Thermal components in the early X-ray afterglows of Gamma...

Thermal components in theearly X-ray afterglows of

Gamma-ray bursts

VLASTA VALAN

Licentiate Thesis in PhysicsStockholm, Sweden 2017

Licentiate Thesis in Physics

Thermal components in the early X-rayafterglows of Gamma-ray bursts

Vlasta Valan

Particle and Astroparticle Physics, Department of Physics,Royal Institute of Technology, SE-106 91 Stockholm, Sweden

Stockholm, Sweden 2017

Cover illustration: Spectrum of GRB 150727A in the time interval 84–96 s fittedwith an absorbed power-law + blackbody model, with residuals shown in the bot-tom panel. In the top panel the total model is shown as a black line, while blueand magenta line are the power-law and blackbody components of the model, re-spectively. Figure produced by Vlasta Valan.

Akademisk avhandling som med tillst̊and av Kungliga Tekniska högskolan i Stock-holm framlägges till offentlig granskning för avläggande av teknologie licentiatexa-men måndagen den 27 oktober 2017 kl 10:15 i sal FD5, AlbaNova Universitetscent-rum, Roslagstullsbacken 21, Stockholm.

Avhandlingen försvaras p̊a engelska.

TRITA-FYS 2017:63ISSN 0280-316XISRN KTH/FYS/--17:63--SE

c⃝ Vlasta Valan, September 2017Printed by Universitetsservice US-AB

Abstract

Gamma-ray bursts (GRBs) are still puzzling scientists even 40 years after theirdiscovery.Questions concerning the nature of the progenitors, the connection withsupernovae and the origin of the high-energy emission are still lacking clear answers.Today, it is known that there are two populations of GRBs: short and long. It isalso known that long GRBs are connected to supernovae (SNe). The emissionobserved from GRBs can be divided into two phases: the prompt emission andthe afterglow. This thesis presents spectral analysis of the early X-ray afterglowof GRBs observed by the Swift satellite. For the majority of GRBs the early X-ray afterglows are well described by an absorbed power-law model. However, thereexists a number of cases where this power-law component fails in fully describing theobserved spectra and an additional blackbody component is needed. In the paperat the end of this thesis, a time-resolved spectral analysis of 74 GRBs observed bythe X-ray telescope on board Swift is presented. Each spectrum is fitted with apower-law and a power-law plus blackbody model. The significance of the addedthermal component is then assessed using Monte Carlo simulations. Six new casesof GRBs with thermal components in their spectra are presented, alongside threepreviously reported cases. The results show that a cocoon surrounding the jet isthe most likely explanation for the thermal emission observed in the majority ofGRBs. In addition, the observed narrow span in radii points to these GRBs beingproduced in similar environments.

iii

Sammanfattning

Gammablixtar (GRBs) har förbryllat forskare ända sedan de upptäcktes. Idagvet vi att de kommer fr̊an utanför v̊ar egen galax och är ett resultat av n̊agonav tv̊a processer: kärnkollaps av väldigt massiva stjärnor, eller en kollision mellantv̊a kompakta objekt (vanligtvis tv̊a neutronstjärnor). De frstnämnda är ocks̊aassocierade med en specifik klass av supernovor, typ Ic bl. Dock är inte alla fr̊agorkring dessa objekt är besvarade. N̊agra av de fr̊agor som återst̊ar är fr̊agan denobserverade str̊alningens ursprung i olika v̊aglängder, samt hur det kommer sig attbara vissa närbelägna och l̊anga GRBs observeras tillsammans med en supernova.Man delar typiskt in str̊alningen fr̊an GRBs i tv̊a olika epoker: prompt emissionoch efterglöd.

Efterglöden är vanligtvis ett resultat av att den relativistiska jeten kolliderarmed det cirkumstellra material som omger ursprungsobjektet. Det observeras idet elektromagnetiska spektrat fr̊an röntgenstr̊alning ända till radiov̊agor. Kop-plingen mellan gammablixtar och supernovor upptäcktes genom analys av optiskaljuskurvor av GRBs medan man letade efter signaturer efter en supernova (en s̊adansignatur ter sig vanligtvis som en bula i ljuskurvan när supernovan blir tydligaremedan efterglöden fr̊an gammablixten avtar).

I den här avhandlingen utforskas den tidiga röntgenstr̊alningen i eftrglöden fr̊anGRBs. Eftersom jordens atmosfär absorberar röntgenstr̊alning fr̊an rymden måstesatelliter användas för att observera efterglöden vid dessa v̊aglängder. Röntgen-spektra fr̊an den tidiga efterglöden beskrivs ofta väl av en enkel, absorberad potens-funktion. Dock existerar det ett f̊atal GRBs för vilka man behöver ytterligare enspektralkomponent, termiska komponenten, för att adekvat beskriva spektret. Detfinns ett flertal olika föreslagna förklaringar till uppkomsten av den termiska kom-ponenten, inklusive att den uppst̊ar genom shock-utbrott i supernovan, fr̊an envarm kokong runt jeten, eller fr̊an jeten sjlv.

Den bifogade artikeln i slutet av avhandlingen presenterar en analys av tidigaröntgenspektra fr̊an teleskopet XRT ombord p̊a satelliten Swift. Jag undersöker 74l̊anga GRBs med kända rödskift och som detekterats under tiden 2011 till 2015.Jag genomför en tidsupplöst analys och studerar de fysikaliska parametrarna fr̊anden termiska komponenten som en funktion av tid. Signifikansen hos den termiskakomponenten evalueras med hjälp av Monte Carlo-simuleringar, användandes 10000 iterationer. I analysen finner vi 6 nya GRBs med en termisk signatur, samt

v

vi

bekräftar existensen av 3 (av totalt 15) tidigare rapporterade fall av termiska kom-ponenter. Vi tolkar resultaten som att den termiska komponenten uppkommer d̊aen varm kokong bryter sig ut fr̊an en tät vind fr̊an stjärnan. Att alla termiska kom-ponenter befinner sig inom ett smalt radiellt intervall fr̊an deras ursprungsstjärnortolkar vi som att alla gammablixtarna har liknande ursprung.

Acknowledgements

I would like to thank my supervisor Josefin Larsson for her time and patience inanswering all my questions. Also, I would like to thank my office mates Husne,Zeynep, Rakhee and Dennis for making the office atmosphere pleasant and warm.I am thankful to Björn for his insights in amazing world of Python, help withlearning Swedish and interesting discussion on any topic one can think of. To Thediand Ramona for all the laughs, the best langos I ever ate and their never-endingoptimism.

I am thankful to my mother for being constant wind in the back. And to all ofmy friends who listened to me speaking about gamma-ray bursts and kept askingquestions no matter how abstract the topic might be to them. Finally, a big thankyou goes to my husband for motivating me and getting the best out of me, evenwhen I thought there is no more energy left in me. You are the best support I couldever have and I am deeply thankful for it.

vii

Contents

Abstract iii

Sammanfattning v

Acknowledgements vii

Contents ix

Outline xi

1 Gamma-ray bursts 11.1 Introduction . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 11.2 Different populations of GRBs and their progenitors . . . . . . . . 21.3 Connection between long GRBs and SNe . . . . . . . . . . . . . . . 4

1.3.1 Ultra-long GRBs and SNe . . . . . . . . . . . . . . . . . . . 51.4 Relativistic jets . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 6

2 Emission phases in GRBs 72.1 Prompt emission . . . . . . . . . . . . . . . . . . . . . . . . . . . . 72.2 Afterglow emission . . . . . . . . . . . . . . . . . . . . . . . . . . . 10

2.2.1 X-ray emission . . . . . . . . . . . . . . . . . . . . . . . . . 11

3 Thermal component in the early X-ray emission 153.1 Possible origins of the thermal component . . . . . . . . . . . . . . 15

3.1.1 Shock breakout . . . . . . . . . . . . . . . . . . . . . . . . . 153.1.2 Cocoon emission . . . . . . . . . . . . . . . . . . . . . . . . 163.1.3 Prompt emission . . . . . . . . . . . . . . . . . . . . . . . . 18

3.2 Thermal components in the context of this thesis . . . . . . . . . . 19

4 Observations and data analysis 214.1 Swift satellite: detectors and data reduction . . . . . . . . . . . . . 21

4.1.1 BAT . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 214.1.2 XRT . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 22

4.2 Data analysis . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 25

ix

x Contents

4.2.1 Bayesian blocks binning . . . . . . . . . . . . . . . . . . . . 26

Summary of the attached papers 29

Author’s contribution to the attached paper 31

List of figures 33

Bibliography 35

Outline

The introductory chapters of this thesis give a general overview of Gamma-raybursts (GRBs), their emission phases and important spectral features that areused in the scientific work presented. It starts with Chapter 1 that is devoted tointroduction about GRBs, their progenitors and the connection between long GRBsand supernovae (SNe). In Chapter 2 I describe the two emission phases of GRBs:prompt and afterglow. The focus is on the afterglow emission and especially theX-ray emission from GRBs. Chapter 3 is devoted to different theories explainingthe origin of the thermal components observed in the early X-ray afterglow of someGRBs. Chapter 4 contains basic information about the instruments used, datareduction and the spectral analysis. The main goal of the thesis is to give betterconstraints on the origin of thermal component. The majority of the work is thenpresented in the form of an article at the end of the thesis, which has been submittedto Monthly Notices of Royal Astronomical Society.

xi

Chapter 1

Gamma-ray bursts

1.1 Introduction

Gamma-ray bursts (GRBs) are the most energetic phenomena discovered so far inthe Universe. They are observed as relatively short pulses of gamma-rays. Theisotropic energies released in this process are of the order 1051–1054 erg in a coupleof seconds up to a couple of minutes. These extreme energies (and in turn extremeluminosities) make them observable to very high distances. In fact, the most distantobserved GRB is GRB 090423, detected at redshift z=8.2 corresponding to ∼ 3 ×1011 ly [1].

These objects were serendipitously discovered in 1967 when the military Velasatellites recorded a peak of gamma-rays [2]. Since the Nuclear treaty was in place atthe time, the data were kept secret for six years before they were published in 1973.The authors excluded Earth and the Sun as sources for these gamma-rays [2]. Bythe 1990s there were more than 100 different theories explaining what these objectsmight be and where they may originate from. With the launch of the ComptonGamma-Ray Observatory (CGRO) 1 in 1991, which carried the instrument Burstand Transient Source Experiment (BATSE) on board, it became clear that theseobjects are isotropically distributed across the sky [3]. Soon thereafter, in 1996,the Beppo-SAX satellite [4] was launched, which discovered that GRBs have long-lasting afterglows at longer wavelengths. It was as a result of these afterglows thatthe redshift of the GRBs could be determined, which confirmed their extragalacticorigin. [5].

1https://heasarc.gsfc.nasa.gov/docs/cgro/cgro.html

1

2 Chapter 1. Gamma-ray bursts

1.2 Different populations of GRBs and theirprogenitors

As CGRO was launched, the number of GRBs observed increased at a rapid pace.This enabled for statistical studies on GRBs. The simplest task was to plot theduration of bursts detected. When plotted on a logarithmic scale the plot showeda bimodal distribution with a separation between the two peaks at around 2 s[6]. Since then GRBs are divided into two different categories: short bursts withduration less than 2 s and long bursts with duration longer than 2 s. In order toquantify the duration, a parameter called T90, defined as the time in which 90% ofthe burst radiation is observed, is used. Today, short and long GRBs are not justseparated by their duration, but are also thought to arise from different progenitors.

