TeV Gamma-ray Astronomy: A Summary

TeV Gamma-ray Astronomy: A Summary

J. Holder Department of Physics and Astronomy and the Bartol Research Institute, University of Delaware, Newark, DE 19716, USA,                                     Tel: +(1) 302 831 2545, Fax: +(1) 302 831 1637

The field of TeV gamma-ray astronomy has produced many exciting results over the last decade. Both the source catalogue, and the range of astrophysical questions which can be addressed, continue to expand. This article presents a topical review of the field, with a focus on the observational results of the imaging atmospheric Cherenkov telescope arrays. The results encompass pulsars and their nebulae, supernova remnants, gamma-ray binary systems, star forming regions and starburst and active galaxies.

Gamma-ray Astronomy, High-energy astrophysics
journal: Astroparticle Physics

1 Introduction

Teraelectronvolt (TeV) astronomy concerns the study of astrophysical sources of gamma-ray photons, with energies in the range between and . The TeV range is one of the most recent windows of the electromagnetic spectrum to be opened for study, beginning with the identification of the first source, the Crab Nebula, in 1989 Weekes and et al. (1989). The results have been impressive, with significant advances in instrumentation leading to the detection of well over 100 sources over the last decade. The goals of TeV astronomy are wide-ranging, but can broadly be described as the study of sites of relativistic particle acceleration in the Universe, both hadronic and leptonic. This encompasses a huge range of size scales and energetics, from the interactions of galaxy clusters, to the magnetospheres of individual pulsars.

Numerous recent reviews of TeV gamma-ray astronomy have been written (e.g. Hinton and Hofmann (2009); Aharonian et al. (2008); Chadwick et al. (2008)), but the field is rapidly evolving. For example, since the extensive review by Hinton & Hofmann in 2009 Hinton and Hofmann (2009), the number of sources has grown from around 80 to more than 120, and a number of new source classes have been identified. This paper therefore aims to provide an update, and to supplement the existing reviews with a summary of the current observational status. The primary focus of the review is on the results from imaging atmospheric Cherenkov telescopes, with a relatively brief discussion of the air shower particle detector experiments. We use the definitions outlined by Aharonian Aharonian (2004) as follows: “high energy”, or “GeV”, astronomy refers to the energy range from to , while “very high energy”, or “TeV”, astronomy, refers to the range from to .

2 A Brief History

The development of ground-based gamma-ray astronomy is closely linked to the study of cosmic rays and cosmic ray air showers. The idea of searching for astrophysical gamma-ray sources at energies was first proposed by Morrison in 1958 Morrison (1958). A prediction of a very high TeV gamma-ray flux from various sources, including the Crab, which might be detectable with an air shower particle array at high altitude, was made by Cocconi in 1959 Cocconi (????). Cherenkov radiation associated with large cosmic ray air showers was first detected by Galbraith & Jelley in 1953 Galbraith and Jelley (1953), and the possibility of using this phenomenon to study gamma-ray initiated showers led to the development of a number of dedicated facilities in the 1960s. This effort was boosted by the apparent detection of a gamma-ray signal from the black-hole binary Cygnus X-3 by both particle air shower arrays and atmospheric Cherenkov detectors. With hindsight, this detection was likely spurious, as were numerous unsubstantiated claims throughout the 1970s, 1980s and into the 1990s.

The field reached a firm experimental footing with the development of the imaging technique, which provides a method of effectively discriminating between gamma-ray initiated showers and the background of cosmic ray showers, based on the morphology of their Cherenkov images, and guided by the results of Monte Carlo simulations Weekes and Turver (????); Hillas (1985). This technique was applied by the Whipple collaboration to detect steady gamma-ray emission from the Crab Nebula using a 10m reflector with a 37-element photomultiplier tube camera in 1989 Weekes and et al. (1989). A number of imaging atmospheric Cherenkov telescopes (IACTs) were subsequently developed around the world (including Durham, CANGAROO, Telescope Array, Crimean Astrophysical Observatory, SHALON, TACTIC), with the northern hemisphere instruments (Whipple, HEGRA and CAT) leading the field. The 1990s saw two particularly important developments: the detection of the first extragalactic sources by the Whipple Collaboration, starting with the nearby blazars Markarian 421 Punch and et al. (1992) and Markarian 501 Quinn and et al. (1996), and the application of the stereo imaging technique by the HEGRA array Daum and et al. (1997). HEGRA consisted of 5 telescopes of modest aperture (), and demonstrated that the combination of Cherenkov image information from multiple telescopes located within the same Cherenkov light pool could dramatically improve the sensitivity of the technique.

