The physics of gamma-ray bursts & relativistic jets
Introduction
This introduction to Gamma-Ray Bursts (GRBs) is meant to provide a brief summary of their main properties so that someone not interested in details can obtain a quick overview, in a few pages, of the main properties of these explosions from the reading of this introduction.
The serendipitous discovery of Gamma-Ray Bursts (GRBs) in the late sixties by the Vela satellites1[1] puzzled astronomers for several decades: GRBs are irregular pulses of gamma-ray radiation (typically lasting for less than one minute), with a non-thermal (broken power-law) spectrum peaking at 10–, and are seen a few times a day at random locations in the sky e.g. [2], [3], [4]. Their spectacular nature, detection at redshift larger than 9 with current generation of instruments, and their connection with supernovae explosions and possibly black-holes formation, have led to a great deal of time and effort invested to their study (e.g. [5], [6], [7], [8], [9], [10], [11], [12], [13]).
The histogram of GRB duration has two distinct peaks. One at 0.3 s and the other at about 30 s, which are separated by a dip at 2 s. Bursts with duration less than 2 s are classified as short-GRBs and those that last for more than 2 s are called long-GRBs. Based only on the two peaks in the duration distribution, and well before anything was known about the distance or physical origin of GRBs, it was suspected that these peaks correspond to two physically distinct progenitors. Recent observations have confirmed that long-GRBs are one possible outcome of the collapses of massive stars (mass ), and that at least some of the short-GRBs arise in the mergers of compact objects in binary systems (perhaps merger of two neutron stars or a neutron star and a black hole). The connection between the classifications based on burst duration and based on distinct physical origins turns out to be more complicated though, and is still not fully understood.
Distances to GRBs were completely uncertain until the launch of Compton-Gamma-Ray-Observatory (CGRO), from space shuttle Atlantis, on 5 April 1991 in a low earth orbit at 450 km (in order to avoid the Van Allen radiation belt that covers km altitude). It carried four instruments that provided a wide energy band coverage of 20 keV–30 GeV (at 17 tons, CGRO, was the heaviest astrophysical payload flown at that time). CGRO established that these bursts are isotropically distributed [4] and their number at the faint end (but well above the instrument threshold) deviates from the expected Euclidean count2 (e.g. [14], [15], [16]). These two discoveries taken together convinced most astronomers that GRBs are located at distances much larger than the size of the local group of galaxies.
The confirmation of the cosmological distance to GRBs was obtained in 1997, when the BeppoSAX satellite, launched on April 30, 1996, provided angular position of bursts to within 4 arc-minutes–more than a factor 20 improvement compared with the Compton Gamma-ray Observatory–which enabled optical and radio astronomers to search for counterparts for these explosions. A rapidly fading X-ray & optical emission (the “afterglow”) accompanying a GRB was found on February 28, 1997, about 8 hours after the detection of the burst, and that led to the determination of redshift for this GRB to be 0.695 [17], [18], [19]. This launched a new era in the study of GRBs which has led to a wealth of new information and a much deeper understanding of these enigmatic explosions (e.g. [20], [21], [22], [8], [9], [12], [13]).
From burst redshift and flux we know that GRBs radiate between 1048 and 1055 ergs, if isotropic. This means that GRBs are the most energetic explosions in the Universe; the luminosity of the brightest bursts rivaling that of the entire Universe at all wavelengths albeit for only a few seconds [23].
Our understanding of GRBs has improved enormously in the last 18 years due to the observations made by several dedicated -ray/X-ray satellites (BeppoSAX, KONUS/Wind, HETE-2, Swift, Integral, AGILE, Fermi) and the follow-up observations carried out by numerous ground-based optical, IR, mm and radio observatories. Much of this progress has been made possible by the monitoring and theoretical modeling of long-lived afterglow emissions following the burst.
We know from breaks in optical & X-ray afterglow lightcurves that GRBs are highly beamed [24], [25], and the true amount of energy release in these explosions is – ergs [26], [27], [28], [29], [30], [31], [32].
The follow-up of GRBs at longer wavelengths (X-ray, and optical) has established that afterglow light-curves often decay as a power-law with time (–) and have a power-law spectrum (). The synchrotron radiation from the external forward-shock–which results from the interaction of GRB-ejecta with the circumburst medium [33], [34], [35], [36]–provides a good fit to the multi-wavelength afterglow data for GRBs (e.g. [37]).
In many cases, the decay of the optical or X-ray afterglow light-curve steepens to at ∼1 day after the burst. The most natural explanation for this steepening (foreseen by Rhoads [24]) is that GRB outflows are not spherical but collimated into narrow jets [25]. As the ejecta is decelerated and the strength of the relativistic beaming diminishes, the edge of the jet becomes visible to the observer. The finite angular extent of the ejecta leads to an achromatic faster decay of optical & X-ray lightcurves. This achromatic transition from a slower to a faster decay of lightcurves is called “jet-break”.