Short GRBs are thought to originate in mergers of two compact objects, likea neutron star–neutron star or a black hole–neutron star [7]. Observations of theafterglows of these bursts placed constraints on their host galaxies. An associationbetween short GRBs and elliptical galaxies was observed, which points to the hostgalaxies of short GRBs being dominated by older stellar population [8]. However,this association is not exclusive and short GRBs are also found in variety of othertypes of galaxies, as well outside their host galaxies. This is consistent with oldprogenitors being ”kicked” out due to the supernova (SN) [9]. However, these areonly indirect observations of short GRB progenitors and we are still lacking a directdetection in order to confirm these theories. The closest to a direct detection wasa kilonova associated with the GRB 130603B [10]. A kilonova (sometimes referredto as macronova) is a day to week-long thermal, SN-like transient powered byradioactive decay of neutron-rich elements [11]. They are predicted to accompanyneutron star–neutron star or black hole–neutron star mergers [12].

Long GRBs are thought to originate in the core-collapse of massive stars. Aconfirmation of this theory came in 1998 with the detection of the GRB 980425together with SN 1998bw [13]. This discovery, together with the discovery of thelink between long GRBs and star-forming regions in galaxies [14], enforced theargument that long GRBs are actually produced in the core-collapse of massivestars. An image illustrating the diversity of host galaxies of long GRBs, as wellas their locations within the galaxies, is shown in Fig. 1.1. The most popularmodel for explaining the production of long GRBs is the collapsar model [15, 16].In this model the progenitor of a GRB should be stripped of H and He and have arapidly rotating stellar core, which points to the progenitors being rotating Wolf–Rayet (WR) stars. However, a problem with this picture is that WR stars exhibitstrong stellar winds, of the order 10−5 − 10−4 M⊙ yr−1 [17], which strips away theangular momentum during the star’s life. The only possibility for this effect notto be severe is that these WR stars are located at low metallicities [18]. On theother hand, the observations of GRB host galaxies are not linking them exclusivelyto low metallicity environments [19] and there is a need for a possible explanationof how these objects are made at high metallicities. Another alternative is thatthese rapidly rotating WR stars are produced in massive close binaries. This idea

1.2. Different populations of GRBs and their progenitors 3

was investigated by Cantiello et al [20] where they tested if mass accretion andquasi-chemically homogeneous evolution can produce a rapidly rotating WR star.A consequence of this model is that a large fraction of GRBs can occur in runawaystars. A runaway star is a star that is moving with extremely high velocity relativeto its surrounding interstellar medium (ISM). The leading theory is that these starsget their high velocities when expelled from a binary system or a cluster due to e.g.a SN explosion [21]. This was further explored by Eldridge et al [22], but up todate observational confirmations of the idea are missing.

Figure 1.1: Host galaxies of long GRBs imaged by the Hubble Space Telescope. Thegreen crosses mark the position of a GRB. The image was produced by NASA, ESA,GOSH Collaboration and STScI and is publicly available for download at the Hub-ble website: http://hubblesite.org/image/1915/news/107-illustrations.

In the collapsar model, regardless of the progenitor, we expect the formation of ablack hole with a short-lived accretion disk. The main source of energy of a GRB isaccretion onto a compact object and this process is one of the most energy-efficientprocesses in the Universe. The assumed accretion onto a compact object is alsothe reason why GRBs are not expected to be produced in the merger of two black


holes. However, the discovery of gravitational waves [23] and Fermi2 observationsof a tentative gamma-ray counterpart to GW150914 [26] may challenge this picture.Other GW observations to date are not accompanied by gamma-ray signals and aclear association of a GW source with a significant gamma-ray signal is needed toconfirm this picture. An alternative theory proposed by Perna et al [27] is basedon a binary system of two massive, low-metallicity stars, in which evolution willlead to the formation of two massive black holes. One of the black-holes will have afallback disk at large radii and this disk will become neutral as it cools and will beable to survive for a long time. As the merger of these black holes begins, accretionat the fallback disk of one of the black-holes will resume, resulting in a short GRBat the end.

An alternative to the collapsar model is a scenario in which a magnetar iscreated. A magnetar is a proto-neutron star with a rotational period of ∼ 1 msand a strong magnetic field of the order 1015 G [28]. This model has been proposedas an explanation for an unusually luminous SN associated with a GRB [29]. Thetotal energy available in the magnetar model is ∼ 1052 erg [30]. However, weare still lacking observations that may confirm or dispute this theory for the totalpopulation of GRBs. Possible observational support for the magnetar model is aplateau feature in the X-ray light curves of GRBs (see Sec. 2.2.1). This featurehas been explained as a consequence of residual magnetic energy. This residualenergy can then inject energy into the outflow at later times [31]. Other possibleexplanations for the plateau in the X-ray light curves are discussed in Sec. 2.2.1.

Recently a new class of GRBs has emerged, the so-called ’ultra-long’ GRBs. Thisclass was suggested by observations of the ’Christmas burst’ GRB 101225A, with agamma-ray emission lasting more than 7000 s [32]. The events like GRB 101225Aare rare and it is still debated whether these bursts are a distinct population ofGRBs with different progenitors than long GRBs [33, 34] or just an extension ofthe same population that makes up the high-end tail of T90 [35, 36]. It was discussedby many authors (see e.g. [36]) that determining the burst’s duration only basedon the value of T90 is ambiguous since this duration varies with energy range. Forthis reason, Levan et al [34] suggested that classification of a GRB as an ultra-longGRB needs to be made based on multi-wavelength criteria. These criteria includea very long duration prompt emission light curve (more than 10000 s in gammaand X-ray wavelengths), short-scale variations during an X-ray plateau phase andvery rapid decay rates (the flux should decay as t−α, where α > 3).

In this thesis I will focus only on long GRBs, including ultra-long GRBs. Oneof the most intriguing aspects of long GRBs is their connection to SNe.

1.3 Connection between long GRBs and SNe

Since 1998, approximately dozen new connections between GRBs and SNe havebeen established. It is striking is that all of the SNe connected to GRBs are of

2Fermi is a gamma-ray telescope launched in 2008 carrying on-board two instruments Gamma-ray Burst Monitor and Large Area Telescope[24, 25]

1.3. Connection between long GRBs and SNe 5

Type Ic with broad lines (Ic bl) in their spectra [37]. Type Ic SNe are SNe thatlack H and He features in their spectra, while the broad lines imply large expansionvelocities. As both GRBs and SNe are produced in the core collapse of massivestars, we would expect to observe a SN accompanying each GRB detected in thelocal Universe. However, in 2006 two nearby GRBs without SNe were discovered:GRB 060505 at z = 0.08913 and GRB 060614 at z = 0.12514[38]. Given theirproximity, they were deeply studied in the optical, but no signature of a SN wasdetected. A more recent addition to the population of nearby long GRBs withno SNe detected in deep observations is GRB 111005A at z = 0.013 [39]. DeepSpitzer observations of this GRB, imply an upper limit for an accompanying SNthat is 20 times less luminous than the previously detected SNe. This demonstratedthat at least some GRBs are either associated with very faint SNe or are produceddifferently [40].

All of the GRBs having an accompanying SNe are low-redshift GRBs, withthe highest spectroscopic redshift being z = 0.55185 [41]. In the beginning ofthe research into the SN-GRB connection it was found that GRBs connected toSNe have Eγ,iso ≤ 1049 erg (see e.g. [13, 42]). This is explained by selection effects,meaning that low-luminosity GRBs are not detected at higher redshifts, while moreluminous GRBs have higher detection rates at higher redshifts [43]. Today, thesituation has changed as GRBs that are connected to SNe with Eγ,iso ∼ 1051 erghave been observed [41].

SNe associated with GRBs are detected by analysing the optical afterglow emis-sion. Approximately, 7–10 days after the GRB detection the optical afterglow willstart changing from a featureless spectrum, characteristic of GRBs, to a spectrumthat exhibits more and more SN features. This is explained with the optical after-glow of the GRB fading, while the SN spectrum is slowly emerging as it reachesmaximum light. The downside of this method is that only a subset of well-localizedGRBs also has a detected optical afterglow. At redshifts z ≳ 0.5 it is impossibleto obtain the spectroscopic signatures of the emerging SN and in these cases theconnection between a GRB and a SN is made from a bump in the afterglow lightcurve that appears 10−30 days after the detection of a GRB. There is currently nochance to detect the SN, by any means, at redshift z > 1. The review by Cano etal. [41] summarizes all important observational aspects of the GRB-SN connectionand the reader is referred to it for further information.

1.3.1 Ultra-long GRBs and SNe

The spectroscopic connection between the ultra-long GRB 111209A and the lumi-nous SN 2011kl [29] posed the question of how these types of events are connected.The SN itself is an unusual SN, being brighter than the regular SN Ic bl connected totypical GRBs, but not as bright as the so-called super-luminous SNe (SLSNe). Thisis so far the only spectroscopically confirmed case of an ultra-long GRB connectedto a SN. All other ultra-long GBRs were either too far away for a SN signature to bedetected (e.g. GRB 121027A at z = 1.774 [34]) or, as is the case for GRB 130925A,the host galaxy contained high amounts of dust, such that the optical extinction


would completely block any possible signature of a SN [44]. The Christmas burstGRB 101225A at z = 0.847 [32] showed a flattening in its optical and near-infraredlight curves (as in GRB 111209A), which has been interpreted as an indicationof additional emission from a rising SN, but the spectroscopic confirmation for aconnection in this case is lacking. The conclusive evidence of a connection betweenultra-long GRBs and SLSNe (or very luminous SNe) is needed in the form of alarger sample of nearby ultra-long GRBs that can be followed up spectroscopicallyin search for SN signatures.

1.4 Relativistic jets

A relativistic jet is defined as a beam of ionized matter traveling close to the speedof light. Jets are known to form in accreting black hole systems. In regions overthe poles of the black hole the in-falling matter can get accelerated to very highvelocities, thus escaping the vicinity of the black-hole in the form of a jet [45].

Even though active galactic nuclei and micro quasars were known to host rela-tivistic jets, the initial models of GRBs were produced in spherical symmetry andused spherical outflows, due to simplicity. In 1997, Rhoads [46] argued that GRBsshould host relativistic jets. The main argument was that if the GRB emission isactually beamed into narrow jets, then the total energy would decrease by a feworders of magnitude compared to the isotropic case [47]. Another argument for theexistence of relativistic jets was the fact that if GRBs were non-relativistic, thenwe would be unable to observe high-energy photons due to the large optical depth,the so called ”compactness problem” [48]. In order for the radiation to escape, arelativistic outflow with high Lorentz factor (Γ > 100) was inferred, implying thatthe outflow is relativistic and beamed [49]. A more direct observational evidencefor the existence of jets in GRBs came with the detection of afterglows and theachromatic breaks (see Sec. 2.2.1) observed in their light curves [50]. From thesebreaks it was possible to infer indirectly the jet opening angles. It was found thatthe opening angles vary between 2◦− 20◦ with a peak around 4◦ [51]. The clear jetbreaks are seen in a minority of GRBs [52].

