Prospects for GRB Science with GLAST

1
Prospects for GRB Science with GLAST

• Jonathan Granot
• University of Hertfordshire
• (Royal Society Wolfson Research Merit Award
  Holder)

Collaborators J. Cohen-Tanugi, E. do Couto e
Silva A. Königl, T. Piran, P. Kumar, D. Eichler,
E. Ramirez-Ruiz, C. Kouveliotou,
MSFC/NSSTC Science Colloquium, August 11, 2008,
Huntsville AL
2
Outline of the Talk

• Short historical overview of Gamma-Ray Bursts
  (important missions, observations, theoretical
  framework)
• Brief outline of GLAST capabilities
• Early afterglow from Swift to GLAST
• Prompt gamma-ray emission
• Emission mechanism, energy budget
• Intrinsic opacity to ?? ? e?e?
• Conclusions

3
GRBs Brief Historical Overview

• 1967 1st detection of a GRB by the Vela
  satellites (serendipitously it was published
  only in 1973)
• In the early years there were many theories, most
  of which invoked a Galactic origin (at some point
  there were more theories than detected GRBs !!!)
• 1991 the launch of CGRO with BATSE lead to
  significant progress in our understanding of GRBs
• BATSE 30 keV 2 MeV, full sky coverage (in
  practice the Earth occulted ½ the sky at any
  time)
• OSSE 50 keV 10 MeV, FOV 3 11
• COMPTEL burst modules 0.1 10 MeV, 2.5 sr
• EGRET 30 MeV 30 GeV, FOV 0.6 sr

4
Isotropic distribution on the sky

• Favors a Cosmological origin over a Galactic
  origin
• An extended Galactic halo was still invoked by
  some

5
Bimodal Distribution Long vs. Short
2 s

• This suggested two distinct classes of bursts

6
The peak count rate distribution

• V/Vmax ? (Cmax/Cmin) -3/2 where V is the volume
  out to the source distance Vmax is the volume
  out to which the same source could be detected
  Cmax peak count rate, Cmin detection limit
• For Euclidean universe constant event rate per
  unit volume, V ? D3 ? F -3/2 ? N(gtC) ? C -3/2
  ?V/Vmax? ½ for any luminosity function
• ?V/Vmax? 0.328 0.012 for the 1st 601 BATSE
  GRBs ? supported a Cosmological origin

7
GRBs Observations - Prompt GRB

• Variable light curve
• Duration 10 -2 103 sec
• Spectrum non-thermal
  ?F? peaks at 0.1-1
  MeV
• Rapid variability, non thermal spectrum z 1 ?
  relativistic source (? ? 100) (compactness
  problem Schmidt 1978 Fenimore et al. 1993
  Woods Loeb 1995)

8
BeppoSAX discovery of afterglow

• Wide Field Camera 40 40, 2 - 30 keV
  ( PDS shielding nearly all sky _at_ 100 - 600
  keV)
• Narrow Field Instruments (1- 0.5) 0.1- 300
  keV
• WFC ? ground ? point NFI ? ground (hours)
• Its abilities led to afterglow detection (1997)
  in X-rays, optical, radio (for long GRBs - LGRBs)
• This led to redshift measurements clear cut
  determination of the distance/energy (LGRBs)
• Afterglow observations provided many new
  constraints on beaming, event rate, external
  density, SN connection, etc.

9
Afterglow Observations pre-Swift(basic features
the model needs to produce)

• X-ray, optical radio emission over (pre-Swift)
  days, weeks months,
  respectively, after GRB
• Light curves power-law decay

Optical
X-ray
Fox et al. (2003)
Piro (1999)
10
Some afterglows show an Achromatic Steepening of
the Light Curve (Jet Break)
Optical light curve of GRB 030329
t-1
Optical light curve of GRB 990510
t-2-t-2.5
(Gorosabel et al. 2006)
(Harrison et al. 1999)
11
Spectrum Linear Polarization

• Spectrum consists of several power law segments
  is well fit by synchrotron emission
• Linear polarization of 1-3 was detected in
  several optical/NIR afterglows ? likely
  synchrotron emission

Spectrum
Linear Polarization
GRB 970508 Spectrum at 12.1 days (Galama et al.
1998)
(Covino et al. 2003)
12
The Size of the Afterglow Image