Despite this progress, the relative scarcity of bright TeV gamma-ray sources ( were identified by 2000) highlighted the necessity for improved instrumentation. Cherenkov wavefront samplers such as CELESTE and STACEE attempted to probe to lower energies, and hence higher gamma-ray fluxes and larger distances, using converted solar farms; however, the difficulty of discriminating gamma-rays from the cosmic ray background using this technique limited its effectiveness. Successful gamma-ray observations using a particle detector were made by the Milagro experiment, which ran from 2000 to 2008, providing a survey of the northern sky with modest sensitivity. Starting with the commissioning of H.E.S.S. in 2003, the new generation of IACTs - H.E.S.S., MAGIC and VERITAS - have provided the required order of magnitude improvement in sensitivity, and firmly established gamma-ray studies as an important astronomical discipline.

3 Current Status and Instrumentation

3.1 Imaging Atmospheric Cherenkov Telescopes

The Cherenkov emission from an air shower forms a column of blue light in the sky, with the maximum emission occuring around above sea level at TeV energies. Cherenkov telescopes used to record these images are essentially rather simple devices, consisting of a large, segmented optical flux collector used to focus the Cherenkov light onto an array of fast photo-detectors. The optical specifications are not terribly strict; an optical point spread function width of typically is adequate. The photo-detector array (usually photomultiplier tubes) comprises a crude camera, covering a few degrees on the sky. A Cherenkov flash triggers read-out of the photo-detectors, with a read-out window defined by the timescale of the arrival of the Cherenkov photons ().

There are currently three major imaging atmospheric Cherenkov telescope systems in operation. H.E.S.S., located in the Khomas Highland of Namibia (N, W, altitude ), consists of four telescopes arranged on a square with 120 side length. Each telescope has a mirror area of and is equipped with a 960 pixel camera covering a field of view. VERITAS, at the Fred Lawrence Whipple Observatory in southern Arizona (N, W, altitude ) has similar characteristics, with 4 telescopes of area and 499-pixel cameras, covering . MAGIC (N, W, altitude ) originally consisted of a single, very large reflector () on the Canary island of La Palma, with a camera. In 2009, a second telescope with the same mirror area was installed at a distance of from the first.

Each of these facilities work in a similar fashion. Cherenkov images of air showers are recorded at a rate of a few hundred Hz, and analyzed offline. The overwhelming majority of these images are due to cosmic ray initiated air showers. Gamma-ray showers can be discriminated from this background based on the image shape and orientation. Gamma-ray images result from purely electromagnetic cascades and appear as narrow, elongated ellipses in the camera plane. The long axis of the ellipse corresponds to the vertical extension of the air shower, and points back towards the source position in the field of view. If multiple telescopes are used to view the same shower, the source position is simply the intersection point of the various image axes (illustrated schematically in Figure 1). Cosmic-ray primaries produce secondaries with large transverse momenta, which initiate sub-showers. The Cherenkov images of cosmic-ray initiated air showers are consequently wider then those with -ray primaries, and form an irregular shape, as opposed to a smooth ellipse. In addition, since the original charged particle has been deflected by galactic magnetic fields before reaching the Earth, cosmic-ray images have no preferred orientation in the camera.

Figure 1: A schematic illustration of an atmospheric Cherenkov telescope array. The primary particle initiates an air shower, resulting in a cone of Cherenkov radiation. Telescopes within the Cherenkov light pool record elliptical images; the intersection of the long axes of these images indicates the arrival direction of the primary, and hence the location of a -ray source in the sky.

Cherenkov light reaching the ground from air showers peaks at optical/UV wavelengths, and so IACTs operate only under clear, dark skies. Both MAGIC and VERITAS have demonstrated that useful observations can be made when the moon is visible above the horizon, but the typical duty cycle of these instruments is still limited to per year (). Given the small field of view of IACTs, regions of the sky containing one or more source candidates are usually targeted for observations. Surveys can only be accomplished slowly, by tiling regions of the sky with overlapping fields-of-view. The sensitivity of the current generation of IACTs is sufficient to detect the Crab Nebula in under a minute, and a source with 1% of the Crab flux ( above ) in . The angular and energy resolution of the technique are energy-dependent, with typical values of and per photon, respectively, at .

