Download Explosive sources of the highest energy radiation

The transient universe • Hill: Detections by Fermi
Explosive sources of the
highest energy radiation
The transient universe Adam Hill reports on the vast variety of transient
systems detected by NASA’s Fermi Gamma-ray Space Telescope.
A
stronomers cannot help but be famil- board Fermi is the Large Area Telescope (LAT),
iar with electromagnetic radiation: it an electron–positron pair production telescope
is our bread and butter, the primary using solid-state silicon trackers and caesium
source of information we have for discovering iodide calorimeters, sensitive to photons from
how the universe ticks. The bulk of the radia- ~20 MeV to greater than 300 GeV (Atwood et al.
tion observed is produced through thermal pro- 2009). Essentially it is a particle physics deteccesses; “hot things” produce it, typically with tor in space, sensitive to photons with energies
a blackbody continuum spectrum. However,
100 billion times greater than optical light.
the extremes of the EM spectrum
Figure 1 shows the sky as seen by
offer the potential to explore the
the Fermi LAT at energies above
Fermi
1 GeV, in which the large highnon-thermal universe. And, in
was launched
contrast to the majority of
energy source population is
in 2008 and has
the EM spectrum, it is only
self-evident.
in the past couple of decades
The primary channels to
discovered around
that sensitive gamma-ray
produce
gamma-rays at these
2000 distinct
telescopes have opened our
energies can be split into hadgamma-ray
eyes to how the sky looks at
ronic and leptonic processes.
sources
the highest possible energies.
In the hadronic scenario, proNASA’s Fermi Gamma-ray Space
tons are accelerated to high energies and then collide with each other,
Telescope is one such facility. The telescope was launched in June 2008 and has been producing neutral pions that subsequently decay
operating successfully for over five years, mak- into two gamma-ray photons or a gamma-ray
ing numerous exciting discoveries including the photon and an electron–positron pair. In the
detection of ~2000 distinct gamma-ray sources leptonic scenario, electrons are the accelerated
(Nolan et al. 2012). The primary instrument on- particles that then inverse Compton scatter off
‘‘
’’
A&G • December 2013 • Vol. 54 1: A map of the gamma-ray sky at energies
above 1 GeV incorporating five years of
survey observations from the Fermi Large
Area Telescope. The map is oriented in
galactic coordinates with the galactic plane
clearly running through the centre of the
image. (NASA/DOE/Fermi LAT Collaboration)
photons (e.g. from a nearby star or the cosmic
microwave background), boosting the photons
to gamma-ray energies. In both scenarios particle acceleration is required, which necessitates
extreme environments such as the jets of a black
hole, colliding shock fronts or extreme magnetic
fields such as those found around pulsars. These
conditions can be produced in a wide variety of
astrophysical objects: from the Sun, to binary
systems in our galaxy, to extragalactic objects
such as gamma-ray bursts.
Flaring high-mass X-ray binaries
There is a small population of high-mass X-ray
binary (HMXB) systems that are known to be
high-energy gamma-ray sources. The majority of these systems are persistent emitters, but
two systems have been observed by Fermi to be
flaring sources that are, most of the time, undetectable by current instruments. Incidentally,
6.31
The transient universe • Hill: Detections by Fermi
2: This diagram illustrates the view from Earth of the LS 2883/PSR B1259-63 binary system as well as key events in the pulsar’s December 2010 periastron
passage. (NASA/Goddard Space Flight Center/Francis Reddy)
these two systems are the only members of
the detected HMXB population for which the
nature of the object is known: the microquasar
Cyg X-3 and LS 2883, a binary system that hosts
the pulsar PSR B1259-63 and a main-sequence
companion.
