1. High-Energy Emission from GRBs: expectations & first results from the Fermi Gamma-ray Space Telescope
Jonathan Granot
University of Hertfordshire
(Royal Society Wolfson Research Merit Award Holder)
on behalf of the Fermi LAT & GMB Collaborations
Nordita, Stockholm, Sweden, May 25, 2009

2. Outline of the Talk: New Space missions help drive progress in GRBs
GRB: some basic properties, theoretical framework
High-energy emission processes & pre-Fermi obs.
Some expectations & hopes from Fermi
Fermi: initial results
GRB081024B: 1st clearly short GRB above 1 GeV
GRB080916C: very bright – a lot of interesting data
Lower limit on the Bulk Lorentz factor
Limit on Lorentz invariance violation
Possible causes of HE delayed onset & longer duration
Conclusions

3. GRBs: Brief Historical Overview
1967: 1st detection of a GRB (published in 1973)
In the early years there were many theories, most of which invoked a Galactic (neutron star) origin
1991: the launch of CGRO with BATSE lead to significant progress in our understanding of GRBs
BATSE: ~ 30 keV – 2 MeV, full sky (~ ½ Earth occ.)
Isotropic dist. on sky: favors a cosmological origin
Bimodal duration distribution: short vs. long GRBs
EGRET: ~ 30 MeV – 30 GeV, FoV ~ 0.6 sr
~2 s

4. BeppoSAX ( ): lead to discovery of afterglow (1997) in X-rays, optical, radio (for long GRBs)
This led to redshift measurements: clear cut determination of the distance/energy (long GRBs) Eγ,iso ~ erg
Afterglow observations provided information on beaming (narrow jets: Eγ ~ 1051 erg), event rate, external density, supernova connection ( long GRB progenitors)
Swift ( ?): autonomously localizes GRBs slews in ~1-2 min) and observed in X-ray + optical/UV
Discovered unexpected behavior of early afterglow
Led to the discovery of afterglow from short GRBs  host galaxies, redshifts, energy, rate, clues for progenitors
Fermi ( ?): may also lead to significant progress in GRB research, like previous major space missions

5. Fermi Gamma-ray Space Telescope (Fermi Era; launched on June 11, 2008):
Fermi GRB Monitor (GBM): 8 keV – 40 MeV (12×NaI 8 – 103 keV, 2×BGO MeV), full sky
Slightly less sensitive than BATSE: expected to detect ~ 200 GRB/yr (≳ 60 in the LAT FoV)
Large Area Telescope (LAT): 20 MeV – 300 GeV FoV ~ 2.4 sr ; up to 40 times the EGRET sensitivity
Its new capabilities are expected to lead to progress in the field, similar to previous major new space missions

6. Prompt GRB Observations (≲ MeV)
Time
Flux
Variable light curve
down to millisecond timescales
Duration: ~10 −2 – 103 sec
Spectrum: non-thermal F peaks at ~ MeV
(well fit by a Band function)
Rapid variability, non thermal spectrum & z ~ 1  relativistic source ( ≳ 100) (compactness problem: Schmidt 1978; Fenimore et al. 1993; Woods & Loeb 1995;…) 
F

7. GRB Theory: Fireball vs. Poynting Flux
Afterglow
*Meszaros & Rees 92,
Katz 94, Sari & Piran 95
Internal Shocks
R ~ cm
Prompt GRB
X-rays
Optical
Radio
† Shemi & Piran 90,
Goodman 86,
Paczynski 86,…
Optical
Radio
Matter dominated†
outflow Ekin ≳ EEM
ejecta
Compact
Source
Reverse
shock*
External
medium
Forward Shock
(Rees & Meszaros 92)
Particle acceleration
 -rays (synchrotron ?)
Poynting flux†† dominated flow EEM ≫ Ekin
X-rays
Optical
Radio
reconnection
(or other EM
instability)
R ~ cm
Magnetic
bubble
†Thompson 94, Usov 94,
Meszaros & Rees 97,
Katz 97,…
Lyutikov & Blandford 02,03

8. GRB: High Energy Emission Processes
Inverse-Compton or Synchrotron-Self Compton (SSC):
Ep,SSC / Ep,syn ~ γe2 , LSSC / Lsyn=Y , Y(1+Y) ~ εradεe /εB
Hadronic processes: photopair production (p + γ  p + e e), proton synchrotron, pion production via p –γ (photopion) interaction or p-p collisions
The neutral pions decay into high energy photons π0  γγ that can pair produce with lower energy photons γγ  e e− - producing a pair cascade
Fermi may help determine the
identity of the dominant emission
mechanism at high & low energies
Most of the radiated energy can
be in the LAT range (energetics)
ν Fν
Ep,syn
Ep,SSC
γe2 Y
SSC
synchrotron ν