The initial opening angle of the jet and its kinetic energy can be obtained by modeling the broadband emission (radio to X-ray) of those GRB afterglows whose light-curve fall-off exhibited a jet-break. From these fits it is found that the opening angle of GRB jets is in the range of –10 degrees, thus the ejecta collimation reduces the required energy budget by a factor relative to the isotropic case; the true amount of energy release for most long duration GRB is found to be erg [24], [25], [26], [27], [28], [32]. The medium within ∼0.1 pc of the burst is found to have a uniform density in many cases, and the density is of the order of a few protons per [37]. This is a surprising result in the light of the evidence that long duration GRBs are produced in the collapse of a massive star–as suggested by Woosley [38], Paczynski [39], MacFadyen and Woosley [40]–where we expect the density to decrease with distance from the center as due to the wind from the progenitor star [41], [42], [43], [44].
It was expected from theoretical considerations that GRB outflows are highly relativistic (e.g. [45], [46], [47], [6]). A direct observational confirmation of this was provided by measurements of radio scintillation for GRB 970508 [48], [20], and “superluminal” motion of the radio afterglow of a relatively nearby burst GRB 030329 [49] where the blastwaves were found to be still mildly relativistic several weeks after the explosion.
The evidence for association of long-duration GRBs (those lasting for more than 2 s) with core collapse SNa comes from two different kinds of observations: (i) GRBs are typically found to be in star forming regions of their host galaxies (e.g. [50], [51], [52], [53]); (ii) For several GRBs, Type Ic supernovae have been detected spectroscopically associated with the GRBs. Most of the SNe-associated GRBs have luminosity significantly lower than typical GRBs,3 e.g. GRB 980425 [56], 030329 [57], [58], 060218 [59], [60], [61], 100316D [62], [63], 101219B [64], and 120422A [65]. However, two nearby high-luminosity GRBs, i.e. 031203 [66] and 130427A [67], [68], are also found to be associated with Type Ic SNe. Additionally, a subset of about a dozen GRBs show at late-times (∼10 days) SNa-like “bump” in the optical data and simultaneously a change in color that is inconsistent with synchrotron emission, and suggests that optical flux from the underlying supernova is starting to overtake the GRB afterglow flux [22], [10].
Significant progress toward answering the long standing question regarding the nature of short duration GRBs (those lasting for less than 2 s) was made possible by the Swift satellite’s more accurate localization of these bursts (3 arcmin vs. a few degrees for Compton-GRO). This led to the discovery that a fraction of these bursts are located in elliptical galaxies, i.e. associated with older stellar population, and were found to be on average less energetic and at a lower redshift [69], [70], [71], [72], [73], [74], [75], [76]. These observations are consistent with the old idea that these bursts originate from neutron star mergers [77], [78]. However, there is no conclusive proof for this model as yet.
The Swift satellite, designed for the study of GRBs and launched in November 2004, has X-ray and UV-optical telescopes on board and provides localization of bursts to within 3 arcminutes. When Swift’s gamma-ray telescope (Burst and Altert Telescope or BAT) detects a burst, the X-ray Telescope (XRT) and the UV-Optical Telescope (UVOT) on board Swift quickly slew to the GRB position within 60–100 s to observe the target, which provides excellent coverage of the transition from the prompt -ray phase to the lower-frequency afterglow emission phase.4 Swift has provided a wealth of puzzling observations [79], [80], [81], and revealed that a variety of physical processes shape the early X-ray afterglow lightcurves [82]. Its XRT has found that for about 50% of GRBs the X-ray flux decays very rapidly after the burst (), followed by a plateau during which the X-ray afterglow flux decrease is much slower () than expected in the standard forward-shock model. The former feature indicates that the -ray prompt radiation and afterglows are produced by two different mechanisms or arise from different outflows while the latter perhaps suggests that the forward shock that powers the afterglow takes a long time (of order several hours) to become a self-similar blast wave with constant energy (another possibility is that the observed X-ray radiation is not produced in the external shock).
Swift has also discovered episodes of a sharp increase in the X-ray flux (flares) minutes to hours after the end of the GRB [83], [84], [85], [86]. The rapid rise time for the X-ray flux, with , rules out the possibility that flares are produced as a result of inhomogeneity in the circumstellar medium where the curvature of the relativistic shock front limits or [87], [88], [89]. This suggests that the central engine in these explosions is active for a time period much longer than the burst duration5 [83], [82], [92], [93].
While the X-ray and optical data for s (time measured from -ray trigger) are consistent with external forward shock emission, the features seen in the X-ray data prior to s are not well understood. Similarly the expected achromatic breaks in the lightcurves (associated with finite jet angle) are seen in some bursts but not others [94], [95], [96], [97], [30], [29], [31].