Cyg X-3 is a well-known, powerful HMXB
with a short orbital period of 4.8 hours. The
system comprises a compact object accreting
matter from a Wolf–Rayet companion star (van
Kerkwijk et al. 1992). It regularly becomes the
brightest radio source among known binary systems, with large flares that are attributed to its
relativistic jets. The system exhibits a complex
X-ray spectrum that fluctuates between two
main states: “soft” and “hard”. The source is
known to flare in radio when entering the “soft”
state with associated relativistic jet ejection
events (Fender et al. 2006). Historically there
were reported detections at energies between
MeV and PeV in the 1970s and 1980s. However,
the detections were typically of low significance
and although some experiments claimed detections, others could not confirm them. Hence,
claims of high energy and very high energy
emission from this microquasar were controversial and highly contested.
The identification of Cyg X-3 as a source of
gamma-rays was put beyond a shadow of a
doubt by the independent detection of flares
6.32
from the source by the Fermi and AGILE missions. The LAT detected a 29σ point source at a
location consistent with that of Cyg X-3 and at
the same time as a radio outburst was detected.
The association of the gamma-ray flaring source
was confirmed through the detection of the
4.8-hour orbital period in the LAT light curve
(Abdo et al. 2009). The gamma-ray light curve
of the source flux indicates that the source is
highly variable. The initial Fermi detection
reports on the identification of two specific
active periods that comprise one or more flares:
11 October to 20 December 2009 and 8 June
to 2 August. Assuming a distance of 7 kpc, the
average luminosity of Cyg X-3 in outburst at
energies above 100 MeV is ~3 × 1036 erg s –1. The
AGILE mission also reported detecting gammaray activity coincident with the location of
Cyg X-3 during these epochs as well as activity
from 16–17 April 2008, before the launch of
Fermi (Tavani et al. 2009).
The correlation between the gamma-ray and
radio emission specifically links the gammaray activity to periods of relativistic ejection
events. In this case, the gamma-ray emission
could be explained by inverse Compton scattering of UV photons from the Wolf–Rayet
star off high-energy electrons. However, this
scenario requires that the emission region is
not too close to the accretion disc, otherwise
the gamma-ray emission would be absorbed
through pair production on soft X-ray photons
coming from the disc.
Gamma-ray emission was again detected
from Cyg X-3 in 2011 during a giant radio flare
(Corbel et al. 2012). However, analysis of this
outburst also identified detectable gamma-ray
activity associated with much weaker radio flaring immediately prior to the system entering
into a radio-quenched (ultrasoft X-ray) state.
This has been interpreted to indicate that transitions into and out of the radio-quenched state
trigger gamma-ray emission, implying a connection to the accretion process and the production of relativistic jets.
LS 2883/PSR B1259-63 is a very different type
of system. The binary consists of a ~47 ms pulsar orbiting a massive Be star in a highly eccentric orbit with a period of ~3.4 years (Johnston
et al. 1994). Unpulsed radio, X-ray and TeV
gamma-ray emission has been observed from
the system during periastron passage when the
pulsar is closest to the Be star and can interact
with the dense circumstellar disc. Although this
source was observed by EGRET in the 1990s,
there was no GeV detection prior to the launch
of Fermi. Consequently the December 2010
periastron passage of this system, the first in
the Fermi era, was hotly anticipated.
While the system was far from periastron
A&G • December 2013 • Vol. 54
The transient universe • Hill: Detections by Fermi
3: Integral flux above 100 MeV as a function
of time during the April 2011 Crab flare;
mean bin duration is 9 minutes. The dotted
line indicates the sum of the 33-month
average fluxes from the inverse Compton
nebula and the pulsar. The dashed line
shows the flux of the average synchrotron
nebula, the inverse Compton nebula and the
pulsar. (Figure 5 from Buehler et al. 2012)
there was no detection with the Fermi LAT. surprise when the Fermi and AGILE missions
From ~28 days prior to periastron through to reported detections of powerful flares from the
18 days after periastron, the LS 2883 system was Crab Nebula at energies above 100 MeV (Abdo
observed to brighten in gamma-rays as expected et al. 2011a, Tavani et al. 2011).
The Crab Nebula is the remnant of a superand yielded a ~5σ detection (Abdo et al. 2011b).