The catalog of TeV sources grew rapidly with the commissioning of H.E.S.S. in the southern hemisphere, which provided the first high sensitivity observations of the densely populated inner Galaxy. It has continued to expand in recent years as MAGIC and VERITAS have come online, and now numbers 130 sources, as listed in the online catalog TeVCat Wakely and Horan (????) 111We note that a small fraction of these 130 sources may be duplicate detections of the same object, while other detections remain contentious or unconfirmed. Full details are available in the individual source annotations in TeVCat, available at http://tevcat.uchicago.edu. Figure 2 shows the locations of these sources in Galactic coordinates.

Figure 2: A map of the catalog of localized sources of TeV gamma-ray emission in Galactic coordinates as of November 2011, provided by the online catalog TeVCat Wakely and Horan (????)

3.2 Particle Detectors

The direct detection of air shower particles using arrays of particle detectors at ground level offers some important advantages over the atmospheric Cherenkov technique. In particular, observations can be made continuously, both day and night, and over the entire viewable sky (a field-of-view of ). These advantages are offset, however, by the rather low sensitivity to point sources, which is primarily due to the difficulty of rejecting the substantial background of cosmic ray initiated air showers. The angular and energy resolution of these detectors are also significantly worse than IACTs. Early claims for emission from binary systems using sparse particle detector arrays were not confirmed by later, more sensitive instruments (e.g. McKay and et al. (1993); Aharonian and et al. (2002)), indicating the need for a new approach to this problem. Milagro, which operated between 2000 and 2008 in northern New Mexico (N, W), was the first successful attempt at this.

The Milagro detector consisted of a large water reservoir () at an altitude of , covered with a light-tight barrier, and instrumented with PMTs. The central reservoir provided high-resolution sampling of air shower particles over a relatively small area (compared to the air shower footprint). In 2004 an array of 175 small outrigger tanks were added, irregularly spread over an area of around the central reservoir. This configuration, coupled with the development of analysis techniques for cosmic ray background discrimination, provided sufficient sensitivity for the first comprehensive survey of the northern TeV sky. The results showed strong detections of the bright, known TeV sources Markarian 421 and the Crab Nebula, along with the detection of three extended sources in the Galactic plane, each with integrated fluxes comparable to the Crab Nebula at . A few less significant source candidates were also identified in the plane, and a reanalysis following the launch of Fermi-LAT also showed fourteen excesses co-located with bright Galactic LAT sources Abdo and et al. (2009). The Milagro results for the region around the Galactic plane are shown in Figure 3.

Figure 3: The Milagro survey of the Galactic plane. The z-axis is the pre-trials statistical significance, with a fixed maximum of . Figure from Abdo et al. Abdo and et al. (2007).

Milagro ceased operation in 2008; however, two large particle detector arrays remain in operation at very high altitude in Tibet. ARGO-YBJ is located at in Yangbajing, and consists of a single layer of resistive plate chambers completely covering an area of . The results of 1265 days of observations were recently presented Cao et al. (2011), showing detections of the Crab Nebula, Markarian 421 and two Milagro sources. The fact that one of the brightest Milagro sources, MGRO J2019+37, is not detected in these observations presents something of a mystery. The Tibet AS air shower array, also at Yangbajing, consists of closely-spaced scintillation detectors covering an area of , and has demonstrated that this technique is also practical for the detection of bright TeV sources Amenomori and et al. (2009).

4 Extragalactic TeV Sources

4.1 Blazars

Approximately 1% of all galaxies host an active nucleus; a central compact region with much higher than normal luminosity. Around 10% of these Active Galactic Nuclei (AGN) exhibit relativistic jets, powered by accretion onto a supermassive black hole. Many of the observational characteristics of AGN can be attributed to the geometry of the system; in particular, the orientation of the jets with respect to the observer. Blazars, which host a jet oriented at an acute angle to the line of sight, are of particular interest for gamma-ray astronomy, as the emission from these objects is dominated by relativistic beaming effects, which dramatically boost the observed photon energies and luminosity.

Figure 4: The location of the 49 AGN (BL Lac, FSRQs and radio galaxies) with detected TeV emission (from TeVCat Wakely and Horan (????)). On the left is the spatial distribution, in Galactic coordinates; The right plot shows the redshift distribution, for 41 of the objects with redshifts listed in the literature.

