1.  -ray radiation mechanism – Fermi data Gamma-ray bursts in the Fermi Era † Work done with Ramesh Narayan & Rodolfo Barniol-Duran Pawan Kumar Outline † Problems with the current paradigm and possible solutions Brief historical introduction

2. For about 30 years we did not even know how far away are these explosions from us – and therefore we were ignorant about the energy release in GRBs.. Picture showing Vela launch using a Tital II-c rocket; the nuclear test ban treaty was signed in 1963. Insert at the top shows Vela at Strategic Air & Space Museum, Nebraska. We have come a long way since the accidental discovery of Gamma-ray bursts (GRBs) by the Vela satellites about 40 year ago. year ago.

3. It is now firmly established – as a result of observations carried out by BeppoSAX, HETE II & Swift – that these explosions are at large (cosmological) distances, Swift launch - Nov 20, 2004 HETE II Oct 2000

4. Stanek et al., Chornock et al. Eracleous et al., Hjorth et al., Kawabata et al. SN 1998bw: local, energetic, core-collapsed Type Ic GRB 030329: z=0.17 1.8 billion light-year (afterglow-subtracted) ‏ GRB 030329 Or SN2003dh Emission lines of CII, OII and OIII and that at least some GRBs are produced when a massive star undergoes collapse (Woosley and Paczynski)‏ According to Mazzali et al. (2003) the energy in this hypernova was 2x10^{52} erg, The velocity & mass of ejecta 25000 km/s and 8 M_sun. Matheson et al. (2003) suggest that SN lightcurve Peaked at about 20 days, and the luminosity of 030329 SN was similar to SN-98BW. GRB 031203 (an x-ray flash) at z=0.105 was also found to be associated with a Sne This was based on both the spectrum and the LC which peaked at ~ 20 days and then Declined -- similar to the spectrum & LC of 98bw -- except that it was brighter than 98bw By about 50%! (Malesani, Tegliaferri et al/astro-ph-0405449 is a good paper). X-ray flash 020903 also seems to show a spectrum like 98bw at 25 days after the burst (Soderberg astro-ph/0502553). Spectrum is like a finger print -- the criminal had Left a finger print astronomers could easily read! 030329 was at a redshift of 0.1685; d a =589 Mpc.

5. Short burst Long burst Kouveliotou et al. (1993) Gamma-ray bursts last for ~10ms – 10 2 s. And the true amount of γ-energy release in these explosions is ~10 48 erg – 10 52 ergs.

6. These explosions are observed for a much longer duration of time – days to months – in x-ray, optical & radio bands. The long lasting emission – following the brief period of γ-ray emission – is called the afterglow. Afterglows are produced when the GRB jet shock-heats medium in the vicinity of the explosion, and accelerates e - s which radiate via synchrotron process. Gehrels, Piro & Leonard: Scientific American, Dec 2002

7. Panaitescu & Kumar (2001) Afterglow theory provides a very good fit Afterglow theory provides a very good fit to the multi-wavelength data for t>0.5 day to the multi-wavelength data for t>0.5 day Afterglow radiation was predicted well before its discovery by Paczynski & Rhoads (1993) and Rees & Meszaros (1992), and extensively investigated by Piran, Sari, Narayan….

8. Prompt  -ray generation mechanism O’Brien et al., 2006 Factor ~ 10 4 drop in flux! Data from the Swift satellite shows very clearly that the engine turns off quickly at the end of γ-ray burst – a few x 10s after the first photons are detected. If the rapid turn-off is due to a fast decline of accretion rate onto the newly formed black-hole, then we can “invert” the observed x-ray lightcurve and determine progenitor star structure.

9. time flux steep fall off prompt GRB emission rapid decline X-ray plateau r ≈ 9  10 9 cm r ~ 1.5  10 11 cm f Ω ~  2  f Ω  2 3  10 10 cm   r -2.5 f    k Progenitor Star Properties Kumar, Narayan & Johnson (2008)

10. 6/11/2008 Fermi 8 KeV to 300 GeV One of the goals for Fermi is to understand γ-ray burst prompt radiation mechanism by observing high energy photons from GRBs. Let us see how Fermi has done… How are γ-rays generated?

11. MeV emission lasts for a shorter time duration than x-ray optical & radio. Radiation Mechanism Prompt γ-rays & x-ray, optical and radio (AG) photons are produced by two different sources. And it turns out that ~MeV emission is shorter lived than GeV! (CGRO/EGRET had already noticed this for a few burts, and Fermi has shown this to be true for just about every burst it has detected.) This suggests different sources for ~MeV and GeV photons.  γ-rays not ES

