UHE neutrino interactions and the subsequent response of the detector to these events are simulated using the dedicated high energy Monte Carlo programs described in the previous chapters of this thesis Details on the actual simulations for UHE neutrinos are given in the next section The proposed strategy to d
171. or reconstruction of SIRENE events using AartStrategy The algorithm depends on many parameters especially on the hit selection criteria and further tuning is re quired to improve its performance with the reconstruction of events generated using SIRENE Given the arguments above it is not unreasonable to find that only about half of the events which survived the trigger selection are well reconstructed using AartStrategy This is sufficient to give an order of magnitude estimate of the per formance of ANTARES in the energy range below 107 GeV However AartStrat egy is not suitable for the reconstruction of neutrino events at energies above 107 GeV For an optimal analysis of such data a dedicated high energy reconstruction algorithm needs to be developed Results with ScanFit In Figure 4 13 the distribution of the error on the re constructed direction is shown as obtained for events simulated with SIRENE and reconstructed with the ScanFit algorithm As can be seen in the figure a first peak is observed around a 0 5 and a second one at a 85 As for Aart Strategy the first peak represents the well reconstructed events while the second peak corresponds to the badly reconstructed events About 63 5 of the triggered events are reconstructed within 10 of the true direction and 50 8 within 1 The ScanFit reconstruction program performs slightly better with SIRENE than AartStrategy More events are well reconstructed This can b
172. ov angle c The photons travelling on a distance r create a photon field shockfront P the Cherenkov angle 6 and can be expressed by the following equation 1 cosc mB 1 2 In the energy range interesting for ANTARES that is above 10 GeV particles are ultra relativistic and pf is close to 1 The refractive index of the seawater at the ANTARES site is 1 35 Thus the Cherenkov angle is approximately 42 5 As the Cherenkov photons are emitted in the form of a cone of light a pre cise reconstruction of particle trajectories requires the measurement of only a few hits in the various photo multiplier tubes PMTs of ANTARES at different space points The light emitted by a track of length dx can be projected upon a surface Sc defined by Sc 27rdx sin Oe with r the distance travelled by the photons This is illustrated in Figure 1 2 According to Equation 1 1 with 1 this last result yields a photon flux observed in the electromagnetic wavefront at a distance r from the track of length dx Sca dN r d r rA2 In the deep Mediterranean Sea at the ANTARES site about 3 5 x 104 Cherenkov photons are emitted at wavelengths between 300 and 600 nm per meter of track 14 1 2 Background signal 1 2 Background signal Atmospheric muons from hadronic showers are produced at 10 20 km height when cosmic ray particles interact with nuclei in the atmosphere These down ward going muons can be mis reconstructed as upward going m
173. owing an AGN like EP 1076 GeV cm sr s power law spectrum at the surface of the Earth As can be seen in the figure the muon energy is not well re constructed for values of the true energy smaller than 100 GeV The algorithm per forms better at high energies above 10 GeV It is less accurate at low energies below 10 GeV where the reconstructed energy is biased towards higher values This behaviour can be attributed to the reconstruction program 68 as several low energy events are not reconstructed The program also does not take into account light from hadronic showers This may play a role in the energy range considered It can be concluded that the energy reconstruction algorithm used with SIRENE has a moderate accuracy for muon energies above 10 GeV which is the energy domain of interest for our present study However a dedicated al gorithm for the reconstruction of the neutrino energy above 107 GeV needs to be developed 106 4 5 Future extentions logi9 Reconstructed muon energy GeV 1 2 3 4 5 6 7 logy True muon energy GeV Figure 4 14 Distribution of the reconstructed muon energy as a function of the true muon energy Muons were generated by ANIS following a muon neutrino AGN like spectrum at the surface of the Earth The strategy described in 97 was used to estimate the muon energy of events simulated by SIRENE 4 5 Future extentions To further improve the performance of SIRENE several extensions
174. partures from the idealised assumptions are found In the case PDFs are a function of the hit time points with large fluctuations of the time resid uals appear mainly due to light scattering as seen in Figure 4 4 To minimise their effects on the fitting procedure another estimator is used the M estimator This is a so called robust estimation which is insensitive to small deviations and irregularities from the idealised assumptions used The following expression is minimised x ti tH Purdy M gri En een with g r a function which ensures that the minimisation of M is also efficient for large residuals The function is chosen to provide the desirable properties in terms of bias and efficiency of the M estimator when the data are truly from the assumed distribution 44 3 Track reconstruction with AartStrategy The algorithm AartStrategy is used for the reconstruction of muon tracks with energies above 50 GeV According to previous work 68 it can safely be used up to an energy of 10 GeV The algorithm is optimised to reconstruct upward going neutrinos and is therefore less suitable for downward going events The code was adapted so it could take into account the photon weights simulated by SIRENE Probability Density Functions To describe the signal AartStrategy benefits from PDFs which have been devel oped for the first reconstruction algorithm 89 used by the ANTARES collabora tion These PDFs are parameter
175. phere dashed SIRENE was used to simulate the UHE signal while MU PAGE was used to generate the atmospheric muons Events with a minimum of five local coincidences at the same storey or a large amplitude on one of the storeys are shown The astrophysical neutrino flux corresponds to an AGN like spectrum The atmospheric spec trum has been extrapolated for energies above 10 GeV with an E spectrum following the primary cosmic ray spectrum at UHE hits from a single muon Further analysis is necessary to improve the separation of the signal and the atmospheric background at the trigger level 5 2 3 First event selection Differences between the ultra high energy UHE signal and the atmospheric muon background can be exploited for data analysis The main differences rely on the direction from which the muons arrive at the detector and the energy of the muon which is estimated by the charge of the hits induced in the photo multiplier tubes PMTs In Figure 5 6 the distribution of the cosine of the zenith angle as a function of the total charge of the hits in an event is shown for both the UHE signal and the background events As can be seen in the figure downward going UHE muons concentrate near the horizon cos 0 whereas the direction of atmospheric muons is mainly vertical cos 1 Even though the charge induced by atmospheric muons is rather large up to about 104 photo electrons the hits from UHE muons give an even l
176. r at the telescope site is defined in 82 The values of the density of the main water elements are shown in Table B 1 Atomic Atomic Atomic Percent Symbol Number Z werent A by weight 1 ee 9 43 1073 2 09 1074 10 87 1074 2 09 1074 1 106 10 2 5 82 10 2 Table B 1 Chemical composition of the sea water at the ANTARES site MMC reads and outputs data in the AMANDA ASCII F2000 compliant format to and from the standard input and output so the events can be passed to the detector simulation The F2k output format can be translated into the ANTARES event ASCII format to process the output of MMC with the ANTARES detector simulation and analyse the Monte Carlo data with the help of the main trigger 142 number of muons per TeV 0 10 20 30 40 50 60 70 80 90 100 energy TeV number of muons per TeV 96 96 i 100 energy TeV Figure B 1 Comparison of the final energy distributions of 10 muons with initial energy 100 TeV which were propagated through 300 meters of ANTARES Water calculated by MMC with an energy cutoff Veut 0 05 solid line Veut 0 01 dashed Veut 10 3 dotted and Veut 1074 dashed dotted On the right is a close up on the left figure 143 Implementation of ANTARES in MMC gt number of muons per TeV 86 88 90 92 94 96 98 100 energy TeV Figure B 2 Same as the right part of Figure B 1 but with the continuous randomisation option which is used to prevent muon spe