One of the foremost unanswered questions about GRBs is the physical mechanism by which prompt -rays–the radiation that triggers detectors on board GRB satellites–are produced. Is the mechanism the popular internal shock model6 [98], the external shock model, or something entirely different? Are -ray photons generated via the synchrotron process or inverse-Compton process, or by a different mechanism? Answers to these questions will help us address some of the most important unsolved problems in GRBs — how is the explosion powered in these bursts? Does the relativistic jet produced in these explosions consist of ordinary baryonic matter, electron–positron pairs, or is the energy primarily in magnetic fields?
The Fermi satellite, a multi-purpose high energy satellite launched in June 2008, has provided useful data extending from to to help answer some of these questions. It has made several important discoveries regarding GRBs [99], [100], [101], [102], [103]: (1) in most cases the high energy photons () are detected with a delay of a few seconds with respect to the lower energy emission (); (2) high energy emission lasts for a time period much longer () than emission below (which lasts for less than 1 min for most GRBs); (3) the broad-band prompt -ray spectra are found in most cases to consist of one peak and power law functions with different indices at low and high energies with a smooth transition from one to the other over a factor in frequency (this is the so called “Band” spectrum), however in a few cases the spectrum has an addition component.
There are several different lines of strong evidence suggesting that the high energy photons () we observe after the prompt phase () are produced in the external forward shock via the synchrotron process [104], [105]. On the other hand the origin of prompt -ray emission, low and high energies, remains a puzzle. Some of the proposed models are: synchrotron and inverse-Compton (IC) radiation processes in internal or external shocks or at sites where magnetic field in Poynting jet is dissipated (e.g. [33], [106], [107], [108]); and photospheric radiation with contribution from multiple IC scatterings (e.g. [109], [110], [111], [112], [113], [114], [115], [116], [117], [118], [119], [120], [121], [55]).
Swift satellite has found GRBs at high redshifts; the highest redshift GRB discovered to date is at when the universe was just 0.52 billion years old or 3.8% of its current age [122]. Swift is capable of detecting bursts of similar intrinsic brightness up to redshift of about 15. Because of their intrinsically simple spectrum and extremely high luminosity, GRBs are expected to offer a unique probe of the end of cosmic dark age when the first stars and galaxies were forming.
This review is organized in the order of the best understood GRB properties discussed first and the least well understood phenomena described last. We start with a brief review of radiation physics, and describe the theory of GRB afterglows which began to be developed even before the first detection of afterglow radiation. We then describe how well the afterglow theory does when confronted with observations. We first consider the late time afterglow observations (these observations starting from roughly half a day after the explosion can last for weeks to months), and what they have taught us about GRBs and the medium in their vicinity. This is followed by early afterglow observation–starting from ∼30 s (2 s) since the burst trigger for long (short) GRBs and spanning a duration of a few hours–and our current understanding of the puzzles they pose. Then the least well understood of all the data–properties of the GRB prompt radiation–and the strengths and weaknesses of various models proposed to explain these observations are reviewed. Next, we take up the properties of the central engine, and describe the two leading models: a new-born hyper-accreting black hole, and a strongly magnetized, rapidly spinning neutron star (magnetar). We then move on to discuss the possible progenitors of GRBs. We also devote a section to discussion of possible neutrino emission from GRBs.
