Different progenitors of short hard gamma-ray bursts
Eleonora Troja, Andrew R. King, Paul T. O'Brien, Nicola Lyons, Giancarlo Cusumano
aa r X i v : . [ a s t r o - ph ] N ov Mon. Not. R. Astron. Soc. , 1– ?? (2007) Printed 24 October 2018 (MN L A TEX style file v2.2)
Different progenitors of short hard gamma–ray bursts
E. Troja , , , A. R. King , P. T. O’Brien , N. Lyons and G. Cusumano Department of Physics and Astronomy, University of Leicester, Leicester, LE1 7RH, UK INAF - Istituto di Astrofisica Spaziale e Fisica Cosmica, Sezione di Palermo, via Ugo la Malfa 153, 90146 Palermo, Italy Dipartimento di Scienze Fisiche ed Astronomiche, Sezione di Astronomia, Universit`a di Palermo, Piazza del Parlamento 1,90134 Palermo, Italy
Accepted – – –. Received – – –.
ABSTRACT
We consider the spatial offsets of short hard gamma–ray bursts (SHBs) fromtheir host galaxies. We show that all SHBs with extended–duration soft emissioncomponents lie very close to their hosts. We suggest that NS-BH binary mergers offera natural explanation for the properties of this extended–duration/low offset group.SHBs with large offsets have no observed extended emission components and are lesslikely to have an optically detected afterglow, properties consistent with NS-NS binarymergers occurring in low density environments.
Key words: gamma rays: bursts; stars: neutron.
In the last few years, the successful
Swift mission(Gehrels et al. 2004) has greatly expanded our knowledgeof gamma-ray burst (GRB) phenomenology. In particular,it has transformed the study of SHBs. The ability to reactrapidly to GRBs triggers led to the first detection of a SHBX-ray afterglow (GRB 050509B; Gehrels et al. 2005), and,a few months later, to the detection of the first SHB opti-cal counterpart (GRB 050709; Fox et al. 2005; Hjorth et al.2005). Accurately pinpointing the afterglow position on thesky can link the SHB to its host galaxy, constraining itsdistance and energetics through the galaxy’s redshift mea-surement. Identifying SHB hosts can also provide a pow-erful insight into the progenitor population and formationhistory. Almost all SHB models invoke close binary systemscontaining at least one neutron star. The mass loss involvedin the supernova forming the neutron star gives the binary asignificant space velocity, depending on its total mass. Thiscan be enhanced if the back reaction (‘kick’) on the neu-tron star is anisotropic. There is ample observational evi-dence (e.g. Wang, Lai, & Han 2006 and references therein)for such anisotropic kicks in both single and binary neutronstars.The analogous inferences for long GRBs (Bloom et al.2002; Fruchter et al. 2006) are well known. Forinstance, only a few important cases show anobserved GRB/supernova (SN) connection (e.g.GRB 060218/SN2006aj; Campana et al. 2006a; Pian et al.2006), but the measured low offsets from the galaxycentres and the preferential location of long GRBs in thebluest regions of these galaxies strengthen the link withmassive stars and their collapse. By contrast, associating SHBs with a host is complicated by the faintness of theirafterglows and their potential origin in NS binaries whichcan travel far from their birth sites before coalescence(Bloom, Sigurdsson, & Pols 1999; Belczynski et al. 2006;Wang et al. 2006; Lee & Ramirez-Ruiz 2007). Finding theabsorption redshifts of SHB afterglows would strengthenthe association with their hosts.Since its launch, in November 2004,
Swift has detected25 GRBs classified as SHBs up to August 2007. In a signif-icant fraction of them ( ∼ γ -rayepisode is followed by a second spectrally softer emissioncomponent, lasting tens of seconds. Despite their long dura-tion, exceeding the canonical cut of 2 s (Kouveliotou et al.1993), these bursts display all the distinctive features of theSHBs class: a first short-hard event with zero spectral lag(Norris & Bonnell 2006); a heterogeneous population of hostgalaxies, in stark contrast to the hosts of long GRBs whichare all late type (Covino et al. 2006; Prochaska et al. 2006);very tight limits on the presence of any accompanying SN,at odds with the standard core-collapse origin of long GRBs(Woosley 1993).In 18 cases out of 25 ( ∼ ∼ HETE-2 (GRB 050709, GRB 060121; Villasenor et al.2005; Donaghy et al. 2006) and
INTEGRAL (GRB 070707;Gotz et al. 2007) satellites. A total of 21 SHBs have arcsec-ond or sub-arcsecond localizations, allowing us to infer theirhosts and estimate their redshifts with some security.In this Letter we report on the full sample of well-localized SHBs and their possible progenitors, focussing c (cid:13) Eleonora Troja et al.