12. Radiation mechanisms IC of sub-MeV seed photons in internal Shock, RS or FS, or hadronic collision or photo-pion process... Katz 1994; Derishev et al. 1999; Bahcall & Meszaros 2000 Dermer & Atoyan 2004; Razzaque & Meszaros 2006 Fan & Piran 2008; Gupta & Zhang 2008; Granot et al. 08… Meszaros & Rees 1994; Pilla & Loeb 1996; Dermer et al. 2000 Wang et al. 2001 & 06; Zhang & Meszaros 2001; Sari & Esin 01’ Granot & Guetta 2003; Piran et al. 2004; Fan et al. 2005 & 08 Beloborodov 2005; Fan & Piran 2006; Galli & Guetta 2008 Pe’er et al. 06; Granot et al. 08; Bosnjak, Daigne & Dubus 09 (~10 2 man-years of work has gone into it) Radiation mechanism ( high energy  -rays ) Fermi/LAT data offers a surprising answer…

13. Abdo et al. (2009) Delayed high energy emission;

14. Abdo et al. 2009 (GRB 080916C) Long lived lightcurve for >10 2 MeV (Abdo et al. 2009) f ν α ν -1.2 t -1.2 α = 1.5β – 0.5 (FS) α = 1.5β – 0.5 (FS)

15. >10 2 MeV data  expected ES flux in the X-ray and optical band (GRB 080916C) We can then compare it with the available X-ray and optical data. Abdo et al. 2009, Greiner et al. 2009, Evans et al. 2009 Long lived lightcurve for >10 2 MeV (Abdo et al. 2009) Kumar & Barniol Duran (2009)

16. Or we can go in the reverse direction… Assuming that the late (>1day) X-ray and optical flux are from ES, calculate the expected flux at 100 MeV at early times And that compares well with the available Fermi data. X-ray Optical > 100MeV 50 - 300keV Abdo et al. 2009, Greiner et al. 2009, Evans et al. 2009 Kumar & Barniol Duran (2009)

17. GRB 090902B: LAT data  expected flux at late times Bissaldi et al. 2009 f ν α ν -1.2 t -1.2 α = 1.5β – 0.5 (FS) α = 1.5β – 0.5 (FS)

18. We can then compare it with the available X-ray and optical data. We can then compare it with the available X-ray and optical data. Bissaldi et al. 2009 Bissaldi et al. 2009; Swenson et al. 2009; Guidorzi et al. 2009; Evans et al. 2007, 2009 GRB 090902B: LAT data  expected flux at late times Using the parameter space determined from the LAT data: What is the expected ES flux in the X-ray and optical band? Using the parameter space determined from the LAT data: What is the expected ES flux in the X-ray and optical band? Kumar & Barniol Duran (2009)

19. Assuming the X-ray, optical and radio flux are from ES: what is the expected flux at > 100 MeV at early time? Assuming the X-ray, optical and radio flux are from ES: what is the expected flux at > 100 MeV at early time? Swenson et al. 2009; Guidorzi et al. 2009; Evans et al. 2007, 2009 GRB 090902B: Late time data  expected LAT flux

20. We can then compare it with the available Fermi data. Swenson et al. 2009; Guidorzi et al. 2009; Evans et al. 2007, 2009 Assuming the X-ray, optical and radio flux are from ES: what is the expected flux at > 100 MeV at early time? Assuming the X-ray, optical and radio flux are from ES: what is the expected flux at > 100 MeV at early time? Bissaldi et al. 2009; Swenson et al. 2009; Guidorzi et al. 2009; Evans et al. 2007, 2009 GRB 090902B: Late time data  expected LAT flux Kumar & Barniol Duran (2009)

21. Ghisellini, Ghirlanda and Nava (2009) t -(3p-2)/4 ≈ t -1.3 t -(3p-2)/4 ≈ t -1.3 (independent of n, ε ) B

22. Using only >100MeV Fermi data Parameter search at t = 50 sec. Magnetic field in shocked fluid: GRB 090902B 30 μG 5 μG only late time x-ray, optical & radio data Parameter search at t = 0.5 day. 30 μG 5 μG ε B is consistent with shock-compressed magnetic field of CSM of a few tens of μG; Similar conclusions found for GRB 080916c & GRB 090510. (Kumar & Barniol Duran, 2009) This suggests either a weak dynamo or no dynamo in relativistic shocks. Similar conclusion found for SNa remnants (Thompson et al. 2009)

23. central engine relativistic outflow Make point 1 ONLY: 2. Central engine is completely hidden from our view so the progress that is being made is via numerical simulation of core collapse and other very interesting works (woosley et al. Quataert et al…)‏ 1. The relativistic jet energy produced in these explosions is dissipated at some distance from the central engine and then a fraction of that energy is radiated away as gamma-rays. CONSIDERING OUR LACK OF UNDERSTANDING OF GRB JET COMPOSITION IT IS BEST TO TREAT JET DISSIPATION AND GAMMA-RAY PRODUCTION SEPARATELY. DO NOT seond more than 30d on this slide. Jet energy dissipation and γ-ray generation External shock radiation central engine  jet   -rays I hope you see the compelling evidence we have that >10 2 MeV photons – and x-ray, optical & radio photons – from GRBs are generated in external shock. How about photons of energy between ~10kev & 0.1 Gev? Emission in this band lasts for <10 2 s, however it carries a good fraction of the total energy release in GRBs. a good fraction of the total energy release in GRBs. And it offers the best link to the GRB central engine.