Section snippets
Radiative processes
We provide in this section a brief overview of a few of the most important radiative processes in GRBs which will be used extensively in this review. There are excellent books that cover this topic in detail such as the monograph by Rybicki and Lightman [123], and books on high energy astrophysics by Longair [124], Krolik [125], Dermer and Menon [126], Kulsrud [127]. This section is no substitute for the extensive coverage of this topic provided in these books. The purpose here is to
Relativistic shocks: basic scalings
One important piece of the GRB theory is a “generic” model that does not depend on the details of the central engine. This is a relativistic blastwave theory that describes interaction between the “fireball”–which moves with Lorentz factor before deceleration & has total “isotropic equivalent” energy –and the circumburst medium (CBM) described by the density profile, . Such a fireball–CBM interaction is inevitable for any type of energetic explosion. A power-law decaying
Afterglow observations and interpretations
Broadband GRB afterglows were predicted before their discoveries [34], [273], [36]. Shortly after the publication of the seminal paper by Mészáros and Rees [36] which provided detailed predictions for the broad-band afterglow based on the external shock model, the first X-ray and optical afterglows were discovered for GRB 970228 [17], [19], and the first radio afterglow was discovered for GRB 970508 [20]. Since then, regular follow-up observations of GRBs have been carried out, and a large
Collisionless shock properties from GRB afterglow observations
GRB afterglows provide a good laboratory for the study of relativistic collisionless shocks. In spite of many years of theoretical work several basic questions regarding collisionless shock remain unanswered. Perhaps foremost amongst these questions are generation of magnetic fields down/up stream of the shock front (), particle acceleration () and the fraction of energy of shocked plasma that is given to electrons (). The calculation of synchrotron radiation from shocked fluid requires
Temporal properties
Observationally, the prompt emission phase of a GRB is conventionally defined as the temporal phase during which sub-MeV emission is detected by the GRB triggering detectors above the background level. Quantitatively, the duration of a burst is defined by the so-called “”: the time interval between the epochs when and of the total fluence is registered by the detector. Such an observation-based definition has some limitations: 1. It depends on the energy band of the detector. A
Progress toward understanding GRB prompt radiation
The origin of the prompt -ray emission from GRBs is not well understood. This is due in large part to our lack of knowledge of jet composition, energy dissipation and particle acceleration mechanisms. A widely used model is the matter-dominated “fireball”, which consists of baryons (primarily protons and neutrons), electron & positron pairs, and photons. A fireball could be produced in cataclysmic events such as mergers of binary neutron stars [78] or collapses of massive stars [38]. The
GRB central engine
Although the progenitors of long- and short- GRBs might be very different objects, the basic nature of the central engine–the mechanism by which highly luminous relativistic jet is produced–is expected to be similar for these bursts. The details of the process can be somewhat different. The discussion below mostly focuses on long duration GRBs, but short GRBs will be discussed whenever noticeable differences with long GRBs exist. For a more detailed discussions of short GRB central engine,
Two physically distinct types of GRBs
Gamma-ray observations led to identification of two phenomenological classes of GRBs in the duration-hardness () plane: long/soft vs. short/hard [3]. The boundary between the two classes is vague. The duration separation line is around 2 s in the BATSE band (30 keV–2 MeV). Long and short GRBs roughly comprise 3/4 and 1/4 of the total population of the BATSE sample, but the short GRB fraction is smaller for other detectors [411], [412], [413], [414], [415]. This is because the duration
High energy neutrinos from GRBs
As energetic, non-thermal photon emitters, GRBs are believed to be efficient cosmic ray accelerators as well. The standard scenario invokes first order Fermi acceleration mechanism in relativistic shocks, both in internal shocks and the external (forward and reverse) shocks. Alternatively, magnetic reconnection can also accelerate cosmic rays to high energies.
The maximum proton energy can reach the ultra-high energy (UHE) range [831], [832], [833]. The maximum energy of the shock accelerated
GRBs from the first stars (pop III stars) and their use for investigating the high redshift universe
The universe was essentially devoid of stars until the redshift of –20, when the first stars were born, and the strong UV radiation from them contributed to the reionization of the universe, and bringing to an end the cosmic dark age (e.g. [859], [860], [861], [862]). A fraction of these stars likely ended their life as GRBs. In this section, we describe how GRBs can be used to study the end of the cosmic dark ages.
According to the cold dark matter (CDM) paradigm of hierarchical structure
Concluding thoughts and future prospects
It has been a long wild ride for people working on gamma-ray bursts to figure out the true nature and origin of these cosmic explosions. We provide a brief summary of things we have learned, and the questions still unanswered, regarding these powerful transient events.
What are the things we know with confidence?
The distance to these events is well established from afterglow observations, and hence, we know the isotropic equivalent of energy release in -rays. The mean redshift for long duration
Acknowledgments
PK thanks Rodolfo Barniol Duran, Jonathan Granot, Milos Milosavljević, Ramesh Narayan, Tsvi Piran, Craig Wheeler for numerous discussions. He is especially indebted to Alin Panaitescu for nearly a decade long collaboration, and countless illuminating discussions.
We are grateful to Tsvi Piran and his group at the Hebrew university for carefully reading a draft of this review and discussing it over a period of about two months, and for providing extensive feedback that helped improve this work.
References (894)
Gamma-ray bursts and the fireball model
Physics Reports
(1999)Radio scintillation of gamma-ray-burst afterglows
New Astron.
(1997)Short-hard gamma-ray bursts
Physics Reports
(2007)- et al.
Observations of Gamma-Ray Bursts of Cosmic Origin
ApJ
(1973) - et al.
BATSE observations of gamma-ray burst spectra. I - Spectral diversity
ApJ
(1993) - et al.
Identification of two classes of gamma-ray bursts
ApJ
(1993) - et al.
Spatial distribution of gamma-ray bursts observed by BATSE
Nature
(1992) - et al.
Gamma-ray bursts
Ann. Rev. Astron. Astrophys.
(1995) Theories of Gamma-Ray Bursts
Ann. Rev. Astron. Astrophys.
(2002)- et al.
Gamma-Ray Bursts: progress problems prospects
Internat. J. Modern Phys. A
(2004)