The most popular theoretical framework for jets in GRBs is the ’fireball’ model[53, 54]. The main idea of the model is that a fireball containing gamma-rays,electron-positron pairs, a small amount of baryons and a large amount of energyis released in a short time-scale. This optically thick, hot plasma overcomes thegravitational pressure (due to radiation pressure) and forms a relativistic jet whichis responsible for the observed gamma-rays [55]. The bulk Lorentz factor of thejet will initially increase as Γ(r) ∼ r/r0, where r0 ∼ 107 cm is the inner radiusof the black-holes accretion disk, and will eventually become constant at a valueΓ ∼ 102 − 103, at the so-called saturation radius [56]. The jet will be thermallydominated in the beginning, but as the baryons carry the bulk of the initial energyat the saturation radius, the jet later becomes kinetic energy dominated [57].

Chapter 2

Emission phases in GRBs

GRB emission is divided into two parts: the prompt emission and the afterglow.The prompt emission is short-lasting, from milliseconds to a couple of minutes andis mainly observed in gamma-rays. On the other hand, the afterglow is longerlasting (it can last up to years) and it is observed from X-rays all the way toradio frequencies. Here I will briefly sketch the physical processes behind these twoemission phases and what we can learn from each of them.

2.1 Prompt emission

The prompt emission is primarily observed in gamma-rays. It can also be detectedas X-ray emission that occurs simultaneously with the gamma-rays. The promptemission is thought to originate within the relativistic jet (for a detailed review ofGRB prompt emission see [58]). Light curves of GRB prompt emission are irregular,complex and diverse as shown in Fig. 2.1. This makes these objects complex toanalyse as no two GRBs are alike and a wide range of spectral properties areobserved.

When fitting the prompt emission spectra of GRBs an empirical function, whichwas proven to often fit the spectra well, is commonly used. This empirical model isa type of smoothly broken power-law, called ”the Band function” [59] and is definedas:

N(E) = A

( E

100keV

)αexp

(− E

Eb

), E < (α− β)Eb[ (α− β)Eb

100keV

]α−βexp(β − α)

( E100keV

)β, E ≥ (α− β)Eb.

(2.1)

This model has four free parameters: the low energy slope α, the high energyslope β, the break energy E0 and a normalization. A schematic figure of the Bandfunction is shown in Fig. 2.2. In the figure the observed peak energy is shown

instead of Eb. These quantities are related as Epk =(2−α)Eb

α−β . Note that this is an

7

8 Chapter 2. Emission phases in GRBs

Figure 2.1: Prompt emission light curves of four different GRBs observed by SwiftBAT. The top two panels represent GRB 111123A and GRB 131030A, while thebottom two panels represent GRB 150727A and GRB 151027A. These four GRBsare analysed in the paper included in this thesis. Note that all curves are uniqueand erratic. The light curves are binned with signal to noise ration equal 5.

empirical model and that conclusions about the physical processes governing theprompt emission are hard to deduce from it. When the Band model is fitted to agreat number of GRBs the resulting distributions of key parameters are very wide.The distributions of the best-fitting parameters using the Band model as well asthe smoothly broken power law (SBPL) 1, power law (PL) and a cut-off power lawor ’the Comptonized’ model (COMP) to GRBs detected by the Fermi satellite isshown in the Fig. 2.3 [60]. The spread in α (also used to describe the slope of thePL and COMP models, as well as the low-energy slope of the SBPL model) is from−2 ≤ α ≤ 0, with a peak around −0.5 (upper left panel in Fig. 2.3). The spread inβ is from −3 ≤ β ≤ −1.8, with a peak around −2 (upper right panel in Fig. 2.3).The bottom panel in Fig. 2.3 show distribution of the break energy Eb, which hasa spread 60 ≤ Eb ≤ 103 keV, with a clear peak at Eb ≈ 130 keV.

As mentioned in Sec. 1, the prompt gamma-ray emission from GRBs is thoughtto originate in the relativistic jet. The thermal emission from a fireball is releasedat the photosphere (the surface of the last scattering) and if the radiation has

1This model differs from the Band model as it has five parameters rather than four. Anotherdifference is the way the two power laws are connected: in the Band model the power laws aresmoothly joined by an exponential roll-over, while in the SBPL the two power laws are joined bya hyperbolic tangent function.

2.1. Prompt emission 9

Figure 2.2: Schematic view of the Band function, with the parameters of the modelindicated.

not been altered the observed spectrum would be a Plank function, which we donot observe in the vast majority of GRBS (however, see [61, 62, 63]). The mostpopular model for producing the observed non-thermal spectrum is the internalshocks model [64]. If we assume that the outflow consists of shells of plasma,the collisions between them will produce shocks and a fraction of the bulk kineticenergy will be transformed into random particle energy and magnetic fields [55].This energy is then radiated via synchrotron emission as non-thermal gamma-rays.

This model can explain the observed peak of the gamma-ray spectra of GRBsand the high-energy part of the Band spectrum above the E0, but it often failsin explaining the observed gamma-ray spectrum below E0 [65]. In particular, itwas shown that the values of α cannot exceed −2/3 as α > −2/3 is violationthe so-called synchrotron line of death [65]. This has led to a number of othertheories for explaining the prompt emission in GRBs. Today, there is a variety ofphysical models used to fit the GRB prompt emission spectra e.g. subphotosphericdissipation [66], external shocks [67] and magnetic reconnection [68].


Figure 2.3: Distribution of best-fitting parameters for the Band model, as well asthe smoothly broken power law (SBPL), power law (PL) and cut-off power law(COMP) from the Fermi catalogue. The figure is from Fig. 3 in [60], reproducedwith permission c⃝ ESO.

2.2 Afterglow emission

The afterglow of GRBs arises when the jet collides with the circum stellar material(CSM) surrounding the progenitor star and sweeps it up into the ISM. The emissionarising from this process is synchrotron emission, which is produced when electronsare accelerated in a magnetic field. The observed afterglow is the consequence of theexpanding shock wave sweeping up more and more material and thus losing energy[69]. When the jet collides with the CSM, forward and reverse shock are formed.The forward shock will have its peak emission in the X-rays, while the reverse shockis traveling slower and will peak at optical/IR wavelengths [70]. The afterglows ofGRBs were first detected by Beppo-SAX in the X-rays and the accurate positionsprovided by Beppo-SAX enabled the discovery of the first optical afterglow [71].Approximately 90% of well-localized GRBs have observed X-ray afterglows, whileoptical and radio afterglows are available for only 50% of the well-localized bursts

2.2. Afterglow emission 11

[48].Even though the afterglow is referred to as one entity, its time evolution and

spectra are different in the different wavelengths in which it is observed. This thesisand the research presented in it, focuses on the X-ray emission of GRBs.

2.2.1 X-ray emission

The origin of the X-ray emission from GRBs is still poorly understood. It is unclearwhether the early part of this X-ray emission represents the true afterglow or thelate-time prompt emission [72]. Examples of light curves of GRBs detected by XRTon board the Swift telescope (see Sec. 4.1.2) are presented in Fig. 2.4.

Figure 2.4: Early X-ray light curves of four different GRBs observed by XRT. Thetop two panels represent GRB 111123A and GRB 121211A, which have erratic lightcurves. The bottom two panels represent GRB 150727A and GRB 151027A, whichhave more smooth light curves.

It was shown that there is a linear correlation between the prompt and afterglowfluxes for bursts detected by Swift [73]. Later, by analysing a sample of GRBsdetected by XRT, four commonly observed phases of X-ray light curves emerged[74]. The emission starts with a steep decay, followed by a plateau, a so-callednormal decay and a post-jet break phase, respectively, as is schematically shown inFig. 2.5.

The steep decay phase was originally thought to corresponds to high-latitudeemission from the jet, which is a part of the prompt emission. This theory issupported by the fact that this emission joins smoothly to the tail of the prompt


Figure 2.5: Different phases of the X-ray light curves of GRBs and the typicalduration of each phase. Phase I corresponds to the initial steep decay, phase IIto the shallow decay (’plateau’), phase III to the normal decay, phase IV to thepost-jet break and phase V marks flares that are observed on top of the X-ray lightcurve. This figure schematically shows the light curve phases identified by Zhanget al (Fig. 1 in [74]).

emission [75]. Later, it was shown by O’Brien et al [72] that pure high-latitudeemission was inconsistent with the observations and it was instead suggested thatthis phase is most likely a combination of the late-time prompt and the afterglowemission. This phase is followed after 100–1000 s by a shallow decay phase - theplateau. Its origin is debated, but the most promising theory is late-time activityof the central engine [74, 76]. This theory is supported by the observations of flaresover-imposed on the early X-ray light curves, which are also thought to representlate central engine activity [77]. The transition from steep-to-shallow indicates thatthe fireball (see Sec. 1.4) has already decelerated when the tail of the GRB promptemission fades. Otherwise, an initially rising light curve peaking at the fireballdeceleration time would be observed. The third phase of the X-ray light curveis the so-called normal decay phase, where the flux decays as t−1 as expected inthe standard forward shock model [78]. A jet break is a predicted observationalsignature of the jet emission. It is achromatic and it can be observed in the afterglow

2.2. Afterglow emission 13

light curve across all frequencies of the afterglow. This break arises when theedge of the jet becomes visible as the relativistic beaming decreases when the jetdecelerates [52]. However, this jet break is not observed in a large fraction of GRBafterglows. Its absence has been attributed to over-simplified theoretical models,which are assuming homogeneous jets with sharp edges and not accounting for thecomplex evolution of the afterglow synchrotron emission spectra. Recent numericalsimulations show that the jet can maintain its structure much longer than previouslythought and thus delay the break [79, 80]. It is important to stress that not allphases of the X-ray emission are observed for all GRBs and all four componentsare visible for less than half of the observed GRBs [78].

The X-ray spectra of GRBs are usually very well described by a power law.However, there exists a number of cases where a thermal component is needed inorder to fully explain these spectra. This is presented in great detail in Chapter 3.

Chapter 3

Thermal component in theearly X-ray emission

The spectra of early X-ray afterglow are usually well described by a simple absorbedpower law [48]. However, in 2006, observations of a low-luminosity, nearby GRB,GRB 060218, showed that this power-law model is not fully explaining the spectrumof certain GRBs [42]. In this case, it was established that an additional componentin form of a thermal component is needed to explain the spectrum. Since then,thermal component is reported to be present in dozen new cases ([81], [32], [82],[83], [84], [85], [86]). Even though the presence of this component in certain burstsis confirmed, the origin of it is still greatly debated. Theories suggest that thiscomponent originates from the SN shock breakout, the cocoon surrounding the jetor the jet itself.

3.1 Possible origins of the thermal component

3.1.1 Shock breakout

Here I will give a basic introduction to the theory of supernova shock breakout andits possible explanation of the thermal component observed in the X-ray spectra ofsome GRBs. For a detailed review of shock breakout theory the reader is referredto the recent review by Waxman and Katz [87].