24. Internal/external shocks, magnetic reconnection etc. A lot of people have worked on this problem Synchrotron, SSC, IC of external photon field, thermal radiation, jitter radiation… Piran et al. ; Rees & Meszaros; Dermer; Thompson; Lyubarsky; Blandford, Lyutikov; Spruit… Papathanassiou & Meszaros, 1996; Sari, Narayan & Piran, 1996 Liang et al. 1996; Ghisellini et al. 2000; Thompson (1994); Lazzati et al. (2000); Medvedev (2000); Meszaros & Rees 1992-2007 Totani 1998; Paczynski & Xu 1994; Zhang & Meszaros 2001…

25. 2. Synchrotron-self-Compton solutions SSC solutions have E e  R 3 and E B  R -4  emission must be produced within a narrow range of R (factor ~ 2) and that narrow range of R (factor ~ 2) and that is highly unlikely for the internal shock is highly unlikely for the internal shock model. model. sharp minimum of E e + E B ; Kumar & Narayan (2009) No reason that jet energy should dissipate at the minimum of E e +E B There is another problem with the SSC solution: Lack of excess flux in the Fermi LAT band; Abdo et al. 2009 (also Ando, Nakar & sari 2008; (also Ando, Nakar & sari 2008; Piran et al. 2008) Piran et al. 2008) Synchrotron: e - s cool rapidly  f    rarely observed) 1. Synchrotron: e - s cool rapidly  f    rarely observed) External shock ruled out (Sari and Piran’s excellent argument) It can’t be internal shock either: (<MeV & GeV behave differently)

26. 1. Thermal radiation + IC Thompson (1994 & 06); Liang et al. 1997; Ghisellini & Celloti 1999; Meszaros & Rees (2001); Daigne & Mochkovitch (2002); Pe’er et al. (2006)… (for prompt  -rays) Problem: we don’t see a thermal component in GRB prompt emission – Ryde (2004, 05) claims to find evidence for thermal spectrum, but Ghirlanda et al. 2007 do not. There are only three possibilities 2. Continuous acceleration of electrons If electrons can be continuously accelerated (as they lose energy to radiation) then many of the problems mentioned earlier disappear (Kumar & McMahon, 2008). However, is it possible to accelerate electrons continuously in shocks (as opposed to impulsively (Transferring energy from at the shock front) ? (Transferring energy from shock heated protons to e - s continuously requires unlikely fine tuning.) 3. Relativistic turbulence Lyutikov & Blandford 03’; Narayan & Kumar 09’; Lazar, Nakar & Piran 09’ Beloborodov (2009)

27. Line of Sight 1/  RsRs tt  2. Relativistic Turbulence Model Variability time = R s (1+z) ———— (2c  2 )  t 2 Synchrotron emission in quiescent part of shell  less variable optical IC scattering of synchrotron off of blobs   -rays (more variable)‏ Lyutikov & Blandford 03’; Narayan & Kumar 09’; Lazar, Nakar & Piran 09’ Consistent solutions for  -rays are found in this case & R s ~R d as suggested by observations. One possibility: (This is a way to get around the argument of Sari and Piran, 1997)

28. GRB 080319B: relativistic turbulence model Kumar & Narayan (2009)‏ The ratio of IC & synchrotron energies is ~10, i.e. Y~ 10. Optical band lies above the synchrotron peak where f  -2.9 and so the total energy in synchrotron photons is larger than optical emission by a factor ~ 10. The total energy in 2 nd IC component is of order the 1 st IC scattering. The 2 nd IC lies in the K-N regime since  i /  >1, and the peak photon energy in e - rest frame is ~ 2 MeV. Focus on Y being order 10 and energy in 2nd IC of order 1st IC

29. Steep decline (~t -5 ) consistent with LAE (not RS)‏ RS emission GRB 080319B: x-ray & optical LCs Kumar & Narayan (2009)

30. Relativistic turbulence model: outstanding issues 2. Not clear how to get asymmetric pulses in GRB lightcurve. 3. Correlation between pulse width and the gap preceding it (reported by Nakar and Piran, 2002; Quilligan et al. 2002). 1. How to generate relativistic turbulence?

31. Relativistic turbulence (Poynting jet) has a number of attractive properties, but there are also a number of difficulties. Relativistic turbulence (Poynting jet) has a number of attractive properties, but there are also a number of difficulties. Summary 2. Generation of sub-MeV γ-rays remains unclear: 2. Generation of sub-MeV γ-rays remains unclear: ✪ The internal-shock model for GRBs + synchrotron ✪ The internal-shock model for GRBs + synchrotron (SSC) radiation mechanisms have severe problems. (SSC) radiation mechanisms have severe problems. 1. High energy  -ray data for a number of bursts — 080916C, 090510, 090902B — observed by Fermi suggest that >10 2 MeV photons are produced in shock heated CSM; ε B ~ shock-compressed CSM field. 1. High energy  -ray data for a number of bursts — 080916C, 090510, 090902B — observed by Fermi suggest that >10 2 MeV photons are produced in shock heated CSM; ε B ~ shock-compressed CSM field. ✪