1. Outline 1)Motivation of neutrino astronomy 2)Design of AMANDA/IceCube 3)Detection method -Data processing -Detection of ultra high energy neutrinos 4)Results

2. Why neutrino astronomy? Directional information of charged particles scrambled Energetic photons absorbed by pair production -γ 1 + γ 2 → e + + e −, for 4E 1 E 2 > (2m e ) 2    PeV limit for gamma astronomy For E> 50 EeV proton astronomy may be feasible as directions are no longer scrambled by the ambient magnetic field of our own galaxy Leaves neutrinos! -But: directional information determined by their gyroradius in the intergalactic magnetic field which is still an issue of speculation,    resolution may be anywhere from sub-degree to nonexistent Neutrons, pions etc. barely live long enough to reach us  same applies to E prot > 50 EeV  absorption length l=50Mpc  GZK Cut-Off

3. Neutrino source candidates Active Galaxy Nuclei (Blazars) Black hole with 10 8 x mass of sun  10 6 LJ extra-galactic – GRB? Supernova remnant Microquasar (SS433 etc.) Crab nebula Black hole with  mass of sun  1 LJ galactic

4. A cosmic accelerator

5. Requirements to generate the high energy cosmic rays Cosmic accelerators R > R gyro = E/B condition yields E max just by dimensional analysis  may not be beams of cosmic accelerators but decay products of remnants from the early Universe  cosmic accelerators must reach efficiencies matching the dimensional limit

6. Cosmic accelerators

7. Predicted neutrino fluxes ––– represents a source where each cosmic ray interacts only once before escaping the object ––– ideal neutrino source where all cosmic rays escape in the form of neutrinos

8. AMANDA/IceCube

10. Design of AMANDA/IceCube

11. 1994: AMANDA A 4 strings, 73 OM 1997: AMANDA B10 10 strings, 302 OM 2000 AMANDA II: 19 Strings, 667 OM

12. 2005-2010: IceCube 80 strings 4800 OM

13. Optical Module PMT achieves 10 9 -fold amplification Time resolution in laboratory 5 ns

15. Detection Method Neutrinos are detected indirectly, following a DIS on a target nucleus –Charge current   p    X   p    X e  N    e  X –Neutral current e,   N    e,   X

16.  triggers two types of events –Charged current   N   X e  N    e  X 1.Cascades –Neutral current e,   N    e,   X

17.  triggers two types of events -Only charged current    N   X 2.Muon tracks  15 m -At E> TeV  and  collinear: - Reconstruction of  trajectory allows identification of direction ! → Main focus on muon events  emitting cerenkov radiation along the track

18. Calorimetry in Ice Absorption length ~100 m Scattering length ~25 m Light is isotropized well before it is absorbed. To first order, sampling is insensitive to geometric position or PMT orientation. Current arrays sample a very small fraction of the total Cerenkov light… ~ 10 -5 for AMANDA and IceCube. PMTs on a string… ~20 m spacing between PMTs String spacing… ~100 m spacing between strings. String spacing determines energy threshold.  Scattered photons complicate  track reconstruction

19. Acoustic Detection Attenuation length of sea water at 15-30 kHz: a few km (light: a few tens of meters) given a large initial signal, huge detection volumes can be achieved. Threshold > 10 PeV Particle cascade → ionization → heat → pressure wave

20. Muon tracks in AMANDA

21. E µ = 10 TeVE µ = 6 PeV Muon tracks in IceCube

22. Cascades in IceCube

23.    decays Double Bang! Cascades in IceCube

24. Elimination of muon background Look for upward going tracks! 1: Cosmic v 2: Atmospheric v 3+4: Atmospheric µ Exception: muons → Traditional way to detect ν in AMANDA: Challenge: AMANDA observes about 3-4 atmospheric neutrinos/day in a atmospheric muon background 10 6 times larger! Virtually all kind of particles filtered by ice layer before reaching detector

25. Elimination of muon background →Neutrino flux is attenuated as it passes through the earth, but survival rate still high at E < PeV Cos  = 1 Cos  = 0 Zenith angel θ

26. Zenith angle distribution of simulated AMANDA triggers generated by CORSICA –– triggers from downgoing cosmic ray muons ….. triggers produced by atmospheric neutrinos → θ highly efficient cutting variable for eliminating µ background Elimination of muon background

27. Data processing at AMANDA Level 0 Event trigger → Raw data Calibration Level 1 a)Hit cleansing b)Line fit by quick analytical calculation c)Tensor of inertia d)Weak cuts Level 2 Iterative likelihood reconstruction Hard cuts Level 3 a)Quality cuts Detector signal → νµ ?νµ ?