As a massive star reaches the end of its life it implodes and then explodes asa SN. In the first stage of the explosion a SN shock wave is formed. The shockwave propagates through the star and will eventually emerge from the star throughthe outer envelope where the optical depth is low. The shock breakout happenswhen the optical depth drops below ≈ v/c, where v is the shock velocity, which isexpected to happen near the surface of the star. The first light we can observe froma SN is actually coming from the shock breaking out. The breakout is expected to

15

16 Chapter 3. Thermal component in the early X-ray emission

produce a flash of X-ray/UV radiation with a blackbody spectrum on a short time-scale (several seconds up to a fraction of an hour) and is followed by UV/opticalemission coming from the expanding cooling envelope on a longer time-scale (orderof a day). If the SN progenitor goes through episodes of enhanced mass-loss fromits surface, prior to explosion, then the shock breakout will not take place at thesurface of the star but rather at a larger radius in the CMS. In this case, thebreakout time scale may be prolonged from hours to days.

There have been reports of shock-break out in SN Type II [88, 89], as well as inthe Type Ibc SN 2008D. SN 2008D was serendipitously discovered as a bright flashof X-rays during Swift observations of the galaxy NGC2770 [90] and it is the onlydirect observation of the SN shock-breakout in a SN Ibc so far. Several authorsanalysed the spectra obtained from this SN and even though it was first classifiedas a normal SN Ic [90], it was later reclassified as Sn Ib due to presence of Helines in the spectra [91]. The progenitor estimates for this SN are in line with theprogenitor being a WN1 star, with an inferred radius of R∗ ∼ 6 × 1011 cm [92].These types of observations are still very rare, due to the short time-scale for shockbreakout in X-rays. Dedicated X-ray surveys are needed in order to increase thesample of SN shock breakouts observed.

GRBs have been linked to SN Type Ic bl (see Sec. 1.3). Their inferred pro-genitor is a WR star, with a typical radius of 1011 cm [93]. These stars are alsoassumed to have strong mass-loss towards the end of their lives, with observationssupporting this theory [94]. There have been theoretical studies investigating SNshock breakout from thick winds suggesting the expected breakout radius of or-der ∼ 1012 cm [95], which is in agreement with observations of SN 2008D [91].According to this model, the expected energy from the shock breakout of a GRBprogenitor is Ebo ∼ 1044 erg and the spectrum is expected to peak in the X-rays[87].

3.1.2 Cocoon emission

As the jet pierces its way through the progenitor star part of its energy is depositedinto a cocoon surrounding it. The cocoon is the shocked jet material accumulatingaround the jet as it makes its way through the star [96, 97, 98]. As the jet emergesfrom the star, the hot plasma composing the cocoon will swiftly escape the star andaccelerate to mildly relativistic speeds. A thermal signature from this process canbe observed in the X-rays. This signature is similar to the one produced by the SNshock breakout but much more energetic with luminosities Lbo ∼ 1047−1050 erg s−1[99]. The schematics of this process is shown in Fig. 3.1.

Nakar and Piran [100] proposed an analytical model for GRB cocoons wherethe cocoon is composed from the inner shocked jet material and the outer shockedstellar material. Three different observational signatures arise from this structureof the cocoon. Cooling shocked stellar material will produce isotropic emission,cooling shocked jet material will produce beamed emission and the interaction of

1A Wolf-Rayet star with dominant lines of ionized N in their spectra

3.1. Possible origins of the thermal component 17

Figure 3.1: Schematic illustration of the cocoon that surrounds the jet. The jet isshown emerging from the progenitor (shown in black), shocked jet material thatrepresents cocoon is shown in yellow.

the relativistic cocoon component with the surrounding medium will give rise toafterglow like emission. They also discuss the detectability of these componentsand conclude that in gamma-rays a typical luminosity ≥ 1051 erg s−1 is expected,a signal that would be easily detectable by Swift. The emission is expected to bequasi-thermal and to peak at ∼ 100 keV. In X-rays a detectable signal is expectedonly if a large fraction of the shocked jet cocoon is deposited in a material withLorentz factor ≥ 10. The luminosity is expected to have its peak in soft X-rays(0.3–10 keV) and to reach up to ∼ 1044 erg s−1. It should be noted that thismodel is highly dependent on the amount of mixing between the shocked jet andthe shocked stellar material.

Recently, numerical simulations that produce light curves and spectra expectedfrom the cocoon were made. Suzuki and Shigeyama [101] performed hydrodynam-ical simulation of a jet that emerges from a massive star surrounded by a CSM.

18 Chapter 3. Thermal component in the early X-ray emission

They considered the jet interaction with the CMS, which the authors assumed to bea steady wind with mass-loss rates ranging from 10−7 M⊙ yr

−1 to 10−3 M⊙ yr−1

and a velocity of 103 km s−1. The authors noted that a denser CSM will reducethe cocoon in size compared to a dilute CSM (see Fig. 2 and Fig. 3 in [101]). Thecocoon from this model can produce thermal X-ray photons with luminosities upto 1045 − 1048 ergs−1. De Colle et al [99] produced light curves by post-processinghydrodynamical simulations. They considered two different progenitors for a GRBand investigated how the stellar material influences the observable signal from thecocoon. The progenitors are two pre-SN stellar models. One of the progenitors isa star with an initial mass of 25 M⊙, which during its life shrank to a final mass of5.45 M⊙ and a final radius of 3×1010 cm. The second progenitor is a star that hadan initial mass of 12 M⊙, a final mass of 9.23 M⊙ and a final radius of 10

11 cm.For both stellar models the authors assumed a stellar wind with a mass-loss of10−5 M⊙ yr

−1 and a velocity of 108 km s−1. In their results De Colle et al [99]showed the influence of the stellar structure as a key parameter influencing the ob-served luminosity from the cocoon (see Fig. 5 in [99]). When the extended stellarmodel is considered (the model with initial mass of 12 M⊙) the luminosity of thecocoon is lower, the cocoon breaks out at later times and the increase in luminosityis slower when compared to the stellar model without an extended envelope. Thespectra are quasi-thermal and peak in the soft X-rays, while the light curves exhibita rapid rise until the peak luminosity of ∼ 1047 erg s−1 and then slowly decay ona time-scale of 100 s [99]. While the results described above give insight into howthe observed signal is affected by the properties of the star and the CMS, it shouldbe noted that the first-principles numerical simulations of cocoons are still lacking.

3.1.3 Prompt emission

As discussed in Chapter 2, section 2.1, a minority of GRBs display (quasi) black-body spectra in the prompt gamma-ray phase. A thermal component is naturallyexpected in the fireball scenario. The main idea is that the jet is initially opaqueand that photospheric emission will inevitably occur [64]. Ryde in 2004 [61] usedBATSE data to identify a number of GRBs with blackbody components in theprompt phase. It was found that the temperature is mildly increasing followed bya decrease as kT ∼ t−1/4 − −2/3 and that there is a strong correlation betweenthe temperature and the flux. In the Fermi era, clear signs of thermal emissionhave been observed in a number of GRBs [61, 62, 63, 102, 103, 104]. In some casesthe thermal component dominates the emission [61, 63], while in others it is seentogether with a dominant non-thermal component [103, 104, 105].

As was discussed in Sec. 2.2.1 the X-ray emission from GRBs is suggested tobe a combination of the ’late prompt’ and the true afterglow emission. Anothersuggested signature of prompt emission in the early X-ray emission are flares, whichhave hard-to-soft evolution as is also observed in the prompt emission [106]. Flaringintervals are commonly excluded from the spectral analysis in studies of thermalX-ray emission, mainly due to the rapid spectral evolution and the fact that thespectra sometimes deviate from power laws [106, 107]. Friis and Watson [85] were

3.2. Thermal components in the context of this thesis 19

first to propose a jet origin for the thermal component observed in the early X-rayafterglow in cases where the thermal component is very luminous and has hightemperature. A jet origin of the thermal component has been suggested for severalbursts (see [86, 108]).

3.2 Thermal components in the context of thisthesis

All three scenarios provide important insights into the physics behind the X-rayemission from GRBs. Therefore, where and how this thermal component arisesenables us to study the properties of the jets and progenitors. This puzzle isextensively addressed in the attached paper, where the common properties of GRBswith thermal components are examined. From our analysis, we infer that the mostprobable scenario to explain the thermal components observed is cocoon emission.Another result of the paper is the small spread of radii derived from the thermalcomponent, implying that these GRBs originate in similar environments and areobserved when they break out of the thick wind of the progenitor stars. The totalsample consists of only nine GRBs with detected thermal components. In orderto draw strong conclusions regarding the progenitors and the origin of the thermalcomponent more GRBs with thermal emission are needed.

Chapter 4

Observations and dataanalysis

In this chapter I will describe the instruments used to obtain spectra, the datareduction procedure and the process of the analysis of the reduced data.

4.1 Swift satellite: detectors and data reduction

The Swift satellite [109] was launched in 2004 and today it is one of the mainfacilities for observing GRBs. The main purpose of the satellite is to detectedGRBs and then to follow them up in X-rays, UV and optical. It is placed in alow Earth orbit with a period of ∼ 96 minutes. The satellite carries 3 instrumentson board: BAT (Burst Alert Telescope) [110], XRT (X-ray telescope) [111] andUVOT (UV/Optical Telescope) [112]. The satellite is shown in Fig. 4.1 togetherwith the positions of each of the instruments. After the satellite triggers on a GRB,it automatically slews to the position derived from the trigger during the next 20–70 s. Spectra and multiwavelength light curves for the duration of the aftergloware publicly available on the Leicester catalogue pages 1. In the next sections I willonly describe the BAT and XRT instruments, since I have not used UVOT datain the analysis. The XRT is described in more detail as it is the main instrumentused for the data analysis in the paper.

4.1.1 BAT

The BAT is the instrument that actually detects a GRB. It is a highly sensitive,large field of view telescope (approximately one steradian) designed with the mainpurpose to monitor a large fraction of the sky for GRBs. BAT uses a coded-aperturemask that was purpose-built for Swift. It provides the burst trigger and the position

1http://www.swift.ac.uk/xrt_live_cat/

21

22 Chapter 4. Observations and data analysis

Figure 4.1: The Swift satellite with its instruments. The figure is from Fig. 2 in[109].

with a precision between 1–4 arcmin [110]. The energy range covered by BAT is15–150 keV with an energy resolution of ∼ 7 keV. The number of bursts detectedby BAT is > 100 yr−1. Apart from triggering on GRBs, BAT is also accumulatingan all-sky hard X-ray survey.

Data reduction of BAT data

The steps for BAT data reduction are described in ”The Swift BAT SoftwareGuide”2. The pre-requirement in order to produce spectra and response files neededfor spectral analysis are constructing the Detector Plane Image (DPI). After thisstep a Fast Fourier Transform (FFT) is applied to deconvolve the detector planefrom the coded mask pattern, which results in an image in sky coordinates. Thenext step is to find and filter out the noisy detectors, which is done using the Cal-ibration Database. Finally, spectra and response files can be made. The routineproducing the spectra takes as input the event file, as well as the start and stoptimes of the observations. After producing the spectra, the response matrix file(RMF) needs to be built in order to fit the spectra. This is done using the batdrm-gen command, which takes as an input calibration files and the spectrum. All thesteps above are semi-automatic, meaning that they need input from the user to theautomatic procedures.

