1. Shih-Hao Wang 王士豪 Graduate Institute of Astrophysics & Leung Center for Cosmology and Particle Astrophysics (LeCosPA), National Taiwan University 1 This talk is based on : S.H. Wang, P. Chen, J. Nam, and M. Huang, JCAP 11 (2013) 062; arXiv:1302.1586

2. Cosmic rays with E > 10 18 eV (UHECR) are observed.  Source (AGN, GRB)?  Acceleration mechanism? Neutrinos can be generated when UHECR interact with photon or matter  CMB photon  GZK cutoff in CR spectrum  Cosmogenic neutrinos! 2 Astrophysical: Cosmogenic: from http://www.physics.utah.edu/~whanlon/spectrum.html UHECR F. Halzen and S. Klein, Physics Today 2008 ~E -2.8 Charged pion decay GZK cutoff at ~5×10 19 eV

3. Being electrically neutral and weakly interacting, UHECNs carry information from their sources  Energy spectrum ▪ CR composition, source distribution, max acceleration energy  Flavor ratio: relative flux between ν e : ν μ : ν τ ▪ Source size & B-field strength  ratio transition with E ν (1:2:0  0:1:0) 3 Kotera et al., JCAP (2010); S. Hümmer et al., Astropar. Phys (2010). decay v.s. synchrotron cooling

4. UHECN is an unique probe with  Extremely high energy (>10 17 eV) pc  Extremely long baseline (>10 Mpc) Flavor ratio at Earth is determined by :  Flavor ratio at source  Flavor transition in propagation  Neutrino oscillation  Non-standard physics? 4 J. Beacom et al., PRL (2003, 2004), PRD (2003); G. Barenboim et al., PRD (2003); D. Hooper et al., PRD (2005).; P. Keraenen et al., PLB (2003)

5. 5 Bustamante et al., PRL (2015); J. Beacom et al., PRL (2003, 2004), PRD (2003); G. Barenboim et al., PRD (2003); D. Hooper et al., PRD (2005).; P. Keraenen et al., PLB (2003) (ν e :ν μ :ν τ ) Flavor ratio at Earth is determined by :  Flavor ratio at source ▪ Source environment  Flavor transition during propagation ▪ Neutrino oscillation ▪ Non-standard physics?  Neutrino decay & mass hierarchy  Sterile neutrinos with small Δm 2  CPT violation, Lorentz violation, etc. f e +f μ +f τ =1 this neutrino decay model is used in our work ν e fraction ν μ fractionν τ fraction

6. ~4000 km Antarctic ice as target Vast :thousand km across, 2-3km deep Quiet: less human activity Clean: radio attenuation length ~1km Askaryan effect ~20% excess in negative charge of shower Coherent radio Cherenkov pulse 37 stations, 3 deployed ~200km 2 coverage NTU in collaboration ~50 UHECN events is expected in 3 year 6 ARA collaboration, Astropar. Phys (2012), ARA collaboration, arXiv:1507.08991; http://ara.wipac.wisc.edu/home ARA 2&3 10 months

7. Antenna Low Noise Amplifier RF over Fiber DAQ 7 Adapted from J. Nam

8. 8

9. MMC: http://dima.lbl.gov/work/MUONPR/ SADE: D. Seckel et al. 9 1.Neutrino generation with given spectrum 2.Neutrino and secondary lepton propagation (with MMC package)  Stochastic E-loss > 1PeV is converted to shower 3.Event detection (with SADE package) Credit: MMC ARA Earth ARA37 Spectrum of radio Cherenkov emission Ray propagation in ice ν μ  μ  showers

10. Assuming initial ratio of 1:1:1 Interaction length at 1 EeV is ~500km Neutrino CC interaction produces charged lepton, making the angular distribution different ν e releases all its energy with short distance τ decay & ν τ regeneration (ν τ  τ  ν τ ) Stochastic energy loss of μ & τ 10 red:ν e green:ν μ blue:ν τ Earth shadowing νeνe νμνμ ντντ μ τ zenith angle (cosθ) Up-going e θ red:ν e green:ν μ blue:ν τ

11. 11 red:ν e green:ν μ blue:ν τ Very rare interaction at shallow depth νeνe νμνμ ντντ zenith angle (cosθ) Down-going Assuming initial ratio of 1:1:1 Interaction length at 1 EeV is ~500km Neutrino CC interaction produces charged lepton, making the angular distribution different ν e releases all its energy with short distance τ decay & ν τ regeneration (ν τ  τ  ν τ ) Stochastic energy loss of μ & τ cutoff due to the orientation of Cherekov cone

12. Assuming initial ratio of 1:1:1 12 red:ν e green:ν μ blue:ν τ Interaction length ~ traveling distance νeνe νμνμ ντντ zenith angle (cosθ) Quasi-horizontal νμνμ ντντ μ τ Assuming initial ratio of 1:1:1 Interaction length at 1 EeV is ~500km Neutrino CC interaction produces charged lepton, making the angular distribution different ν e releases all its energy with short distance τ decay & ν τ regeneration (ν τ  τ  ν τ ) Stochastic energy loss of μ & τ

13. υ e has higher detection efficiency due to electron energy loss ν τ distribution is the most different τ decay & ν τ regeneration (ν τ  τ  ν τ ). υ e and υ μ have similar shapes, extra constraint is required  υ μ -ν τ symmetry is assumed (i.e. they are of equal ratio). Initial ratio 1:1:1  detected ratio 0.59:0.15:0.26 13 By fitting event angular distribution, UHECN flavor ratio can be extracted red:ν e green:ν μ blue:ν τ Assuming initial ratio of 1:1:1 Comparison of curve shape