Table 1.
SHBs sample properties. Putative host Angular ProjectedGRB T z R P chance Afterglow offset Error offset Error Refs.(s) (mag) (arcsec) (arcsec) (kpc) (kpc)(1) (2) (3) (4) (5) (6) (7) (8) (9) (10) (11)050509B . . . 0.03 [0.01] 0.225 16.8 5.0 × − X – – 17.87 3.40 64 12 1–3050709. . . . . 130 [7] a × − X,O – 1.30 0.10 3.57 0.27 4–6050724. . . . . 152 [9] 0.258 19.8 1.0 × − X,O,R 0.64 0.02 2.57 0.08 7–9051210. . . . . 1.4 [0.2] > × − X – – 2.80 2.90 <
50 – 3, 10, 11051221A . . . 1.27 [0.05] 0.546 22.0 2.4 × − X,O,R 0.12 0.04 0.76 0.25 12, 13051227. . . . . 110 [10] – 25.6 2.0 × − X,O – 0.05 0.02 < b > × − X,O – 0.32 0.10 < < × − X,O – 0.40 0.56 < c < × − X – – 16.33 3.70 70 16 3, 19060505. . . . . 4 [1] 0.089 17.9 1.0 × − X,O – 4.53 0.32 7.45 0.53 20, 21060614. . . . . 103 [5] 0.125 22.5 6.0 × − X,O – 0.50 – 1.10 – 22–24060801. . . . . 0.5 [0.1] 1.131 23.0 4.1 × − X – – 2.39 2.40 19.7 19.8 3, 11061006. . . . . 130 [10] 0.438 23.7 1.8 × − X,O – 0.32 0.50 1.8 2.8 11, 26061201. . . . . 0.8 [0.1] 0.111 19.0 3.8 × − X,O – 17.00 0.20 33.9 0.4 27061210. . . . . 85 [5] 0.410 21.1 4.7 × − X – – 1.99 1.80 10.7 9.7 3, 11, 28061217. . . . . 0.30 [0.05] 0.827 23.4 3.9 × − X – – 7.41 3.80 55 28 3, 11, 29070724A . . . 0.40 [0.04] 0.457 ∼ d ∼ × − X – – 0.72 2.10 4 12 3, 30
Notes :
Col. (1): GRB name; Col. (2): T duration and its error in the 15-350 keV energy band; Col. (3): Redshift of the putativehost galaxy; Col. (4): Observed R magnitude of the putative host galaxy; Col. (5): Probability that the association is a chance ofcoincidence; Col. (6): Detection of the GRB counterpart in different energy band (X - X-ray; O - optical; R - radio); Col. (7)-(8):Angular offset between the afterglow position and the associated galaxy centroid, and its error, respectively; Col. (9) and (10):Projected physical offset and its error, respectively; Col. (11): Reference to publications of the presented data. Refs.: (1) Gehrels et al. 2005; (2) Bloom et al. 2006; (3) Butler 2007 (4) Hjorth et al. 2005; (5) Fox et al. 2005;(6) Villasenor et al. 2005; (7) Campana et al. 2006b; (8) Berger et al. 2005; (9) Prochaska et al. 2006; (10) La Parola et al. 2006;(11) Berger et al. 2007; (12) Burrows et al. 2006; (13) Soderberg et al. 2006; (14) Sakamoto et al. 2007; (15) Donaghy et al. 2006;(16) de Ugarte Postigo et al. 2006; (17) Levan et al. 2006; (18) Roming et al. 2006; (19) Bloom et al. 2007; (20) Ofek et al. 2007;(21) Levesque & Kewley 2007; (22) Gal-Yam et al. 2006; (23) Gehrels et al. 2006; (24) Mangano et al. 2007; (25) Sato et al. 2006;(26) Malesani et al. 2006; (27) Marshall et al. 2006; (28) Cannizzo et al. 2006; (29) Ziaeepour et al. 2006; (30) Ziaeepour et al.2007. a Hete-2 trigger. The duration is given in the 2-25 keV energy band. b Hete-2 trigger. The duration is given in the 30-400 keV energy band. Donaghy et al. (2006) detected a faint and long-lasting softbump of emission at a significance level of ∼ σ . c A faint ( R =26 mag) object (S2 in Bloom et al. 2007) has been proposed as the high redshift host galaxy. The measured angularoffset is 4.2 ± chance ∼ ±
30 kpc at z ∼ d We assume R − I ∼ on their spatial distribution with respect to their putativehosts. We also estimate the prompt γ -ray and X-ray after-glow energetics of the available sample. The paper is or-ganized as follows: in § §
