INTEGRAL Y6+ The New Soft Gamma-Ray Sky New Observational Results with Integral: GRB Pietro Ubertini, Neutron Stars & Gamma Ray Bursts Alexandia, April , Page 1 P. Ubertini, A. Corsi, S. Foley, A. Bazzano, et al., on behalf of the IBIS Survey Team

INTEGRAL Instruments IBIS mask Optical Monitor IBIS Detectors JEM-X JEM-X Masks SPI GRB within the instruments FoV: 17keV-10 MeV GRB out of the FoV through IBIS and SPI shields: E>~200 keV Real time data transmission

INTEGRAL GRBs INTEGRAL has detected 62 GRBs since launch in October 2002 up to January 2009 ~1/month 1st INTEGRAL catalogue of 46 GRBs published (Foley et al. A&A, 2008) 4 INTEGRAL GRBs have confirmed redshifts: GRB – z = GRB – z = GRB050502a – z = 3.79 GRB050525a – z = Pietro Ubertini, Neutron Stars & Gamma Ray Bursts Alexandia, April , Page 3

Spatial Distribution of INTEGRAL GRBs INTEGRAL exposure map in Galactic coordinates from Oct 2002 to July 2007 Distribution of INTEGRAL GRBs in Galactic coordinates, dominated by the exposure Pietro Ubertini, Neutron Stars & Gamma Ray Bursts Alexandia, April , Page 4

GRB : 1 st GRB observed by INTEGRAL On Nov. 25th, 2002, the satellite was set up for a special observation with the PICsIT layer in a non–standard photon–by–photon mode, reduced number of channels, and most of the satellite telemetry allocation: we new IBIS was able to see GRBs!. During this test, at 17:58:30 UTC GRB occurred in the partially coded field of view of IBIS (about 7.3 deg off–axis), and lasted about 24 s. Malaguti et al SPI light-curve (in Crab units) obtained from the detector count rates in the energy range 0.02–8 MeV; time starts from 17:58:00 UT. Pietro Ubertini, Neutron Stars & Gamma Ray Bursts Alexandria, Egypt, April , Page 5

Pietro Ubertini, Neutron Stars & Gamma Ray Bursts Alexandria, Egypt, April , Page 6 One month larer, on December 19, 2002, during the Performance and Verification Phase GRB has been detected and localized in real time with the INTEGRAL Burst Alert System (IBAS). - IBIS/ISGRI light curve of in the 15–500 keV band (upper panel), -SPI light curve of GRB (middle) -A clear hard-to-soft spectral evolution with time was observed 1 st GRB : 1 st Gamma-Ray Burst localized in real time with IBAS

GRB : 1 st detailed time-resolved spectroscopy of a faint GRB Gotz et al IBIS/ISGRI light curve a) 15–50 keV b) 50–300 keV c) 15–500 keV. The 6 data gaps are caused by satellite telemetry saturation. 4 peaks can be identified at 15, 40, 55, 85 s. d) Spectral variation of the GRB with time. BATSE in brighter GRBs (e.g. Preece et al. 1998) or by BeppoSAX GRBs (e.g. Frontera et al. 2000; 2003). The fluence (7exp-6 erg cm -2 ) was a factor of 10 smaller than BATSE bright bursts, indicating that such spectral behavior applied also to faint Spectroscopy: consistent with the hard-to-soft evolution observed by BATSE in brighter GRBs (e.g. Preece et al. 1998) or by BeppoSAX GRBs (e.g. Frontera et al. 2000; 2003). The fluence (7exp-6 erg cm -2 ) was a factor of 10 smaller than BATSE bright bursts, indicating that such spectral behavior applied also to faint GRBs. Pietro Ubertini, Neutron Stars & Gamma Ray Bursts Alexandria, Egypt, April , Page 7

GRB a): 1 st quick localization by Integral… Quick localization by INTEGRAL Burst Alert System: discovery of X-ray and optical afterglow. about 8 hr later by XMM-Newton (was a record at that time!). The GRB lasted about 20 s, and the X-ray afterglow was detected about 8 hr later by XMM-Newton (was a record at that time!). (see e.g. Piro 2002), Absorbed power-law fit to the afterglow spectrum: tentative evidence for a Fe emission line (see e.g. Piro 2002), at 1.67 keV. The implied z ∼ 3 was consistent with the z derived from the absorption. Mereghetti et al But the existence of Fe lines in GRBs remains an open issue: EPIC PN best-fit spectra of GRB afterglow

