Gamma-Ray Astronomy E N CYC LO PE D IA O F AS T R O N O MY AN D AS T R O PHYS I C S
Gamma-Ray Astronomy
Gamma-ray photons are the most energetic ones of the
whole electromagnetic spectrum. Like all photons with
wavelengths shorter than 10 nm, they are absorbed by
the Earth’s atmosphere. Detection of γ -ray emission
from space requires then the use of stratospheric balloons,
rockets and satellites, which became feasible only with the
advent of the space era.
The borderline between the spectral regions of x-rays
and γ -rays is sometimes placed at an energy of E ∼
30 keV (wavelength λ ∼ 4.13 × 10−9
cm or a frequency
of ν ∼ 7.2 × 1018
Hz). At lower energies, radiation from
astrophysical sources is dominated by thermal emission
(i.e. resulting from radiation in equilibrium with matter,
in optically thick media), while at higher energies nonthermal
emission prevails. Alternatively, the x–γ -ray
borderline may be placed at 511 keV (the energy equivalent
of the electron rest mass); above this ‘limit’ electrons
behave relativistically. In any case, the use of large
photon collectors becomes impossible above a few tens
of keV: photons have wavelengths smaller than typical
interatomic distances in solids and cannot be reflected by
any kind of mirror (contrary to what happens at longer
wavelengths).
The γ -ray domain extends over 12 decades in energy,
as much as all other spectral regions put together.
However, above a few tens of GeV, the number of photons
received from even the brightest sources is so small
that their observation during the limited lifetime of a
satellite is no longer possible (see VERY-HIGH-ENERGY GAMMARAY
SOURCES). Those very high energy (VHE) photons
are detected through their interaction with the Earth’s
atmosphere, producing cascades of secondary particles.
Astronomy in the γ -ray domain faces several serious
handicaps.
• The impossibility of using large collectors to focus the
γ -ray photons implies that the detector itself should be
as large as possible.
• A detector placed outside the Earth’s atmosphere is
constantly bombarded by energetic cosmic ray particles,
coming from the outer space or trapped in the Van Allen
radiation belts. The interactions of these particles with
the detector induces very high backgrounds, which are
difficult to attenuate with passive or active shielding. As
a result, the sensitivity of γ -ray detectors increases only
as the square root of their collecting area.
• Astrophysical sources emit generally very small numbers
of γ -ray photons. Statistically significant detections
then require very long exposures, which are not
always compatible with the generally short lifetime (a
few years) of an observatory in space.
These difficulties explain why γ -ray astronomy is
the last window of the electromagnetic spectrum to be
opened to our observation of the universe. However, it
has undergone a spectacular development in the 1990s
and has grown to a mature astrophysical discipline. It
offers information that is impossible to obtain in other
wavelengths, allowing us to probe the close surroundings
of a black hole or the interiors of a supernova, to identify
nuclear isotopes and to obtain clues to the origin and
acceleration of cosmic rays, either in our Galaxy or in the
very active central regions of remote galaxies.
A short history of γ -ray astronomy
The opening of the γ -ray window was preceded by a
relatively long period of theoretical studies concerning
the γ -emissivity of potential astrophysical sources. Most
of these ideas were put forward by scientists working in
fields other than astronomy (particle and nuclear physics
and cosmic ray physics). Following the detection of
atomic hydrogen in the Galaxy (through its characteristic
21 cm line; see RADIOASTRONOMY) Hayakawa realized in
1952 that the interaction of COSMIC RAYS with the hydrogen
atoms of the interstellar medium should give rise to γ -ray
emission. Six years later, Morisson suggested two known
radiosources (the CRAB NEBULA and CYGNUS A) as possible
γ -ray emitters. Thus, theoreticians realized quite early
that relativistic particles could generate photon emission
in both extremes of the electromagnetic spectrum.
Progress in observational γ -ray astronomy has been
too slow to test those ideas, because of the smallness of
early detectors and underestimated background fluxes.