3. We discuss our findingsand their implication for SHBs progenitors in §
4. A sum-mary of our conclusions is given in §
5. Throughout thepaper we have adopted a standard cosmology with Hubbleconstant H =71 km s − Mpc − and parameters Ω Λ = 0 . M = 0 .
27 (Spergel et al. 2007).
We include in our analysis GRBs whose prompt emissionfollows the original classification (T < >> . ′′ radius) SHBs, we excluded six other burstssince their hosts and distance scales are not constrained(GRB 050813, GRB 070429B, GRB 070707, GRB 070714B,GRB 070729, GRB 070809).In addition two bursts, GRB 060505 and GRB 060614(Fynbo et al. 2006; Gehrels et al. 2006), which display sev-eral features of the SHBs class, were considered and com-pared to the sample.Table 1 lists the properties of our sample of bursts andtheir putative hosts. In each case we give the probability,P chance , that the proposed association is a chance coinci-dence (col. 5). If no value is given in the literature, we simplyestimated it as the probability that a galaxy of magnitude R Left panel: projected physical offsets as a function of the burst duration (T ) in the γ -ray band. The vertical dashed linemarks the canonical temporal division between long and short hard bursts. The horizontal dot-dashed line reports the median offset fora sample of long GRBs with known redshift (from Bloom et al. 2002). Right panel: Offsets histogram for the same sample of long GRBs. redshift. When the galaxy centroid lies within the error cir-cle position (e.g. GRB 061006), then the GRB cross sectionis determined by the size of the uncertainty region. Oth-erwise, if the galaxy is well outside the position circle (e.g.GRB 061217), it is determined by the angular offset (col. 7).We used the results of Hogg et al. (1997) and Huang et al.(2001) to calculate the galaxy sky-density in the R -band.The derived values listed in col. 5 reflect the difficulty ofidentifying SHB hosts. These result either from poor local-izations or large offsets (e.g. GRB 061217). The chance of aspurious association obviously increases when only an X-rayposition is available, as several galaxies lie within or close tothe X-ray error circle. In those cases, the guiding criterionis usually the object brightness, favouring the associationwith the brightest galaxy. Interestingly, the probability that4 associations out of 17 are spurious is ∼ × − , and indeedthe chance of 4 or more misidentifications is well below the3 σ confidence level.The quoted errors are mainly due to the GRB local-izations, usually pinpointed within a 90% confidence levelerror circle. We caution that the offset is a positive-definedquantity, thus the associated uncertainties do not properlyreflect a probability distribution, especially in cases of neg-ligible offsets (see Bloom et al. 2002). Fig. 1 presents the projected galactocentric offset of SHBsas a function of the burst duration in the γ -ray band (ob-server frame). For comparison, the median offset value forlong bursts ( ∼ ∼ γ -ray and the X-ray energies are calculatedin the 15–150 keV and the 0.3-10 keV bands respectively.To refer our results to the same rest frame energy band wederived a k-correction from the burst spectral parameters(see references in Tab. 1).The γ -ray energy radiated during the short hard spikeand over the total T are reported in the