9. High energy emission from GRBs: Pre-Fermi era
Little known about GRB emission above ~100 MeV
EGRET detected only a few GRBs, most notably:
GRB940217: GeV photons were detected up to 90 minutes after the GRB trigger
GRB941017: distinct high- energy spectral component (up to 200 MeV), with a different temporal evolution & at least 3 times more energy
AGILE recently observed GRB080514B and detected photons up to a few 100 MeV lasting somewhat longer than the soft gamma-rays
GRB940217
(Hurley et al. 94)
EGRET
GRB941017
(Gonzalez et al. 03)
GRB080514B
(Giuliani et al. 08)
-18 to 14 sec
14 to 47 sec
47 to 80 sec
GRB940217: long lived emission
GRB941017: search for extra component (could help distinguishing leptonic / hadronic model)
AGILE
sec
sec
BATSE - LAD EGRET - TASC

10. High-Energy emission from GRBs: Expectations & Hopes for Fermi
Prompt GRB emission:
Detecting a distinct high-energy spectral component
 teach us about emission mechanism & energetics
Detecting  opacity signatures (internal / external-EBL)
Constraining Lorentz Invariance Violation (LIV)
First several hours after the GRB:
HE afterglow emission (SSC from the forward shock)
IC from the forward-reverse shock system near ~ tdec
Pair echo: TeV (GRB) + CIB γγ  e e− EC on CMB
IC involving the external shock & late source activity
Hadronic induced cascades in the early afterglow

11. Fermi GRB detections: GBM: 160 GRBs so far (18% are short)
Detection rate: ~ GRB/yr
A fair fraction are in LAT FoV
Automated repoint enabled
LAT detections: 8 in 1st 10 months
GRB080825C: >10 events above 100 MeV
GRB080916C:
>10 events above 1 GeV and >140 events above 100 MeV
GRB081024B: first short GRB with >1 GeV emission (090510)
8 + 2 more possible detections
080916C
081215A
080825C
090217
081024B
LAT FoV
cosθ

12. Breaking News: GRB
Short-hard GRB (T90 ~ 0.5 s, Epeak ~ 4-5 MeV)
Detected by Swift, Fermi, Suzaku, AGILE, K-W
AGILE: >5 σ detection of >100 MeV γ-rays, during the prompt phase & a “delayed” phase (GCN 9343)
Fermi LAT: >50 γ’s above 100 MeV (>10 γ’s above 1 GeV) in 1st sec.; >150 γ’s above 100 MeV (>20 above 1 GeV) in 1st min. (GCN 9350)
X-ray (Swift XRT) & optical afterglow
Spectroscopic redshift: z =  Eγ,iso ~ 4×1052 erg (GCN 9350)
Very interesting GRB: I can’t say much more now

13. GRB080825C: the 1st LAT GRB
The 1st LAT events coincide with the 2nd GBM peak
The high-energy emission lasts longer: highest energy photon arrives when the GBM emission is very weak
PRELIMINARY!
Fluence(50-300): 2.5x10^-5 ergs.cm-2

14. GRB081024B: 1st short GRB >1 GeV
1st LAT events coincide with 2nd GBM peak (delayed HE onset)
The HE emission lasts longer than low-energies
Single Band spectrum!
Lower limit on Γ from pair opacity constraints:
Γmin(z = 0.1) ≈ 150 Γmin(z = 3.0) ≈ 900
(best so far for a short GRB, but z is uncertain)
Fluence (50-300): 3.4x10^-7 ergs.cm-2
PRELIMINARY!