The first γ -ray emission of extraterrestrial origin was
detected during the solar flare of March 1958. However, it
was only 10 yr later that γ -ray emitters outside our solar
system were detected: the US satellite OSO-3 (launched
in 1961) recorded a few hundred photons of energy
>50 MeV from the inner Galaxy, while a balloon-borne
experiment detected photons in the 100–300 keV range
from the direction of the Crab nebula. A few months later,
radioastronomers discovered the CRAB PULSAR; subsequent
analysis of the γ -emission from the Crab revealed a
periodicity similar to that in the radio band. During a
solar flare in August 1972, a spectrometer aboard the US
satellite OSO-7 recorded the first γ -ray spectrum of the
Sun, revealing several lines of nuclear origin, as well as
the electron–positron annihilation line at 511 keV. Also,
the following year a US team reported the detection
of γ -ray bursts of extraterrestrial origin by the military
satellites of the Vela series (launched to monitor the Earth’s
atmosphere for nuclear explosions).
After these early discoveries, the γ -ray sky was
explored in some detail in the 1970s by the US SAS-
2 and the European COS-B satellites, which provided
the first high-energy images of the Milky Way, detailed
observations of the first known γ -ray pulsars and the
first detection of extragalactic γ -ray sources. Extrasolar
γ -ray spectroscopy started in the same period, with
the discovery of the electron–positron annihilation line
from the galactic centre direction by a balloon-borne
spectrometer. This discovery was confirmed in the
early 1980s by a γ spectrometer aboard the US HEAO-3
satellite, which discovered also the first extrasolar γ -ray
Copyright © Nature Publishing Group 2001
Brunel Road, Houndmills, Basingstoke, Hampshire, RG21 6XS, UK Registered No. 785998
and Institute of Physics Publishing 2001
Dirac House, Temple Back, Bristol, BS1 6BE, UK 1
Gamma-Ray Astronomy E N CYC LO PE D IA O F AS T R O N O MY AN D AS T R O PHYS I C S
line emission of nuclear origin, namely the 1.8 MeV line of
the radioactive nucleus 26
Al.
The pace of discoveries slowed down in the 1980s, as a
result mainly of the lack of dedicated γ -ray satellites. The
explosion of the supernova SN1987A in the nearby galaxy
of the Large Magellanic Cloud on February 23 1987 offered
the γ -ray community an unprecedented opportunity: the
SMM (Solar Maximum Mission) satellite, as well as several
balloon borne instruments, detected γ -ray lines from
the decay of radioactive 56
Co (predicted long ago to be
the source of power for the late optical light curves of
supernovae). Finally, the early 1990s constitute the golden
era of γ -ray astronomy, thanks to the discoveries of the
Soviet–French SIGMA mission (see GRANAT) and the US
COMPTON GAMMA RAY OBSERVATORY (CGRO).
Emission processes of γ -ray photons
Nature creates γ -ray photons in a variety of ways. If the
temperature of the environment is high enough (>108
K),
γ -photons are produced as thermal emission of the
astrophysical PLASMA and they generally have a blackbody
spectrum. The interiors of stars in their advanced
evolutionary stages reach such high temperatures, but
they cannot be observed, because the star’s dense material
makes it opaque to photons of all wavelengths.
The close neighborhood of a strongly gravitating
body, such as a black hole accreting matter from a
companion star, may also reach high temperatures, as the
gravitational energy of accreted matter turns into heat.
The detection of γ -ray photons produced in this way is
a unique tool for studying the interactions of compact
objects with their environment as well as their energetics.
However, astrophysical γ -rays are mainly produced
by various non-thermal processes, involving in general
the interactions of high-energy particles. For instance,
high-energy electrons radiate γ -rays by interacting
electromagnetically with nuclei, photons or intense
magnetic fields; the corresponding processes are known
as bremsstrahlung, inverse Compton emission and
synchroton emission, respectively, and they can obviously
take place also in a high-temperature plasma. In addition
nuclear interactions of protons with energies higher than
∼1.2 GeV produce secondary unstable particles, pions and
mesons; γ -ray photons are produced by the decay of some
of these particles, the most prolific producers being the π0
pions. The resulting γ -ray spectrum has a characteristic
maximum around 70 MeV (corresponding to half the restmass
energy of π0
).