177. r of the virtual sphere is oriented 146 perpendicular to the direction of the arriving Cherenkov photons The orienta tion depends on the position and direction of each source of Cherenkov photons Once the photons have been tracked towards the modules of the telescope their position direction and arrival time are recorded on this optimally oriented sur face This gives each ABSTRACT MODULE a list of recorded photons inside the volume and on the surface delimited by the virtual sphere see the definition of the class MODULE HIT The class ABSTRACT MODULE derives its data from the classes GEOMETRY and IDENTIFIER MODULE The class MODULE defines a module of the telescope It consists of a container housing a set of one or more photo multiplier tubes The class MOD ULE is derived from the class ABSTRACT MODULE The characteristics of the glass and the gel are implemented with their thickness but also their absorption length as a function of the wavelength TELESCOPE The class TELESCOPE defines a neutrino telescope which con sists of positioned virtual spheres called ABSTRACT MODULEs IDENTIFIER The class IDENTIFIER gives a unique identifier to every module of the telescope POSITION The class POSITION is an utility class used to define a cartesian vector position in three dimensions DIRECTION The class DIRECTION is an utility class to define a three dimen sional orientation with spherical polar coordinates GEOMETRY The class GEOM
178. racking the particles of interest can be found in several ANTARES inter nal notes 90 84 and PhD theses 55 68 For a realistic simulation of the angular resolution of the telescope light scat tering cannot be neglected This is included in both SIRENE and KM3 The lat ter program uses the package MUSIC see Chapter 3 to propagate the muons in straight lines through the telescope Electromagnetic showers EM showers are randomly generated along the path when the muon energy loss exceeds the Cherenkov threshold 0 3 GeV Since the contribution of the Cherenkov photons from charged hadrons can be substantial in the low part of the energy range cov ered by ANTARES GEASIM is used in combination with KM3 to add the light emitted from hadronic showers As in SIRENE light scattering in sea water is neglected for EM showers and hadronic showers The models implemented in GEASIM for light scattering and the wavelength dependence of the refractive in dex are the same as the ones in SIRENE While SIRENE is designed to process neutrino events of the highest energies Ey lt 10 GeV the existing ANTARES detector simulation programs cannot be used above 107 GeV For the study of the ultra high energy cosmic neutrinos it is thus necessary to use SIRENE However in order to validate the new detector 92 4 3 Comparison with the ANTARES detector simulations simulation program SIRENE has been compared to KM3 with GEASIM at lower neutrino ene
179. relative time offsets of the PMTs The optical beacon system of ANTARES that is used for time calibration is described in Section 2 2 The relative time calibration method is based on the determination of the time offsets of the various PMTs in response to a flash of the Optical Beacon device of ANTARES This procedure has been tested using the MILOM prototype instrumentation line of ANTARES 19 All critical components of ANTARES have been properly tested and calibrated 29 The ANTARES data acquisition system before deployment but their performance might change during operation A di agnostic tool is therefore required to monitor whether components are function ing as expected or if they are degraded in their performance during the opera tions This is the purpose of PhAntOM the data quality monitoring system of ANTARES which is described in Section 2 3 2 1 Data Acquisition The purpose of the data acquisition DAQ system is to acquire and process the data produced by the photo multiplier tubes PMTs Both the arrival times of the Cherenkov photons on the PMTs and the amount of light deposited is recorded by the DAQ system The data are digitised and subsequently transmitted to shore for off line analysis Contrary to other Cherenkov neutrino detectors AMANDA IceCube Baikal the data is not filtered by the front end chips off shore but on shore by trigger software running on a set of computers This is the concept of all data to
180. rgies 4 3 2 Simulation results Event rates To compare the output of SIRENE with that of KM3 GEASIM Neen 10 up ward going muon neutrinos were generated with GENHEN assuming an AGN like E 106 GeV m sr s power law spectrum at the surface of the Earth in the energy range 10 10 GeV The produced muon secondaries were propa gated towards the default Can of ANTARES using the lepton propagator MUM Both SIRENE and KM3 GEASIM were used to process the output file of GEN HEN On output only those events were selected which produce detectable light In Figure 4 8 an example of the corresponding energy spectra of the simulated neutrino events which could be detected is drawn Only charged current CC neutrino interaction events are shown 5 E o Pa r Lt io 2 Ebert ae i r A4 ona Ul 15 10 Cl if L L L k L L L L jk if L L if L l L L L 1 2 3 4 5 6 7 log Muon energy GeV 10 Figure 4 8 Simulated energy spectra of events producing detectable secondaries in ANTARES SIRENE solid and KM3 dashed were used with the same input file con taining muon tracks generated by GENHEN The Cherenkov photons were simulated within the default Can of ANTARES As can be seen in Figure 4 8 the two distributions are similar in size and shape in particular for energies above 104 GeV As SIRENE has been
181. rgy GeV Figure 3 10 Differential neutrino nucleon cross sections for the deep inelastic DIS channel of the charged current CC interaction for the CTEQ6 parametrization solid line and the ANIS results based on the CTEQ5 parametrization dashed Both ANIS and GENHEN generate neutrinos which follow a power law spec trum E 7 with y the spectral index at the surface of Earth While in GENHEN fluxes are described in the unit GeV m sr s71 ANIS uses GeV cm sr sb In order to compare event rates produced with ANIS and GENHEN fluxes and atmospheric weights in GENHEN have been scaled with a factor 104 to con vert square meters into square centimeters GENHEN generates either neutrinos or anti neutrinos for each run while ANIS simulates 50 neutrinos and 50 anti neutrinos For ease of comparison the ANIS code has been modified such that either neutrinos or anti neutrinos are exclusively produced similar to GENHEN GENHEN generates either charged current CC or neutral current NC neu trino interactions with nuclei for each run while ANIS simulates both interaction channels Only CC or NC events have therefore been selected with ANIS af ter production for comparison with GENHEN It is also possible to combine two runs of GENHEN with CC and NC events and to renormalize them in terms of generated numbers of events so they can be compared with ANIS In this thesis the first method was chosen 5In this comparison the known
182. ric muons Both signal and background can induce a total charge up to 10 p e A loose cut is placed at 104 p e to reject events with low charge values which are mostly entirely due to atmospheric background 5 2 4 Expected rates and neutrino flux limit After selecting events with a charge larger than 104 photo electrons p e and a zenith angle such that cos gt 0 6 the remainining simulated event rates per year have been determined Figure 5 9 shows the event rates as a function of the estimated total charge of the hits top and the cosine of the zenith angle bottom after the cuts have been applied As can be seen in the figure the selection criteria reduce the number of background events but may require further optimisation In order to determine an experimental limit on a flux the maximum number of signal events needs to be known as a function of the number of observed events and expected background after all selection cuts have been applied The upper limit on a source flux Ey is thus calculated as Hoo Nops Np T 5 1 P Ev oo Ev 118 5 2 Data selection and analysis a a ooOaa OoaonanogcA oA ooi mi 0 a O a OooadnAoao oo a noOoadaanoau aA oo a Oo jOeoooooeae OOO0Oeg0ooo0e ao ao 0 0 DO a Hon E AD a Aa oOo oO es 200 ooo Mmoooandanonodaa Oo a