During this period the average >100 MeV flux nova explosion reported by Chinese astronoof the system was 0.9 × 10 –10 erg cm –2 s –1. What mers in AD 1054. At the heart of the nebula is
was completely unexpected was that 30 days a pulsar that loses rotational energy through a
after periastron passage the system exhibited a relativistic electron/positron wind that injects
much more powerful flare, peaking at ~10 times energetic particles into the nebula. The Crab
higher than the integrated average flux observed Nebula can be detected at wavelengths from
during the first disc passage. This flare contin- radio all they way up to TeV energies; below
ued for seven weeks, with an average
400 MeV the spectrum is dominated
flux of 4.4 × 10 –10 erg cm–2 s–1. Figure
by synchrotron emission, whereas
This
2 illustrates the system geometry
the highest energy radiation is
produced via inverse Compand the location of the pulsar
is an exciting
during the two gamma-ray
result as it breaks the ton scattering of photons off
active periods.
relativistic particles.
axiom that pulsars
The gamma-ray emission
Since the original reports of
are steady gammafrom the system is believed
flares by Fermi and AGILE,
ray sources
to be produced in the shock
the Crab has been observed
to flare at high energies on
front formed from the interacaverage once a year. Remarktion of the pulsar wind with the
stellar wind of the Be companion.
ably, there is no sign of simultaneous
While the detection of gamma-ray emission
variability at any other wavelengths despite
by Fermi was anticipated, the highly variable numerous multi-wavelength follow-up probehaviour was not predicted in any model of grammes. The brightest reported Crab flare
gamma-ray emission expected from this system. to date occurred in April 2011 (Buehler et al.
The unexpected, strong flare seen at GeV ener- 2012); the source was seen to double its flux in
gies was not detected at any other wavelength the space of 8 hours and reached a peak flare
despite a highly organized multi-wavelength flux of 1.86 × 10 –10 erg cm –2 s –1, a factor of 30
campaign, and it took place 30 days after peri- increase in the average value of the nebula flux
astron passage and after the pulsar had passed (a factor of 7 when including both pulsar and
through the dense equatorial disc for the second nebula flux)! Figure 3 shows the dramatic rise in
time. Possible explanations for the behaviour the Crab flux during the April 2011 flare. Upon
include anisotropy of the gamma-ray emission, further analysis it has been shown the gammaan abrupt change in the physical environment ray emission from the pulsar during the flare is
or the emitting region of the emergence of a new constant, indicating that it is some part of the
nebula that is variable.
emission component (Abdo et al. 2011b).
Isolated rotation-powered pulsars observed
One of the most amazing results from these
at high energies are renowned for their steady, flares, other than the complete lack of detectnon-variable emission. And in the case of the ability at other wavelengths is the rapid timeCrab Nebula and pulsar it is one of the earli- scales of variability and the associated power of
est and best-studied high-energy objects whose the emission. From simple causality arguments,
constancy has been relied upon as a “standard a doubling time of eight hours implies a comcandle” to calibrate generations of high-energy pact emission region of less than 2.8 × 10 –4 pc
space missions. Consequently, it came as a huge and the isotropic power at the peak of the flare
‘‘
’’
A&G • December 2013 • Vol. 54 is ~4 × 1036 erg s –1, which corresponds to ~1%
of the total spin-down power of the Crab pulsar, the power source for the nebula emission.
Despite the observation of multiple gamma-ray
flares of the Crab, there is still no clear picture
of the origin of this behaviour and the observed
gamma-ray properties present great challenges
to current theoretical models of particle acceleration within the nebula.
Recent studies by the LAT of the pulsar
PSR J2021+4026 in the Gamma Cygni region
have discovered that this system is also variable (Allafort et al. 2013). The variability
is much less spectacular than in the case of
the Crab, however, and in this instance the
variability is associated with the pulsar itself.
During the analysis of 52 months of Fermi
observations, the researchers unveiled a sudden decrease in flux above 100 MeV; around 16
October 2011, the pulsar flux decreased from
8.33 × 10 –10 erg cm –2 s –1 by ~20% over less than
a week. At the same time, the spin-down rate
of the pulsar increased by ~4%. Allafort et al.