4.1.2 XRT

The XRT is a sensitive, X-ray CCD detector, which is designed to measure fluxes,spectra and light curves of GRBs and other X-ray sources. The basic principleof operation of an X-ray CCDs is as follows: the incoming X-ray is stopped bythe energy converter which produces a number of visible light photons or electricalcharges. For a detailed description of X-ray CCDs and their components, the readeris referred to [113] and their Figs. 1 and 13. The CCD in XRT has a thin filter

2https://swift.gsfc.nasa.gov/analysis/bat_swguide_v6_3.pdf

4.1. Swift satellite: detectors and data reduction 23

to protect it against optical light, as the CCD is also sensitive to optical light. Inaddition, the CCD is protected from sunlight by a shutter, which will automaticallyclose if the CCD is pointed to a direction within 10◦ of the Sun. The XRT coversthe energy range 0.2–10 keV and it generally starts observing ∼ 100 s after theBAT trigger. The field of view is 23.6 arcmin2, which is much narrower than theBAT field of view [111]. The XRT effective area is shown in Fig. 4.2.

Figure 4.2: Plot of the effective area of the XRT. The figure is taken from https://swift.gsfc.nasa.gov/about_swift/xrt_desc.html

The background in the instrument is a combination of the instrumental back-ground (electronic noise) and the astrophysical background. The astrophysicalbackground is composed of the Cosmic X-ray background and various particles,such as cosmic rays and solar protons. The electronic noise component of thebackground is important at low energies, but since it is concentrated in a few badpixels it is easily filtered out by standard software. The astrophysical background,especially the particle induced one, is significant at higher energies [114]. The back-ground spectra are extracted away from the source to account for this contribution.The background in the XRT is generally very low compared to the count rate of aGRB, typically 0.01% of the total count rate (see Fig. 4.3).

The instrument has two main modes of operation: Photon-counting (PC) andWindowed Timing (WT) mode. WT mode has a 1.8 ms time resolution and it ismainly used when the count rate is high, generally in the beginning of the XRTobservations. The PC mode has a worse time resolution of 2.5 s and it is generallyused when the count rate is lower. The XRT will automatically change betweendifferent observing modes as the X-ray emission fades. Should the emission becomebrighter at later times, the XRT will automatically change between modes in thereverse direction. In this thesis I am mainly using WT mode data, due to a better


Figure 4.3: Plot of the XRT background (lower curves) compared to the GRBspectra for time resolved spectra of GRB 111123A. Different colors correspond todifferent time-resolved spectra produced using the Bayesian blocks algorithm (Sec.4.2.1).

time-resolution as well as better statistics in the data. These data appear as asingle strip of data in the CCD image, oriented at the spacecraft roll angle sinceWT data, unlike PC data, are read out in one dimension.

Data reduction of XRT WT data

The resulting files that can be downloaded from the Leicester Swift catalogue3 areevent files that contain information about each individual photon detected. Apartfrom the event files, attitude files which contain the attitude determined from thesatellite star trackers are available. Before starting actual data reduction, it isessential to check if the data exhibit pile-up, an effect in CCD cameras when thereis a significant probability that two or more photons will be registered as a singleevent. For the WT mode data, pile-up occurs if the light curve has more than100 counts s−1. The simplest way to deal with pile-up is described in Appendix Aof [115]. The authors selected 5 time intervals during which the observed sourcecount rate was < 100, 100–200, 200–300, 300-400 and > 400 counts s−1. By fittingthese spectra with an absorbed power law they demonstrated that the source isaffected by moderate pile-up for 100–300 counts s−1. 1 pixel needs to be excludedfrom the extraction region to remove the pile-up. The source is strongly affected bypile up for 300–400 counts s−1, where 2 pixels need to be excluded, and for > 400counts s−1 4 pixels need to be excluded from the extraction region. These pixels

3http://www.swift.ac.uk/xrt_live_cat/

4.2. Data analysis 25

are excluded from the centre of Point Spread Function (PSF) since it is where thecount rate is the highest.

The detailed instructions how to perform data reduction are given in the ”XRTAnalysys Guide”4. From the WT image, in the first step, source and backgroundregions are produced using the xselect tool and ds9 environment. Then, these regionfiles, together with event and attitude files as well as a possible time binning filesare fed to the xrtproducts pipeline, which produces spectral and background files,together with response files. The XRT has two types of response files: the RMF andthe Ancillary Response Files (ARF). These two files contain information about theXRT effective area, which is made up of three components: the mirror effective area,the filter transmission and the CCD quantum efficiency. The quantum efficiency isincluded in the RMFs, while the mirror effective area and the filter transmission areincluded in the ARFs. The response files also contain information about channel-energy relations. Both the RMF and the ARF files are needed in order to fit theXRT spectra. The final step in the data reduction process is to flag the bad columns.These bad columns are a consequence of Swift being hit by a micrometeorite in 2005.Since then a small number of columns is being vetoed to prevent saturation of thetelemetry. For the WT data, columns 0–29 are marked as bad, corresponding tothe energy range 0.2–0.3 keV.

Since Swift was launched 13 years ago the time it spent in orbit took its toll onthe instruments on board. Currently there are various calibration issues with XRTdata at energies below ∼ 0.6 keV. These issues include redistribution problems atsoft energies in absorbed sources, resulting in a deviation from the true spectralshape. These issues and their impact on the XRT data analysed in the energyrange 0.3–10 keV are discussed in the attached paper.

4.2 Data analysis

The spectral files are by default grouped so that one count occupies one bin. Asthe data are Poisson distributed, this type of grouping allows for the use of Cashstatistics [116]. However, since this statistic cannot directly be interpreted as agoodness-of-fit, it is common practice in the field when using XRT data to groupthem such that a minimum of 20 counts occupies one bin, which in turn allows forthe usage of χ2 statistics. The need for this particular grouping of the data arisesfrom the fact that χ2 statistics is strictly defined for Gaussian-distributed dataand when the number of counts per bin exceeds 20 the deviation from Gaussianitybecomes small [117]. In this thesis, the data were binned so that a minimum of 20counts occupies one bin and χ2 statistics was used.

Spectra were fitted with the standard software package XSPEC [118]. The soft-ware uses spectral, background and response files. In order to actually fit the modelto data, a procedure called forward folding is used [119]. This procedure requires amodel to be predefined before a predicted count spectrum can be produced. What

4http://www.swift.ac.uk/analysis/xrt/index.php


this essentially means is that the spectral shapes cannot be described without firstassuming a model. This model is then convolved with the response file, which re-sults in model’s count prediction that is then compared to the detector counts. Theresults are the best-fitting parameters from the model that give the best agreementwith the data. This leads to results being highly dependent on the assumptionsof the model. Another caveat when fitting a model in XSPEC is the fact thatthe fitting algorithm is local and not global, which can result in the fitting processgetting stuck in a local minimum without finding the global minimum. The fittingalgorithm used by XSPEC is the Levenberg-Marquardt algorithm [120].

In this thesis, I explore which of two models (simple power law or power law+ blackbody) better describes the data. In order to asses the significance of theadded component (the blackbody) Monte Carlo simulations were performed. Inthis procedure, a large number of spectra (10 000) are simulated from the best-fitting parameters of the power-law model and then the two models are re-fitted tothese fake spectra. The resulting distributions of χ2 are then compared to the onesfrom real data to determine the significance of the added component. The maindownside of this procedure is that it is computationally expensive. XSPEC has abuilt-in procedure for assessing which model fits data better called an F-test, butthis procedure is not valid when assessing the significance of an added component[121].

4.2.1 Bayesian blocks binning

100 150 200 250 300 350 400 450Time since BAT trigger (s)

0.0

0.5

1.0

1.5

2.0

2.5

3.0

3.5

4.0

Fx

10−

9(e

rgcm

−2

s−1)

Figure 4.4: Light curve of GRB 111225A binned with Bayesian blocks.

In order to perform a comprehensive spectral analysis, a time-resolved analysisof the GRB spectra is needed. This is due to the fact that in the time-averagedspectrum, spectral evolution can occur that can alter the results and thus our un-derstanding of the physical processes responsible for the production of the spectra.

4.2. Data analysis 27

Another reason for choosing a time-resolved analysis is that it is important to followthe evolution of the physical parameters with time.

In this analysis, I chose to use the Bayesian blocks binning method [122]. Itsmain purpose is to capture local temporal variations in the count rate with Bayesianprobability theory. This in practice means that, unlike constant exposure time, theblocks may vary in duration. This method is non-parametric, i.e. the binning isdetermined by the data itself rather than some pre-setup condition (like signal-to-noise ratio). While this method traces the evolution of the flux well, some bins mayend up with poor statistics due to low count rates. A light curve of GRB 111225Abinned using Bayesian blocks is shown in Fig. 4.4 as an illustration of the method.

List of appended Papers

Paper IValan et al. (2017), Thermal components in theearly X-ray afterglows of GRBs: likely cocoonemission and constraints on the progenitors

In this paper we analyzed 74 Swift GRBs in search for a thermal component inthe X-ray spectra. All previously reported GRBs with such components were alsoanalysed. We found a significant thermal component in 6 GRBs from our sampleand confirmed 3 of the previously reported cases. By analyzing their commonproperties we conclude that the cocoon surrounding the jet is a likely explanationfor the origin of the blackbody component in the majority of cases. In addition, weconclude that these GRBs originate in similar environments as the radii inferredfrom the thermal component span a narrow range.

29

Author contribution

Paper I

I performed the majority of the work in this paper, including data reduction, dataanalysis and production of all the figures in the paper. The Fermi data were ana-lyzed by Björn Ahlgren. The manuscript was written by me, except Sec. 7.3 whichwas written by Björn Ahlgren and Sec. 7.4 written by Josefin Larsson. All workwas supervised by Josefin Larsson.

31

List of Figures

1.1 Host galaxies of long GRBs . . . . . . . . . . . . . . . . . . . . . . 3

2.1 Prompt emission light curves of four different GRBs observed bySwift BAT . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 8

2.2 Schematic view of the Band function . . . . . . . . . . . . . . . . . 92.3 Distribution of Band model parameters . . . . . . . . . . . . . . . 102.4 Early X-ray light curves of four different GRBs observed by XRT . 112.5 Different phases of the X-ray light curve of GRBs . . . . . . . . . . 12

3.1 Shematic illustration of the cocoon that surrounds the jet . . . . . 17

4.1 The Swift satellite with its instruments . . . . . . . . . . . . . . . . 224.2 Plot of the effective area of the XRT . . . . . . . . . . . . . . . . . 234.3 Plot of the XRT background compared to the GRB spectra . . . . 244.4 Light curve of GRB 111225A binned with Bayesian blocks. . . . . 26

33

Bibliography

[1] N. R. Tanvir, D. B. Fox, A. J. Levan, E. Berger, K. Wiersema, J. P. U.Fynbo, A. Cucchiara, T. Krühler, N. Gehrels, J. S. Bloom, J. Greiner, P. A.Evans, E. Rol, F. Olivares, J. Hjorth, P. Jakobsson, J. Farihi, R. Willingale,R. L. C. Starling, S. B. Cenko, D. Perley, J. R. Maund, J. Duke, R. A. M. J.Wijers, A. J. Adamson, A. Allan, M. N. Bremer, D. N. Burrows, A. J. Castro-Tirado, B. Cavanagh, A. de Ugarte Postigo, M. A. Dopita, T. A. Fatkhullin,A. S. Fruchter, R. J. Foley, J. Gorosabel, J. Kennea, T. Kerr, S. Klose, H. A.Krimm, V. N. Komarova, S. R. Kulkarni, A. S. Moskvitin, C. G. Mundell,T. Naylor, K. Page, B. E. Penprase, M. Perri, P. Podsiadlowski, K. Roth,R. E. Rutledge, T. Sakamoto, P. Schady, B. P. Schmidt, A. M. Soderberg,J. Sollerman, A. W. Stephens, G. Stratta, T. N. Ukwatta, D. Watson, E. Wes-tra, T. Wold, and C. Wolf. A γ-ray burst at a redshift of z˜8.2. Nature, 461:1254–1257, October 2009. doi: 10.1038/nature08459.

