Thank you for coming to this conference, and thank you for listening in English. I have many contributors to this work, coming out of several GRB projects done with Dan Perley, Pawan Kumar, Kevin Hurley, and many others.
In this talk, I will discuss almost exclusively Long GRBs, usually lasting greater than 2 s, with a softer spectrum than shorter ones.
[After introduction, ask if everyone is familiar with this diagram and tell them they may stop me if they are unfamiliar with anything.] GRBs are the universe’s most powerful cosmic explosions, visible back to when the universe was very different, beyond even the first stars were they to occur there, though they are already among the most distant objects measured. From X-ray, gamma-ray observations, and from observations of the aftermath of these explosions, we have models of many aspects of GRBs... But we know our simple models of emission from GRBs are wrong, that is, we cannot even connect our theories to what we measure! But this is low-hanging fruit, a great opportunity to make a really important discovery about the nature of GRBs.
I want to emphasize I am talking about the prompt emission in this talk.
I want to emphasize I am talking about the prompt emission in this talk.
Only about a dozen GRB have been measured well in the optical during prompt emission.
The “classic” optical flash, sometimes called Early External Reverse Shock, was long ago predicted by Meszaros & Rees, and Sari & Piran. The idea here is that the physics of the reverse shock gives electrons less energy, so while forward external shock is still building in X, the reverse shock could be radiating in opt. So, yes, it can be complicated, but there are different shapes, decay slopes, and correlation to help us identify the prompt emission. The point is that only the prompt emission is dominated by the physics of the jet, only this can tell us about the physics of the actual gamma ray burst.
This reference model, the Band function, absolutely does not explain optical emission, and without the optical measurements, you may be fooling yourself that X-gamma components are all there is.
Please note these plots are not my work.
This extremely long burst had UVOT data.GUiriec fits thermal + two other PLs and finds the HE component fits the optical the best.
Originally there was the ISS; it’s not consistent with data, and probably unphysical.We see slopes all over the place, and some thermal components But here’s something that is telling: Extrapolation to optical is often off my orders of magnitude. What’s going on there???
[CLICK ONCE TO START MOVIE] In the minute it takes me to start talking about the challenge of discovering the GRB Emission Mechanism, most GRBs will have been over, and, most optical telescopes would be unable to acquire the GRB until tens of minutes later. So, we have good measurements in the gamma bands, but not in the optical. The animated trace at bottom shows the gamma activity, over in about 50 s here, in this long GRB, the brightest ever measured. This is the only reason there is a good optical light curve to show you; measurements like this are less than once per decade events. The optical emission is very very hard to measure during the gamma emission; usually optical meaurements are later, and come from a completely different process, called the afterglow.
The first challenge to this problem is to build a telescope fast enough to catch this emission, usually much much fainter than in this exceptionally bright GRB. [ADVANCE] The second challenge is to measure it in at least two points on the spectrum, to determine the location of any spectral breaks, or measure the shape of any potentially independent optical component. [ADVANCE FOUR TIMES TO MAKE THE TELESCOPE DIAGRAM APPEAR] The way we have chosen to meet this challenge is to build a fixed telescope with a rapid beam-steering mirror. The SWIFT and other instruments scan for GRBs, and can alert a ground-based telescope rapidly. If we can respond within a few seconds, we can be the first instrument on target. This beam-steering solution is proposed because rapidly moving and damping a simple disc-shaped mirror should be easy and scalable to rather large mirrors. We know well how to stiffen such a simple shape. The second feature of this telescope is two cameras that simultaneously measure different bands by means of a beam splitter. NO other instrument does this and can respond fast enough to measure prompt emission, that is, emission during the gamma burst. If you put these two capabilities together, then you will be the first to have a catalog of GRB with measured emission mechanisms... a paper that will surely get a lot of attention and references! We want to call this the KRMT. [ADVANCE] Here is an example of such a beam-steered telescope, a small one built by our scientists at IEU for a space mission which was unable to fly.