The first extragalactic source discovered at TeV energies was Markarian 421 Punch and et al. (1992), a blazar of the BL Lacertae sub-class. The extragalactic TeV catalog now comprises objects, and continues to increase steadily (Figure 4). Blazar SEDs show a double-peaked structure in a representation of their spectral energy distribution (SED), with the lower frequency peak usually attributed to synchrotron emission of energetic electrons, and the higher frequency peak to inverse Compton. BL Lac objects are further classified as low-, intermediate- or high-frequency peaked, according to the location of the peak of their synchrotron emission. The majority () of the known TeV blazars are high-frequency peaked objects, in part because of inherent biases in the target selection: initially, objects were chosen based primarily upon their radio and X-ray spectral properties (e.g. Costamante and Ghisellini (2002)). More recently, the TeV observatories have expanded their selection criteria, using additional guidance in the form of Fermi-LAT results, and multi-wavelength observation triggers. This has broadened the catalog to include examples of intermediate- and low-frequency peaked objects. The overall data quality has also improved markedly since the launch of Fermi: Figure 5 shows a recent compilation of spectral measurements for the bright TeV blazar, Markarian 501 Abdo and et al. (2011), taken during an extensive multiwavelength campaign in 2009.

The mechanisms which drive the high energy emission from blazars remain poorly understood, and a full discussion is beyond the scope of this review. Briefly; in leptonic scenarios, a population of electrons is accelerated to TeV energies, typically through Fermi acceleration by shocks in the AGN jet. These electrons then cool by radiating X-ray synchrotron photons. TeV emission results from inverse Compton interactions of the electrons with either their self-generated synchrotron photons, or an external photon field. The strong correlation between X-ray and TeV emission which is often observed provides evidence for a common origin such as this, although counter-examples do exist Krawczynski and et al. (2004). Another class of models has hadrons as the primary particle population, which can then produce TeV gamma-rays through subsequent interactions with target matter or photon fields. Hadronic models are less favoured, typically, in part because the cooling times for the relevant processes are long, making rapid variability difficult to explain. One exception to this is the case of proton synchrotron emission, which may provide a plausible alternative, in which the emission results from extremely high energy protons in highly magnetized (), compact regions of the jet Aharonian (2000).

Figure 5: Extensive multifrequency measurements showing the spectral energy distribution of Markarian 501 for observations in 2009. Emission from the host galaxy is clearly visible at infrared/optical frequencies. The VERITAS data are divided to show both the average spectrum (red circles), and the spectrum during a 3-day flare (green triangles). Figure from Abdo et al. Abdo and et al. (2011)

Many of the AGN detected at TeV energies exhibit extreme variability. The timescales can range from years to minutes, and the observed flux can change by more than an order of magnitude. Figure 6 shows the H.E.S.S. lightcurve from July 2006 for one of the most extreme examples, the BL Lac object PKS 2155-304 Abdo and et al. (2007); H.E.S.S. Collaboration (2010). Such rapid variability can be used to place constraints on the size of the emission region, which depend upon the Doppler factor, . itself is constrained by the requirement that the emission region should be transparent to gamma-rays (e.g. Gaidos and et al. (1996)). Extremely rapid TeV gamma-ray variability of distant blazars can also be used to place limits on the energy-dependent violation of Lorentz invariance Biller and et al. (1999); H.E.S.S. Collaboration (2011), as predicted in some models of quantum gravity.

Figure 6: Integrated flux () versus time for H.E.S.S. observations of PKS 2155-304 on MJD 53944 (28 July, 2006). The data are binned in 1-minute intervals, and the horizontal dashed line shows the steady flux from the Crab Nebula for comparison. Figure from Aharonian and et al. (2007).

Related to the BL Lacertae objects are Flat Spectrum Radio Quasars (FSRQs). These are characterized primarily by their intense UV emission, associated with an accretion disk, strong broad emission lines in the optical spectrum, and infra-red emission associated with a dusty torus. FSRQs are similar to low-frequency peaked BL Lacs, in that the X-ray emission is dominated by the inverse Compton peak of the SED. Despite this, inverse Compton emission can extend up to TeV energies, particularly during intense flaring episodes, and three FSRQs have recently been detected by IACTs (PKS 1222+21 Aleksić and et al. (2011) , PKS 1510-089 Wagner and H.E.S.S. collaboration (2010) and 3C279 Aleksić and et al. (2011)).

A unique additional case is the TeV detection of IC 310, a “head-tail” radio galaxy in the Perseus cluster, possibly hosting a low-luminosity BL Lac nucleus Rector et al. (1999). Head-tail radio galaxies display a distinctive radio morphology, consisting of a bright “head” and a fainter “tail”, which is believed to be the result of their rapid motion with respect to the intracluster medium. The source was originally identified as a VHE emitter in an analysis of the highest energy () Fermi photons by Neronov et al. Neronov et al. (2010), and then subsequently detected from the ground by MAGIC Aleksić and et al. (2010). Neronov et al. considered the intriguing possibility that the emission might originate with particles accelerated at the bow shock formed by the interaction between the relativistic outflow from the galaxy and the intracluster medium. This scenario is ruled out by the detection of variability in the TeV flux by MAGIC, and the more familiar BL Lac mechanisms are now favoured.

