The Tale of IceCube and GRB080319B or the Mystic Magic of TestDAQ Alexander Kappes UW-Madison...
-
Upload
dorothy-hunter -
Category
Documents
-
view
218 -
download
0
Transcript of The Tale of IceCube and GRB080319B or the Mystic Magic of TestDAQ Alexander Kappes UW-Madison...
The Tale of IceCube and GRB080319Bor the Mystic Magic of TestDAQ
Alexander Kappes
UW-Madison
IceCube Collaboration Meeting
May 3, 2008, Madison
May 3, 2008 Alexander Kappes, IceCube Meeting, Madison 2
The Event: GRB080319B
Triggered Swift on March 19, 06:12:49 UTC
Position: RA = 217.9º, Dec = +36.3º
T90 > 60 s (main emission within first 60 s)
Brightest (optical) GRB ever observed: Fluence (15-150 keV; 60 s) = 8110-6 erg/cm2
Isotropic energy = 1.31054 erg
Optical afterglow saturated several UVOT pixels
Redshift z = 0.94 (DA = 1.6 Gpc, light travel time 7.5 Gyr)
Remark: 3 other bursts on that day (normally 2 per week)+ all in northern sky; IceCube data for 2
May 3, 2008 Alexander Kappes, IceCube Meeting, Madison 3
GRB080319B: Swift Instruments
XRT (X-ray) UVOT (ultra violet)
May 3, 2008 Alexander Kappes, IceCube Meeting, Madison 4
GRB080319B: Optical Brightness
Visible to the naked eye for ~10 s !!!
Animated GIF:http://grb.fuw.edu.pl/pi/index.html
May 3, 2008 Alexander Kappes, IceCube Meeting, Madison 5
And what about IceCube?
... but fortunately the detector was actually running (weekly runs): (Francis: “we were running in maintenance mode”)
IC9 TestDAQ run (~6 min) ( this talk)(last run until 110685 was started)
Full 2 hour AMANDA TWR run ( see following talk by Martijn) (running outside of ANVIL control)
GRB080319B occurred 06:12:49 UTC (main burst ~60 s)
May 3, 2008 Alexander Kappes, IceCube Meeting, Madison 6
IceCube Data (IC9)
GRB time: 08-03-19 – 06:12:49 GRB duration:
main burst 06:12:49 – 06:13:49
Run containing GRB: SPS-TESTDAQ02_run0002484_1-LocalCoincidence-02-2006-Config-IceCube9-InIce-Amanda-V6-Weekly
duration: 08-03-19 – 06:09:14.6 (first event, 214 s before main burst)08-03-19 – 06:15:51.2 (last event, 122 s after end of main
burst) # events: 40492 102 Hz (after standard multiplicity 8 trigger)
Konus data (-ray)IceCube data
GRB direction
May 3, 2008 Alexander Kappes, IceCube Meeting, Madison 7
Estimated Events in IceCube
Standard fireball model (Guetta et al) IC9 point source effective area @ dec = 35º (© Chad Finley)
(averaged over azimuth; angular cut loosened from 2.5º to 5º)
Expected # of obs. events ~0.1, but large fluctuations in neutrino flux possible
gamma spectrum neutrino spectrum
IC9 eff. area # events in IC9
May 3, 2008 Alexander Kappes, IceCube Meeting, Madison 8
Background Calculation
Precision of background estimate crucial:with 0.1 expected signal events we have
~10% chance to observe 1 event~0.5% chance to observe 2 events . . .