(2013) interpret these results as indicating that
the observed change in the pulsed gamma-ray
emission is driven by changes in the emission
beaming and speculate that it may be precipitated by a shift in the magnetic field structure.
This is an exciting result as it breaks the axiom
that pulsars are steady gamma-ray sources.
The discovery of gamma-ray novae
Astronomers have observed novae for thousands of years, but it is only in more modern
times that the processes behind these cosmic
eruptions have become clear. Classical novae
(CNe) are a subset of cataclysmic variables,
binary systems which host a white dwarf that
accretes material from the secondary star. In the
vast majority of these systems the orbital period
ranges between 1.4 and 8 hours and the secondary star is a low-mass, main sequence star; there
are a few longer-period systems in which for
the secondary star to fill its Roche lobe it has to
have evolved off the main sequence (Bode 2012).
The white dwarf accretes material off the
secondary star, building up H-rich material on
6.33
The transient universe • Hill: Detections by Fermi
4: Japanese amateur astronomers discovered
Nova Cygni 2010 in an image taken at 19:08 UT
on 10 March (4:08 a.m. Japan Standard Time, 11
March). The erupting star (circled) was 10 times
brighter than in an image taken several days
earlier. The nova reached a peak brightness
of magnitude 6.9, just below the threshold
of naked-eye visibility. (K Nishiyama and F
Kabashima/H Maehara, Kyoto Univ.)
5: Fermi’s Large Area Telescope saw no sign of
a nova in 19 days of data prior to 10 March (left),
but the eruption of Nova Cygni 2010 is obvious
in data from the following 19 days (right). The
images show the rate of gamma-rays with
energies greater than 100 million eV (100 MeV);
brighter colours indicate higher rates. (NASA/
DOE/Fermi LAT Collaboration)
its surface. Once sufficient material has accumulated for the critical pressure/temperature
to be achieved at the base of the accreted envelope, then a thermonuclear runaway explosion is triggered, producing a CN event. The
outburst increases the luminosity to ~10 000
solar luminosities and large amounts of mass
are ejected at high velocities (Bode 2012), typically ~10 –5 –10 –4 solar masses at velocities of
~102 –103 km s –1. The inter-outburst period of
CN explosions is believed to be 1000–10 000
years. Conversely, recurrent novae (RNe) have
inter-burst timescales of the order of decades
and may host more evolved red giant companions. Only ten such systems have been identified
within our galaxy.
Gamma-ray line emission at ~MeV energies
has long been predicted from novae as a consequence of the decay of radioactive elements
produced in the nova explosion (Hernanz 2013).
However, prior to the discovery of GeV emission
from V407 Cyg by the Fermi LAT in 2010 (Abdo
et al. 2010), there had been no predictions of
continuum emission at such high energies.
On 10 March 2010, Japanese amateur astronomers reported the discovery of a new 8th magnitude nova in the Cygnus constellation (figure
4). The nova was identified as originating from
the known symbiotic binary V407 Cyg. This
6.34
system comprises a hot white dwarf accreting
from a Mira-type variable red giant and consequently the white dwarf is embedded in a particularly dusty environment generated by the
heavy wind of the donor star.
The discovery of a classical nova event in this
system was completely unexpected and was compounded by the surprising discovery of gammaray emission above 100 MeV from the nova by
the LAT (Abdo et al. 2010). The gamma-rays
were detectable for about two weeks after the
optical nova onset, with an average spectral
energy distribution in the form of a power
law with an exponential cutoff and flux above
100 MeV of 4.4 × 10 –7 ph cm–2 s –1 (figure 5).
A natural explanation for the origin of the
gamma-ray emission appeared to be a result of
the nova ejecta shell expanding outwards and
colliding with the stellar wind of the red giant
companion and forming a shock front where
particles could be accelerated to high energies.
The line connecting the white dwarf with the
donor star contains the largest local density
enhancement attributable to the red giant wind
in which the nova shell can sweep up material
in a fashion similar to what is ascribed in supernova explosion models. Taking the estimated
parameters of the system (wind density, ejecta
velocity, etc) and the measured spectral energy
distribution, it was shown that this model could
feasibly produce gamma-ray emission through
the decay of neutral pions produced in proton collisions or inverse Compton scattering
off accelerated electrons (Abdo et al. 2010).