GRB b):…and evidence for an IC component The NIR/optical (solid line) and X-ray (dotted line) afterglow spectrum at 0.87 days since trigger. The NIR/optical spectral index is consistent with the X-ray one, but they dot match each other’s extrapolations, similarly to e.g. GRB (Harrison et al. 2001), GRB (in’t Zand et al. 2001), GRB (Corsi et al. 2005), GRB (Chandra et al. 2008). IC component Synchrotron component Castro-Tirado et al Pietro Ubertini, Neutron Stars & Gamma Ray Bursts Alexandria, Egypt, April , Page 9

GRB a): an extremely hard burst INTEGRAL detects GRBs in two different ways from 17 kev up to 10 MeV : - a small number of events fall in the FoV of IBIS and SPI; Compton mode of IBIS - a significantly larger number occurs outside of the FoV but can be monitored by the SPI-ACS. For some of these bursts it is also possible to perform a more detailed localization analysis using the Compton mode of IBIS, provided the burst is strong and spectrally hard, as in this case. Marcinkowski et al.2003 Pietro Ubertini, Neutron Stars & Gamma Ray Bursts Alexandria, Egypt, April , Page 10

GRB b): an extremely hard burst Combined ISGRI and Compton mode data. Broken power law model in each case: precursor (left), peak part (center) and tail (right). The peak spectrum is very hard: low energy νFν spectrum below 400 keV rises with index ≈+3.5 and above this energy is still positive ≈+0.3, so E peak >1.1 MeV. Clear contradiction with the synchrotron model of GRBs which predicts that there is a strict upper limit on the low energy spectral index of −2/3 (Preece et al. 1998). The low energy spectral slope is consistent within the error bars with the jitter synchrotron model of GRBs (Medvedev 2000). Pietro Ubertini, Neutron Stars & Gamma Ray Bursts Alexandria, Egypt, April , Page 11

GRB : unusually low luminosity, nearby burst On 2003 December 3 at 22:01:28 UTC, IBIS detected a pulse of 40 s duration, with a simple profile. The spectrum was also typical, with a single power law model that constrained E peak >190 keV. (4±1)x10 49 erg at z=0.1 The burst fluence in the 20–200 keV band implied an isotropic energy of (4±1)x10 49 erg at z=0.1. Is sub-energetic as GRB , associated with the nearby (z=0.0085) SN 1998bw, that had E iso <10 48 erg and violates the E iso –E peak relation, that would predict E peak <10 keV, as GRB sub-energetic in the  -ray band The 2 nearest long GRBs are clearly sub-energetic in the  -ray band, and their proximity (and hence implied abundance) makes it of great interest to understand their origin and relation to the more distant cosmological GRBs. Searches for associated GW signals were perform by LIGO in coincidence with this GRB (Abott et al., 2005) Sazanov et al., Soderberg et al., 2003, worth 2 Nature papers

Pietro Ubertini, Neutron Stars & Gamma Ray Bursts Alexandria, Egypt, April , Page 13 (5 and 10 days after the burst) confirm the presence of the afterglow. 0.3<z<0.7 Observed with IBIS+JEM-X: light curve and spectrum are consistent with other XRFs; Chandra data taken in two epochs (5 and 10 days after the burst) confirm the presence of the afterglow. The 1” X-ray afterglow localization enabled to discover the host galaxy of this XRF at 0.3<z<0.7 (D'Avanzo et al. 2004). ; is more an X- ray rich GRB rather than an XRF Taking into account z: it matches the E peak -E iso correlation; is more an X- ray rich GRB rather than an XRF (as instead are and ). GRB : a (false XRF) and X-ray rich GRB observed by Integral

GRB a a): intense and ultra-long burst enable studies GRB a a): intense and ultra-long burst enable polarisation studies - Brightest burst localized by INTEGRAL. Peak flux: 1.84 ×10 −5 ergs cm −2 s −1 (20 keV– 8MeV, 1 s integration); - T 90 duration of ∼ 186 s ( ∼ 20 keV–8MeV). - The intense burst occurred about ∼ 250 s after the precursor and the long delay enabled optical and near infrared telescopes to observe the prompt emission. McBreen et al Pietro Ubertini, Neutron Stars & Gamma Ray Bursts Alexandria, Egypt, April , Page 14