• Quenching of diffractive scintillations after
  30 days in the radio afterglow of GRB 970508 ? R?
  1017 cm
• The radio afterglow of GRB 030329 was
  (marginally) resolved directly using the VLBA
  (Taylor et al. 04,05)

Indirect Scintillation
Direct VLBA
spectal slope 4.8-8.4 GHz
GRB 970508
Light Curve 8.4 GHz
(Frail et al. 2000)
GRB 030309 (z 0.17) VLBA _at_ 1.4, 8.4 GHz
(Taylor et al. 2005)
(Waxman et al. 1998)
13
GRB Theory Fireball vs. Poynting Flux
Afterglow
Meszaros Rees 92, Katz 94, Sari Piran 95
Prompt GRB
X-rays Optical Radio
Shemi Piran 90, Goodman 86, Paczynski 86,
Optical Radio
Matter dominated outflow Ekin ? EEM
ejecta
Reverse shock
External medium
Forward Shock (Rees Meszaros 92)
Particle acceleration ? synchrotron ?-rays
Poynting flux dominated flow EEM Ekin
X-rays Optical Radio
reconnection (or other EM instability) R
1016-1017 cm
Magnetic bubble
Thopson 94, Usov 94, Meszaros Rees 97, Katz
97,
Lyutikov Blandford 02,03
14
Afterglow Theory Dynamics1. A spherical outflow

• A compact source ejects a relativistic outflow
• Dissipation within the outflow causes the prompt
  GRB
• a relativistic forward shock sweeps up external
  medium

• The outflow is decelerated by a reverse shock
• When most of the energy is transferred to the
  shocked external medium the flow approaches
  self-similarity (Blandford McKee 1976)
• Finally the flow becomes Newtonian (Sedov-Taylor)

forward shock
1
CD
2
reverse shock
3
4
source
1. Unperturbed ext. medium 2. Shocked external
medium 3. Shocked ejecta 4. Freely expanding
ejecta
15
Emission Synchrotron Radiation

• Relativistic electrons gyrating in a magnetic
  field
• The electrons are presumably shock-accelerated to
  a power-law distribution dN/d?e ? ?e-p (?e gt ?m)
• Convenient parameterization of our ignorance the
  electrons the magnetic field are assumed to
  hold fractions ?e ?B of the internal energy
• Individual electron P? ? ?1/3 _at_ ? lt ?syn ?
  ?B?e2
• Break frequencies ?m ?syn(?m), ?c ?syn(?c),
  ?a
• Synchrotron-self Compton may also be relevant

16
Spectra Light Curves
F? ? ta ?ß
(Sari, Piran Narayan 1998)
(JG Sari 2002)
17
Origin of Different Emission Components

• The long lived afterglow emission lasting days,
  weeks, months in the X-ray, optical radio is
  attributed to the forward shock
• the reverse shock is believed to produce the
  optical flash and radio flare emission, whose
  polarization probes B-field structure in outflow

The simplest spherical model was very successful
in explaining afterglow observations during the
first 2 years after the detection of afterglow
in 1997
Radio Flare (Kulkarni et al. 99)
Optical Flash (Sari Piran 1999)
GRB 990123
18
Complications variants of basic
modelmotivation both theoretical observational

• Radiative losses (Blandford McKee 67 Cohen,
  Piran Sari 98 Panaitescu Meszaros 98
  Meszaros, Rees Wijers 98)
• Were expected theoretically in the early
  afterglow
• invoked to reduce the high prompt ?-ray
  efficiency
• Wind-like external density ?R-2 (Chevalier Li
  00)
• Motivation expected for massive star progenitor
• Jets narrowly collimated outflow (Rhoads 97,
  99)
• Motivation in analogy to other relativistic
  sources reduces total energy output in ?-rays
• Predicted a jet break which was soon observed

19
(Long) GRB SN (Type Ic) Connection

• Firmly established the connection between long
  GRBs and core collapse Supernovae (in 2003
  circumstantial or less conclusive evidence exited
  earlier)
• Supports the Collapsar model, in which a BH is
  formed during the collapse of a massive star

(Hjorth et al. 2003)
20
The Swift Era

• Burst Alert Telescope (BAT) sensitive coded
  mask
• Energy 15-150 keV (imaging), ? 350 keV
  (otherwise)
• FOV 1.4 sr , angular position accuracy 4
• Triggers autonomous slewing of the spacecraft
• X-Ray Telescope (XRT) 0.2-10 keV, FOV 23,
  typical angular position accuracy a few
  arcseconds
• Ultra-Violet Optical Telescope 24 mag in 103 s
• Detects 100 GRB/yr X-ray afterglow for most
• Discovered unexpected behavior of early afterglow
• Led to the discovery of afterglow from short GRBs
  ? host galaxies, redshifts, energy, rate,
  progenitors?