Understanding Prompt Emission
from Gamma-Ray Bursts:
Fast, Early, and Multi-Colored
Energetic Cosmos Laboratory, Nazarbayev University & UC Berkeley
ECL Collaborators: George Smoot, Eric Linder,
I. What is a GRB?
• GRB stands for “gamma ray burst”
• This talk is almost exclusively LONG type GRBs
Long Gamma-Ray Bursts(GRBs)
• 15-200 keV Swift
BAT light curves
in every way
• LCs by Kas McLean
• Most Common GRB
• Associated with Star-
forming regions, SNe
• Typical z ~ 2
• LGRBs > 2 s, softer
spectrum than others
More GRB Basics
• Most energetic events in the universe
‣ Distant: z = 8.2 (090423); zphoto= 9.4 (090429b)
• Can be seen to z~12 with large detectors
• Gamma-Ray Bursts (GRB) last msec – hr.
• Measured up to GeV
• Long Type GRB sometimes massive star collapse
‣ GRB 980425 ==> SN1998BW
Gamma-Ray Bursts: Cosmic Mystery
• We have models of: Explosion, Jets, Gravitational
But the emission mechanism remains unknown!
• The Universe’s most energetic explosions
• Bright: can be seen beyond even first stars
~0-102 s ≳102 s
GRB Cartoon Picture
• GRB are so bright and variable because of relativistic be
Log Time ->
Most GRB Have Optical Afterglows
Prompt X/𝛾-ray light curves
bright (wide field instruments),
LGRB typically ~ 40-100 s
OPT AG MEASUREMENTS
COMMON - Hundreds
observed, dozens per year.
(Flares ignored here)
Physics well understood to
be interaction of a jet with
[Early or prompt phase not
measured for most GRBs]
Optical Prompt Measurements
Difficult & Rare
• Hundreds of GRBs have been observed in X/𝛾 bands.
• Hundreds of GRB afterglows, the interaction of the
blast and surrounding ISM, have been observed in
almost every band.
– This is NOT the burst (jet) emission.
• Prompt emission ≲ 102 s.
• Typical optical telescopes
require ≳102 s to point...
prompt optical observations rare
Not Observable by
Typical Optical Telescopes
Current Prompt Optical Observations
• Conventional Telescopes Too Slow
• Wide-Field Instruments
– Great Successes! (e.g. Pi of the Sky, Raptor)
– LIMITED SENSITIVITY~ 10th mag
• Medium-Field Fast instruments
– Great Successes! Polarization measurements!
– Limited Sensitivity -
• e.g. ROTSEIII - 45 cm - R ~ 16.9 mag 10 s ~ 3/ yr.
• e.g. MASTER-NET - 40 cm - 12-14 mag 10 s w/polarization
– NO OPERATIONAL SIMULTANEOUS MULTI-COLOR
• Note that filter wheels are useless for this rapidly-varying source
Either of first two could be
sub-dominant, and not
Log Time ->
But note, at Least 3 Opt Peaks Possible
Meszaros & Rees
(Flares ignored here)
-external forward shock
• 080319A was in same part of
the sky just before, so many
instruments were open,
– Lucky! Prompt optical emission finished
in ~ 100 s
– most telescopes cannot open or
point in less than minutes.
• Incredibly Bright!
– Nearly 5th mag!
– Amazing light curve by TORTORA,
vidicon instrument (Molinari+06)
– Detection by Pi-of-the-sky
Above instruments not sensitive to any
but most exceptionally bright bursts.
Best Prompt Light Curve: 080319B
080319B Light Curve
• Two-component jet proposed, 1 (𝜞~103) for ultra-
bright prompt optical, second (low 𝜞) for afterglow,
consistent with decay slope breaks and mis-matches
Time-Resolved Optical Data
• Such rich
burst in > 10
Spectrum in 3 time periods
Just one optical
point, doesn’t fit!!!
What about in here?
Fall steep or
Rich data here;
Spectra Commonly fit by
Band function, 2 PL with
10 s integration about:
Green: T0+3 s
Blue: T0+ 17 s
Red: To+32 s
• Uncorrelated 𝛾, Opt
• Opt >> 𝛾 (same as 080319b)
• Vestrand+14: Reverse Shock dom-
inates first ~ 50s (shock propagating
backwards toward jet origin; decay
but… non-unique fit , several parts
• ==> baryon-dominated jet
(reverse shock traveling into a
magnetic jet produces weak Optical*)
• Note optical spectrum not available to
* Zhang & Kobayashi 2005; Narayan et al. 2011; Giannios et al. 2008
Two more famous prompt optical
• Other famous cases
• Some might have 𝛾
optical, some not.