All these thermal and non-thermal processes give rise
to the emission of γ -rays with a continuous spectrum of
energy. There are also processes producing γ -rays with
discrete energies, involving in general the de-excitation
of atomic nuclei. Just as the more familiar optical lines
are caused by the electron transitions in atoms, nuclear
lines result from the much more energetic transitions
(typically, in the MeV range) between energy levels of
the atomic nuclei. This makes γ -ray spectroscopy an
invaluable diagnostic tool for nuclear astrophysics, since
it allows for an unambiguous identification of isotopic
species (radiation originating from electrons enables one
in general, to identify chemical elements, but not isotopes).
The excited nuclear states are populated either
through the decay of an unstable nucleus or through highenergy
nuclear interactions. In the former case, we obtain
information about the synthesis of radioactive nuclei in
various sites (mostly explosive ones such as supernovae)
and the physics of these sites. In the latter case, we
learn about the nature of the energetic particles (e.g.
their composition, spectrum, acceleration mechanism etc).
γ -ray lines from the de-excitation of nuclei are routinely
observed in the Sun’s photosphere during solar eruptions;
in some cases, the nuclei are produced through neutron
captures, in a medium sufficiently dense to favor these
reactions, but also sufficiently thin to allow γ -rays to
escape.
Finally, the annihilation of an electron and a positron
also gives rise to emission in the γ -ray domain. Positrons
may be produced in astrophysical sites through the decay
of π+
pions or through the β+
decay of radioactive
nuclei. Depending on the conditions they may annihilate
directly with electrons or combine first with them to form
positronium. In the former case (at high temperatures,
T >106
K) two γ -ray photons of energy 0.511 MeV are
produced. In the latter case, the resulting emission
depends on the state of positronium: in 25% of the cases it
is formed in the para (singlet, 1
S0) state and two photons of
0.511MeVareagainemitted; in75%ofthecasesitisformed
in the ortho (triplet, 3
S1) state and a continuous spectrum
of three photons, each with an energy of ≤0.511 MeV, is
then produced.
The γ -ray sky
The poor resolving power of current γ -ray instruments is
one of the main problems in γ -ray astronomy (see GAMMARAY
TELESCOPES). Observations in this energy range usually
define ‘error boxes’, regions of the sky where the source
most probably lies. Instruments launched in the 1990s
provide error boxes of ∼1 arc min at low energies (the
SIGMA telescope) and ∼5—10 arc min at higher energies
(the EGRET instrument aboard CGRO), to be compared
with, for instance, ∼0.1 arc sec in the visible. Identifying
a source in these conditions is much more difficult than
at other wavelengths. In the case of a variable source,
the time-scale of variability sets an upper limit to the
dimensions of the source (which cannot be larger than the
distance light travels during this time-scale).
The sources of the γ -ray sky are extremely varied:
some of them appear as ‘points’ (i.e. with angular
dimensions smaller than the resolution of γ -detectors) and
others as extended sources, both in our Galaxy and in the
remote universe.
Stars on the main sequence (i.e. burning hydrogen
in their cores) do not, in principle, radiate γ -rays and
only the Sun has been observed to do so sporadically.
During their advanced evolutionary stages, some stars
may eject radioactive isotopes through their powerful
Copyright © Nature Publishing Group 2001
Brunel Road, Houndmills, Basingstoke, Hampshire, RG21 6XS, UK Registered No. 785998
and Institute of Physics Publishing 2001
Dirac House, Temple Back, Bristol, BS1 6BE, UK 2
Gamma-Ray Astronomy E N CYC LO PE D IA O F AS T R O N O MY AN D AS T R O PHYS I C S
stellar winds and become γ -ray emitters (red giants,
Wolf–Rayet stars). However, radioactivity is mostly
produced during stellar explosions (supernovae and, to
a lesser extent, novae). On the other hand, stellar
remnants may produce copious amounts of γ -rays in
several cases: rotating neutron stars may convert a fraction
of their rotational energy into high-energy photons (γ -ray
pulsars), while accreting black holes convert the energy
of accreted matter into low-energy γ -rays in their vicinity
and present a steady or variable emission. Some of these
objects produce jets of high-energy particles which interact
with their environment to produce γ -rays. Finally, the
emission mechanism of the mysterious γ -ray bursts is not
understood yet, although it—most probably—involves
compact objects (neutron stars and black holes).