21
Gamma-ray Large Area Space Telescope (GLAST Era
launched on June 11, 2008)

• GLAST Burst Monitor (GBM) 10 keV 25 MeV
  (12NaI 10 103 keV, 2BGO 0.15-25 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

22
LAT performancecompared to EGRET

• More than 40 times the sensitivity of EGRET
• Large Energy range 20 MeV to gt300 GeV
• Optimized Point Spread Function
• (0.35o _at_ 1 GeV)
• Wide Field of View
• (2.4 sr)
• Good Energy Resolution
• (DE/E 10)

23
GRB High Energy Emission Processes

• Inverse-Compton or Synchrotron-Self Compton
  (SSC)
• Ep,SSC/Ep,syn max(?m,?c)2, LSSC/LsynY,
  Y(1Y) eradee/eB
• Hadronic processes photopair production (p ?
  ? p e? e?), proton synchrotron, pion
  production via p-p collisions or p ?
  (photopion) interaction
• The neutral pions decay p0 ? ?? into high energy
  photons that can pair produce with lower energy
  photons ?? ? e? e- -producing a pair cascade
• GLAST 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)
• in the LAT range

High energy photons (gt50 MeV)
24
EGRET Observations of GRBs

• EGRET detected only a few high-energy bursts

• The observed properties were
• different between those cases
• GRB A distinct high energy spectral component has
  been observed in the prompt phase of one EGRET
  GRB - 941017
• GRB 940217 delayed emission

GRB 941017 (Gonzàlez et al 03)
GRB 940217
25
Prompt High Energy Emission in GRB 941017

• The high-energy spectral component (? 3 MeV) last
  longer (200 s) than the sub-MeV component (with
  T90 77 s) and shows much less temporal
  variability
• Hadronic cascades? (Gonzalez et al. 2003)

• More likely inverse- Compton emission from
  forward-reverse shock system (JG Guetta 03)

Low Energy lt 3 MeV T90 77 sec Epeak 0.5 MeV
Where is the high-energy peak? Is there a
cut-off? Internal or external shocks? Are
hadrons involved? Time dependent photon index?
How common is this behavior? We Need GLAST
data!!
High Energy gt 3 MeV dN/dE E-1 Duration 200s
(Gonzalez et al 2003)
26
GRB 940217 Delayed High Energy Emission

• The origin of the delayed emission is not clear
• Afterglow SSC emission?
• Hadronic processes?
• Late time flaring activity?
• Interaction with the CIB?
• GLAST may help study the different possible
  mechanisms

GRB 940217 (Hurley 1994)
27
Early X-ray Afterglows from Swift
flat part t0-t-1
Post jet break
usual decay t-1-t-1.5
Tail of prompt emission
rapid decay t-5-t-3
tjet
102.5 s
104 s
(Vaughan et al. 2006)
(Obrien et al. 2006)
28
Possible Explanations for Early Flat Decay

• Energy injection into afterglow (Nousek et al.
  06)
• I. Continuous relativistic wind L? t-0.5
  (magnetar?)
• II. Slower material ejected during the prompt GRB
• gradually catches up the decelerating afterglow
  shock
• Afterglow efficiency increases with time (varying
  shock micro-physics parameters JG, Königl
  Piran 06)
• Observer outside emitting region (JG Eichler 06)

(JG, Ramirez-Ruiz Perna 05)
29
Possible Explanations for Early Flat Decay

• Energy injection into afterglow (Nousek et al.
  06)
• I. Continuous relativistic wind L? t-0.5
  (magnetar?)
• II. Slower material ejected during the prompt GRB
• gradually catches up the decelerating afterglow
  shock
• Afterglow efficiency increases with time (varying
  shock micro-physics parameters JG, Königl
  Piran 06)
• Observer outside emitting region (JG Eichler
  06)
• Two component jet

wide jet ?0 20-50
narrow jet ?0 gt 100
?w
?n
(JG, Königl Piran 06)
observer
tdec ? ?0-2(4-k)/(3-4) for ?ext ? r-k ? tdec,n
tdec,w
30
Implications for ?-ray Efficiency