Figure 4 also shows the distribution of measured redshifts for 41 TeV AGN. Some of these measurements are rather uncertain, since BL Lac optical spectra, by definition, do not contain strong emission lines. The most distant object detected is 3C279 Aleksić and et al. (2011), with a relatively modest redshift of . The population is truncated at large distances due to the absorption of TeV gamma-rays by electron-positron pair production with the low energy photons of the extragalactic background light (EBL). This effect is energy dependent, and can strongly modify the observed VHE spectra of extragalactic sources. While this limits the observation of more distant TeV sources, it also provides a mechanism by which to infer the intensity of the EBL, using reasonable assumptions about the intrinsic TeV spectra at the source Stecker et al. (1992). The EBL provides a calorimetric measure of the complete history of star and galaxy formation in the Universe, but is difficult to measure directly, due to the presence of bright local foreground sources of emission, such as zodiacal light. Presently, all of the TeV blazar measurements are consistent with a relatively low level of EBL, with the constraints derived from VHE measurements now approaching the lower limits derived from galaxy counts Aharonian and et al. (2006); MAGIC Collaboration (2008); Orr et al. (2011).

TeV blazar observations have also been suggested as probes of other physical phenomena, such as the acceleration and propagation of ultra-high energy cosmic rays Aharonian et al. (1994); Gabici and Aharonian (2007), or, more speculatively, the production of axion-like particles Hooper and Serpico (2007); Domínguez et al. (2011). Various authors have also discussed the possibility that TeV observations may be used to measure or constrain the strength of the intergalactic magnetic field (IGMF) (e.g. Aharonian et al. (1994); Tavecchio et al. (2011)). Temporal, spectral and spatial signatures of the IGMF are all possible; however, the fact that blazars are intrinsically variable gamma-ray sources limits the power of this technique. Accounting for this, Dermer et al. derive a lower limit of Dermer and et al. (2011).

4.2 Radio Galaxies

As described above, the TeV fluxes from blazars are dramatically enhanced by the effects of Doppler boosting. Nearby radio galaxies, in which the jet is not directly oriented towards the line-of-sight, provide an alternative method by which to investigate the particle acceleration and gamma-ray emission from relativistic outflows in AGN. The advantage of studying such objects lies in the fact that the jets can be resolved from radio to X-ray wavelengths, allowing the possibility of correlating the gamma-ray emission state with observed changes in the jet structure. Three radio galaxies have been identified as TeV emitters: M 87, Centaurus A and NGC 1275.

M 87 is the most well studied of these, and was first reported as a gamma-ray source by the HEGRA collaboration Aharonian and et al. (2003), with subsequent confirmation by H.E.S.S. Aharonian and et al. (2006), VERITAS Acciari and et al. (2008) and MAGIC Albert and et al. (2008). M 87 is a giant radio galaxy at a distance of , displaying a prominent misaligned jet, with an orientation angle of to the line-of-sight. The mass of the central black hole is estimated to be  M. The TeV source is strongly variable, and has undergone three episodes of enhanced emission in 2005, 2008 and 2010 (Figure 7). The results are summarized by Abramoswski et al Abramowski and et al. (2012). Causality arguments use the shortest variability timescale of around to place strong constraints on the size of the TeV emission region, corresponding to only a few Schwarzschild radii. The TeV emission region cannot be directly resolved with IACTs, but correlations with contemporaneous X-ray and radio observations provide some clues to its location. Two structures are of particular interest: the inner region close to the central black hole (the “core”), and HST-1, a bright jet feature first resolved in the optical band by the Hubble Space Telescope. HST-1 underwent a multi-year flare in radio, optical and X-rays, peaking around the time when the first short-term variability was detected at VHE energies Aharonian and et al. (2006). In contrast to this, the 2008 VHE flare was accompanied by enhanced radio and X-ray fluxes from the core region Acciari and et al. (2009). The 2010 VHE flare showed no enhanced radio emission from the core, although an enhanced X-ray flux was observed 3 days after the VHE peak. The results, therefore, remain somewhat ambiguous, and the possibility remains that the observed VHE flares may have different origins, or that the total VHE emission may be the sum of multiple components. Given the burdensome observing requirements for instruments with a limited duty cycle, M 87 represents the best example of the importance of data-sharing and coordinated observing planning between the various IACTs.