1 event in on-time window (60 s) significant @ 4 requires calculated background to be 610-5
Options: For background estimate purely from data
we need at least 1 event in given time period T
T > 60s / 610-5 ≈ 12 days
Fit function to low statistics background map from Testdaq data statistics even in total map very low, systematics,
doesn’t work with unbinnned likelihood
Use of MC not enough statistics, systematics
. . . and use IC9 2006 dataset (137 days). . . and use IC9 2006 dataset (137 days)
May 3, 2008 Alexander Kappes, IceCube Meeting, Madison 9
IC9 TestDAQ vs. 2006 RealDAQ Data
IC9 TestDAQ data is taken with same settings as in 2006 (e.g. OM voltages)
However, there are differences: Simplified time window for local coincidences (800 ns) in TestDAQ
(signal travel time not corrected) TestDAQ reads out only one ATWD (slightly more deadtime) . . . ?
Requires careful comparisonsRequires careful comparisons
May 3, 2008 Alexander Kappes, IceCube Meeting, Madison 10
Data Processing
TestDAQ: fat-reader with SMT 8 and LC-Span 1
“Simulation” of 2006 online muon filter:
zenith (linefit || dipolefit) > 70º && Nchan ≥ 10
std-processing (Level1 + 2) as for IC22
2006 IC9 data (thanks to Martin Merck): Same data as for 2006 atm. muon and point source analysis
std-processing (Level1 + 2) as for IC22
Required some tweaking as not all entries in database filled in 2006
(e.g. some droop correction values not set)
FilterMask not contained in current Level2 files
differences at low cut levels in following plots
May 3, 2008 Alexander Kappes, IceCube Meeting, Madison 11
IC9 TestDAQ - 2006 RealDAQ Comparison
Mainly down-goingmuons from HE-filter
(special 1 hour IC9 TestDAQ run 1 week after GRB)
May 3, 2008 Alexander Kappes, IceCube Meeting, Madison 15
IC9 TestDAQ - 2006 RealDAQ Comparison
37% signal efficiency
May 3, 2008 Alexander Kappes, IceCube Meeting, Madison 16
Analysis Method
Unbinned loglikelihood method (similar to Chad’s point source analysis)
Track information in PDFs: Relative position to GRB together with full paraboloid error ellipse Relative time to GRB
Optimization of quality cuts on Ndir and paraboloid error(already used in IC9 point source analysis)
ontime
May 3, 2008 Alexander Kappes, IceCube Meeting, Madison 17
Signal Simulation
Generation of muon neutrinos from direction of GRB with modified neutrino-generator (Chad Finley)
allows specification of direction and time period
generates neutrinos randomly within time period
Sample source events (Poisson sampling) from generated neutrinos weighted with calculated GRB spectrum
May 3, 2008 Alexander Kappes, IceCube Meeting, Madison 18
Current Limitations
Only 26 out of 137 days of 2006 data for background
Ndir cut on Pandel track removes high energy events in future cut on MPE Ndir
Differences between TestDAQ and RealDAQ
May 3, 2008 Alexander Kappes, IceCube Meeting, Madison 19
Background Map for PDF (preliminary)
GRB position
Ndir 6, paraErr ≤ 6
May 3, 2008 Alexander Kappes, IceCube Meeting, Madison 20
MDF Calculation (preliminary)
Cuts: Ndir 6, paraErr ≤ 6 Significance = 2.510-3 (3), Power = 0.5
Expected signal events = 0.16 MDF = 6.8 Probability to discover signal ~7%
10,000 randomizedbackground samples
Likelihoodrequired mean
# signal evts
sig
nal
eve
nts
May 3, 2008 Alexander Kappes, IceCube Meeting, Madison 21
MDF Optimization (preliminary)
Optimize MDF varying cuts on
Ndir_C and paraboloid error
Best MDF:
Ndir 6, paraErr ≤ 6
May 3, 2008 Alexander Kappes, IceCube Meeting, Madison 22
To Do List and Time Schedule
Process data/MC to level3 (MPE)
Investigate differences between TestDAQ and RealDAQ
Varify stability of GRB TesDAQ run
Check that paraboloid error gives reasonable error estimate
also for IC9
Do checks on absolute timing and pointing
Systematics (differences TestDAQ - RealDAQ, uncertainty on
background estimation etc.)
Time schedule: Unblinding proposal until end of May