Pietro Ubertini, Neutron Stars & Gamma Ray Bursts Alexandria, Egypt, April , Page 15 GRB a b): polarization studies Azimuthal distributions of the flux in the different time intervals. Chance prob. of a non-polarized signal reported in each panel. Light curve of GRB A. The analyzed intervals, are shown with dashed lines. P8 is omitted for clarity. Measuring the polarization of the prompt GRB can significantly improve our understanding of both the emission mechanisms as well as the underlying engine driving the explosion. 200–800 keV The technique was to use the IBIS telescope on board the INTEGRAL to measure the polarization of the prompt gamma-ray emission of the long and bright GRB A in the 200–800 keV energy band. No polarization signal found integrating over the whole first peak, and the upper limit is 4%. On the other hand, a modulated signal is seen in the second peak corresponding to 43 ± 25%. Integrating over smaller portions of the GRB, give highly polarized signals, especially in P8, P9 and P30 (Gotz et al. 2009, but see also McGlynn et al. 2007, Kalemci et al. 2007). Gotz et al. 2009

GRB a (II): polarization studies Measuring the polarization of the prompt GRB emission can significantly improve our understanding of both the GRB emission mechanisms as well as the underlying engine driving the explosion. SPI has the capability to detect the signature of polarised emission from a bright γ–ray source. Polarisation can be measured using multiple events scattered into adjacent detectors because the Compton scatter angle depends on the polarisation of the incoming photon. GRB a (McGlynn et al. 2007): degree of linear polarisation in the brightest pulse of duration 66 s: −30 % at an angle of −11 deg (100–350 keV). In the brightest 12 s of the GRB: −40 % at an angle of −14 deg (100–350 keV). A systematic effect that could mimic the weak polarisation signal could not be definitively excluded. However, this case demonstrated the effectiveness of using SPI as a polarimeter in intense GRBs. McGlynn et al. 2007, Kalemci et al. 2007

Pietro Ubertini, Neutron Stars & Gamma Ray Bursts Alexandria, Egypt, April , Page 17 INTEGRAL/BATSE T 90(%of γ ph) Distribution 3 short INTEGRAL GRBs global properties INTEGRAL detects Proportionally less short bursts than BATSE

INTEGRAL/Swift Photon Index Distribution apparently no big difference Pietro Ubertini, Neutron Stars & Gamma Ray Bursts Alexandia, April , Page 18

INTEGRAL/Swift Peak Flux Distribution INTEGRAL detects proportionally more faint : why? GRBs than Swift: why? Pietro Ubertini, Neutron Stars & Gamma Ray Bursts Alexandia, April , Page 19

Spectral Lags Many GRBs show time delay between the arrival time of high and low energy photons  more energetic photons tend to arrive earlier Cross-correlation analysis of light-curves in different energy bands is used to determine this time lag (  ) ‏ Anticorrelation observed between lag and luminosity in long-duration GRBs: L peak ~ 1.3x10 53 (  /0.01) erg sec -1 (Norris et al. 2000) ‏  Can in principle use spectral lag as a GRB distance indicator Pietro Ubertini, Neutron Stars & Gamma Ray Bursts Alexandia, April , Page 20

INTEGRAL Spectral Lag Distribution Lags determined for 30 GRBs in the sample between the 25-50keV and keV energy bands No negative lags observed (i.e. low energy photons leading high energy photons) ‏ the are 12 Long-lag GRBs with τ > 0.75 seconds  = 0.75s τ = 0.75 s

IBIS (20–200 keV) with12 long-lag GRBs Cumulative logN-log P distribution of the 55 GRBs detected by IBIS (20–200 keV) with the small subset of 12 long-lag GRBs shown separately. The distribution is biased by the lower sensitivity of IBIS at large off-axis angles, but they looks different. New Observational Results with Integral

INTEGRAL GRB Distribution in supergalactic Coordinates INTEGRAL exposure map and GRB distribution in Supergalactic Coordinates Supergalactic plane – plane containing local superclusters of galaxies, web of filaments and sheets rather than an isolated pancake structure, superclusters evident out to ~400 Mpc 10/12 Long-lag GRBs within ± 30 o of Supergalactic plane Quadrupole Moment = ± for long-lag GRBs Pietro Ubertini, Neutron Stars & Gamma Ray Bursts Alexandia, April , Page 23 Foley et al A&A