• ?? E?/E0, ??/(1-??) ?f ? E?/Ek(t), f
  Ek(t)/Ek,0
• ? 1 from the X-ray afterglow flux at t 10 hr

31
Implications for ?-ray Efficiency

• ?? E?/E0, ??/(1-??) ?f ? E?/Ek(t), f
  Ek(t)/Ek,0
• ? 1 from the X-ray afterglow flux at t 10 hr
• f ? 10 if flat decay is energy injection ?? ?
  0.9
• If the flat decay phase is due to an increase in
  the afterglow efficiency then f 1 ?? 0.5
• If also Ek(t 10 hr) is underestimated (e.g., ?e
  0.1 instead of 1) then possibly ? 0.1 ??
  0.1
• ? a typical afterglow kinetic energy ? 1052 erg
  (? 1053 erg) for a uniform (structured) jet
• GLAST might find a larger E? ? higher ??
• Models differ in GLAST range (SSC componet)

32
GLAST may help distinguish between the different
possibl explanations

• Energy injection long lived reverse shock
• The reverse shock is highly (mildly) relativistic
  in Type I (II) energy injection
• ? different inverse-Compton emission is expected
  (in both cases 4 IC components ff fr rf rr)
• Afterglow efficiency increases with time most of
  the energy could potentially be radiated in LAT
  range
• Observer outside emitting region SSC from the
  external shock with similar shallow decay phase

33
X-ray Flares prolonged source activity?

• Short time scale (?t t) Large amplitude (?F ?
  F) rule out an afterglow origin

• They are most likely due to long lived central
  source activity (late time fallback?)
• Late localized dissipation events within the
  outflow?

(Nousek et al. 2006)
(Krimm, JG, et al. 2006)
34
X-ray Flares

• Temporal spectral properties similar to prompt
  GRB
• The emission site mechanism is similarly
  uncertain
• GLAST observations can help solve such questions
  (SSC component,
  opacity effects)
• 4 IC components are
• predicted, as there are
• 2 emission regions

The largest flare so far in the X-ray
afterglow (Falcone et al. 2006)
X-ray flare
afterglow
35
The Compactness Problem

• The large ?-ray flux implies huge luminosities
  for cosmological GRBs, Liso 1050 - 1053 erg/s
• For Newtonian sources short variability time ?t
  ? small source R lt c?t e Epheak /mec2 1 ?
  large fraction of ?s can pair produce (?? ?
  e?e-)
• ???(e) sTnph(1/e)R, nph(1/e) L1/e/4pR2mec3 ?
• ???(e) sTL1/e/4pmec3R ? 1014 L1/e,51(?t / 1
  ms)-1
• Such a huge ??? would produce a thermal spectrum
  ? inconsistent with the observed high energy tail

36
Solution Relativistic Motion ? 1

• Source can be larger R lt G2c?t (factor G-2 in
  ???)
• ?? ? e?e- threshold e1e2 ? G2 (G2(1-a),
  Le?e1-a)
• Factor of 1- cos?12 G-2 in ??? expression
  (G-2)
• Altogether ??? is reduced by a factor of G2(1a)
  and since a 2-3, ??? lt 1 typically implies G ?
  100
• ??? sTG-2aL1/e/4pmec3R ? sTG-2(a1)L1/e/4pmec4?t

37
Opacity Buildup in Impulsive Relativistic
sources Motivation(JG, Cohen-Tanugi do Couto
e Silva 2008)

• Opacity effects are expected to be important in
  GLAST LAT energy range ( 20 MeV - 300 GeV)
• Above some photon energy ?1, ??? gt 1 the
  spectrum is expected to cut off exponentially
• Lack of such a cutoff up to an observed photon
  energy ?max ? ? ? 100L0,52(?max)?-1/ R131/2?
  where ? Eph/mec2 and L? L0?1-?
• This was used to put a lower limit on assuming R
  ?2c?t where ?t observed variability time

38
Motivation (2)

• Observing the high energy cutoff due to ??? will
  determine ?2?R (instead of just a lower limit)
• Some sources are highly variable, suggesting
  impulsive emission (GRBs, flares in Blazars,)
• We consider the opacity to pair production
  (?? ? e?e-) within the source (flaring region)
• Together with an independent estimate of ? this
  can determine R and check if indeed R ?2c?t
• Initially there is no photon field the opacity
  builds-up with time ? even ? gt ?1(steady state)
  photons can initially escape, as long as ?1(t) gt
  ?
• ? a distinct temporal spectral signature