The Sun
Since the first detection of its γ -ray emission, in 1958,
the Sun has been regularly monitored by US, Soviet and
Japanese satellites (OSO, HEAO, Prognoz, SMM, Hinotori,
Yohkoh, CGRO). γ -rays are emitted continuously, but
particularly during a brief period of intense activity,
associated with the 11 yr cycle of solar flares (see SOLAR
FLARES: GAMMA RAYS). During this period the energy stored
in unstable magnetic configurations of the solar surface
layers is suddenly liberated; particles are accelerated to
high energies and their interactions produce γ -rays. The
study of the γ -ray emission provides information on
the acceleration sites, their physical conditions and their
composition.
The low-energy part of the continuum γ -ray
spectrum of the Sun results from bremsstrahlung
emission of accelerated electrons interacting with the solar
atmosphere. At higher energies (100 MeV) part of the
emission is due to the decay of π0
s produced by the
interactions of high-energy accelerated protons.
The de-excitation of atomic nuclei (excited by highenergy
collisions) gives rise to prompt γ -ray lines, which
may be narrow (when ambient nuclei are excited by
energetic protons and α particles) or broad (when fast
nuclei are excited by ambient protons and α particles).
The most intense narrow lines are those at 4.438 MeV
and 6.129 MeV, resulting from the de-excitation of the
abundant 12
C and 16
O nuclei, respectively; the lines
of several other nuclei (15
N, 20
Ne, 24
Mg, 28
Si and
56
Fe) have also been detected (figure 1). The relative
intensities of these lines provide information on the
isotopic composition of the Sun’s surface.
The most intense solar γ -ray line, at an energy of
2.223 MeV, results from the de-excitation of deuterium
nuclei (D), produced by neutron captures on protons; since
free neutrons are unstable (they decay to protons within
10 min) the ones involved in the formation of deuterium
are ejected by accelerated heavy nuclei. The ejected
neutrons have to slow down to the thermal velocities of
the ambient medium before being captured by the protons
and thus they give rise to delayed γ -ray line emission.
The delay time (about 100 s) provides information on the
Figure 1. γ -ray spectrum of a solar flare modeled by Murphy
et al (1991).
ambient density (∼1017
particles cm−3
) and the location of
the emission site in the solar photosphere (at a depth of
∼300 km).
The second most intense line observed in the
solar γ -ray spectrum results from the annihilation of
positrons with electrons. Positrons are produced by
β+
decay of short-lived radioactive nuclei and of π+
pions (both produced in high-energy interactions). Before
annihilating, positrons have to slow down and their γ -ray
emission is delayed (as in the case of neutrons).
Stellar explosions and remnants
Radioactive nuclei are produced by thermonuclear
reactions in high-density environments, which are opaque
to γ -rays. The photons of the radioactive decays interact
with the ambient electrons and lose energy via Compton
scattering; the medium is heated and the energy is
ultimately radiated at longer wavelengths. Only in the
case of a site violently expanding after an explosion
(SUPERNOVAE, NOVAE) or suffering mass loss through a stellar
wind (WOLF–RAYET STARS or stars in the asymptotic giant
branch stage) can the γ -ray photons escape, since the
opacity decreases with decreasing density.
The characteristic energy of escaping γ -ray lines
enables one to identify isotopic species, while their
intensity provides information on the amounts that have
been synthesized and on the physical conditions of the
corresponding stellar zones. The shape of the lines may
give information on the structure (velocity and density
profiles) of the stellar ejecta. These properties make
γ -ray spectroscopy a unique tool for nuclear and stellar
astrophysics.
These theoretical ideas, put forward mostly by
D D Clayton in the 1960s, were spectacularly confirmed
Copyright © Nature Publishing Group 2001
Brunel Road, Houndmills, Basingstoke, Hampshire, RG21 6XS, UK Registered No. 785998
and Institute of Physics Publishing 2001
Dirac House, Temple Back, Bristol, BS1 6BE, UK 3
Gamma-Ray Astronomy E N CYC LO PE D IA O F AS T R O N O MY AN D AS T R O PHYS I C S
in 1987: the 0.847 MeV and 1.238 MeV lines of the decay
of 56
Co were observed in SUPERNOVA 1987A, a massive star
(20M ) that exploded in the nearby galaxy of the Large
Magellanic Cloud. Their detection confirmed that 56
Fe, the
most strongly bound stable nucleus in nature, is produced
in the form of unstable 56
Ni (through the decay chain 56
Ni
−→ 56
Co −→ 56
Fe). It also confirmed that the observed late
optical luminosity of supernovae (decreasing by a factor
of ∼2 every 75 days) is due to the energy released by the
radioactive decay of 56
Co (which has a half-life of 77 days).