28. AMANDA Triggering System SWAMP shapes and amplifies incoming electrical signals from OM Direct signal to trigger unit, delayed signal to ADC/TDC In case tigger criteria are met infomation of all ADC and TDC are read and stored by DAQ Each event contains following entries: –GPS time, met trigger criteria –For each hit: LE, TOT, ADC value

29. Level 0: Event trigger Multiciply trigger constitute energy trigger Multiciply trigger: minimum number N of hit OM within a time frame T String trigger: minimum number N of hit adjacent OM of one string within a time frame T String trigger lowers energy threshold, hence increases efficiency for low energy events Multiplicity threshold N was raised in line with the increasing number of OM in order to maintain constant trigger rate ~ 100 Hz

30. Operating Mode of Multiplicity Trigger (analog!)

31. Level 1: hit cleansing Rejection of the following hits: Uncalibrated ADC below 50mV Calibrated ADC below 0.1 and above 1000 photo electrons Hit time beyond a timeframe between 0 and 4500 ns TOT beyond a timeframe between 125 and 5000 ns Hits without further hits within 500ns in a vicinity of 100m Non-primary hits

32. Maximum likelihood algorithm of level 2 requires initial values Line fit algorithm produces an initial track based on the hit times with an optional ADC weight Unphysical model ignores geometry of cerenkov cone and the possibility of scattered photons Quick analytical calculation allows for realtime processing Line fit yields a rough guesses for velocity v and zenith angle θ that is used for preselection of events: Θ > 50°…70°, v > 0.1…0.3 m/ns Cuts are very weak, yet have proven efficient for elimination atmospheric muon background and misreconstructed events Level 1: line fit

33. Discrimination between… Muon track Electromagnetic or hadronic cascade Level 1: tensor of inertia

34. „Inertial tensor“ of responding OM, weighted by ADC 3 real eigenvalues: I 1 > I 2 > I 3 Cascade event: punctual distribution: I 1 / I 3 ~1 Track event requires upper limit for ratio: I 1 / I 3 ~0 Ratio I 1 / I 3 is used to eliminate cascade events Again weak but efficient cut Level 1: tensor of inertia

35. I 1 / I 3 ~ 1 I 1 / I 3 ~ 0 E µ = 10 TeV E e = 375 TeV Level 1: tensor of inertia

36. Level 2: likelihood reconstruction Reconstruction of an event can be generalized to the problem of estimating a set of track parameters, given a set of experimental values Parameters determined by maximizing likelihood L Values recorded by AMANDA: time t i, duration TOT i, peak ADC i of each PMT signal Using initial estimates generated by line fit algorithm

37. Hit times give most relevant information Dominant effect on photon arrival times is scattering in ice Simplest time likelihood function: p i is the pdf of observing the measured time t res,i for given values of parameters direct hit: t res ~ 0 Level 2: likelihood reconstruction Algorithm yields better value for zenith angle θ → selection criteria for hard cut

38. Level 3: quality cuts Possible quality cuts are the following: Zenith angle mismatch between two types of fits Likelihood L of the fit Sphericity of photon distributions Track length Number of unscattered photons with t res ~ 0 “Smoothness” of photon distribution along the track Background events which passed zenith angle cuts can still be misreconstructed events, e.g. muons undergoing catastrophic energy loss through bremsstrahlung/decay or that are coincident with other muons

39. Level 3: quality cuts Example 1: track length Short track length = more likely to be background

40. The smoothness is a measure of how regular the photon density is distributed along the track. A well reconstructed muon track is more likely to have a high smoothness. High Low Level 3: quality cuts Example 2: smoothness

41. Level 3: quality cuts Efficiency

42. Ultra High Energy  Detection Background At E > EeV, atmospheric are negligible Atmospheric charm production (c → µ), significant theoretical uncertainty, but becomes dominant at E µ ~10 EeV Atmospheric multi-muon events may mimic higher energy events The latter two backgrounds are angular dependent predicted neutrino fluxes

43. Ultra High Energy  Detection Signal → Earth opaque above a few PeV EeV PeV TeV Signal concentrated around the horizon →

44. Relative Muon Flux from Isotropic 10 EeV Neutrino Flux Absorption of the v 's Limited overburden for v  µ conversion ?? Ultra High Energy  Detection Signal

45. Signal predominantly from horizon Atmospherical background closer to zenith Rejection of atmospherical background straightforward Ultra High Energy  Detection In AMANDA

46.  This background for EeV events can be vetoed by detecting the fringe of the coincident horizontal air shower in an array of water Cerenkov detectors Penetrating muon bundle in shower core Incident cosmic-ray nucleus Threshold ~ 10 17 eV to veto this background Ultra High Energy nm Detection In IceCube with shower veto

47. AMANDA results Atmospheric neutrinos (  ): 60/day Atmospheric muons: 8.6*10 6 /day Lifetime: 135 days Observed Data Pred. Neutrinos Triggered (Level 0) 1,200,000,0004574 Reconstructed upgoing (Level 1+2) 5000571 Pass Quality Cuts (Q ≥ 7) (Level 2) 204273

48. Neutrino sky as seen by AMANDA No clustering observed → No evidence for extraterrestrial neutrinos...

49. Sources The AMANDA Collaboration: Search for Extraterrestrial Point Sources of Neutrinos with AMANDA-II The AMANDA Collaboration: Muon Track Reconstruction and Data Selection Techniques in AMANDA The AMANDA Collaboration: Observation of High Energy Atmospheric Neutrinos with the Antarctic Muon and Neutrino Detector Array Francis Halzen and Dan Hooper: High-energy Neutrino Astronomy: The Cosmic Ray Connection Thomas Feser: Triggerstudien am AMANDA-Detektor Thomas Becka: Entwicklung einer Echtzeit-Datenfilterung für das Neutrinoteleskop AMANDA Steven W. Barwick: Physics and Operation of the AMANDA-II High Energy Neutrino Telescope

50. End