top panel andin the middle panel of Fig. 2, respectively. Bursts with ex-tended emission are on average more energetic than burstswith T < α ∼ β ∼ F ν, t ∝ ν − α t − β ). The normalizations were determined bythe upper limits from Swift /XRT observations. Filled sym-bols indicate those bursts with a detected optical counter-part, empty symbols those lacking an optical detection.Even given the small number of SHBs detected sofar, it is clear that large offset bursts (GRB 050509B, c (cid:13) , 1– ?? Eleonora Troja et al. Figure 2. Prompt and afterglow energetics (source rest frame)as functions of the projected physical offset. Top panel : Isotropicenergy (15-150 keV) released over the initial short hard γ -rayevent only (spike). Middle panel : : Isotropic energy (15-150 keV)over the T duration. Bursts with a long lasting emission areenclosed by the dashed ellipse. Bottom panel : Isotropic energy(0.3-10 keV) released over the temporal range 400 s–500 ks. Filledand empty symbols indicate GRBs with and without a detectedoptical afterglow, respectively. GRB 060502B, GRB 061201 and GRB 061217) lie on thelower part of the bottom panel of Fig. 2, while small off-set bursts instead have on average more energetic X-ray af-terglows and a much higher chance of a detectable opticalafterglow. Malesani et al. (2007) noticed that optical coun-terparts of SHBs with extended emission are more frequentlydetected. Our Figure 2 suggests that this is an enviromentalproperty, since these bursts seem to happen closer to theirhosts, and hence presumably in denser interstellar environ-ments. As shown in Fig. 1, short GRBs with measured offsets ap-pear qualitatively divided into two groups. The group withextended durations all lie very close to their hosts, while thegroup with short duration have a mean offset a factor of 15larger. Though the low statistic does not allow us to firmlyassess that the proposed groups belong to two distinct offsetdistributions, a Kolmogorov-Smirnov test, ran on the cur-rent sample of bursts, excludes at the 2 σ confidence levelthat they are drawn from the same distribution. Further-more, we point out that the two groups are characterizedby very different observational features, which are hard toexplain if they proginate from the same parent population.The two groups (extended duration/small offset, shortduration/large offset) have similar redshift distributions (seeTab 1, col. 2). Accepting the usual arguments that the short duration/large offset group are probably NS–NS mergers,we then have four a priori possibilities for explaining theextended duration/small offset group. These are: a differentclass of NS+NS mergers, NS+massive WD mergers, collap-sars, and NS+BH mergers. We consider these in turn. The obvious possibility here is ultracompact NS–NS binaries, which for suitable binary kick velocities v kick ∼ 100 km s − can produce rather small offsets fromcertain types of host (cf. Belczynski et al. 2006, Fig. 3).The problem here is that the initial (pre–afterglow) NS–NSmerger process should be exactly the same as for NS–NSbinaries starting from wider separations. Yet the smalloffset group have rather distinct features (e.g. a promptextended tail of emission, a higher energetic budget) whichcannot result from environmental effects. This group has the desirable properties (King et al. 2007)of extended duration and no supernovae, but is likely tohave similar merger times and kicks as the standard NS–NSgroup. It therefore cannot explain the small offsets. Collapsars offer a simple explanation of the small offsets,but have other problems. In particular one would havechange the model (e. g. Fryer et al. 2007) to explain boththe very different light curves and the lack of supenovae inthe extended duration/small offset group. Moreover at leastone observed member of this group is hosted by an ellipti-cal galaxy (GRB 050724; Berger et al. 2005; Malesani et al.2007), which is hard to reconcile with a collapsar origin. Low offsets are expected for NS–BH mergers on two quitegeneral grounds. First, there is mounting observational ev-idence that at least some black holes do not receive natalkicks. Mirabel & Rodrigues (2003) show that the 10 M ⊙ BH binary Cyg X–1 has a peculiar velocity of < 10 km s − ,and Dhawan et al. (2007) show that the kick in the BH bi-nary GRS 1915+105 was probably similarly small. Thesemay therefore be examples of direct collapse to a black hole(Fryer & Kalogera 2001). (Direct collapse to a neutron staris not possible, as this has far lower entropy than its progen-itor, unlike a black hole.) Second, the gravitational radiationmerger times t GR for BH–NS and NS–NS binaries of a giveninitial separation scale as ∼ ( M BH /M NS ) − ∼ . 01 for typ-ical masses M BH = 14 M ⊙ , M NS = 1 . ⊙ . Together thesetwo effects show that some BH–NS binaries would move verylittle before merging to produce a short GRB.The advantage of the latter explanation of the low off-sets is of course that it offers natural interpretations of thepeculiar features of the group of SHBs with extended emis-sion. Rosswog (2007) proposed that if a significant fractionof the shredded NS is not immediately accreted, but remainsin bound orbits around the central object, the fallback ac-cretion of the NS remnants can inject power up to late times c (cid:13) , 1– ?? rogenitors of short hard gamma–ray bursts ( . . Swift /XRTfollow-up observations (79 s, 92 s, 68 s, 78 s, and 62 s af-ter the bursts, respectively). We speculate that the expectedlow density of the intergalactic environment may explain thefaint X-ray afterglows, placing these X-ray dark bursts in theupper-left side of Fig. 1. However, other mechanisms, relatedto the microphysics of the shocks and the initial Lorentzfactor, could suppress the early X-ray emission (see Nakar2007). Also, a magnetar origin, as debated for GRB 050906(Levan et al. 2007) and GRB 050925, might explain the lackof detection. The offset distribution of SHBs displays several interest-ing features suggesting two types of progenitor. Most strik-ingly we found that SHBs with extended soft emission(T ∼ 100 s) tend to remain close to their host galaxies. NS–BH mergers naturally account for these properties, althoughother explanations are still possible. SHBs with large offsetshave properties consistent with NS–NS mergers occurring inlow density environments. ACKNOWLEDGMENTS This work is supported at the University of Leicester bythe Science and Technology Facilities Council, and at INAFby funding from ASI on grant number I/R/039/04 and byCOFIN MIUR grant prot. number 2005025417. ET acknowl-edges the support of the Marie Curie Spartan exchange