Finally, the appearance of the γ -ray lines about 6 months
earlier than expected from theoretical models suggested
that some of the 56
Ni nuclei, produced in the dense inner
zones of the star, had been mixed into the outer layers.
The extraordinary opportunity offered by SN1987A
was due to its proximity; indeed, it was the first supernova
visible to the naked eye in the past four centuries,
since supernovae are mostly detected in distant galaxies.
Moreover, theory suggests that exploding massive stars
(characterized as supernovae of type II or SNII) produce
moderate amounts of 56
Ni (∼0.1M ) and remain opaque
to γ -rays for a long time (∼1 yr, i.e. until the radioactivity
of 56
Co has dropped to low levels). Supernovae of
another class, SNIa, are much brighter γ -ray line emitters.
They are white dwarfs in binary systems, undergoing
a thermonuclear explosion, which is triggered by the
accretion of mass from their companion star; they produce
∼10 times more 56
Ni and become transparent to γ -rays
much earlier than SNII, because of their smaller envelope
mass. Unfortunately, observations show that SNIa are
less frequent than SNII and their occurrence in a closeby
galaxy is unexpected, on statistical grounds. The first
SNIa to be (marginally) detected in the light of the 56
Co
lines was SN1991T in the Virgo cluster of galaxies, at an
estimated distance of 55 million light- years (by CGRO).
Future γ -ray line observations of extragalactic SNIa will
provide important information on the physics of those
poorly understood objects.
Long-lived radioactive nuclei can emit characteristic
γ -ray lines hundreds or thousands of years after the
supernova explosion, while they are evolving in an
environment transparent to γ -rays. The 1.156 line of 44
Ti
(half-life 60 yr) has been detected by CGRO in the remnant
of Cas-A, presumably a massive star that exploded about
300 yr ago, at a distance of 7000 light-years from the
Earth. The 44
Ti emission may help to reveal remnants
of recent supernova explosions that are unobserved in
otherwavelengths, becauseofobscurationbysurrounding
material (i.e. dense molecular clouds or layers of dust).
Finally, for tens of thousands of years after the
explosion the SUPERNOVA REMNANT may radiate high-energy
γ -rays, produced by accelerated electrons and protons
interacting with the ambient medium. The Crab remnant,
observed since the late 1960s, as well as five other sources
detected by CGRO, belongs to this class of γ -ray emitters.
γ -ray pulsars
PULSARS are rapidly rotating neutron stars, first detected in
the late 1960s through their periodic radioemission. More
than 550 radio pulsars are known to day, with periods
ranging from a few ms to several seconds. Before the
launch of CGRO, only the Crab and VELA PULSARS were
known to emit periodically in the γ -ray domain. Since
then five more objects have been added to the list and all
of them (except Geminga) emit also in the radiofrequencies
(figure 2).
The fact that only a small fraction of the radio
pulsars emits in γ -rays suggests that a special environment
is required for γ -rays to be produced. The required
energy is presumably provided by the rotational energy
of the neutron star, which slows down in the process;
however, the emission mechanism as well as the role
of the magnetic field and of the pulsar environment are
not well understood. A hint may lie in the fact that
the younger pulsars are more active in the low-energy
(∼1 MeV) domain than the older ones, which have in
general weaker magnetic fields and emit essentially in the
>100 MeV range.
GEMINGA is the only known pulsar with no detectable
radio emission. Many years after its detection in γ -rays
(by SAS-2), it represented an enigma, since it had not a
counterpart in any other wavelength. Its period (∼0.237 s)
was first measured in x-rays by the German satellite
ROSAT and confirmed in γ -rays by CGRO. Its distance,
evaluated in the early 1990s to ∼100 light-years, makes it
the closest known neutron star.
