Erik Strahler Ph.D. Thesis Defense 6/26/2009 Searches for Neutrinos from Gamma Ray Bursts with the AMANDA-II and IceCube Detectors Erik Strahler Ph.D. Thesis Defense 6/26/2009

Outline Motivation Neutrinos from Gamma Ray Bursts Detection AMANDA-II Analysis IceCube Analysis Discussion and Outlook

Why Neutrinos? Neutrinos from GRBs Magnetic Fields Absorption source Charged particles are deflected, photon can be absorbed in opaque materials or highly suppressed at high energy due to CMB photon pair production. Neutrinos are neutral and low cross-section. Also a signature of hadronic interactions. source

The Cosmic Ray Connection Sources of the highest energy cosmic rays are unknown. Need either very large B fields or spatial extent in order for acceleration mechanisms to work. GRBs are one of the few source candidates.

GRBs : A History Nuclear tensions led to the Vela satellite system Discovery in 1969 of short duration, highly energetic bursts of gamma rays coming from space 70’s and 80’s – a Dark Age for GRB physics Many theories without observational backing 1991 – Launch of CGRO BATSE begins large scale observation of GRBs Comets colliding in the Oort cloud, neutron star collapse in the galaxy. Thought must be galactic due to energy requirements.

GRB Observations Isotropic Highly Variable Extremely Energetic X-ray/Optical Afterglows So what are GRBs? Extremely high energy, widely time variable on short time scales, short duration. Isotropic -> cosmological -> very very high energies

Compactness Problem dt ≈ 10-2 s implies Rs = cdt ≈ 3000 km g’s may pair produce if Given typical GRB parameters  tgg ≈ 1013 At odds with observed non-thermal photon spectra Relativistic emission: E  Eobs / G, Rs G2cdt With G≈102, gives tgg < 1 Average optical depth. FD^2 comes from total burst energy E^-a observed spectrum. Fraction that can pair produce drops by Gamma^-2a Yields Tau = 10^13 / Gamma^(4+2a) (Tau goes like 1/R^2)

Fireball Model Relativistically expanding plasma of e+, e-, g, (p?) Initially optically thick G increases with time, until tgg < 1 Kinetic Energy dissipated via shocks Internal between shells with varying G Prompt emission External with ISM Afterglow emssion

Photon Emission Shocks accelerate electrons which then produce photons Synchrotron Inverse compton Empirically fit Band function GRB030329

Prompt GRB Neutrinos Assume shock accelerated protons with same spectrum as electrons (~ E-2) A fraction fp ~ 0.20 of the proton energy is given to the pion  En ~ 0.05 Ep Traces the observed photon spectrum Further break due to p and m synchrotron losses

see Razzaque, Meszaros, and Waxman: Phys.Rev. D68 (2003) 083001 Precursor Neutrinos Before prompt emission, jet must drive through stellar envelope Possibility for ‘choked’ GRBs pp and pg interactions with thermalized jet photons, cold stellar nucleons and jet protons. Photon interactions highly suppressed due to magnetic fields and subsequent synchrotron radiation see Razzaque, Meszaros, and Waxman: Phys.Rev. D68 (2003) 083001

Afterglow Neutrinos Same processes as prompt spectrum Lower energy afterglow photon field Neutrinos up to 1019 eV Large shock radii reduce synchrotron loss effect

Neutrino Spectra Gamma emission from Syncrotron, Inverse Compton off of accererated charged particles. Precursor emission from shocks in the fireball before it becomes optically thin. Oscillations between source and earth lead to ratio 1:1:1

Swift Satellite Launched in November 2004 BAT (Burst Alert Telescope) 1.4 sr field-of-view ~100 bursts / yr. Slew within 20-75s 0.3-5.0 arcsec positioning XRT, UVOT Afterglow measurements Coded mask is transparent above 150 KeV

Neutrino Detection charged current Cherenkov effect

Detector Concept Array of optical sensors High energy muon emits Cherenkov light Record light arrival times, amplitudes

Antarctic Physics AMANDA-II station south pole IceCube

Neutrino Detectors Detector 19 strings 677 optical modules Operated 2000-2009 dfasdfsadf 59 1450-2450 m deployment. Depth block atmospheric background. Neutrinos interact in the ice to produce muons/electrons. Muons track through detector giving off cherenkov light.