39
Simple (yet rich) Semi-Analytic Model

• Ultra-relativistic (? 1) spherical thin (?
  R/?2) shell emits in a finite interval R0 R
  R0?R
• Isotropic emission in the shell co-moving frame
• For simplicity ? 2 ? R-m, L? ? (?)1-?Rb is
  assumed while the formalism is more general

Corresponds to a single flare/spike in light curve
expanding shell
gg ee-
turns off
turns on
40
Calculation of the observed Flux

• Flux calculation integration over the equal
  arrival time surface of photons to the observer
• The photon field is calculated at all space
  time
• The pair-production optical depth is calculated
  by integrating along the trajectory of each photon

equal arrival time surface
? photon arrival time to observer ? emission
angle from the l.o.s. t emission time (in lab
frame)
41
Calculating the ?? ? ee- Optical Depth

equal arrival times surface of photons to
the observer (EATS)
the shell emits a test photon
photon front
Rt,0
R(t1)
qt,0
R0
Expanding spherical ultra-relativistic shell
observer at infinity
radius where the GRB source turns on
t0
?t,0 ?(Rt,0)
42
Calculating the ?? ? ee? Optical Depth
At each point along the test photon trajectory
the local photon field is calculated by
integrating along the equal arrival time surface
to that space-time point EATS-II
43
Results Light Curves Instantaneous Spectra
Time of instantaneous spectrum
Time integrated spectrum
one dynamical time
T0 time when first photon reaches the observer
at infinity
1 GeV
44
Time Integrated Spectrum Power law High Energy
Tail
GBM
LAT
GLAST
300 GeV
8 keV
1 MeV
1 GeV
25 MeV
45
Temporal signatureHigh energy photons, above
the break in time integrated spectrum escape
mainly near the onset of aflare or spike in the
light curve
gg ee-
source opaque to ?-rays
high energy photons reach the observer near the
onset of the flare / spike in light curve
?-rays escape freely
The opacity builds-up saturates on a dynamical
time scale
Theoretical Calculations
46
Conclusions

• Like previous major relevant space missions,
  GLAST is also expected to significantly advance
  the GRB field
• Early Afterglow
• May help find the cause of the shallow decay
  phase
• May help find the origin of the X-ray flares
• Prompt GRB emission
• May determine emission mechanism (soft hard)
• Will better determine the total radiated energy
• Opacity effects constrain R, G ? composition (e?
  / p / B)
• ?? ? e?e? opacity has distinct observable
  signatures
• GLAST may find surprising new things (more fun)

47
Validity of the Model Assumptions

• Thin Shell in internal shocks tcool tdynamic ?
  thin cooling layer behind the shock
• Spherical geometry reasonably valid in GRBs
  should not qualitatively affect the main results
• Power law emission spectrum only marginally
  valid for GRBs ? will be generalized
• Neglecting external opacity valid for GRBs not
  so clear how valid in Blazar flares (but can be
  distinguished by lack of ??? time dependence)
• Single spike/flare reasonably valid for spikes
  after quiescent period vicinity to previous
  spike or flare would effect manly high energies ?
  ?1

48
Why is there an exponential cutoff in the
spectrum of a (quasi-) steady source?

• If the emission and absorption are in the same
  region (e.g. by the same material), then photons
  can escape only from a thin layer of width R/?
  at the edge of the emitting region Lesc
  Lemit/?
• For ?? ? e?e? attenuation occurs also outside of
  the emitting region ? ?2 ?1 ? nphR for steady
  sources ? exponential cutoff
• This assumes a uniform nph
• in emission region ? requires
• reasonably localized emission
• Holds for a relativistic source

Photon 1
R
Photon 2
49
The Relativistic Self Similar Regime
?12 1
c/3
The particle velocities are randomized at the
shock kinetic energy (bulk motion) turns into
internal energy
1. Upstream Ordered Velocities
2. Downstream Random Velocities
Shock Front

• The internal energy per unit rest mass is
  (?12-1)c2
• ? E ? ?2Mc2 where ? ? ?12 if E ? const (no
  energy gains or losses) ? ? M-1/2 ? R(k-3)/2 for
  ?1 ? R-k and M(ltR) ? R3-k

50
LAT Performance