Accreting compact objects
NEUTRON STARS and BLACK HOLES are the end products of
the evolution of massive stars and they may be found in
binary systems (as happens with about half of the stars we
observe in the Galaxy). Matter from the companion star
may then be accreted onto the compact object, forming
around it an ACCRETION DISK. As matter spirals in, heat
is generated at the expense of gravitational energy and
the inner disk may reach high temperatures. According
to theory, disks around neutron stars may never reach
temperatures higher than T ∼ 108
K (corresponding to
thermal emission in the x-rays), whereas disks around
black holes may become much hotter (T ∼ 109
K) and
radiate in the low-energy γ -ray domain, up to a few
hundreds of keV. γ -ray observations would allow then
neutron stars to be distinguished from stellar black holes.
Observations with SIGMA and CGRO allowed a few
black hole candidates in our Galaxy to be identified, on the
basis of their γ -ray signature. Some of the sources show
steady emission, while others present a variable behaviour
(presumably as a result of variation in the accretion rate or
of instabilities developed in their accretion disk). Atypical
example of this class of transient sources (or γ -ray novae)
is the source GRS 1124-68, alias Nova Muscae. Detected
in January 1991, it has been observed in radio, x-rays and
γ -rays, its luminosity in the last of these energy ranges
exceeding that of the Crab (which is the strongest γ -ray
source in the sky).
An extremely important manifestation of variable
γ -ray activity is the excess emission detected in the 300–
600 keV range from the source 1E1740.7-2942, located
Copyright © Nature Publishing Group 2001
Brunel Road, Houndmills, Basingstoke, Hampshire, RG21 6XS, UK Registered No. 785998
and Institute of Physics Publishing 2001
Dirac House, Temple Back, Bristol, BS1 6BE, UK 4
Gamma-Ray Astronomy E N CYC LO PE D IA O F AS T R O N O MY AN D AS T R O PHYS I C S
Figure 2. The folded light curves (intensity as a function of time) of various γ -ray pulsars at different energies (from radiowaves to
γ -rays). A straight line means no known pulses. Explaining the variety of pulse shapes and energy dependences is a challenge to
theorists.
in the galactic center region; observed for 2 days in
October 1990 by SIGMA, this ‘paroxysmic’ activity has
been interpreted in terms of electron–positron annihilation
in the vicinity of a stellar mass black hole. Follow up radioobservations
of the source revealed radio-emitting jets,
presumably due to synchrotron emission of accelerated
electrons. This behaviour (compact ‘erupting’ γ -ray
source + radio jets moving at relativistic velocities) has also
been observed in a few other sites in our Galaxy. These
objects form the family of ‘micro-quasars’ (by analogy
with the much more powerful and distant QUASARS) and
represent ideal laboratories for the study of gravitational
and relativistic astrophysics.
Diffuse γ -ray emission in the galaxy
Some of the ‘products’ of stellar activity may pervade
the interstellar medium and give rise to diffuse γ -ray
emission. This is the case of cosmic rays (energetic
particles, acceleratedbysupernovaexplosions), long-lived
radionuclei (decaying away from their production sites)
and positrons.
Diffuse γ -ray emission from the Galaxy was first
detected in 1968 by OSO-7 and galactic maps were
provided by COS-B and CGRO. The emission correlates,
in general, with features of the galaxy known from
observations in other wavelengths: the molecular ‘ring’
at a galactocentric distance of 10 000–15 000 light-years,
the spiral arms or a few molecular clouds, i.e. sites of
rich gaseous content and active star formation. The
analysis of the local γ -ray spectrum shows that at low
energies(1–30 MeV)itresultsmainlyfrombremsstrahlung
radiation of cosmic ray electrons; at higher energies, the
emission of π0
s dominates. The LARGE MAGELLANIC CLOUD is
the only other normal galaxy where γ -ray emission of this
type has been detected.
During its million-year lifetime, radioactive 26
Al
from more than 10 000 Wolf–Rayet stars and supernova
explosions has been accumulated in the Galaxy’s disk. Its
characteristic 1.8 MeV line has been detected by HEAO-
3 in 1984 and mapped by CGRO as diffuse emission
along the galactic plane (figure 3). The HEAO-3 detection
revealed that stars continue to enrich the Galaxy with their
nucleosynthesis products. The galactic 26
Al map reveals,
better than any other tracer, where the sites of current
nucleosynthesis lie in the Milky Way.