Figure 7: VHE light curves of the three flares from M87 observed in 2005, 2008 and 2010, showing integral fluxes above an energy of . Figure from Abramowski and et al. (2012).

The closest active galaxy, Centaurus A, at a distance of , was identified as a VHE source in a deep, 120 hour exposure by H.E.S.S. Aharonian and et al. (2009). Cen A is among the faintest VHE sources detected, with a flux of 0.8% of the Crab Nebula above . The emission is steady, although the variability is not strongly constrained as a result of the low flux level. As with M 87, numerous sites for the production of the TeV emission have been suggested, from the immediate vicinity of the central black hole (with a mass of  M), to the AGN jet, or even beyond Stawarz et al. (2006). The final, and most recent, addition to the known TeV radio galaxies is NGC 1275, identified by MAGIC Lombardi and et al. (2011). NGC 1275 is the central galaxy of the Perseus cluster, at a distance of . The VHE emission has a flux of of the Crab Nebula above and exhibits a very soft spectrum, with a power-law index of Hildebrand and et al. (2011), indicating a sharp turnover from the measured Fermi spectrum at lower energies, where the index is Brown and Adams (2011).

4.3 Starburst Galaxies

Starburst galaxies are those which exhibit an extremely high rate of star formation, sometimes triggered by interaction with another galaxy. High cosmic-ray and gas densities in the starburst region make these objects promising targets for gamma-ray observations, with emission predicted to result from the interactions of hadronic cosmic rays in the dense gas. TeV emission has now been identified from two starburst galaxies: M 82 VERITAS Collaboration (2009) and NGC 253 Acero and et al. (2009).

M 82 is a bright galaxy located at a distance of approximately , with an active starburst region at its centre. The star formation rate in this region is approximately 10 times that of the Milky Way, with an estimated supernova rate of 0.1 to 0.3 per year. A deep VERITAS exposure () in 2008-2009 resulted in a detection of gamma-ray emission from M 82 with a flux of above the 700 energy threshold of the analysis. NGC 253 lies at a distance of , and also has a central, compact () starburst region. The supernova rate in this region is estimated at per year. TeV emission was detected by H.E.S.S. with an integrated flux above of Acero and et al. (2009).

The emission from both M 82 and NGC 253 is consistent with the predictions of models based on the acceleration and propagation of cosmic rays in the starburst core (e.g. Lacki et al. (2011)), assuming that they act as efficient “proton calorimeters” (i.e. cosmic rays lose the majority of their energy to collisions). In this case, the estimated cosmic ray density in the starburst region is 2-3 orders of magnitude larger than that of the Milky Way. An alternative explanation is proposed by Mannheim et al. Mannheim et al. (2010), who suggest that the TeV luminosity is consistent with the combined emission from a large population of pulsar wind nebulae, which result from the elevated supernova rate. More accurate TeV spectra, and observation of other starburst classes, such as ultra-luminous infrared galaxies, should provide more insight in the future.

4.4 The Large Magellenic Cloud

Galaxies of the Local Group are also of interest to TeV observatories, although the predicted fluxes due to cosmic ray acceleration and propagation lie below the current instrumental sensitivity in the TeV range. Given their proximity, the possibility arises of detectable emission from individual objects, or from localized regions, particularly in the Milky Way’s satellite galaxies. H.E.S.S. has recently identified the first such object in the Large Magellenic Cloud (LMC), at a distance of Komin and et al. (2011). An unresolved source with a flux of of the Crab Nebula ( between 1 and 10 TeV) was detected, consistent with the location of PSR J0357-6910. This object is the most powerful pulsar known, with a spin-down energy of Ė. Given the positional coincidence, and based on comparisons with similar objects within our Galaxy, it seems likely that the TeV emission is due to inverse Compton emission from electrons in the pulsar’s wind nebula interacting with a strong infrared target photon field. This is in contrast to the extended GeV emission which has been observed from the LMC by Fermi-LAT, which is attributed to cosmic ray acceleration and interactions in the massive star forming region of 30 Doradus Abdo and et al. (2010).

4.5 Other Extragalactic TeV targets

TeV observations have also been used to place important constraints on the gamma-ray emission from numerous undetected extragalactic source classes, including galaxy clusters and potential sources of ultra-high energy cosmic rays. Here we summarize two of the most important: gamma-ray bursts, and the predicted sites of dense regions of dark matter particles.