14. Due to random sampling, sometimes fitting gives unphysical f e  fail Generate ensemble of data sets to get the success probability depend on number of events N obs and angular resolution Δθ With Δθ =6 ° and N obs = 200, the success probability is about 85%, 60%, and 50%, respectively 14 ~50% ~60% Δθ =0° (red), 2° (green), 4° (blue), 6° (black) Pseudo- experiment data ~85% Standard 1:1:1 (f e =1/3) Decay with NH 6:1:1 (f e =0.75) ~50% Decay with IH 0:1:1 (f e =0) ~60%

15. Spread of extracted f e from expected one depend on number of events N obs and angular resolution Δθ With Δθ =6 ° and N obs = 200, the resolution (~68%CL) is about 0.3, 0.4, and 0.15, respectively 15 ~0.15 ~0.3 Δθ =0° (red), 2° (green), 4° (blue), 6° (black) ~0.4 Standard 1:1:1 (f e =1/3) Decay with NH 6:1:1 (f e =0.75) Decay with IH 0:1:1 (f e =0) Preliminary constraint on neutrino decay can be set.

16. Measuring UHECN flavor ratio is important in both astrophysics and particle physics By fitting the direction distribution of neutrino events, a preliminary constraint on its flavor ratio can be set for the planned ARA37 configuration. Additional information (e.g. event topology, Cherenkov pulse characteristics) can be exploited to improve the resolution.  Identification of ν e CC events.  Discrimination between electromagnetic and hadronic showers 16 J. Alvarez-Muniz et al. Phys. Rev. D (1999) J. Alvarez-Muniz et al. Phys. Rev. D (2000)

17. 17 http://www.southpolestation.com/

18. 18

19. ν e Fraction f e =[0,1] expected f e = 1/3 (standard), 0.75 (decay w/ NH), 0 (decay w/ IH) Due to random sampling, extracted f e may deviate from expected one Generate ensemble of data sets to get the spread of extracted f e depend on number of events N and angular resolution Δθ With Δθ =6 ° and N = 200, the resolution (~68%CL) is about 0.3, 0.4, and 0.15, respectively preliminary constraint on neutrino decay 19 0:1:1 (decay with IH) 6:1:1 (decay with NH) 1:1:1 (standard) ~0.15 ~0.4 ~0.3 Δθ =0° (red), 2° (green), 4° (blue), 6° (black)

20. What’s Ultra-high energy cosmic neutrino (UHECN)? What’s important about UHECN flavor composition? Detecting UHECN with ARA observatory at South Pole Constraining UHECN flavor ratio with event angular distribution Summery and future work 20

21. Up-going Neutrino CC interaction produces charged lepton Different energy-loss properties of charged leptons in matter lead to different event angular distribution ν τ distribution is the most different τ decay ν τ regeneration (ν τ  τ  ν τ ). υ e and υ μ have similar shapes, extra constraint is required  υ μ -ν τ symmetry is assumed (i.e. they are of equal ratio). Initial ratio 1:1:1  detected ratio 0.59:0.15:0.26 21 By fitting event angular distribution, UHECN flavor ratio can be extracted red:ν e green:ν μ blue:ν τ up-goinghorizontal down-going Assuming initial ratio of 1:1:1 Earth shadowing traveling distance ~ interaction length cutoff due to the orientation of Cherekov cone νeνe νμνμ ντντ μ

22. 22 F. Halzen and S. Klein, Physics Today 2008

23. 23 Hillas plot Size Magnetic field Neutrino source type S. Hümmer et al., Astropar. Phys. 34, 205 (2010).

24.  Neutrinos change flavor during propagation.  The flavor eigenstate is not identical to the mass eigenstate: where U αi ‘s are element of 3×3 mixing matrix.  For cosmic neutrino, The neutrino flux at the source Φ ν0 and that arriving on Earth Φ ν can be related by 24 flavor eigenstate α = e, μ,τmass eigenstate i = 1,2,3 H. Athar, M. Jezabek, and O. Yosuda, Phys. Rev. D 62, 103007 (2000). R.L. Awasthi and S. Choubey, Phys. Rev. D 76, 113002 (2007). (tribimaximal mixing)

25. Neutrinos interact with matter via CC or NC interactions and induce showers.  Increase in negative charges, and decrease in positive ones, developing into ~20% negative charge excess In dense medium, the shower size is compact, emitting coherent Cherenkov radiation at radio frequencies up to few GHz. 25 G.A. Askaryan, JETP Lett. 14, 441 (1962). D. Saltzberg et al., Phys. Rev. Lett. 86, 2802 (2001); P.W. Gorham et al., Phys. Rev. D 72, 023002 (2005); P.W. Gorham et al., Phys. Rev. Lett. 99, 171101 (2007);

26. 26 Cross sections from R. Gandhi et al., Phys. Rev. D(1998) muon Charged leptons ionization loss Bremsstrahlung pair production photonuclear interaction μ, τ decay Neutrinos CC NC

27. 27 The reconstructed e event ratio is a maximum likelihood estimator, and it can be shown that it is asymptotic to a Gaussian distribution as N obs increases. So the resolution R is an approximation to the standard deviation σ of the Gaussian. the spread of its distribution result from finite number of events and imperfect angular resolution. ν e fractionassumed expected event ratio σ Probability distribution of fitted ν e fraction for given N obs and Δθ

28. 28