Balloon-borne spectrometers detected in the 1970s the
electron–positron annihilation line (at 0.511 MeV) from
the Galactic center direction. For some time it was
Copyright © Nature Publishing Group 2001
Brunel Road, Houndmills, Basingstoke, Hampshire, RG21 6XS, UK Registered No. 785998
and Institute of Physics Publishing 2001
Dirac House, Temple Back, Bristol, BS1 6BE, UK 5
Gamma-Ray Astronomy E N CYC LO PE D IA O F AS T R O N O MY AN D AS T R O PHYS I C S
COMPTEL 1.8 MeV, 5 Years Observing Time
Uwe Oberlack
PhD Thesis 1997
0306090
120
150
180
210
240
270300330
-60
-30
0
30
60
Figure 3. Map of the Galaxy in 1.8 MeV (26Al line). Data from COMPTEL on CGRO.
thought that this emission was variable, but this variability
was presumably due to the different properties of the
instruments used. CGRO established the existence of
diffuse (and steady) 0.511 MeV emission in the Milky
Way, with ∼80% coming from the galactic bulge and the
remaining ∼20% from the disk. Positrons produced by the
β+
decay of radionuclides, such as 26
Al, 56
Co and 44
Sc, are
(most probably) at the origin of this galactic emission.
γ -ray bursts
Cosmic γ -ray bursts (GRBs), discovered in the late 1960s
by the Vela satellites, are flashes of high-energy radiation
which appear at random times from random directions,
briefly dominating the γ -ray sky (from a few milliseconds
to a few seconds) and then fading away. For more than a
quarter of a century their nature remained a mystery: their
non-thermal γ -ray emission presented no regularities at
all, they did not seem to radiate in other wavelengths and
it was impossible to evaluate their distances or to associate
any known object with them.
The situation improved considerably in the 1990s.
The BATSE instrument aboard CGRO detected about 1
burst per day for several years and established that their
distribution in the sky is isotropic (figure 4). Such a
distribution is naturally obtained from sources located
at cosmological distances (several 108
–109
light-years); in
that case, their observed frequency implies about one burst
per galaxy per million years, i.e. they represent events
much rarer than supernovae. The Italian–Dutch x-ray
satellite BeppoSAX allowed for an accurate location of
GRBs, and in several cases relatively long-lived emission
(‘afterglow’) at x-ray, optical and radio wavelengths was
detected. In at least one case the afterglow optical spectra
showed redshifted absorption lines due to the presence of
remote intervening gas clouds, supporting the idea that
the GRB is at a cosmological distance.
To be detectable at large distances, GRBs must radiate
huge amounts of energy, about 1053
erg on average.
Collisions involving neutron stars and/or black holes may
produce such energies, mainly in the form of neutrinos,
but the exact mechanism of the energy conversion to
γ -rays is not elucidated yet. ‘Fireball’ models, involving
material expanding at relativistic energies, seem the most
promising at present.
Active galactic nuclei and blazars
The first detection of an extragalactic γ -ray source,
the BRIGHT QUASAR 3C 273, was reported in 1978 by a
team working with the COS-B satellite. Several other
objects were subsequently detected in the low-energy
γ -ray band (<1 MeV) and identified with the highly
variable, moderately distant (redshift z < 0.05) ACTIVE
GALACTIC NUCLEI (AGNs). In the 1990s, the EGRET
instrument aboard CGRO detected high-energy γ -ray
emission (>100 MeV) from several dozens of powerful and
distant (z > 2) sources, a class of AGNs now known as
BLAZARS; these objects radiate in their ‘active’ phase more
energy in high-energy γ -rays (1045
erg s−1
on average) than
in all other wavelengths.