183. ronic showers ANTARES internal note ANTARES Software 2002 015 83 R Mirani Parametrisation of em showers in the ANTARES detector volume Master s thesis Universiteit van Amsterdam 2002 84 J Brunner Geasim http antares in2p3 fr internal software geasim html 85 S Bottai and L Perrone Simulation of uhe muons propagation for geant3 Nucl Instrum Meth A 459 319 325 2001 86 J A Aguillar et al Transmission of light in deep sea water at the site of the ANTARES neutrino telescope Astropart Phys 23 131 155 2005 87 X Quan and E S Fry Applied Optics 34 Iss 18 3477 3480 1995 88 L A Kuzmichev On the velocity of light signals in the deep underwater neutrino experiments 2000 hep ex 0005036 89 F Hubaut Optimisation et caract risation des performances d un t lescope sous marin neutrinos pour le projet ANTARES PhD thesis Universit de la M diterran e Marseille France April 1999 90 S Navas and L Thompson Km3 user guide and reference manual ANTARES internal note ANTARES SOFT 2002 006 91 P Amram et al The ANTARES optical module Nucl Instr and Methods A 484 369 2002 92 P Kooijman On the angular acceptance of the optical module ANTARES internal note ANTARES PHYS 2007 002 93 J A Aguilar et al Study of large hemispherical photomultiplier tubes for the ANTARES neutrino telescope Nuclear Instruments and Methods in Physics Research A 555 132 141 2005
184. ront of the generated photon field hits the optical module OM located in Fom at time tpi The distance the photons travelled to reach the OM is The distance of closest approach between the muon track and the center of the OM isr Pena and the time tstart at Fstart In the following paragraphs it is explained how these quantities are calculated Photon path length The path length of the photon field is the distance be tween the OM and the position Femit on the track emitting the light According to Figure 4 3 it can simply be calculated as Ty sin Oc l Fom femit where r is the distance of closest approach between the track and the center of the OM which is the distance perpendicular to the muon direction For Cherenkov 83 Detector simulation and Track reconstruction radiation see Chapter 1 the sinus of the emission angle can be expressed as 1 i Sars eae en sinc 1 cosc pen with B v c the velocity of the particle v expressed as a fraction of the speed of light in vacuum c Since the muons are relativistic particles 6 1 the path length of the photons in sea water can thus be calculated as Incident photon angle The cosine of the angle of incidence of the photon field on the OM can be approximated by the dot product between the direction of the emitted photons and the direction of the OM axis The maximum efficiency of the OM is obtained when the photon field is parallel to the front of the O
185. s 98 99 since the Romeyer approach does not depend on the telescope geometry The corresponding pro gram has been modified to take into account the photon weights simulated by SIRENE The quality of the reconstruction algorithm has been verified for ener gies up to 10 GeV 97 68 The mean muon energy loss and thus the muon energy in this approach is estimated by dE 1 NPMT A 0 AD mes gt dx NpMT i l Leco 1 exp lreco ee 4 12 with Ag inci the angular acceptance of the optical module OM as a function of the angle of incidence binc Lreco the photon path length from the reconstructed track and Later the absorption length in sea water The constant Ajo is the sum of the recorded amplitudes and D is the part of the track which is contained in the Can of the telescope An empirical fit is made to determine the muon energy estimate using the average muon energy loss This function does not extend to low energies as below 100 GeV the estimate of the average muon energy loss is almost independent of the muon energy Therefore the energy reconstruction algorithm cannot be used for energies smaller than E 100 GeV 4 4 6 Performance of the Reconstruction algorithms Track Reconstruction of simulated events To evaluate the performance of the full simulation chain Neen 10 upward going muon neutrinos were generated with GENHEN assuming an AGN like E 106 GeV cm sr s power law spectrum at the surf
186. s from astrophysical and atmospheric muons but also from the decay of K and bioluminescence see Chapter 2 While hits from high energy muons are related in time and position as a consequence of the properties of Cherenkov light emission hits due to K and bioluminescence are uncorrelated The trigger software is used to separate these isolated hits from the ones created by astrophysical and atmospheric muons In this work a back ground rate of 70 kHz due to K and bioluminescence has been added to the ultra high energy UHE signal and the atmospheric background see Chapter 5 for details A three dimensional trigger trigger 3D is used to search for time correlated hits A hit is in local coincidence if it is within 20 ns of another hit on a different optical module OM at the same storey These coincident hits are referred to as L1 events Only events with a sufficient number of correlated hits are selected for further analysis while the others are being discarded This is motivated by the assumption that high energy muons induce multiple hits in the telescope while uncorrelated background will mainly produce single hits Since a muon track is defined by five independent parameters a minimum of five local coincidences 5L1 is required Coinciding hits on the same photo multiplier tube PMT will generally result in a single hit with a large charge of typically 2 5 photo electrons p e or more Consequently events with
187. s position direction energy and time can thus be treated as a source of Cherenkov photons in the same way as a muon The total energy E q of the Cherenkov light radiated by the EM showers can be obtained in the simulation from an exponentiation of the muon energy Ey En 1 Eraa Kinin hoc min with 7 a random number uniformly distributed between 0 1 The distribution ranges between Knin and E since the muon can transfer substantial amounts or even almost all of its energy to the particles in the shower In the code an EM shower is created by drawing a random number R uniformly distributed in the 76 4 1 SIRENE range 0 1 which needs to fulfill the following requirement 3 R lt 1 o ge This expression is the muon Bremsstrahlung energy loss 53 with v E na En the relative energy transfer In order to save computing time a parametrisation is used replacing the in dividual particle tracking for the simulation of the EM shower by results of pre vious studies on EM showers parametrisations in ANTARES 83 The amount of light emitted by the EM showers and the corresponding angular distribution are smeared using empirical fits to GEANT4 detector simulation results Since at the energies considered all EM showers are identical only electrons are taken into account for the parametrisation The total number of Cherenkov photons emitted depends on the energy Er mo of the initial particle which initiated the EM shower
188. s which have been developed for Aart Strategy 68 Description of the procedure The algorithm used by the ScanFit reconstruction program has been modified to take into account the photon weights simulated by SIRENE According to previ ous work 21 the ScanFit algorithm can be safely used up to an energy of 107 GeV A preselection of the hits is made by scanning over the angular space pa rameters Correlations between hits are searched for in about 170 pre defined directions requiring the standard 1D Trigger conditions see Chapter 2 A set of hits is obtained for each candidate direction A linear prefit of the hits allows to give a first estimate of the best track positions and times The problem is lin ear since the direction is fixed and M estimator PDFs are used which do not take 100 4 4 Reconstruction background hits into account A chi squared estimate is constructed based on the previously selected hits to determine the final muon parameters in each of the remaining directions Finally the track candidates are ordered according to the number of selected hits and the chi square values To improve the results a final fit is thus made on all hits using the complete PDFs with background hits 4 4 5 Energy Reconstruction The algorithm used to calculate the energy of the muon is similar for both Aart Strategy and ScanFit and was first developed by Alain Romeyer 97 The algo rithm was chosen in favor of two alternative method