[2] R. W. Klebesadel, I. B. Strong, and R. A. Olson. Observations of gamma-raybursts of cosmic origin. ApJL, 182:L85, June 1973. doi: 10.1086/181225.

[3] M. S. Briggs, W. S. Paciesas, G. N. Pendleton, C. A. Meegan, G. J. Fishman,J. M. Horack, M. N. Brock, C. Kouveliotou, D. H. Hartmann, and J. Hakkila.Batse observations of the large-scale isotropy of gamma-ray bursts. ApJ, 459:40, March 1996. doi: 10.1086/176867.

[4] Boella, G., Butler, R. C., Perola, G. C., Piro, L., Scarsi, L., and Bleeker,J. A.M. Bepposax, the wide band mission for x-ray astronomy. Astron.Astrophys. Suppl. Ser., 122(2):299–307, 1997. doi: 10.1051/aas:1997136. URLhttps://doi.org/10.1051/aas:1997136.

[5] M. R. Metzger, S. G. Djorgovski, S. R. Kulkarni, C. C. Steidel, K. L. Adel-berger, D. A. Frail, E. Costa, and F. Frontera. Spectral constraints on theredshift of the optical counterpart to the γ-ray burst of 8 may 1997. Nature,387:878–880, June 1997. doi: 10.1038/43132.

[6] C. Kouveliotou, C. A. Meegan, G. J. Fishman, N. P. Bhat, M. S. Briggs, T. M.Koshut, W. S. Paciesas, and G. N. Pendleton. Identification of two classes ofgamma-ray bursts. ApJL, 413:L101–L104, August 1993. doi: 10.1086/186969.

35

36 BIBLIOGRAPHY

[7] E. Berger, P. A. Price, S. B. Cenko, A. Gal-Yam, A. M. Soderberg, M. Kasli-wal, D. C. Leonard, P. B. Cameron, D. A. Frail, S. R. Kulkarni, D. C. Mur-phy, W. Krzeminski, T. Piran, B. L. Lee, K. C. Roth, D.-S. Moon, D. B.Fox, F. A. Harrison, S. E. Persson, B. P. Schmidt, B. E. Penprase, J. Rich,B. A. Peterson, and L. L. Cowie. The afterglow and elliptical host galaxy ofthe short γ-ray burst grb 050724. Nature, 438:988–990, December 2005. doi:10.1038/nature04238.

[8] E. Berger. The host galaxies of short-duration gamma-ray bursts: Luminosi-ties, metallicities, and star-formation rates. ApJ, 690:231–237, January 2009.doi: 10.1088/0004-637X/690/1/231.

[9] P. D’Avanzo. Short gamma-ray bursts: A review. Journal of High EnergyAstrophysics, 7:73–80, September 2015. doi: 10.1016/j.jheap.2015.07.002.

[10] N. R. Tanvir, A. J. Levan, A. S. Fruchter, J. Hjorth, R. A. Hounsell,K. Wiersema, and R. L. Tunnicliffe. A ‘kilonova’ associated with the short-duration γ-ray burst grb 130603b. Nature, 500:547–549, August 2013. doi:10.1038/nature12505.

[11] L.-X. Li and B. Paczyński. Transient Events from Neutron Star Mergers.ApJL, 507:L59–L62, November 1998. doi: 10.1086/311680.

[12] B. D. Metzger. Kilonovae. Living Reviews in Relativity, 20:3, May 2017. doi:10.1007/s41114-017-0006-z.

[13] T. J. Galama, P. M. Vreeswijk, J. van Paradijs, C. Kouveliotou, T. Au-gusteijn, H. Böhnhardt, J. P. Brewer, V. Doublier, J.-F. Gonzalez, B. Lei-bundgut, C. Lidman, O. R. Hainaut, F. Patat, J. Heise, J. in’t Zand, K. Hur-ley, P. J. Groot, R. G. Strom, P. A. Mazzali, K. Iwamoto, K. Nomoto,H. Umeda, T. Nakamura, T. R. Young, T. Suzuki, T. Shigeyama, T. Koshut,M. Kippen, C. Robinson, P. de Wildt, R. A. M. J. Wijers, N. Tanvir,J. Greiner, E. Pian, E. Palazzi, F. Frontera, N. Masetti, L. Nicastro, M. Fe-roci, E. Costa, L. Piro, B. A. Peterson, C. Tinney, B. Boyle, R. Cannon,R. Stathakis, E. Sadler, M. C. Begam, and P. Ianna. An unusual supernovain the error box of the γ-ray burst of 25 april 1998. Nature, 395:670–672,October 1998. doi: 10.1038/27150.

[14] B. Paczyński. Are gamma-ray bursts in star-forming regions? ApJL, 494:L45–L48, February 1998. doi: 10.1086/311148.

[15] A. I. MacFadyen and S. E. Woosley. Collapsars: Gamma-ray bursts andexplosions in “failed supernovae”. ApJ, 524:262–289, October 1999. doi:10.1086/307790.

[16] S. E. Woosley. Gamma-ray bursts from stellar mass accretion disks aroundblack holes. ApJ, 405:273–277, March 1993. doi: 10.1086/172359.

BIBLIOGRAPHY 37

[17] P. A. Crowther, L. Dessart, D. J. Hillier, J. B. Abbott, and A. W. Fullerton.Stellar and wind properties of lmc wc4 stars. a metallicity dependence forwolf-rayet mass-loss rates. A&A, 392:653–669, September 2002. doi: 10.1051/0004-6361:20020941.

[18] J. S. Vink. Gamma-ray burst progenitors and the population of rotating wolf-rayet stars. Philosophical Transactions of the Royal Society of London SeriesA, 371:20120237–20120237, April 2013. doi: 10.1098/rsta.2012.0237.

[19] E. M. Levesque, A. M. Soderberg, L. J. Kewley, and E. Berger. No CorrelationBetween Host Galaxy Metallicity and Gamma-ray Energy Release for Long-duration Gamma-ray Bursts. ApJ, 725:1337–1341, December 2010. doi: 10.1088/0004-637X/725/1/1337.

[20] M. Cantiello, S.-C. Yoon, N. Langer, and M. Livio. Binary star progenitorsof long gamma-ray bursts. A&A, 465:L29–L33, April 2007. doi: 10.1051/0004-6361:20077115.

[21] A. Blaauw. On the origin of the o- and b-type stars with high velocities (the“run-away” stars), and some related problems. Bulletin of the AstronomicalInstitutes of the Netherlands, 15:265, May 1961.

[22] J. J. Eldridge, N. Langer, and C. A. Tout. Runaway stars as progenitors ofsupernovae and gamma-ray bursts. MNRAS, 414:3501–3520, July 2011. doi:10.1111/j.1365-2966.2011.18650.x.

[23] B. P. Abbott and et al. Gw150914: The advanced ligo detectors in the eraof first discoveries. Physical Review Letters, 116(13):131103, April 2016. doi:10.1103/PhysRevLett.116.131103.

[24] C. Meegan, G. Lichti, P. N. Bhat, E. Bissaldi, M. S. Briggs, V. Con-naughton, R. Diehl, G. Fishman, J. Greiner, A. S. Hoover, A. J. van derHorst, A. von Kienlin, R. M. Kippen, C. Kouveliotou, S. McBreen, W. S.Paciesas, R. Preece, H. Steinle, M. S. Wallace, R. B. Wilson, and C. Wilson-Hodge. The Fermi Gamma-ray Burst Monitor. ApJ, 702:791-804, September2009. doi: 10.1088/0004-637X/702/1/791.

[25] W. B. Atwood, A. A. Abdo, M. Ackermann, W. Althouse, B. Anderson,M. Axelsson, L. Baldini, J. Ballet, D. L. Band, G. Barbiellini, and et al. TheLarge Area Telescope on the Fermi Gamma-Ray Space Telescope Mission.ApJ, 697:1071–1102, June 2009. doi: 10.1088/0004-637X/697/2/1071.

[26] V. Connaughton, E. Burns, A. Goldstein, L. Blackburn, M. S. Briggs, B.-B. Zhang, J. Camp, N. Christensen, C. M. Hui, P. Jenke, T. Littenberg,J. E. McEnery, J. Racusin, P. Shawhan, L. Singer, J. Veitch, C. A. Wilson-Hodge, P. N. Bhat, E. Bissaldi, W. Cleveland, G. Fitzpatrick, M. M. Giles,M. H. Gibby, A. von Kienlin, R. M. Kippen, S. McBreen, B. Mailyan, C. A.Meegan, W. S. Paciesas, R. D. Preece, O. J. Roberts, L. Sparke, M. Stanbro,

38 BIBLIOGRAPHY

K. Toelge, and P. Veres. Fermi gbm observations of ligo gravitational-waveevent gw150914. ApJL, 826:L6, July 2016. doi: 10.3847/2041-8205/826/1/L6.

[27] R. Perna, D. Lazzati, and B. Giacomazzo. Short gamma-ray bursts fromthe merger of two black holes. ApJL, 821:L18, April 2016. doi: 10.3847/2041-8205/821/1/L18.

[28] V. V. Usov. Millisecond pulsars with extremely strong magnetic fields as acosmological source of gamma-ray bursts. Nature, 357:472–474, June 1992.doi: 10.1038/357472a0.

[29] J. Greiner, P. A. Mazzali, D. A. Kann, T. Krühler, E. Pian, S. Prentice,F. Olivares E., A. Rossi, S. Klose, S. Taubenberger, F. Knust, P. M. J.Afonso, C. Ashall, J. Bolmer, C. Delvaux, R. Diehl, J. Elliott, R. Filgas,J. P. U. Fynbo, J. F. Graham, A. N. Guelbenzu, S. Kobayashi, G. Leloudas,S. Savaglio, P. Schady, S. Schmidl, T. Schweyer, V. Sudilovsky, M. Tanga,A. C. Updike, H. van Eerten, and K. Varela. A very luminous magnetar-powered supernova associated with an ultra-long γ-ray burst. Nature, 523:189–192, July 2015. doi: 10.1038/nature14579.