– Look at 990123: Could
this just be a delay, like
080319b, but longer?
• Really need better
optical time resolution
MOST GRBs Extinguished!
• Most GRBs have little optical emission (30/77 UVOT)
– BUT VIRTUALLY ALL GRBs HAVE IR EMISSION1
• Median extinction AV~0.35 mag2; range 0.5 - 5 mag1
• If you cannot study extinguished GRB, you may have
some kind of bias against the most active star-
• If you can detect extinguished GRB, you will detect
many more, ~ 1.6X more than UVOT3!
1 - Perley et al. 2009; 2- Covino et al. 2013; 3-Grossan et al. 2014,
110205A - 260 s burst
23Guiriec+16: ApJL 831, L8;
• Guiriec fits 3 components:
nTh1, nTh2 = PL w/expo
Th = Thermal (photo-
• Optical is key; claims
related to high energy
• Very long, bright in Swift, Suzaku/WAM (MeV)
• Fits suggest photospheric emission
• Prompt UVOT (very rare!) resembles WAM (MeV)
110205A - Guiriec Model
• NO optical spectral data here!!
• Fit is plausible
…But look at
-huge gap to optical!
-huge band from just one
point in optical!
• -> Need optical Spectrum
for more convincing fit.
Guiriec+16: ApJL 831, L8; https://arxiv.org/abs/1606.07193
• “Standard” Internal Synchrotron Shock Model1 (ISS; log slope +1/3)
– Equipartition roughly gives correct 𝜈 f 𝜈 peak energy(2)
– Most observations inconsistent; may be unphysical(1)
• Either multiple or variable slopes, components/mechanisms required
– Log Slopes 20-200 keV have broad distribution, ~0.1±0.35
– Thermal photospheric component pretty clear in some GRB
– Extrapolation to optical off (+ or -) by orders of mag
• More recent fits explore Maxwellian vs. PL e– N(E),
still disagree whether synchrotron acceptable or not. (3)
• Conclusion: Not just heterogeneous, but also no consensus.
1. Rees & Me ́sza ́ros 1994, see Piran 2005
2. Ghisellini, Celotti, Lazzati, 2000 MNRAS 313,1 . Note they state that correct time-averaging gives slope -1/2, inconsistent
2. Burgess arXiv 1705.05718 vs. Axelsson & Borgonovo 2015 MNRAS447,3150; Yu et al. 2015 A&A, 583, A129
• Optical Slope carries important information about
• Break Frequency encodes information about the physics
of the blast, including remission and electron energy.
Shen & Zhang 2009
Prompt Optical Spectrum, Light
Curve Add Important Information
• Test Extrapolation of 20-200 keV component
– Optical Brighter than Extrapolation: Must be an added component
– Optical Fainter: Must be break between optical & X
• Important Tool: 𝛾 / Opt Correlation
– correlated variability indicates same location, mechanism, and
– Want instrument with high time resolution!
• Spectral Slope within Optical
– Indicates mechanism of optical component
• Self-absorption frequency1, 𝜈a
– If synchrotron: Gives B, e- thermal Lorentz factor, radius of emission.
– If photospheric: thermal Lorentz factor (energy dissipation of jet)
28(1) Shen & Zhang 2009 MNRAS 398,1936
SIMPLE EXPERIMENTAL GOAL
For burst sample covering wide range of properties,
• measure optical-X/𝛾 light curve correlations
• measure broad-band optical spectral shape, including
Not Possible with Current Instruments
• limited by time resolution
• limited by filter-wheel non-simultaneous colors
IV. How to Observe GRBs in
the Prompt Phase
How to Observe GRBs in Prompt Phase
Swift Monitors GRBs in γ rays
Sends alert via GCN (internet) in ~ 2 s,
elescope Rapidly Points
– Fast (< 10s response)
– Simultaneous “blue”, “red”, IR cameras
• Faster Telescopes (direct-drive, computer-control of
• EMCCDs - allow high time resolution, with high QE
and negligible electronic noise penalty
Modified 700mm Telescope
• Can point anywhere in
sky in < 8 s
– Test data on sky 10/16
• CDK design with Two
– Can operate a triggered
search with one instrument,
have another used for other
Try movie IMG_1254.MOV
Multi-Channel Simultaneous Observations
• Separate waveband to each camera via dichroics
• High time resolution without noise penalty via
Electron Multiplied CCD cameras
CAD design rendering
Burst Simultaneous Three-Channel Instrument
• B,R Bands: EMCCD Cams +
Blue, Red Filter
• H-band Camera
• H2RG (HgCdTe) sensor,
cryostat + Lyot Stop + LPT
Cooler (no consumables)
– 3rd channel to identifies
absorption frequency or curvature
– allows study of extinguished
bursts (most GRB extinguished;
Prochaska+09, ApJL 691, L27)
Additional mirrors not shown for clarity.