189. se hits are referred to as LO hits A LO trigger is generated internally within the ARS when the PMT output sig nal reaches the low amplitude threshold designed to select single photo electron SPE hits see Section 2 1 1 This causes the charge the time stamp TS and the time to voltage converter TVC value of each accepted SPE hit to be temporarily stored in a pipeline memory The L1 trigger is based on local coincidences within a storey and corresponds to a local trigger search This trigger level is used to discriminate between uncor related background and muon signals A local coincidence consists of at least two 31 The ANTARES data acquisition system hits on two different PMTs of the same storey within a time window of typically 20 ns As a consequence the L1 trigger is built out of a time coincidence between at least two LO triggers from the same storey The time window compensates the internal delays specific to each optical module OM The L1 trigger software also searches for single hits with a large charge These large hits are selected when the PMT output signal exceeds the high voltage threshold of the ARS that is typically equivalent to 3 photon electrons p e It is unlikely that these hits are produced by single photons from uncorrelated background light but could correspond to two or more hits on the same PMT created by a muon traversing the telescope The L1 trigger software sends a readout request to each hit OM of a loc
190. sented onte Carlo simulations are of prime importance when studying the performance of a Cherenkov neutrino telescope as it enables the evaluation of the response of the telescope to neutrino interac y tions Two distinct stages of simulation can be recog nised the neutrino event generation and the de tector response simulation Event generators accu y rately model the neutrino interactions in the me dia surrounding the Cherenkov neutrino telescope of interest in the energy range covered by the tele scope The resulting secondary particles such as muons are propagated through the surrounding media towards the telescope using separate computer programs or subroutines Neutrino event generation is the subject of the present chapter In the second stage the detector response to the Cherenkov light produced by the secondary leptons within the instrumented volume of the telescope is simulated This second part is discussed in Chapter 4 The chapter is organised as follows In Section 3 1 the general concepts of neutrino event generators are explained In 3 2 the neutrino generator GENHEN or GENerator of High Energy Neutrinos is presented The neutrino genera tor ANIS or All Neutrino Interaction Simulation is the subject of Section 3 3 When using ANIS the produced secondary leptons need to be propagated to wards the neutrino telescope using the lepton propagation code MMC or Monte Carlo Muon Code MMC is also descri
191. ses The combined digital information on the time and the charge of the PMT constitute a single photon electron SPE hit A field programmable gate array FPGA buffers the hits from each ARS into so called data frames A frame contains all hits produced by a particular ARS within a predefined time interval of typically 13 ms Time information is provided by the same clock signal as used 30 2 1 Data Acquisition for timestamping of the PMT output signals by the ARS The digitized data pro duced by the ARS chips are transmitted to shore via the master control module MLCM by means of an optical fibre system Each processor in the DAQ system uses the Ethernet protocol for communication The local control modules LCMs are operating at a data transmission rate of 100 Mbit s The data of the six ARS chip of each local control module LCM is sent via the optical fibres to the master LCM MLCM of the corresponding sector In the MLCM the Ethernet links from all five LCMs in the sector are merged At the string controle module SCM the optical signals from the six MLCMs of each detection line are collected and sent to shore via the junction box JB through the main electro optical cable MEOC At the shore the links are connected to a 1 Gbit Ethernet switch All data frames are sent as separate packages to shore where they are pro cessed by a set of computers which constitute the ANTARES on line PC farm The frames from all ARS chips toget
192. shore ADTS It allows the DAQ system to handle the large amount of data created by the telescope without loosing information 2 1 1 Readout of the data A Cherenkov photon arriving at the photo cathode inside a photo multiplier tube PMT can liberate an electron with a probability equal to the PMT s quantum efficiency The electron is referred to as a photon electron p e It induces an amplified electrical signal on the anode of the PMT When the amplitude of the electrical pulse exceeds a threshold the signal is digitized by the front end Ana logue Ring Sampler ARS chip of the PMT readout system A voltage threshold is applied to exclude small signals due to electric noise in the PMT Its value is typically equivalent to 30 of the signal induced by a single photon electron hit At the moment of threshold crossing the ARS provides a time stamp TS using a local clock an amplitude to voltage converter AVC value and a time to voltage converter TVC to interpolate between clock pulses It also starts the integration of the anode current over a programmable time interval of typically 25 ns yield ing an estimate of the charge Time reference is provided by the clock distribution system After digitisation of the PMT s analogue signal the ARS is inactive for a pe riod of approximately 250 ns To reduce the associated dead time effects on the data taking each PMT is read out by two ARS chips which process consecutively analogue pul
193. sists of a number of optical beacons several LED Beacons and a laser beacon which can emit short and intense light pulses few nanoseconds in the part of the spectrum to which ANTARES is sensitive 350 550 nm The LED beacons are fixed onto the optical module frames OMFs at four different loca tions on each detection line of the telescope They can illuminate up to 10 storeys of each neighbour detection line The laser beacon is located on the bottom string socket BSS of the instrumentation line It can illuminate a large part of the bot tom half of the telescope This offers an important redundancy for the time cali bration of the photo multiplier tubes PMTs located on the first storeys of every string which are less well illuminated by the LED beacons 34 2 2 Relative time calibration Besides the laser beacon the prototype instrumentation line MILOM con tained two additional LED beacons attached to its bottom and top storeys see Chapter 1 This set up was used to evaluate the time calibration of ANTARES ahead of the full production of the twelve detection lines The beacons are flashed independently with a frequency of a few hertz emitting a light pulse when receiv ing a signal from the on shore clock The time of emission of the generated light pulses is being recorded by an internal PMT in each optical beacon 2 2 2 Relative time calibration of the MILOM A dedicated optical beacon trigger is used to find correlations bet