Gamma-Ray Bursts: GRBs are the signatures of brief, extremely energetic explosions which occur at cosmological distances. They are observationally divided into short and long classes, which are presumably the result of different progenitor systems. Long duration GRBs are generally ascribed to the collapse of massive, rapidly rotating stars into black holes. The origin of the short bursts is less certain, although neutron star - neutron star merger events are among the favoured candidates. The search for emission has been a long-running goal of both the IACTs and particle detectors (see Williams and et al. (2009) for a review). Given the brief duration of GRB emission, the wide field-of-view of particle detectors is particularly important in this regard, although observations so far have been hampered by limited sensitivity. IACTs, conversely, must be re-pointed rapidly on receipt of an alert. The detection of delayed high energy emission by Fermi-LAT , lasting hundreds to thousands of seconds longer than the sub-MeV emission, has provided additional impetus to the search Abdo and et al. (2009a, b), and IACTs now regularly target burst locations within of the burst alert. The task is difficult, since the burst must also be at a small enough redshift such that the high energy emission is not completely suppressed by photon-photon pair production (). No convincing signals have been detected as yet (e.g. Acciari and et al. (2010); Aleksić and et al. (2010); Aharonian and et al. (2009); Acciari and et al. (2011)), although predictions based on the brightest bursts observed by the LAT indicate that the potential for detecting TeV emission associated with a GRB is promising, assuming no intrinsic spectral cut-off of the high energy emission Aune and for the VERITAS Collaboration (2011).

Extragalactic Dark Matter: The search for the self-annihilation signature of dark matter particles in astrophysical objects is wide-ranging, and complementary to direct detection techniques on the Earth (see Porter et al. (2011) for an excellent review). As discussed later in this review, the centre of our own Galaxy is a natural target for dark matter searches, and provides the most stringent limits to date Abramowski and et al. (2011). Objects outside of our own Galaxy are also worthy of investigation, however, and are potentially much less affected by contamination from unknown astrophysical background sources (supernova remnants, pulsar wind nebulae, etc.). Dwarf spheroidal galaxies of the local group are among the most promising of these, due to their proximity and their presumed large dark matter content. The Sloane Digital Sky Survey has more than doubled the known population of dwarf spheroidals, providing additional targets for the TeV searches. No signals have been detected as yet, despite deep exposures on a number of objects Vivier and for the VERITAS collaboration (2011); Aleksić and et al. (2011); H. E. S. S. Collaboration (2011); Aharonian and et al. (2009, 2008). Figure  8 shows upper limits on the annihilation cross-section derived for various annihilation channels using VERITAS observations of the Segue I dwarf spheroidal.

Figure 8: Upper limits at the 95% confidence level on the velocity-weighted annihilation neutralino cross-section for different annihilation channels, based on VERITAS observations of the Segue I dwarf spheroidal galaxy. The dark band represents the typical range of predictions. Figure from Vivier and for the VERITAS collaboration (2011)

5 Galactic TeV Sources

There are presently known TeV sources within our Galaxy, as indicated either by their association with known Galactic sources at other wavelengths, or by their location in the Galactic plane - a particularly compelling argument when coupled with a resolvable angular extent. The majority of these sources were identified as TeV emitters during the H.E.S.S. survey of the inner Galaxy Aharonian and et al. (2006). H.E.S.S. is the only IACT currently operating in the southern hemisphere, which allows it to view the inner Galaxy at high elevation, and hence with a low energy threshold and good sensitivity. H.E.S.S. was the first instrument with sufficient sensitivity to observe sources with of the Crab Nebula flux in this region, and the results have been revelatory - the Galactic plane is densely populated with TeV sources, primarily clustered within the inner in Galactic longitude (Figure 9). The most recent survey results consist of 2300 hours of observations, allowing the detection of over 50 sources within the range to and to Gast and et al. (2011). Observations of the outer Galactic regions by VERITAS, MAGIC, Milagro and ARGO-YBJ have revealed a less densely populated sky, but containing some unique objects of particular interest for TeV studies. Many Galactic TeV sources are extended, allowing detailed studies of source morphology and spatially resolved spectra, while others are time variable and/or periodic. The various source classes are discussed in some detail below.

