1. The Crab Light Curve and Spectra from GBM: An Update
Gary Case
La Sierra University
IACHEC, Pune, India, 29 February 2016

2. The Fermi Satellite Gamma-ray Burst Monitor (GBM)
12 NaI detectors
12.5 cm diameter x 1.25 cm thick
8 keV - 1 MeV
2 BGO detectors
150 keV - 40 MeV
12.5 cm diameter x 12.5 cm thick
All GBM detectors are non-imaging
LAT
GBM
Sodium Iodide
(NaI)
Detector
This analysis uses the GBM instrument on Fermi. We currently only use the NaI and not the BGO, though we have the capability to incorporate the BGO.
GBM
Bismuth Germanate
(BGO)
Detector

3. Earth Occultation
Question: How do you measure the intensity of a source if your detector doesn’t know where the photon came from?
Answer: Earth occultation technique θ
β = 25.6°
Count rate
Time
Visible
Not visible
Occultation
step
Since the NaI are non-imaging, we use the Earth Occultation Technique to get fluxes of individual sources. When source goes behind the earth, the count rate in the detector drops creating a “setting” step. When the source reappears, the count rate goes back up creating a “rise” step.

4. GBM Earth Occultation Technique
Current input catalog includes 244 sources, primarily recently active X-ray binaries, the Crab, AGN, SGRs, CVs, and the Sun
Calculate occultation times and center each step in four minute window for each detector and each energy band (8 energy bands in CTIME data, 16 bins for CSPEC)
Generate source model: assumed spectrum convolved with changing detector response and atmospheric transmission
Fit data to source model, plus source models for interfering sources, and quadratic background
120+ sources detected <100 keV, 9 sources detected >100 keV
Advantages of GBM monitoring:
Continuous monitoring
No solar pointing constraints
Useful response up to ~400 keV
Bin number (2.048-second bins)
Counts
Using a catalog of known locations, we predict the occultation times for each source. Data in 4 minute windows centered on the occultation of the source of interest are then fitted with a model consisting of a quadratic background plus a predicted source rate (from convolving an assumed spectrum with the detector response) combined with an energy dependent atmospheric transmission model for each source occulting with in the window. Bright interfering sources are included out to detector angles of 40, 60, or 90 degrees, depending on how bright the interfering source is. GBM advantages over other missions – no solar constraints, coverage above Swift/BAT (up to 400 keV as limited by current software). GBM does see the Sun in Earth occultation when it is flaring.

5. What the Crab has Been Up to Lately…
Light curves for each instrument are normalized to its average rate from MJD
This plot only goes until the end of The keV band has recovered to its pre-decline level, while the keV band has only recovered to a little over 50% of its pre-decline level.
(Thanks to Colleen Wilson-Hodge)
keV band has recovered to pre-decline level
15-50 keV band has only increased ~50% of the way back to pre-decline level

6. Fermi/GBM: Crab Spectra
Complicated by the fact the response is constantly changing
Use CSPEC data binned into 16 logarithmically-spaced channels from keV, though lowest 3 bins ignored for now (possible internal calibration issue)
Spectra fit for Epochs E-K for energy range keV
Minimum 40 days integrated, centered on middle of epoch
I’ve used the CSPEC data binned into 16 logarithmically spaced bins from keV, though I ignore the lowest 3 bins do to a potential internal calibration issue with GBM. I have fit spectra for each of the 7 epochs after August 2008 where there were near simultaneous observations of the Crab. I integrated a minimum of 40 days, centered on the epoch as defined by Lorenzo.
Epoch
Begin date
Begin day (MJD)
End date
End day (MJD) E
54700
54740 F
55045
55085 G
55258
55303 H
55444
55484 I
55604
55647 J
56173
56213 K
56912
56952

7. Fermi/GBM: Crab Spectra
Use absorbed (wabs) broken power law
Included a 5% systematic error (the systematic errors for earth occultation are relatively large, though they have not been studied in detail)
Epoch
PhoIndx1
EB (keV)
PhoIndx2
Norm.
χ2 (DOF)
Red. χ2
F20-100 F E
2.078 −
141 −48 +35
2.76 −
9.38 −
421 (452)
0.93
1.798
0.942 F
2.103 −
113 −
2.36 −
9.78 −
709 (620)
1.14
1.707
0.914 G
2.033 −
127 −46 +66
2.55 −
7.37 −
810 (884)
0.92
1.677
0.933 H
2.062 −
89 −23 +45
2.37 −
8.31 −
760 (848)
0.90
1.694
0.875 I
2.063 −
141 −
2.90 −0.52 +∞
8.23 −
766 (728)
1.04
1.674
0.869 J
2.058 −
97 −24 +30
2.43 −
8.10 −
962 (896)
1.07
0.865 K
2.128 −
> 400
10.63 −
186 (212)
0.88
1.687
0.968
I fit the data to an absorbed broken power law model, as this is the canonical model being used. These fits were done using the chi-squared fit statistic because when I used cstat, the error routine in Xspec didn’t always converge. I included a 5% systematic error, though this might be an overestimation. The systematic error for the earth occultation technique (especially for GBM) are relatively large, though we have not yet studied this quantitatively for the CSPEC data. Systematic errors were estimated for the CTIME data used in our catalog paper.
Fluxes in units of 10-8 ergs/cm2/s

8. Conclusions
Spectra generated for the Crab for 7 epochs from August 2008 to October 2014 using GBM.
As more epochs are defined, GBM spectra can be generated relatively easily.
The GBM spectral results generally agree with the other instruments, though the GBM spectra are a little harder below the break and a little softer above the break.
GBM will continue to monitor the Crab.
I can easily fit spectra for observing epochs as they are defined. The main point is that the GBM results are generally in agreement with the other instruments, though the GBM spectra are little bit harder below the break and a little softer above the break.