194. sit the ANTARES group at CPPM in Marseille for several months and I would like to thank everyone there for their warm wel come Thank you especially to Paschal Coyle and John Carr for supervising my research during my stay I enjoyed a lot working with the instrumentation line and testing the time calibration software in the dark room Thank you to all my colleagues and friends at Nikhef for your help your support and most of all for your good company Since I joined the ANTARES group it has been changing over time and I would like to thank each one of you Garmt Aart Bram Mieke and the new members Jelena Salvatore Ana and Patrick A big hug especially for Gordon Guus Eleonora and Dimitris I wish you all the best and good courage with finishing your PhD theses Ronald alias Mister Wolf my officemate thank you for always solving my problems and being such a kind friend Do not worry I never was angry I just wanted you 151 SIRENE software implementation to give me some more MNM s 21 Groetjes and tanti bacci to my dear housemates Maaike and Jacopo I had the best of time with you and our friends in Amsterdam I already miss the fabu lous pasta dishes of Restorante La Maison and La Tratoria di Angelo of course the nights of MMORPGs playing the Fridays at Melkweg the salsa dancing at Cantinero and so many more memories Miruna and Bram thank you so much for always being there for me espe cially durin
195. spheric muons dashed at the trigger level The astrophysical neutrino flux corresponds to an AGN like spectrum The atmospheric spectrum has been extrapolated to higher charge values assuming an E spectrum following the primary cosmic ray spectrum at UHE 120 5 2 Data selection and analysis u 10 E so C oO a ba S C D L a N bs g a a 10 E D C gt ls H 10 1 L L L l L L L L L 1 0 8 0 6 0 4 0 2 0 Cosine of the Zenith angle Number of events per year 1 2 3 4 5 6 7 log i go imiten total charge p e Figure 5 9 Distributions of the cosine of the muon zenith angle top and the total charge of the hits bottom induced in the telescope by atmospheric muon background dashed and cosmic downward going UHE neutrinos solid Only those events which survived both cuts mentioned in the text are shown 121 High Energy Simulations Results where i99 Nons Np is the 90 confidence interval as a function of the number of observed events Ny and the expected background N The variable N is the total number of signal events The event rates shown in Figure 5 9 need to be integrated to determine the number of expected signal events Ns and background events N above a given value of the total charge of the hits and the cosine of the zenith angle These integrated distributions are shown for the cosine of the muon zenith angle in Figure 5 10 top and for the total charge of the hits in Fi
196. sumed astrophysical neutrino flux 5 2 1 Track energy estimate As can be seen in Figure 5 2 the muon energy spectrum expected from an AGN like neutrino flux at the detector extends to larger values than that of the atmo spheric muons Thanks to the steeper slope of the atmospheric spectrum the signal over background ratio improves with energy In principle it is thus possi ble to search for an excess of astrophysical neutrinos at higher energies Since for ANTARES no reconstruction program is available for ultra high energy muons the reconstructed muon energy cannot be easily used to distinguish muons in duced by cosmic neutrino interactions from atmospheric muons On the other hand the muon energy can be estimated 89 from the muon energy loss on its way to and through the telescope see Chapter 5 In fact the amount of light emitted by the muons and the associated electromagnetic showers provides an indirect measure of the muon energy It should be realized though that the num ber of photons hitting the photo multiplier tubes PMTs in the telescope depends strongly on the position of the track relative to the telescope see Chapter 5 The amount of light observed in the telescope thus needs to be combined with posi tion and direction information to estimate the muon energy without having to rely ona full geometrical reconstruction As discussed in Chapter 3 the trigger software simulates the digitisation of the hits by the front e
197. t Q lt oo For an isoscalar nucleus of mass M i e a nucleus with an equal number of protons and neutrons N ae the leading order electroweak theory predicts the differential cross section for neutrino nucleon DIS scattering as described by do _ 2G7ME dxdy 7 My 2 2 2 ZOE ONE 3 7 P MZ xq x Q xq x Q 1 y where My is the mass of the intermediate boson W and Gr 1 16632 x 10 GeV is the Fermi constant This differential cross section is expressed per nu cleon and differential in terms of the Bjorken scaling variables x and y It is spe cific for neutrinos for anti neutrinos q x Q and q x Q need to be exchanged The hadronic part of the neutrino nucleon interaction leads to the appearance of the parton distribution functions q x Q and q x Q in Equation 3 7 The PDFs represent the probability density for finding a particle with a certain longi tudinal momentum fraction x and momentum transfer Q The differential cross sections in Equation 3 7 depend on both the PDFs of quarks q x Q and anti quarks q x Q The known PDFs are obtained by using experimental data 41 Experimentally determined PDFs are available from various groups worldwide such as the CTEQ 42 MRS MRST MSTW 43 H1 44 and ZEUS 45 46 col laborations 50 3 1 Event Generators and Propagators 3 1 2 Lepton propagators Muon energy losses in matter at high energies Only high energy muons originating from neutrino
198. t Sheldon Glashow who discovered this phenomenon A detailed discussion of this process can be found in 35 Tau neutrinos in CC interactions with nuclei produce tau leptons which can decay again into tau neutrinos The secondary tau neutrinos may on their turn interact with nuclei This is the tau neutrino generation chain The decay of taus is often simulated in neutrino generators using the package TAUOLA 36 see Sections 3 2 and 3 3 Predominant interaction The three neutrino flavors are usually assumed to be produced in their astrophys ical sources in the ratio ve vy v7 1 2 1 While propagating from their source over cosmological distances neutrinos oscillate between the three flavors This results in a ratio when arriving on Earth of ve vy v7 1 1 1 37 For muon neu trinos and muon anti neutrinos the detection probability is higher due to the long absorption length of muons produced in the charged current CC interac tions The primary interaction of interest for the ANTARES telescope is therefore the CC muon neutrino scattering with nuclei 46 3 1 Event Generators and Propagators Various scattering processes must be taken into account while generating neu trino interactions with matter deep inelastic scattering DIS quasi elastic nu cleon QE scattering several nucleon N and A resonance RES channels with mass values below 2 GeV In the QE channel of the CC muon neutrino interac tion with nuclei vat NSS
199. tectors at ground level The measured cosmic ray spectrum is shown in Figure 1 As can be seen in the figure the spectrum steepens at an energy of about 3 x 101 eV which is referred to as the knee and flattens near 3 x 10 8 eV at the ankle The flux of cosmic rays is well described by a broken power law of the form p E E 7 0 1 Very high energy Cosmic Rays 104 JACEE MGU TienShan Tibet07 Akeno CASA MIA Hegra Flys Eye Agasa HiRes HiRes2 Auger SD Auger hybrid Kascade E27F E GeV m gt Ww Pe GBeoe oerBeded dd a 1013 1014 1015 1016 1017 1018 1019 1020 E eV Figure 1 The measured cosmic ray spectrum between 1013 and 10 eV The shaded area shows the range of direct cosmic ray spectrum measurements 3 with an energy dependent spectral index y such that 2 7 ifE lt 3x10 eV y 3 0 if 10P eV lt E lt 10 eV 27 if E gt 10 eV The shape of the measured spectrum constraints possible models of cosmic ray production It is usually assumed that cosmic rays are produced and acceler ated by sources both within and outside the Galaxy The acceleration mechanism is based on initial suggestions by Enrico Fermi 4 First order Fermi acceleration or diffusive shock acceleration occurs when cosmic ray particles are repeatedly reflected by fast moving magnetic fields in the interstellar medium By bouncing back and forth in the magnetic field randomly lets some of the cosmic ray parti