More detailed modelling of the environment
has suggested that leptonic processes may be
dominant (Martin and Dubus 2013). The direct
link between the dense local wind environment
and the gamma-ray production mechanism and
the rarity of symbiotic and RS Oph-type RNe
systems led to the suggestion that gamma-ray
novae would be exceptionally rare events (Nelson et al. 2012).
From 16–30 June 2012, Fermi identified a
new gamma-ray source, Fermi J1750-3243,
which was not consistent with any of the
known 2FGL catalogue sources (Nolan et al.
2012). The location of the new LAT transient
was consistent with the report of a newly discovered optical nova, MOA 2012 BLG-320
(Nova Sco 2012), which had entered into optical outburst 1.77–2.15 June 2012 (Cheung et al.
2012b). It appeared that Fermi had discovered
another gamma-ray nova. However, this system appeared to be more like traditional CNe
systems, with no indication of a dense stellar
environment, and its behaviour at other wavelengths was quite different to that observed in
A&G • December 2013 • Vol. 54
The transient universe • Hill: Detections by Fermi
6: These images show how the sky looks at
gamma-ray energies above 100 MeV with a
view centred on the north galactic pole. The
first frame shows the sky during a three-hour
interval prior to GRB 130427A. The second frame
shows a three-hour interval starting 2.5 hours
before the burst, and ending 30 minutes into the
event. (NASA/DOE/Fermi LAT Collaboration)
V407 Cyg (Hill et al. 2012).
On 22 June 2012, another new unidentified gamma-ray source was detected by the
LAT – Fermi J0639+058 (Cheung et al. 2012a)
– but its proximity to the Sun prohibited follow-up observations at other wavelengths.
In August 2012, a late-stage optical nova,
Nova Mon 2012, was discovered within the
LAT error circle. The first optical spectra were
taken about 55 days after the gamma-ray peak
and identified the white dwarf as being a member of the ONe class (Shore et al. 2013), i.e. the
more massive class of white dwarfs. This system
has also been associated with the more common
CNe systems and has been identified as having a tight binary orbital period of ~7.1 hours
(Osborne et al. 2013).
In August 2013, Fermi discovered gamma-rays
from Nova Del 2013 (Hays et al. 2013), increasing the gamma-ray nova sample to four distinct
objects. The LAT light curves of the novae are
remarkably similar and indicate that each object
lasts as a gamma-ray source for ~2–3 weeks.
However, there are also several distinct differences in their more general multi-wavelength
characteristics. Most notable is that, with the
exception of V407 Cyg, all the other gamma-ray
novae appear to be CNe in the traditional sense
and so the gamma-ray production mechanism
invoked to explain the emission in V407 Cyg
cannot be applied; it is likely that there is
another mechanism in action in these systems.
This raises the potential for many other CNe
and RNe to be gamma-ray bright, although it
is equally evident that numerous other optically detected CNe have not been reported by
the LAT (there have been more than 40 reported
optical novae since Fermi launched). Only with
further study and analysis will we be able to
identify the gamma-ray production channels
and determine what fraction of the novae population of the galaxy can produce emission at
these high energies.
Gamma-ray bursts
Gamma-ray bursts (GRBs) are in many respects
the epitome of transient astronomy. They are
A&G • December 2013 • Vol. 54 rapid flaring events typically lasting from less
than one second to tens of seconds and are
the most luminous explosions in the universe,
detectable across the electromagnetic spectrum.
There are two main progenitor models used to
explain GRBs. “Long” GRBs, those with durations longer than 2 s, are believed to be the result
of the collapse of a massive star when it exhausts
its nuclear fuel. As the core collapses into a
black hole, jets of material shoot outward at
nearly the speed of light. “Short” GRBs, those
with durations less than 2 s, are believed to be a
result of the merger of a pair of neutron stars or
a neutron star with a black hole. Both types of
system are extragalactic in origin.