pro-gram at the University of Leicester. ARK acknowledges aRoyal–Society–Wolfson Research Merit Award. NL acknowl-edges support from an STFC studentship. REFERENCES Belczynski, K., Perna, R., Bulik, T., Kalogera, V., Ivanova,N., & Lamb, D. Q. 2006, ApJ, 648, 1110Berger, E., et al. 2007, ApJ, 664, 1000Berger, E., et al. 2005, Nature, 438, 988Bloom, J. S., Kulkarni, S. R., & Djorgovski, S. G. 2002,AJ, 123, 1111Bloom, J. S., et al. 2007, ApJ, 654, 878Bloom, J. S., et al. 2006, ApJ, 638, 354Bloom, J. S., Sigurdsson, S., & Pols, O. R. 1999, MNRAS,305, 763Burrows, D. N., et al. 2006, ApJ, 653, 468 Butler, N. R. 2007, AJ, 133, 1027Campana, S., et al. 2006a, Nature, 442, 1008Campana, S., et al. 2006b, A&A, 454, 113Cannizzo, J. K., et al. 2006, GCNR, 20, 1Covino, S., et al. 2006, A&A, 447, L5de Ugarte Postigo, A., et al. 2006, ApJ, 648, L83Dhawan, V., Mirabel, I. F., Rib´o, M., & Rodrigues, I. 2007,ApJ, 668, 430Donaghy, T. Q., et al. 2006, arXiv: astro-ph/0605570Fox, D. B., et al. 2005, Nature, 437, 845Fruchter, A. S., et al. 2006, Nature, 441, 463Fryer, C. L., Hungerford, A. L., & Young, P. A. 2007, ApJ,662, L55Fryer, C. L. & Kalogera, V. 2001, ApJ, 554, 548Fynbo, J. P. U., et al. 2006, Nature, 444, 1047Gal-Yam, A., et al. 2006, Nature, 444, 1053Gehrels, N., et al. 2004, ApJ, 611, 1005Gehrels, N., et al. 2006, Nature, 444, 1044Gehrels, N., et al. 2005, Nature, 437, 851Gotz, D., Beckmann, V., Mereghetti, S., & Paizis, A. 2007,GRB Coordinate Network, 6608Hjorth, J., et al. 2005, Nature, 437, 859Hogg, D. W., Pahre, M. A., McCarthy, J. K., Cohen, J. G.,Blandford, R., Smail, I., & Soifer, B. T. 1997, MNRAS,288, 404Huang, J.-S., et al. 2001, A&A, 368, 787King, A., Olsson, E., & Davies, M. B. 2007, MNRAS, 374,L34Kouveliotou, C., Meegan, C. A., Fishman, G. J., Bhat,N. P., Briggs, M. S., Koshut, T. M., Paciesas, W. S., &Pendleton, G. N. 1993, ApJ, 413, L101La Parola, V., et al. 2006, A&A, 454, 753Lee, W. H. & Ramirez-Ruiz, E. 2007, New Journal ofPhysics, 9, 17Levan, A. J., et al. 2006, ApJ, 648, L9Levan, A. J., et al. 2007, MNRAS, accepted, arXiv: astro-ph/0705.1705Levesque, E. M. & Kewley, L. J. 2007, ApJ, 667, L121Malesani, D., et al. 2007, A&A, 473, 77Malesani, D., Stella, L., D’Avanzo, P., Covino, S., Jehin,E., Lidman, C., Naef, D., & Schady, P. 2006, GRB Coor-dinates Network, 5718Mangano, V., et al. 2007, A&A, 470, 105Marshall, F., Perri, M., Stratta, G., Barthelmy, S. D.,Krimm, H., Burrows, D. N., Roming, P., & Gehrels, N.2006, GCNR, 18, 1Mirabel, I. F. & Rodrigues, I. 2003, Science, 300, 1119Nakar, E. 2007, Phys. Rep., 442, 166Norris, J. P. & Bonnell, J. T. 2006, ApJ, 643, 266Ofek, E. O., et al. 2007, ApJ, 662, 1129Pian, E., et al. 2006, Nature, 442, 1011Prochaska, J. X., et al. 2006, ApJ, 642, 989Roming, P. W. A., et al. 2006, ApJ, 651, 985Rosswog, S. 2007, MNRAS, 376, L48Sakamoto, T., et al. 2007, ArXiv e-prints, 707Sato, G., et al. 2006, GRB Coordinates Network, 5381Soderberg, A. M., et al. 2006, ApJ, 650, 261Spergel, D. N., et al. 2007, ApJS, 170, 377Villasenor, J. S., et al. 2005, Nature, 437, 855Wang, C., Lai, D., & Han, J. L. 2006, ApJ, 639, 1007Woosley, S. E. 1993, ApJ, 405, 273Ziaeepour, H., et al. 2006, GCNR, 21, 2 c (cid:13) , 1– ?? Eleonora Troja et al. Ziaeepour, H., Barthelmy, S. D., Parsons, A., Page, K. L.,De Pasquale, M., & Schady, P. 2007, GCNR, 74, 2 c (cid:13) , 1–, 1–