200. tes the Glashow resonance using an internal routine 62 53 Neutrino event generators 3 2 2 Generation method Geometry The instrumented volume of ANTARES is represented as a cylinder centered on the center of gravity of the telescope and containing all the photo multiplier tubes PMTs A larger cylinder referred to as the Can encompasses the instrumented volume of the detector The Can defines the volume within which Cherenkov light is produced in the Monte Carlo detector simulation when evaluating the response of the telescope The Can is surrounded by a third cylinder the Gen eration Volume in which neutrino interactions with nuclei and atomic electrons in the matter surrounding the telescope are simulated The Generation Volume therefore represents the neutrino interaction volume This volume Veen corre sponds to the Can expanded with the maximum lepton range in the appropriate medium rock or sea water for the maximum value of the energy range Emax to be generated The range Rl with p e T is determined for the simu lated neutrino topology assuming that the lepton takes all the neutrino energy 54 In the downwards direction Veen is determined by the maximum lepton range in rock Rmax rock The lower bound of the Can only extends to the seabed as leptons below this point won t be able to produce detectable light Horizontally Vee is calculated using the maximum lepton range in sea water Rinax water x cos Omax with
201. the estimated angular resolution of 0 2 22 3 The ARS TVC calibration ARS TVC decoding The Analogue Ring Sampler ARS chip provides sub nanosecond time informa tion as binary data a 24 bit time stamp TS and a 8 bit Time to Voltage Converter TVC The Histogrammer module of the package PhAntOM cf 2 3 was used to make TVC distributions An exemple of the TVC spectrum is shown in figure 2 2 The ARS has two TVCs which are used alternately to interpolate between con secutive clock pulses The TVC response is assumed to be linear TVCs operate in pairs to correct for the dead time the electronics need to come back from the maximum TVC value to the minimum As can be seen in figure 2 3 the TVC pe 35 The ANTARES data acquisition system OM 1 OM 2 OM 3 7000 s 0 30 soe s 0 35 6000 Counts 5000 5000 4000 4000 3000 3000 2000 2000 1000 1000 4 2 0 2 4 4 2 0 2 4 4 2 0 2 4 Dt ns Figure 2 1 Distribution of measured time differences from OMs of the MILOM and the Led Beacon used as reference The results of a Gaussian fit are shown Data were taken from the run 13440 with the MILOM riod has a positive offset with respect to the clock signal In order to correct for this offset both the time stamp and TVC information are needed Decoding of the time stamp and the TVC values are implemented in the trigger software 25 of ANTARES For each TVC value the offset has been deter
202. the telescope The solid circles represent the hits that were used for the fit The crosses represent other single photon hits occuring during the same time interval in the telescope They consist of random background 21 22 result of the reconstruction of an atmospheric downward going muon is shown Data have been taken continuously for more than two years now While the num ber of detection lines deployed was increasing the efficiency of the detection was increasing as well With five lines connected it was possible to determine the first neutrino event candidates and estimate the performance of the telescope It can be expressed in terms of sensitivity or upper flux limit that indicates the mini mum detectable signal from an astrophysical source The sensitivity of ANTARES after one year of observation with five detection lines is shown in Figure 1 10 The 26 1 4 Development and Construction full telescope is operational since May 30 of 2008 The search for cosmic neu trino sources has now started ya ANTARES 5 Lines binned 140 days y MACRO Source List 2299 5 days c o Super K Source List 1645 9 days 4 AMANDA I unbinned 199 3 days 5 10 AMANDA Source List 985 5 days 5 E L L i Lu 10 10 80 60 40 20 0 20 40 60 80 6 Figure 1 10 Estimated sensitivity to point like sources with the ANTARES 5 Line data Background has been estimated from Monte Carlo simulations Comparis
203. tion Vector lt PMT EXTEND gt Vector lt Abstract Module gt derived from Figure C 1 Schematic overview of the classes of the detector simulation program SIRENE All currently implemented classes are shown Dark grey boxes represent classes used for the geometry of the telescope Light grey boxes represent classes used to represent the tracking of the secondary particles and the produced Cherenkov photons PMT The class PMT defines a photo multiplier tube with its main character istics such as its effective area time transit TT time transit spread TTS am plitude spread and the quantum efficiency of its cathode These parameters can be read from external cards so the program does not need to be recompiled if a different type of PMT has to be used The default values correspond to the ANTARES PMT purchased from Hamamatsu 91 PMT EXTEND The class PMT EXTEND provides the orientation and position of a photo multiplier tube inside the module It is derived from the classes PMT and GEOMETRY ABSTRACT MODULE The class ABSTRACT MODULE defines a virtual sphere with a given radius This class is used to track the Cherenkov photons towards the modules of the telescope and constitutes a pure artificial object of the simula tion which does not represent a physical object The radius of the virtual sphere can be chosen such that it envelops one or more modules of the telescope The surface of the cross section through the cente
204. tion In the present analysis only downward going events will thus be considered The detector simulation program SIRENE see Chapter 5 has been used to determine the photon field and the hits in the telescope In Figure 5 2 the event rates are shown with a solid line as a function of the muon energy for those events which could produce a detectable signal In the figure only downward 3The energy spectrum of cosmic rays primaries follows an E spectrum for energies below 1015 eV the knee and above 10 eV the ankle In between these two limits it is proportional to an E spectrum Since this work focuses on the highest energies the cosmic ray energy spectrum has been approximated with an E energy spectrum 111 High Energy Simulations Results Event rates per year 1 2 3 4 5 6 7 8 9 10 11 log F Muon energy GeV Figure 5 2 Energy distribution of muons from downward going ultra high energy UHE astrophysical neutrino solid and cosmic ray dashed interactions in the atmo sphere Only events which can reach the detector are shown The UHE events have been simulated using ANIS assuming an AGN like spectrum The atmospheric muons have been simulated using MUPAGE The atmospheric spectrum has been extrapolated beyond 500 TeV assuming an E spectrum going events have been used and the ANTARES twelve line geometry is assumed 51 2 Atmospheric muon background simulation The spectrum of atmospheric muon
205. to be taken care of when analysing reconstructed muon trajectories for neutrino source searches are discussed in Section 1 2 The design and architecture of the ANTARES telescope are described in detail in Section 1 3 12 1 1 Neutrino detection principle The history of the construction of the telescope and its present status are pre sented in Section 1 4 1 1 Neutrino detection principle The ANTARES telescope is based on the original idea of Moisei Aleksandrovich Markov 17 who proposed in 1960 the deep sea or ice as a suitable site for the construction of a large neutrino telescope Open water neutrino telescopes use the Earth as a shield to filter all particles except neutrinos and detect Cherenkov light emitted by neutrino induced muons or neutrino induced electromagnetic and hadronic showers Neutrino charged current interactions in the water sur rounding the detector and in the rock below the detector give rise to upward going charged secondary particles such as muons and electrons depending on the flavor of the neutrino The relativistic secondary particles moving through the seawater faster than the speed of light in that medium will emit photons in the blue wavelength range These photons are known as Cherenkov light named after the Russian scientist Pavel Alekseyevich Cherenkov who discovered this phenomenon in 1934 Cherenkov radiation occurs when a charged particle ex ceeds the speed of light in a dielectric and optically tra