15. Γmin : The Compactness Problem: (Schmidt 1978; Fenimore et al
Γmin : The Compactness Problem: (Schmidt 1978; Fenimore et al. 1993; Woods & Loeb 1995;…)
The large γ-ray flux implies huge luminosities for cosmological GRBs, Liso ~ erg/s
For sources at rest: short variability time Δt  small source R < cΔt & ε = Eph /mec2 ~ 1  large fraction of ’s can pair produce (  ee−)
(ε) ~ σTnph(1/ε)R, nph(1/ε) ~ L1/ε/4πR2mec3 
(ε) ~ σTL1/ε /4πmec3R ≳ 1014 L1/ε,51(Δt / 1 ms)−1
Such a huge  would produce a thermal spectrum  inconsistent with the observed high energy tail

16. Solution: Relativistic Motion  ≫1
Source can be larger: R ≲ Γ2cΔt (factor Γ−2 in )
angular time
from a thin
spherical shell
(also the radial time
For emission over ΔR ~ R)
Factor of 1− cosθ12 ~ Γ−2 in  expression (Γ−2)
  ee− threshold: ε1ε2 ≳ Γ2 (Γ2(1+β), Lε ε1+β)
Altogether  is reduced by a factor of Γ2(1−β) and since -β ~ 2-3,  < 1 typically implies Γ > Γmin ~ 100
 ~ σTΓ2βL1/ε/4πmec3R ≳ Γ−2(1−β)σTL1/ε/4πmec4Δt R
R(1− cos θ) ≈ R/2Γ2 ~ cΔt
1/Γ R
1/Γ
photon
observer R

17. GRB080916C: detection, localization, follow-up obs.
GBM on-ground position error: ±1° (68% stat.) + 2-3° syst.
T90 = 66 s, several peaks in lightcurve
LAT on-ground position: statistical error: 0.09° (0.13°) at 68% (90%) + systematic error < 0.1° (preliminary)
GROND optical/NIR follow-up obs. Resulted in a positional accuracy of 0.5” & a redshift of z = 4.35 ± 0.15
Swift X-ray afterglow from T0+17 hr
GROND position
(Greiner et al. 08, submitted)
GROND SED of the GRB afterglow
(GCN report 166.1)
Swift XRT lightcurve

18. GRB080916C: multi-detector light curve
8 keV – 260 keV
260 keV – 5 MeV
LAT raw
LAT > 100 MeV
LAT > 1 GeV T0
First 3 lightcurves are background subtracted
The LAT can be used as a counter to maximize the rate and to study time structures > tens of MeV
>3000 LAT photons in the 1st 100 seconds
The first low-energy peak is not observed at LAT energies
Spectroscopy: LAT event selection >100 MeV  145 events made it
5 time bins (a to e)
14 events > 1 GeV

19. GRB080916C: multi-detector light curve
Most of the emission in the 2nd peak occurs later at higher energies
This is clear evidence of spectral evolution
The delay of the HE emission seems to be a common feature of the GRBs observed by the LAT so far.

20. GRB080916C: time resolved spectroscopy
Main LAT peak
(time bin ‘b’)
Time resolved spectroscopy over 6 decades in energy!!! (10 keV – 10 GeV)
Consistent with a Band function: a single dominant spectral component
No strong evidence for an additional spectral component
Alpha ± 0.02
Beta ± 0.03
Epeak ± 142 keV
Amp ± photons/s-cm2-keV
REDUCED CHISQ = 0.963, PROB = 0.698
time bin ‘d’
A likelihood ratio test in bin ‘d’: the probability of having no additional HE spectral component is 1% (5 bins/trials)
Possible pair production (  ee−) of HE photons with the EBL leaves this probability from unchanged to 0.03% depending on the model chosen.
Band + PL

21. GRB080916C: time resolved spectroscopy
The Stecker et al. model/s would imply  ~ 3-4  > 3 σ for distinct HE spectral component that carries significant energy
For other EBL models  ≪ 1: weak evidence for an extra HE spectral component (5% chance probability for no HE- component accounting for the 5 trials / bins)
EBL model predictions for z = 4.35
time bin ‘d’
A likelihood ratio test in bin ‘d’: the probability of having no additional HE spectral component is 1% (5 bins/trials)
Possible pair production (  ee−) of HE photons with the EBL leaves this probability from unchanged to 0.03% depending on the model chosen.
Band + PL

22. GRB080916C: Spectral Evolution
Band function fits
Soft  hard  soft Epeak evolution
Low (α) & high (β) energy photon indexes change significantly only between time bins ‘a’ and ‘b’
a b c d e