[30] N. Bucciantini. Magnetars and gamma ray bursts. In P. Roming, N. Kawai,and E. Pian, editors, Death of Massive Stars: Supernovae and Gamma-RayBursts, volume 279 of IAU Symposium, pages 289–296, September 2012. doi:10.1017/S1743921312013075.

[31] B. D. Metzger, D. Giannios, T. A. Thompson, N. Bucciantini, andE. Quataert. The protomagnetar model for gamma-ray bursts. MNRAS,413:2031–2056, May 2011. doi: 10.1111/j.1365-2966.2011.18280.x.

[32] C. C. Thöne, A. de Ugarte Postigo, C. L. Fryer, K. L. Page, J. Gorosabel,M. A. Aloy, D. A. Perley, C. Kouveliotou, H. T. Janka, P. Mimica, J. L.Racusin, H. Krimm, J. Cummings, S. R. Oates, S. T. Holland, M. H. Siegel,M. de Pasquale, E. Sonbas, M. Im, W.-K. Park, D. A. Kann, S. Guziy, L. H.Garćıa, A. Llorente, K. Bundy, C. Choi, H. Jeong, H. Korhonen, P. Kubànek,J. Lim, A. Moskvitin, T. Muñoz-Darias, S. Pak, and I. Parrish. The unusualγ-ray burst GRB 101225A from a helium star/neutron star merger at redshift0.33. Nature, 480:72–74, December 2011. doi: 10.1038/nature10611.

[33] B. Gendre, G. Stratta, J. L. Atteia, S. Basa, M. Boër, D. M. Coward, S. Cu-tini, V. D’Elia, E. J. Howell, A. Klotz, and L. Piro. The Ultra-long Gamma-Ray Burst 111209A: The Collapse of a Blue Supergiant? ApJ, 766:30, March2013. doi: 10.1088/0004-637X/766/1/30.

[34] A. J. Levan, N. R. Tanvir, R. L. C. Starling, K. Wiersema, K. L. Page,D. A. Perley, S. Schulze, G. A. Wynn, R. Chornock, J. Hjorth, S. B.Cenko, A. S. Fruchter, P. T. O’Brien, G. C. Brown, R. L. Tunnicliffe,D. Malesani, P. Jakobsson, D. Watson, E. Berger, D. Bersier, B. E. Cobb,

BIBLIOGRAPHY 39

S. Covino, A. Cucchiara, A. de Ugarte Postigo, D. B. Fox, A. Gal-Yam,P. Goldoni, J. Gorosabel, L. Kaper, T. Krühler, R. Karjalainen, J. P. Os-borne, E. Pian, R. Sánchez-Ramı́rez, B. Schmidt, I. Skillen, G. Tagliaferri,C. Thöne, O. Vaduvescu, R. A. M. J. Wijers, and B. A. Zauderer. A NewPopulation of Ultra-long Duration Gamma-Ray Bursts. ApJ, 781:13, January2014. doi: 10.1088/0004-637X/781/1/13.

[35] F. J. Virgili, C. G. Mundell, V. Pal’shin, C. Guidorzi, R. Margutti, A. Me-landri, R. Harrison, S. Kobayashi, R. Chornock, A. Henden, A. C. Updike,S. B. Cenko, N. R. Tanvir, I. A. Steele, A. Cucchiara, A. Gomboc, A. Levan,Z. Cano, C. J. Mottram, N. R. Clay, D. Bersier, D. Kopač, J. Japelj, A. V.Filippenko, W. Li, D. Svinkin, S. Golenetskii, D. H. Hartmann, P. A. Milne,G. Williams, P. T. O’Brien, D. B. Fox, and E. Berger. GRB 091024A andthe Nature of Ultra-long Gamma-Ray Bursts. ApJ, 778:54, November 2013.doi: 10.1088/0004-637X/778/1/54.

[36] B.-B. Zhang, B. Zhang, K. Murase, V. Connaughton, and M. S. Briggs. HowLong does a Burst Burst? ApJ, 787:66, May 2014. doi: 10.1088/0004-637X/787/1/66.

[37] P. A. Mazzali, S. Valenti, M. Della Valle, G. Chincarini, D. N. Sauer,S. Benetti, E. Pian, T. Piran, V. D’Elia, N. Elias-Rosa, R. Margutti, F. Pa-sotti, L. A. Antonelli, F. Bufano, S. Campana, E. Cappellaro, S. Covino,P. D’Avanzo, F. Fiore, D. Fugazza, R. Gilmozzi, D. Hunter, K. Maguire,E. Maiorano, P. Marziani, N. Masetti, F. Mirabel, H. Navasardyan,K. Nomoto, E. Palazzi, A. Pastorello, N. Panagia, L. J. Pellizza, R. Sari,S. Smartt, G. Tagliaferri, M. Tanaka, S. Taubenberger, N. Tominaga, C. Trun-dle, and M. Turatto. The metamorphosis of supernova sn 2008d/xrf 080109:A link between supernovae and grbs/hypernovae. Science, 321:1185, August2008. doi: 10.1126/science.1158088.

[38] J. P. U. Fynbo, D. Watson, C. C. Thöne, J. Sollerman, J. S. Bloom, T. M.Davis, J. Hjorth, P. Jakobsson, U. G. Jørgensen, J. F. Graham, A. S. Fruchter,D. Bersier, L. Kewley, A. Cassan, J. M. Castro Cerón, S. Foley, J. Goros-abel, T. C. Hinse, K. D. Horne, B. L. Jensen, S. Klose, D. Kocevski, J.-B.Marquette, D. Perley, E. Ramirez-Ruiz, M. D. Stritzinger, P. M. Vreeswijk,R. A. M. Wijers, K. G. Woller, D. Xu, and M. Zub. No supernovae associatedwith two long-duration γ-ray bursts. Nature, 444:1047–1049, December 2006.doi: 10.1038/nature05375.

[39] M. J. Micha lowski, D. Xu, J. Stevens, A. Levan, J. Yang, Z. Paragi, A. Kam-ble, H. Dannerbauer, A. J. van der Horst, L. Shao, D. Crosby, G. Gen-tile, E. Stanway, K. Wiersema, J. P. U. Fynbo, N. R. Tanvir, P. Kamphuis,and M. Garrett. The second closest gamma-ray burst: sub-luminous GRB111005A with no supernova in a super-solar metallicity environment. ArXive-prints, October 2016.

40 BIBLIOGRAPHY

[40] M. Della Valle, G. Chincarini, N. Panagia, G. Tagliaferri, D. Malesani,V. Testa, D. Fugazza, S. Campana, S. Covino, V. Mangano, L. A. An-tonelli, P. D’Avanzo, K. Hurley, I. F. Mirabel, L. J. Pellizza, S. Pira-nomonte, and L. Stella. An enigmatic long-lasting γ-ray burst not accom-panied by a bright supernova. Nature, 444:1050–1052, December 2006. doi:10.1038/nature05374.

[41] Z. Cano, S.-Q. Wang, Z.-G. Dai, and X.-F. Wu. The Observer’s Guide to theGamma-Ray Burst Supernova Connection. Advances in Astronomy, 2017:8929054, 2017. doi: 10.1155/2017/8929054.

[42] S. Campana, V. Mangano, A. J. Blustin, P. Brown, D. N. Burrows,G. Chincarini, J. R. Cummings, G. Cusumano, M. Della Valle, D. Male-sani, P. Mészáros, J. A. Nousek, M. Page, T. Sakamoto, E. Waxman,B. Zhang, Z. G. Dai, N. Gehrels, S. Immler, F. E. Marshall, K. O. Mason,A. Moretti, P. T. O’Brien, J. P. Osborne, K. L. Page, P. Romano, P. W. A.Roming, G. Tagliaferri, L. R. Cominsky, P. Giommi, O. Godet, J. A. Kennea,H. Krimm, L. Angelini, S. D. Barthelmy, P. T. Boyd, D. M. Palmer, A. A.Wells, and N. E. White. The association of grb 060218 with a supernova andthe evolution of the shock wave. Nature, 442:1008–1010, August 2006. doi:10.1038/nature04892.

[43] A. M. Soderberg, S. R. Kulkarni, E. Nakar, E. Berger, P. B. Cameron, D. B.Fox, D. Frail, A. Gal-Yam, R. Sari, S. B. Cenko, M. Kasliwal, R. A. Chevalier,T. Piran, P. A. Price, B. P. Schmidt, G. Pooley, D.-S. Moon, B. E. Penprase,E. Ofek, A. Rau, N. Gehrels, J. A. Nousek, D. N. Burrows, S. E. Persson,and P. J. McCarthy. Relativistic ejecta from X-ray flash XRF 060218 andthe rate of cosmic explosions. Nature, 442:1014–1017, August 2006. doi:10.1038/nature05087.

[44] J. Greiner, H.-F. Yu, T. Krühler, D. D. Frederiks, A. Beloborodov, P. N.Bhat, J. Bolmer, H. van Eerten, R. L. Aptekar, J. Elliott, S. V. Golenetskii,J. F. Graham, K. Hurley, D. A. Kann, S. Klose, A. Nicuesa Guelbenzu,A. Rau, P. Schady, S. Schmidl, V. Sudilovsky, D. S. Svinkin, M. Tanga, M. V.Ulanov, K. Varela, A. von Kienlin, and X.-L. Zhang. GROND coverage ofthe main peak of gamma-ray burst 130925A. A&A, 568:A75, August 2014.doi: 10.1051/0004-6361/201424250.

[45] D. L. Meier, S. Koide, and Y. Uchida. Magnetohydrodynamic Productionof Relativistic Jets. Science, 291:84–92, January 2001. doi: 10.1126/science.291.5501.84.

[46] J. E. Rhoads. How to Tell a Jet from a Balloon: A Proposed Test for Beamingin Gamma-Ray Bursts. ApJL, 487:L1–L4, September 1997. doi: 10.1086/310876.

BIBLIOGRAPHY 41

[47] F. A. Harrison, J. S. Bloom, D. A. Frail, R. Sari, S. R. Kulkarni, S. G.Djorgovski, T. Axelrod, J. Mould, B. P. Schmidt, M. H. Wieringa, R. M.Wark, R. Subrahmanyan, D. McConnell, P. J. McCarthy, B. E. Schaefer,R. G. McMahon, R. O. Markze, E. Firth, P. Soffitta, and L. Amati. Opticaland Radio Observations of the Afterglow from GRB 990510: Evidence for aJet. ApJL, 523:L121–L124, October 1999. doi: 10.1086/312282.

[48] T. Piran. The physics of gamma-ray bursts. Reviews of Modern Physics, 76:1143–1210, October 2004. doi: 10.1103/RevModPhys.76.1143.

[49] Y. Lithwick and R. Sari. Lower Limits on Lorentz Factors in Gamma-RayBursts. ApJ, 555:540–545, July 2001. doi: 10.1086/321455.

[50] R. Sari, T. Piran, and J. P. Halpern. Jets in Gamma-Ray Bursts. ApJL, 519:L17–L20, July 1999. doi: 10.1086/312109.