206. tween well reconstructed error on the reconstructed angle lt 1 and the badly reconstructed events error a gt 45 99 Detector simulation and Track reconstruction 4 4 4 Track reconstruction with ScanFit Probability Density Functions The PDF of hit i which describes the signal in ScanFit consists of a pure Gaussian distribution of a given width g given by ia 1 i th fi ti Se 7 rz where is the vector of the muon parameters This approximation is useful when applied with least square fitting methods as it simplifies the optimisation procedures Since maxima are not affected by monotone transformations one can take the logarithm of the associated likelihood to turn the product of individual PDFs into a sum The likelihood function L f can be expressed as an equivalent logarithmic likelihood 20 ne Ee i t t L t _ Inf t In2z Ino 16 Din fil D As the first two terms of the root mean square are constant the maximisation of the likelihood depends on the minimisation of the last term of the sum only In stead of using a chi square distribution for the minimisation preference is given to an M estimator in order to reduce the influence of outliers The function g r of the M estimator method is taken to be t t 1 t t ain ee i When accounting for background hits ScanFit uses the complete PDFs with back ground hit
207. uce a non negligible error considering the high resolution of neutrino telescopes such as ANTARES 88 55 79 Detector simulation and Track reconstruction m 0 224 p ee E en ee ase velocity E 0 222 des gt 8 0 22 0 gt 0 218 group velocity 0 216 OF pressure correction 024 ema 0 212 350 400 450 500 550 600 wavelength nm Figure 4 1 Comparison of measurements of the group velocity of light in sea water at the ANTARES site and model prediction The phase velocity is also shown 86 The group velocity of the photons in the medium and the index of refraction Nmedium are related by gredium c 1 etn 8 Nmedium dr Nmedium This equation is used to define the group index of refraction medium n Ng A A dn 1 HAN which will be used together with the parametrisation given by Equation 4 5 to determine the path length and the arrival time of the photons at the OMs The wavelength dependence of the group velocity and consequently the index of refraction group of light induces a spread in the arrival times of the Cherenkov photons at the OMs Light absorption The absorption length Lijegiym of the transparent medium sur rounding the telescope is the mean free path of a photon before undergoing a 80 4 1 SIRENE non elastic interaction in the medium It is wavelength dependent For sea water at the ANTARES site the absorption length is parametrised by the model de
208. uch as ANTARES are optimized for the detection of upward going neu trinos in the energy range from 10 to 1016 eV it needs to be investigated whether downward going neutrinos of energies in excess of 1016 eV can be observed This 133 High Energy Simulations Results is of relevance not only for the ANTARES telescope but also for the future cubic kilometer scale detector KM3NeT Moreover the present study will provide in formation for the optimisation of the geometry of the KM3NeT telescope and the technologies required A non localised flux of high energy neutrinos has been considered in order to enhance the probability of detection As muon neutrinos are more easy to detect only this neutrino flavor has been taken into account in the data analysis The diffuse flux of muon neutrinos from AGN has been modelled by a generic E spectrum with units of 10 6GeV icm s isr1 2 The performance of ANTARES relies on the timing resolution of the signals recorded in the photo sensors of the telescope and the proper operation of the various mechanical and electronic components of the telescope A time calibra tion method and a diagnostic tool to determine if components are functioning as espected have been developed The time calibration method has been tested with the MILOM prototype instrumentation line of ANTARES to provide preliminary information on the time calibration of the entire telescope The method consists of the determination of the time
209. ue TVC value rer re LAS EDT PE CAE er eae Cn TO ec 0 50 100 150 Figure 2 2 Top raw TVC spectra Bottom corresponding integrated TVC spectra and fits as mentionned in the text Data were taken with the run 11236 with the MILOM is exactly one clock period Any change of these two extrema will result in the change of the TVC slope To determine the slope we calculate the integral of the raw TVC spectrum The integral of the TVC spectrum from figure 2 2 top is also shown in figure 2 2 bottom As can be seen from the figure the integrated TVC signal is indeed a straight line as expected from the transfer function equation 2 1 The fit function corresponds to a straight line between the two points TV Cin 0 and TV Cmax A where A is the normalisation of the histogram The fit result is shown in Figure 2 2 bottom The fit describes the observed distribution well An internal ANTARES note 30 is dedicated to the calibration procedure of the TVC of the ARS Runs 11236 through 11262 have been used to study the time stability of the ARS TVC Data in these runs were collected with the Instrumentation Line shortly after its deployment The evolution in time of the variables TVC jj and TV Cmax of the two ramps of the TVC are shown in figure 2 4 As can be seen in the figure the ARS TVC dynamic ranges appear to be stable at least for the period of time during which the runs 11236 through 11262 were taken The data missing in the different histo
210. uning is required to improve the performance of the reconstruction programs in conjunction with SIRENE However the program that is used for energy reconstruction was found to perform well with SIRENE in the energy range below 10 eV None of the available reconstruction algorithms was found suitable at ultra high energies and a different approach has been envisaged to estimate the ANTARES sensitivity in this domain When searching for neutrinos in the ultra high energy domain the back ground consists of atmospheric muons which can reach the detector They result from interactions of cosmic ray primaries in the Earth s atmosphere and can be mis interpreted as secondary leptons from cosmic neutrinos The program MU 135 High Energy Simulations Results PAGE 100 has been used to simulate the atmospheric muons impinging on the ANTARES detector surface MUPAGE is a parametrisation of the atmospheric muon flux based on HEMAS 101 simulating single and multiple underwa ter muons at the depth of the detector A dedicated data analysis method has been developed in order to reduce the atmospheric muon background while pre serving sensitivity to the ultra high energy signal The method relies on differ ences between the signal and the atmospheric muon background regarding the energy of the muons and the direction from which they arrive at the detector The amount of light emitted by the muons and the associated electromagnetic showers provide
211. uons from cos mic muon neutrino interactions in the matter below the telescope and represent a serious background for neutrino astronomy To simplify the discrimination between up ward going muons and the much higher flux of downward going muons the large background of atmospheric muons and muon bundles induced by cosmic ray atmospheric showers needs to be minimized This can be realized by installing the telescope at a deep site where a large layer of seawater will act as a shield At a depth of 2475 m the expected background rate in ANTARES from atmospheric cosmic ray muons is about 30 Hz To improve the filtering of the atmospheric background the photo multiplier tubes PMTs of the telescope are oriented 45 with respect to the vertical downward direction thus focusing their apperture on the upward going muons This specific orientation towards the seabed also prevents bio fouling to be a serious problem The signal loss due to bio deposition and sedimentation has been measured 18 to be less than 2 per year The expected rate of atmospheric muons at the ANTARES site is small com pared to the effect of the B decay of potassium K in sea water which contributes 30 kHz to the single rate of each PMT It is also rather insignificant with respect to the rates due to occasional bio luminescence caused by living organisms around the telescope which can peak during short bursts up to MHz rates and is depen dent on the strengh of the sea current