Final design Prof. Spitas NU Engineering- in progress.
Optics Design Complete
• Ray Trace By L. Scherr
• Electron Multiplied CCD
– Each pixel’s signal is multiplied before
read by up to 5000X
– Effective read noise ~ 10-2 e-
– Many frames co-added with negligible
• No penalty for sub-second time resolution
• Same quantum efficiency as CCD
• ~300 ms/frame 1024X1024 13 µm pixels
• We estimate > 6 mag dynamic range in our
• 5-sigma Sensitivities
B R H
• 700 mm aperture telescope, high-QE sensors, good
sampling of PSF, KPNO-like sky
• Actual performance will vary.
Can we detect prompt emission?
• Distribution of prompt
brightness not well known-
measurements not uniform
• We do have a uniform early
time (afterglow) sample from
UVOT, t~ 110-170 s. Gives
sensitivity required to detect
prompt optical at least as
bright as afterglow onset.
Conversion of W to R assumes: source log slope -0.75, median from Covino+13 Av=0.35 mag
at source, Av=0.08 mag MW, z= 1.8, sample median.
Brightness at end of prompt phase
• Our Telescope
will go much
fainter than any
afterglow in 10 s.
• R ≤ 20 will detect much fainter than
any afterglow in 10 s.
Estimated Annual Detection Rate
• For a conservative, realistic rate (including clouds,
etc.) we scale from ROTSE-III values, and from
Early Brightness distribution
– Not an accurate detection number, but, an accurate rate of
measurements much more sensitive than early afterglow
• We expect our IR channel to boost detection rate by
factor of 1.6 due to extinction
• Uncertainty dominated by weather!
Detections Upper Limits Total
Optical 4.3 8.7 13
IR 6.9 6.1 13
Color-Color gives Slope, 𝜈a
• Different slopes separate well on color-color plane
• If between our bands, break frequency, 𝜈a, determined.
White = no 𝜈a feature
Green = 𝜈a @ c/1.3 µm
Red= 𝜈a @ c/1.0 µm
1.0 µm feature
1.3 µm feature
Photometry techniques, calibration,
now being optimized by M.
Kistubaev, NU masters student
• LGRB associated with dusty star forming regions
• GRB expected to vaporize dust throughout typical star
• Typical cloud size ~ 10's of light sec
• Time-dependent extinction measurement would
• confirm calculations of dust density, evaporation, probe local
• Solves excess gas absorption problem - Too much X-ray absorption
for blue, low-extinction afterglow(3,4,5)
• Need time-dependent spectral slope starting earlier than
most previous measurements
t ---> 60s
(1)Waxman, E., & Draine, B. 2000, ApJ, 537, 796
(2) Perna, R., Lazzati, D., & Fiore, F. 2003, ApJ, 585, 775
(3) Galama& Wijers 2001, ApJL, 549, L 209 ; (4)Stratta+04, ApJ, 608, 846
(5)Schady+07, MNRAS, 377, 273; Perley+09, AJ, 138, 1690
• Given extinction curve, band ratio gives extinction
• Intrinsic slope given by final, unchanging colors
• Extinction in individual star system at high-z for first time!
(SMC Extinction; source log slope = –0.75; NGRG
Grossan et al. 2014, PASP, 126, 885
Summary: 1 Experiment, 3 Big Topics
1. Measurement of prompt optical-IR broad-band slopes,
never done before, will be a substantial advance in
understanding GRB emission mechanisms.
2. Measurement of the SSC absorption frequency, 𝜈a, never
done before, will give new information about the physical
conditions in GRB Jets, including radius of emission,
magnetic field strength, electron energies, etc.
3. Measurement of dynamic dust extinction, never done
before, will solve the excess X-ray absorption problem,
and allow study of dust around a single star, separate
from galactic dust, in star-forming regions at z> 1.