[51] D. A. Frail, S. R. Kulkarni, R. Sari, S. G. Djorgovski, J. S. Bloom, T. J.Galama, D. E. Reichart, E. Berger, F. A. Harrison, P. A. Price, S. A. Yost,A. Diercks, R. W. Goodrich, and F. Chaffee. Beaming in gamma-ray bursts:Evidence for a standard energy reservoir. The Astrophysical Journal Let-ters, 562(1):L55, 2001. URL http://stacks.iop.org/1538-4357/562/i=1/a=L55.

[52] J. L. Racusin, E. W. Liang, D. N. Burrows, A. Falcone, T. Sakamoto, B. B.Zhang, B. Zhang, P. Evans, and J. Osborne. Jet Breaks and Energetics ofSwift Gamma-Ray Burst X-Ray Afterglows. ApJ, 698:43–74, June 2009. doi:10.1088/0004-637X/698/1/43.

[53] B. Paczynski. Gamma-ray bursters at cosmological distances. ApJL, 308:L43–L46, September 1986. doi: 10.1086/184740.

[54] M. J. Rees and P. Meszaros. Relativistic fireballs - Energy conversion andtime-scales. MNRAS, 258:41P–43P, September 1992. doi: 10.1093/mnras/258.1.41P.

[55] P. Meszaros, P. Laguna, and M. J. Rees. Gasdynamics of relativisticallyexpanding gamma-ray burst sources - Kinematics, energetics, magnetic fields,and efficiency. ApJ, 415:181–190, September 1993. doi: 10.1086/173154.

[56] P. Mészáros and M. J. Rees. Steep Slopes and Preferred Breaks in Gamma-Ray Burst Spectra: The Role of Photospheres and Comptonization. ApJ,530:292–298, February 2000. doi: 10.1086/308371.

[57] A. Shemi and T. Piran. The appearance of cosmic fireballs. ApJL, 365:L55–L58, December 1990. doi: 10.1086/185887.

[58] A. Pe’er. Physics of gamma-ray bursts prompt emission. Advances in As-tronomy, 2015:907321, 2015. doi: 10.1155/2015/907321.

42 BIBLIOGRAPHY

[59] D. Band, J. Matteson, L. Ford, B. Schaefer, D. Palmer, B. Teegarden,T. Cline, M. Briggs, W. Paciesas, G. Pendleton, G. Fishman, C. Kouveliotou,C. Meegan, R. Wilson, and P. Lestrade. Batse observations of gamma-rayburst spectra. i - spectral diversity. ApJ, 413:281–292, August 1993. doi:10.1086/172995.

[60] H.-F. Yu, R. D. Preece, J. Greiner, P. Narayana Bhat, E. Bissaldi, M. S.Briggs, W. H. Cleveland, V. Connaughton, A. Goldstein, A. von Kienlin,C. Kouveliotou, B. Mailyan, C. A. Meegan, W. S. Paciesas, A. Rau, O. J.Roberts, P. Veres, C. Wilson-Hodge, B.-B. Zhang, and H. J. van Eerten. TheFermi GBM gamma-ray burst time-resolved spectral catalog: brightest burstsin the first four years. A&A, 588:A135, April 2016. doi: 10.1051/0004-6361/201527509.

[61] F. Ryde. The Cooling Behavior of Thermal Pulses in Gamma-Ray Bursts.ApJ, 614:827–846, October 2004. doi: 10.1086/423782.

[62] G. Ghirlanda, A. Pescalli, and G. Ghisellini. Photospheric emission through-out GRB 100507 detected by Fermi. MNRAS, 432:3237–3244, July 2013. doi:10.1093/mnras/stt681.

[63] J. Larsson, J. L. Racusin, and J. M. Burgess. Evidence for Jet LaunchingClose to the Black Hole in GRB 101219b A Fermi GRB Dominated by Ther-mal Emission. ApJL, 800:L34, February 2015. doi: 10.1088/2041-8205/800/2/L34.

[64] M. J. Rees and P. Meszaros. Unsteady outflow models for cosmologicalgamma-ray bursts. ApJL, 430:L93–L96, August 1994. doi: 10.1086/187446.

[65] R. D. Preece, M. S. Briggs, R. S. Mallozzi, G. N. Pendleton, W. S. Paciesas,and D. L. Band. The Synchrotron Shock Model Confronts a “Line of Death”in the BATSE Gamma-Ray Burst Data. ApJL, 506:L23–L26, October 1998.doi: 10.1086/311644.

[66] B. Ahlgren, J. Larsson, T. Nymark, F. Ryde, and A. Pe’er. Confronting grbprompt emission with a model for subphotospheric dissipation. MNRAS, 454:L31–L35, November 2015. doi: 10.1093/mnrasl/slv114.

[67] J. M. Burgess, D. Bégué, F. Ryde, N. Omodei, A. Pe’er, J. L. Racusin, andA. Cucchiara. An external shock origin of grb 141028a. ApJ, 822:63, May2016. doi: 10.3847/0004-637X/822/2/63.

[68] B. Zhang and H. Yan. The internal-collision-induced magnetic reconnectionand turbulence (icmart) model of gamma-ray bursts. ApJ, 726:90, January2011. doi: 10.1088/0004-637X/726/2/90.

[69] R. Sari. The Observed Size and Shape of Gamma-Ray Burst Afterglow. ApJL,494:L49–L52, February 1998. doi: 10.1086/311160.

BIBLIOGRAPHY 43

[70] R. Sari and T. Piran. Predictions for the Very Early Afterglow and theOptical Flash. ApJ, 520:641–649, August 1999. doi: 10.1086/307508.

[71] J. van Paradijs, P. J. Groot, T. Galama, C. Kouveliotou, R. G. Strom, J. Telt-ing, R. G. M. Rutten, G. J. Fishman, C. A. Meegan, M. Pettini, N. Tanvir,J. Bloom, H. Pedersen, H. U. Nørdgaard-Nielsen, M. Linden-Vørnle, J. Mel-nick, G. van der Steene, M. Bremer, R. Naber, J. Heise, J. in’t Zand, E. Costa,M. Feroci, L. Piro, F. Frontera, G. Zavattini, L. Nicastro, E. Palazzi, K. Ben-nett, L. Hanlon, and A. Parmar. Transient optical emission from the errorbox of the γ-ray burst of 28 february 1997. Nature, 386:686–689, April 1997.doi: 10.1038/386686a0.

[72] P. T. O’Brien, R. Willingale, J. Osborne, M. R. Goad, K. L. Page, S. Vaughan,E. Rol, A. Beardmore, O. Godet, C. P. Hurkett, A. Wells, B. Zhang,S. Kobayashi, D. N. Burrows, J. A. Nousek, J. A. Kennea, A. Falcone,D. Grupe, N. Gehrels, S. Barthelmy, J. Cannizzo, J. Cummings, J. E. Hill,H. Krimm, G. Chincarini, G. Tagliaferri, S. Campana, A. Moretti, P. Giommi,M. Perri, V. Mangano, and V. LaParola. The early x-ray emission from grbs.ApJ, 647:1213–1237, August 2006. doi: 10.1086/505457.

[73] N. Gehrels, S. D. Barthelmy, D. N. Burrows, J. K. Cannizzo, G. Chincarini,E. Fenimore, C. Kouveliotou, P. O’Brien, D. M. Palmer, J. Racusin, P. W. A.Roming, T. Sakamoto, J. Tueller, R. A. M. J. Wijers, and B. Zhang. Corre-lations of prompt and afterglow emission in swift long and short gamma-raybursts. ApJ, 689:1161-1172, December 2008. doi: 10.1086/592766.

[74] B. Zhang, Y. Z. Fan, J. Dyks, S. Kobayashi, P. Mészáros, D. N. Burrows, J. A.Nousek, and N. Gehrels. Physical Processes Shaping Gamma-Ray Burst X-Ray Afterglow Light Curves: Theoretical Implications from the Swift X-RayTelescope Observations. ApJ, 642:354–370, May 2006. doi: 10.1086/500723.

[75] S. D. Barthelmy, J. K. Cannizzo, N. Gehrels, G. Cusumano, V. Mangano,P. T. O’Brien, S. Vaughan, B. Zhang, D. N. Burrows, S. Campana, G. Chin-carini, M. R. Goad, C. Kouveliotou, P. Kumar, P. Mészáros, J. A. Nousek,J. P. Osborne, A. Panaitescu, J. N. Reeves, T. Sakamoto, G. Tagliaferri, andR. A. M. J. Wijers. Discovery of an Afterglow Extension of the PromptPhase of Two Gamma-Ray Bursts Observed by Swift. ApJL, 635:L133–L136,December 2005. doi: 10.1086/499432.

[76] H. van Eerten. Self-similar relativistic blast waves with energy injection.MNRAS, 442:3495–3510, August 2014. doi: 10.1093/mnras/stu1025.

[77] D. N. Burrows, A. Falcone, G. Chincarini, D. Morris, P. Romano, J. E. Hill,O. Godet, A. Moretti, H. Krimm, J. P. Osborne, J. Racusin, V. Mangano,K. Page, M. Perri, M. Stroh, and Swift XRT Team. X-ray flares in early grbafterglows. Philosophical Transactions of the Royal Society of London SeriesA, 365:1213–1226, May 2007. doi: 10.1098/rsta.2006.1970.

44 BIBLIOGRAPHY

[78] P. A. Evans, A. P. Beardmore, K. L. Page, J. P. Osborne, P. T. O’Brien,R. Willingale, R. L. C. Starling, D. N. Burrows, O. Godet, L. Vetere,J. Racusin, M. R. Goad, K. Wiersema, L. Angelini, M. Capalbi, G. Chincar-ini, N. Gehrels, J. A. Kennea, R. Margutti, D. C. Morris, C. J. Mountford,C. Pagani, M. Perri, P. Romano, and N. Tanvir. Methods and results of anautomatic analysis of a complete sample of swift-xrt observations of grbs. MN-RAS, 397:1177–1201, August 2009. doi: 10.1111/j.1365-2966.2009.14913.x.

[79] H. J. van Eerten, Z. Meliani, R. A. M. J. Wijers, and R. Keppens. Jetsimulations and gamma-ray burst afterglow jet breaks. MNRAS, 410:2016–2024, January 2011. doi: 10.1111/j.1365-2966.2010.17582.x.

[80] H. J. van Eerten and A. I. MacFadyen. Observational Implications of Gamma-Ray Burst Afterglow Jet Simulations and Numerical Light Curve Calcula-tions. ApJ, 751:155, June 2012. doi: 10.1088/0004-637X/751/2/155.

[81] K. L. Page, R. L. C. Starling, G. Fitzpatrick, S. B. Pandey, J. P. Osborne,P. Schady, S. McBreen, S. Campana, T. N. Ukwatta, C. Pagani, A. P. Beard-more, and P. A. Evans. GRB 090618: detection of thermal X-ray emissionfrom a bright gamma-ray burst. MNRAS, 416:2078–2089, September 2011.doi: 10.1111/j.1365-2966.2011.19183.x.

[82] R. L. C. Starling, K. Wiersema, A. J. Levan, T. Sakamoto, D. Bersier,P. Goldoni, S. R.

Thermal components in the early X-ray afterglows of Gamma...

Documents

Transcript of Thermal components in the early X-ray afterglows of Gamma...