212. ur N 3 1 the final nucleon N acquires some momentum but remains intact The nucleon and A resonance channels are given by Vina N ES pr N A SL hadrons 3 2 In this case a nucleon resonance N or Delta resonance A is produced that decays into hadrons High energy neutrino generators often neglect the QE and RES channels since their contribution is only significant at low energies E lt 10 GeV Hence the DIS neutrino nucleon CC scattering is the dominant interaction of neutrinos and anti neutrinos in conventional matter in the energy range covered by ANTARES E gt 10 GeV The DIS channel of the CC muon neutrino interaction with nuclei can be expressed as hadronisation gt vaN Spo X u hadrons 3 3 reflecting the disintegration of the struck nucleon N into partonic fragments X that hadronize into various hadrons Figure 3 1 shows the corresponding Feyn man diagram The outgoing muon is scattered at the interaction vertex over a certain angle The average angular difference between the initial neutrino direc tion and the emerging muon direction 24 is bound by the following inequality Ou lt 3 4 E TeV where E represents the neutrino energy For muons below 10 TeV the angular resolution of the neutrino telescope ANTARES is dominated by the v u scat tering angle described by Equation 3 4 Above 10 TeV detector effects such as the water quality or the timing resolution relative and absolute
213. water at the ANTARES site In Figure B 1 the distribution of the final energy of the muons that crossed 300 meters of Antares water with a fixed initial energy equal to 100 TeV is shown Several values of Veut were used to determine which relative energy cutoff would give the most accu rate results For higher vet the muon spectra are not continuous but present a peak at the energy Epeak which is separated from the distribution by at least the value of Veut X Epeak This is a known behaviour of MMC In order to be able to use MMC with high vet a continuous randomisation feature was introduced The results when applied to the ANTARES water are shown on Figure B 2 As can be seen on the figure the distributions have become continuous but con verge faster than those without the continuous randomisation option In order to bring the number of separate energy loss events to a reasonable value and there fore to increase the running speed when using MMC with ANTARES we will use a rather high value of the relative energy cutoff veur 0 05 with the continuous randomisation feature The salinity of the water has been measured to be 38 44 0 02 g kg at the ANTARES site whereas it was cited as 35 0 kg in 115 for sea water in gen eral Therefore the chemical composition of the water at the ANTARES site is the same as the water defined in 115 but with different densities for the elements containing salt The chemical composition of the sea wate
214. ween the di rect hits generated by the optical beacon and the other hits The trigger software searches for frames with an identifier equal to six which corresponds to an ARS of the Optical Beacon This ARS is taken as a reference The calibrated hit time is calculated from the calibrated time to voltage converter TVC and the corrected time stamp TS values For more information on the trigger software we refer to 24 and 25 The method that was applied to calibrate the TVC of the Ana logue Ring Sample ARS chip is described in Section 2 3 Triggered SPE hits from physics events are selected by the optical beacon trigger and used to make his tograms of the time differences between the ARSs of the optical modules and an ARS of the optical beacon In figure 2 1 an example of a histogram of the time differences between a particular ARS of an optical module and an ARS of the Optical Beacon is shown As can be seen from the figure the distribution shows a clear peak centered on a value close to zero The fit of a Gaussian distribution with mean u and variance o on top of a flat background distribution is shown in each of the three cases The fit describes the data well Typically the mean is less than 0 20 ns and the variance is about 0 30 ns This result meets the requirements of the ANTARES telescope according to which all electronics and calibration sys tems should contribute less than 0 50 ns to the overall timing resolution in order to reach
215. ximum likelihood minimum chi square or M estimate ap proaches In those cases the model prediction can be linearised by only consid ering signal hits and no background hits In the last step of the reconstruction procedure only after the iterative process has converged towards an unique so lution background hits are usually taken into account to give a more precise esti mate of the muon track geometry The PDF that describes the background hits is a constant function of the hit time corresponding to a known rate of uncorrelated bioluminescence and K decay hits usually 70 kHz The relative contribution of the signal and background hits is determined from the hit amplitude A the photon path length and the incident angle inc of the photons on the optical modules OMs The PDF of the hit residual r can 96 4 4 Reconstruction be expressed as a weighted sum of both the PDFs of signal and background hits such that 1 s b filrilli binc i Ai NTF NER li Ones A f 5 li 8incir Ai 4 11 where N is the number of signal hits and N is the total number of hits in the event Both N8 and N depend on Jj inci and Aj The number of signal and background hits is calculated from an empirical fit to simulated events This ensures that the PDF is normalised for any value of 1 0 and Aj 4 4 2 Linear least square methods Maximum Likelihood Principle The estimate is given by the values that maximise the likelihood L i
216. y to their source Thanks to their particular characteristics neutrinos constitute an ideal high energy probe to observe distant astrophysical objects and provide in formation on the dynamics of the most energetic phenomena of the Universe The study of the detectability of very high energy cosmic neutrinos is the central topic of this thesis Such neutrinos are expected to be produced in the interac tions of high energy cosmic rays with ambient matter and or photons 2 Mod els of cosmic ray production are therefore presented in Section 0 1 The shape of the observed cosmic ray energy spectrum is also discussed The problem of the Greisen Zatsepin Kuz min limit and its consequences on the possible observa tion of ultra high energy cosmic rays is described in Section 0 2 The most likely sources of very high energy neutrinos active galactic nuclei AGN are discussed in Section 0 3 Finally the main questions addressed by this thesis are outlined in Section 0 4 0 1 Very high energy Cosmic Rays Cosmic rays are atomic nuclei predominantly protons and electrons which bom bard the Earth from beyond the atmosphere with very high energies The lowest energy cosmic rays are detected directly by experiments on board of satellites or high altitude balloons before they are absorbed in the atmosphere High energy cosmic rays however are detected indirectly through the extensive air showers they produce which can be observed by an array of particle de
217. y experts from IFREMER In stitut Francais de Recherche pour Exploration de la MER to deploy the different components of the tele scope in the deep Mediterranean Sea Seismologists from G osciences Azur and Guralp System benefit from the ANTARES infras tructure to record in real time deep sea seismic activity data occurring both in the vicinity of the telescope or anywhere else in the world ANTARES is an open water neutrino telescope composed of photo sensors optimised for the detection of Cherenkov light emitted by upward going muons resulting from high energy cosmic muon neutrinos interacting with atomic nuclei in the deep water or seabed Unlike high energy photons that can be absorbed before reaching the Earth or protons which can be deflected by magnetic fields neutrinos can travel over very large distances without any disruption straight from their origin Neutrinos can provide new knowledge about astrophysical processes alongside photons and cosmic ray particles By observing neutrinos ANTARES is able to explore new features of the Universe over a large range of distances and energies A neutrino telescope can study for instance the processes involved in the most 11 The ANTARES telescope Amsterdam i NGT Moscow Groningers gt o Erlangen IFREMER __ paris Brest x Mulhouse Sacla Strasbourg Marseille oo Bologna en i oma IFREMER Pisa es Toulon Bari Val ncia Catan

Ultra-high energy neutrino simulations

Contents

Download Pdf Manuals

Related Search

Related Contents