The Fermi LAT has detected several GRBs
since its launch in 2008, but none as spectacular as GRB 130427A, which erupted on 27
April 2013. This GRB is associated with Type
Ic supernova SN 2013cq, at a redshift of z ~ 0.34
(Xu et al. 2013), placing it in the closest 5% of
bursts. This close proximity makes it an exceptionally bright source of gamma-rays; it was
detectable by the LAT for almost an entire day.
Figure 6 shows how bright the GRB appeared
to the LAT. Not only was it the longest GRB
to be detected by the LAT, but it also holds the
record for the highest energy photon detected
at 94 GeV. The brightness and long duration
of detectability of the GRB allowed many telescopes across all wavelengths to observe it while
active. As the results are published in coming
months it is hoped that there will be many exciting discoveries reported.
Summary
It is clear that the universe is a violent and variable place at high energies. A wide variety of
object classes are capable of variations on a
range of timescales from seconds to months
and involve tremendous amounts of power and
extreme environmental conditions that test
and extend our understanding of fundamental physics. Presented here is only a taste of the
diverse range of astrophysical systems capable
of producing variable levels of the most energetic high-energy radiation; others include solar
flares, terrestrial gamma-ray flashes (gammarays produced in thunderstorms in the Earth’s
atmosphere), and the vast population of flaring
active galactic nuclei. These are all sources of
gamma-rays that Fermi has and is observing.
The Fermi mission is answering many questions
about the high energy universe and raising new
ones that challenge our understanding of how
high-energy particle physics processes operate
in astrophysical environments. ●
A B Hill, KIPAC, Dept of Physics and SLAC
National Accelerator Laboratory, Stanford
University, USA and School of Physics &
Astronomy, University of Southampton, UK;
[email protected]. A B Hill acknowledges
support by a Marie Curie International Outgoing
Fellowship within the 7th European Community
Framework Programme (FP7/2007-2013) under
grant agreement no. 275861.
References
Abdo A A et al. 2009 Science 326 1512.
Abdo A A et al. 2010 Science 329 817.
Abdo A A et al. 2011a Science 331 739.
Abdo A A et al. 2011b Astrophys. J. 736 11.
Allafort A et al. 2013 Astrophys. J. Letts 777 2.
Atwood W et al. 2009 Astrophys. J. 697 1071.
Bode M 2012 Proc. 11th Asian–Pacific Regional IAU
Meeting 2011 NARIT Conference Series Vol. 1 eds S
Komonjinda, Y Kovalev and D Ruffolo arXiv:1111.4941.
Buehler R et al. 2012 Astrophys. J. 749 26.
Cheung C C et al. 2012a Astronomers Telegram
#4224.
Cheung C C et al. 2012b Astronomers Telegram #4284.
Corbel S et al. 2012 Mon. Not. R. Astron. Soc. 421 2947.
Fender R P et al. 2006 Mon. Not. R. Astron. Soc. 369 603.
Hernanz M 2013 Proc. Stella Novae: Future and Past
Decades eds P A Woudt and V A R M Ribeiro ASPCS
arXiv:1305.0769.
Hays E et al. 2013 Astronomers Telegram #5302.
Hill A B et al. 2012 on behalf of the Fermi–LAT collaboration in Proc. 4th Fermi Symposium Monterey CA
arXiv:1304.2427.
Johnston S et al. 1994 Mon. Not. R. Astron. Soc. 268 430.
van Kerkwijk M H et al. 1992 Nature 355 703.
Martin P and Dubus G 2013 Astron. Astrophys. 551 37.
Nelson T et al. 2012 Astrophys. J. 748 43.
Nolan P et al. 2012 Astrophys. J. Supp. 199 31.
Osborne J P et al. 2013 Astronomers Telegram #4727.
Shore S et al. 2013 Astron. Astrophys. 553 123.
Tavani M et al. 2009 Nature 462 620.
Tavani M et al. 2011 Science 331 736.
Xu D et al. 2013 Astrophys. J. in press arXiv:1305.6832.
6.35