Reconstruction Arrival times / Intensities Spatial Resolution AMANDA: 2-3 degrees IceCube: < 1 degree Energy Resolution 0.3 in ln(Ereco/Etrue) 1.5 – 2.5 degrees. ~x2 in energy .3-.4 in log(E) vs. .1-.2 in log(E)

Detection Challenges I Down-going muons from CR showers misreconstructed as up-going

Detection Challenges II Up-going atmospheric neutrinos from CR showers on other side of Earth Isotropically distributed Softer energy spectrum than GRB signal

Current Status Previous searches for neutrino-induced muons (Achterberg A. et al. 2008, ApJ, 674, 357) and cascades (Achterberg A. et al. 2007, ApJ, 664, 397) saw no events and set upper limits on GRB neutrino fluxes

AMANDA-II Analysis 2005-2006 data sample contains 85 northern hemisphere bursts that pass all detector stability criteria. Largest sample in the post-BATSE era. Search for prompt phase emission assuming Waxman-Bahcall spectra for all bursts

Upgoing filter, rescaled to GRB on-time GRB Advantages Time, Position of predicted signal known. Drastic reduction in background rate Use of data for background estimation Removes many sources of systematic error Reduces reliance on simulation Upgoing filter (reconstructed track > 80 degrees), 386 days livetime Upgoing filter, rescaled to GRB on-time ~10M 1703.9

Event Selection Angular offset from GRB position Angular resolution of tracks Timing coincidence with gamma emission Likelihood ratio of upgoing reconstruction vs. downgoing reconstruction Data Downgoing Muons Atm. Neutrinos GRB Neutrinos

Final Sample

0 events in the GRB emission windows survive final cuts Results Expected 90% C.L. Limit Upgoing filter Final Cut Level 2005-2006 data 1703.9 0.00087 2005-2006 GRB signal 0.501 0.166 Discovery Potential 1 event: 3s 2 events: 5s Observed 0 events in the GRB emission windows survive final cuts 05-06 Limit: 14.7 * Prompt flux

IceCube-22 Analysis Analysis of 22 String Configuration (6/07 – 4/08) 41 Northern Sky GRBs that pass data quality considerations prompt, precursor, and extended window searches Different Approach Incorporate individual burst information Unbinned analysis, using likelihood ratio test, and incorporating event energy

Burst Modeling

Event Selection ~77 M events all-sky, all-year ~1 event on-time

Likelihood Method Probability that each event belongs to signal or background distribution Test statistic is Likelihood ratio Maximize LLH ratio by varying ns Likelihood function: Null hypothesis: Likelihood Ratio:

Probability Distributions

108 background randomizations Sensitivity 108 background randomizations

Improvement Complementary binned analysis on same data flux factor Complementary binned analysis on same data Factor ~1.8 more powerful at 5s, P=0.5

IC-22 Results Test statistic consistant with the null hypothesis for all searches Checks with looser cuts reveal no signal-like events

Summary While none of the analyses are able to place constraints on the leading models of neutrino emission, we have: Demonstrated that highly significant discoveries of transient sources can be made with only 1-2 observed events Shown that significant gains are achieved by considering the energy information of individual events in a likelihood analysis Set a precedent for the modeling of fluxes on a per burst basis for stacked GRB searches

Outlook Fermi satellite began operations in July 2008 Triple observation rate IC-40 data run complete GRB analysis underway ~2 events expected (WB) IC-86 complete in 2011 Discovery within one year of operations (~150 GRBS)

Conclusion Continued aggregation of GRBs pushes limits closer to model predictions Rapid increases in detector effective areas and satellite observation rates mean we will very soon see astrophysical high energy neutrinos from Gamma-ray bursts or will set meaningful limits on such emission

The End

Systematics Remove many sources of error by using off-time data to estimate the background Ice Properties Optical Module Efficiency Seasonal Atmospheric Variation error mag. Cross-section ±3% Muon propagation ±2% Reconstruction bias +5% Ice simulation ±5% Timing resolution OM efficiency +6% Background rate <1% Total +10%, -7%

Fermi Acceleration