23. GRB080916C: Bulk Lorentz factor ≳ 900
Robust + highest lower limit on Γ from opacity constraints: Γmin≈ 890 ± 20 (bin ‘b’, for t = 2 s) & Γmin ≈ 600 (bin ‘d’)
Δt = 0.5 s
bin-b
INTEGRAL light curve
(Greiner et al. 09))
GBM
NaI
Δt = 2 s
bin-d Δt = 20 s
1st HE (> 100 MeV) detection of a GRB with known redshift
Our Γmin is more robust than before: it doesn’t assume the spectrum extends beyond highest energy detected photon
Under this conservative assumption: Γmin ≲ (1+z)Eph,max/mec2 ≈ 200(1+z)(Eph,max / 100 MeV) so that a high Γmin requires the observed spectrum to reach a sufficiently high energy Eph,max

24. Other Interesting Results for GRB080916C:
Large fluence (2.4×10-4 erg/cm2) & redshift (z = 4.35 ± 0.15)  record breaking apparent isotropic energy release Eγ,iso ≈ 8.8×1054 erg ≈ 4.9 Mc2  suggests strong beaming (jet)
Single dominant emission mechanism: if synchrotron, SSC is expected, and can avoid detection if Epeak,SSC ≫ 10 GeV (γe ≫ 100), or if Y ≈ εe/εB ≲ (constrains other emission mechanisms as well)
Epeak,syn
γe2
Epeak,SSC
ν Fν Y
possible SSC
synchrotron (?)
~10 GeV ν

25. Limits on Lorentz Invariance Violation
Some QG models violate Lorentz invariance: vph(Eph) ≠ c
A high-energy photon Eh would arrive after (or possibly before in some models) a low-energy photon El emitted together
GRB080916C: highest energy photon (13 GeV) arrived 16.5 s after low-energy photons started arriving (= the GRB trigger)  a conservative lower limit: MQG,1 > (1.50 ± 0.20)×1018 GeV/c2
min MQG
(GeV/c2)
1016
1017
1018
1015
1.8x1015
Pulsar
(Kaaret 99)
0.9x1016
1.8x1017
0.2x1018
4x1016
GRB
(Ellis 06)
(Boggs 04)
AGN
(Biller 98)
(Albert 08)
GRB080916C
Planck mass
1019
1.5x1018
1.2x1019
(Jacob & Piran 2008)
n = 1,2 for linear and quadratic Lorentz invariance violation, respectively

26. GRB080916C: temporally extended emission
Most LAT detected GRBs show significant HE emission lasting after the low-energy emission becomes (almost) undetectable (originally detected by EGRET; Hurley et al. 94)
GRB080916C: the HE emission that extends ~103 sec beyond the detectable keV-MeV emission
Possible origins:
Afterglow SSC emission (though no spectral hardening, time gap, or synchrotron/SSC valley in the spectrum are observed)
IC scattering of late X-ray flare photons by afterglow electrons
Long lived cascade induced by ultra-relativistic ions
Pair echo: TeV + EBL   ee−, & the ee− IC scatter the CMB

27. Delayed onset of HE emission: Possible Causes
1. The 1st and 2nd peaks are emitted from distinct physical regions (e.g. different colliding shell in the internal shocks model)
Unclear why a similar behaviour occurs in all 3 LAT GRBs (if it is random then some GRBs should have a reverse order)
2. opacity effects don’t work well as there is no cutoff or steepening of the spectrum at high-energies
3. Hadronic origin: time to accelerate protons & develop pair cascade, if the high-energy emission is of hadronic origin
Two distinct spectral components (leptonic at low-energies & hadronic at high-energies) & hard to soft evolution are expected (but not seen); hard to explain sharpness of 1st LAT peak + coincidence with 2nd GBM peak

28. Fermi LAT GRB detection rate
~ 8-9 GRB/yr with >10 photons above 100 MeV
~ GRB/yr with >10 photons above above 1 GeV
Somewhat below estimates based on a Band spectrum for a bright GRB
BATSE sample
Suggests: most GRBs
don’t have significant
excess (HE component)
or deficit (cutoff) in the
LAT energy range w.r.t
the extrapolation of a
Band spectrum from
lower energies

29. Conclusions: Expectations: First results:
Like previous major relevant space missions, Fermi is expected to significantly advance the GRB field
Particularly: prompt emission mechanism & region, energy budget, early afterglow, opacity effects
First results:
~ 9-10 LAT GRBs/yr suggest that most GRBs do not strongly deviate from a Band spectrum in LAT range
Spectra: consistent with single dominant component
1st 3 LAT GRBs show later onset & longer duration of the high-energy emission, relative to low energies
short & long GRBs seem to have similar HE spectra
GRB080916C:  ≳ 900, MQG,1 > 1.3×1018 GeV/c2