Long-Lag GRBs: distance Scale & Rate of Low-Luminosity GRBs A number of low-luminosity GRBs at low redshift detected e.g. GRB (36 Mpc,  =2.8s) and GRB (145 Mpc,  =66s) ‏ Weak BATSE bursts are correlated with galaxies out to ~150 Mpc (Chapman et al. 2007) ‏ 8 Long-lag GRBs in the partially coded field of view of IBIS (0.1sr) ‏ Assume 2 are at high redshift Adopt a distance of 250 Mpc for the remainder All-sky rate ~2500 Gpc -3 yr -1 for these GRB with a large uncertainty due to distance 3 factor This exceeds the upper limit of 300 Gpc -3 yr -1 of Type 1b/c SNe which produce GRBs, assuming that all low-luminosity GRBs produce a SN (However not all GRBs produce SNe, e.g. low-luminosity GRB060505) - Galaxy clusters may play a role – new progenitor? Pietro Ubertini, Neutron Stars & Gamma Ray Bursts Alexandia, April , Page 24

Distribution of Swift GRBs in Supergalactic Coordinates No obvious concentration of long-lag GRBs with SG plane Quadrupole Moment = ± The distribution of Swift GRBs in supergalactic coordinates. Short-lag GRBs are shown as dark green diamonds and the 17 long-lag GRBs are identified by purple circles. There is no obvious anisotropy for long-lag Swift GRBs with respect with the supergalactic plane, why?...

IBIS vs BAT sensitivity to weak GRBs The answer could be that Integral has a factor of 2-4 better sensitivity to weak GRBs and an efficiency extending up to several MeV. The detection sensitivity of a number of γ-ray missions, shown as the peak flux threshold (1–1000 keV) to a GRB with a given Epeak.

INTEGRAL GRB Summary 1 62 GRB INTEGRAL has detected 62 GRB up to January 2009 and provides localisations for ~ 1 GRB / month low-luminosityhigh-luminosity detects proportionally more faint GRBs than Swift and appears to probe a low-luminosity population distinct from the high-luminosity one 12/30 GRBs for which a spectral lag was measured have long lags (  > 0.75 s) ‏. 40% of the INTEGRAL In comparison, the 149 Swift GRBs with a measured lag, 12% have long lags, compared with 40% of the INTEGRAL sample, the median peak flux of the 17 Swift long-lag GRBs is 1.71 ph cm -2 s -1 a factor of 3 times higher than for INTEGRAL long-lag bursts Long-lag GRBs have low peak fluxes, long slow pulses faint optical and X-ray afterglows and appear to be associated with the Supergalactic plane, and appear to be distinct from high-luminosity population (Foley et al., 2008). Pietro Ubertini, Neutron Stars & Gamma Ray Bursts Alexandia, April , Page 27

INTEGRAL GRB Summary 2 Some of these bursts could be produced by the collapse of a massive star without a supernova. Alternatively, they could result from a different progenitor, such as the merger of two white dwarfs or a white dwarf with a neutron star or black hole, possibly in the cluster environment without a host galaxy. & no narrow gamma-ray lines (SPI) Nor broad (IBIS)! Finally, INTEGRAL detects a large proportion of faint, long-lag GRBs that are inferred to be local. The sensitivity of IBIS is such that it can detect very faint GRBs, allowing INTEGRAL to probe the population of low-luminosity GRBs with long lags. This population appears to be distinct from that of high-luminosity GRBs and dominates the local GRB population. Pietro Ubertini, Neutron Stars & Gamma Ray Bursts Alexandia, April , Page 28

Thanks! Pietro Ubertini, Neutron Stars & Gamma Ray Bursts Alexandria, Egypt, April , Page 29

Lag-Luminosity Relation Long-lag GRBs when a distance of 250 Mpc is adopted INTEGRAL GRBs with z shown in open circles Other low-luminosity GRBs that do not fit on the relation

IBIS vs BAT sensitivity to weak GRBs EXIST will provide a larger area and better on-axis sensitivity with a SeXI X-Ray follow-up capability

Thanks! Pietro Ubertini, Neutron Stars & Gamma Ray Bursts Alexandria, Egypt, April , Page 32