The short time-scale of variability of AGNs (a few
days to weeks for blazars), combined with their high
γ -luminosities, suggests that accretion onto a massive
black hole ((106
–109
)M ) is the most probable power
source. This idea could explain, in a unified framework,
both the low-energy γ -emission of AGNs and the highenergy
γ -emission of blazars. In the former case, the
γ -rays are emitted from the hot inner regions of the
accretion disk surrounding the black hole (as is the case
with accreting stellar mass black holes in our Galaxy); in
Copyright © Nature Publishing Group 2001
Brunel Road, Houndmills, Basingstoke, Hampshire, RG21 6XS, UK Registered No. 785998
and Institute of Physics Publishing 2001
Dirac House, Temple Back, Bristol, BS1 6BE, UK 6
Gamma-Ray Astronomy E N CYC LO PE D IA O F AS T R O N O MY AN D AS T R O PHYS I C S
Figure 4. Isotropic distribution of GRBs in the sky.
Figure 5. Diffuse extragalactic γ -ray background. Data
compiled by Paul and Laurent (1998).
the latter case γ -rays are emitted by relativistic particles
moving in diametrically opposite highly collimated jets
(the intensity of the emission being enhanced along the
direction of motion). This unified scheme seems able to
explain not only the γ -ray properties ofAGNs and blazars,
but also their observed features in other wavelengths.
Diffuse extragalactic background
Since the early days of γ -ray astronomy, observations
established the presence of a uniform, diffuse, isotropic
γ -ray emission in the sky, most probably of extragalactic
origin (figure 5). The most natural explanation for this
diffuse γ -ray background involves the contribution of
numerous extragalactic high-energy sources, unresolved
by current instruments. In the low-energy range
(<300 keV), the observed γ -ray background spectrum
can be explained by the collective emission of Seyfert
galaxies (a class of AGNs), while in the high-energy
range (∼100 MeV) it is the collective emission of blazars
that prevails; indeed, in both cases, the observed spectra
of resolved individual sources compare favourably with
the observed diffuse spectrum. Finally, the CGRO
found no excess diffuse emission in the intermediate
range of ∼1 MeV, contrary to all previous reports
(obtained by instruments with underestimated and
difficult to assess instrumental ‘noise’). Despite that,
this intermediate spectral region is not satisfactorily
explained by the combination of Seyferts and blazars and
another source seems to be required; the superposition of
redshifted nuclear γ -ray lines from remote SNIa (mostly
from the decay of 56
Co) seems the most promising
candidate.
Bibliography
Most of the results of γ -ray astronomy in the 1990s can
be found in the Proceedings of the CGRO and INTEGRAL
symposia, held alternatively every two years.
A good recent overview (in French!) is:
Paul J and Laurent Ph 1998 Astronomie Gamma Spatiale
(London: Gordon and Breach)
Some other selected references are as follows.
• The Sun:
Murphy R, Ramaty R, Kozlovsky B and Reames D 1991
Astrophys. J., 371 793
Copyright © Nature Publishing Group 2001
Brunel Road, Houndmills, Basingstoke, Hampshire, RG21 6XS, UK Registered No. 785998
and Institute of Physics Publishing 2001
Dirac House, Temple Back, Bristol, BS1 6BE, UK 7
Gamma-Ray Astronomy E N CYC LO PE D IA O F AS T R O N O MY AN D AS T R O PHYS I C S
• Accreting compact objects:
Mandrou P et al 1994 Astrophys. J. Suppl. 92 343
• GRBs:
Fishman G and Meagan C 1995 Ann. Rev. Astron. Astrophys.
33 415
Meszaros P and Rees M 1993 Astrophys. J. 405 278
• Radioactivity from stellar explosions:
Diehl R and Timmes F 1998 Publ. Astron. Soc. Pacific 110
637
• Diffuse γ -ray line emission in the Galaxy:
Prantzos N and Diehl R 1996 Phys. Rep. 261 1
• Diffuse γ -ray continuum emission in the Galaxy:
Strong A and Mattox J 1996 Astron. & Astrophys. 308 L21
• AGNs:
Dermer C 1995 The Gamma-Ray Sky with CGRO and SIGMA
ed M Signore, P Salati and G Vedrenne (Dordrecht:
Kluwer) p 39
• Diffuse extragalactic γ -ray background:
Zdziarski A 1996 Mon. Not. Royal Astron. Soc. 281 L9
Nikos Prantzos
Copyright © Nature Publishing Group 2001
Brunel Road, Houndmills, Basingstoke, Hampshire, RG21 6XS, UK Registered No. 785998
and Institute of Physics Publishing 2001
Dirac House, Temple Back, Bristol, BS1 6BE, UK 8