Figure 9: Significance map of the Galactic Plane from the original H.E.S.S. survey in 2004, based on of observations. The complete survey now extends over the range from to and to , and comprises of observations Gast and et al. (2011). Figure from Aharonian and et al. (2006)

5.1 The Galactic Centre and Ridge

A TeV source at the location of the Galactic Centre has been reported by various IACTs Tsuchiya and et al. (2004); Kosack and et al. (2004); Albert and et al. (2006); Aharonian and et al. (2006a). Determining the nature of this source is a difficult task, due to the complexity of the region, which includes multiple different potential counterparts. The most detailed studies have been performed by H.E.S.S., which reveal that the emission is dominated by a bright central source, HESS J1745-290, lying close to the central supermassive black hole, Sgr A. An additional, fainter, component is also seen, which extends in both directions along the Galactic plane Aharonian and et al. (2006b). The extended component is spatially correlated with a complex of giant molecular clouds in the central of the Milky Way, and the TeV emission can be attributed to the decay of neutral pions produced in the interactions of hadronic cosmic rays with material in the clouds. The central source is point-like, steady and exhibits a curved power-law spectrum Aharonian and et al. (2009). Its location with respect to three of the most likely counterparts is shown in Figure 10. This study reveals that the source centroid is displaced from the radio centroid of the supernova remnant Sgr A East, excluding this object with high probability as the dominant source of the VHE gamma-ray emission, and leaving Sgr Aand the pulsar wind nebula G359.95-0.04 as the most likely counterparts Acero and et al. (2010).

The Galactic Centre is also a prime candidate region in which to search for gamma-ray emission due to dark matter particle self-annihilation. The analysis is complicated, however, because of the high background due to astrophysical sources. An analysis by H.E.S.S. using an optimized background subtraction technique shows no hint of a residual dark matter gamma-ray flux at a projected distance of from the Galactic Centre Abramowski and et al. (2011).

Figure 10: 90 cm VLA radio flux density map of the innermost of the Galactic Centre, showing emission from the SNR Sgr A East. Black contours denote radio flux levels; the centre of the SNR is marked by the white square, and the positions of Sgr A and G359.95-0.04 are given by the cross hairs and the black triangle, respectively. The 68% CL total error contour of the best-fit centroid position of HESS J1745-290 is given by the white circle. Figure from Acero and et al. (2010): see that paper for full details.

5.2 The Crab Nebula and Pulsar

The Crab is the nearby () remnant of a historical supernova explosion, observed in 1054 A.D. There is no detected shell, and the broadband emission below is dominated by a bright synchrotron nebula, powered by a central pulsar (PSR B0531+21). PSR B0531+21 is the most energetic pulsar in our Galaxy, with a pulse period of , and a spin-down power of . The Crab Nebula and Pulsar hold a unique place in the development of TeV astronomy: the birth of the field as an astronomical discipline can be traced to the detection of the Crab Nebula TeV source by Weekes et al. using the Whipple telescope, in the first application of the imaging atmospheric Cherenkov technique Weekes and et al. (1989). Subsequently, the Crab has acted as a bright, standard candle for TeV observatories.

The SED of the non-thermal nebula emission displays two components (Figure 11). The dominant, low frequency component is explained by synchrotron radiation of high energy electrons spiraling in the magnetic field of the nebula Kennel and Coroniti (1984); de Jager and et al. (1996). The higher frequency component is attributed to inverse Compton scattering of lower energy photons by these electrons, including microwave background photons, far infrared and the electron-synchrotron photons themselves. The electron population reaches energies of at least , through acceleration occuring both in a relativistic particle outflow driven by the spin-down energy of the pulsar, and in shocks where this outflow encounters the surrounding nebula. The highest energy particles likely require an alternative explanation for their origin, such as direct acceleration in intense electric fields associated with the pulsar itself Abdo and et al. (2011). Observations of the synchrotron nebula from radio to X-ray wavelengths provide high resolution imaging of the emission region; however, the synchrotron data alone only contain information concerning the product of the magnetic field strength and the relativistic electron density. Since the inverse Compton component is independent of the magnetic field strength, the combined SED allows an estimate of the Nebula magnetic field, which is now constrained to be between 100 and Abdo and et al. (2010).

The search for a VHE component to the pulsed emission from the Crab has been long and, until recently, fruitless. Despite discouraging model predictions, and the detection of spectral cut-offs below in other pulsars, the fact that no super-exponential cut-off was observed by EGRET in the Crab Pulsar spectrum initially provided some encouragement for a continued search by IACTs Kuiper and et al. (2001). Fermi-LAT subsequently extended the GeV spectrum and measured a sharp spectral cut-off at Abdo and et al. (2010). A campaign by MAGIC, using a specially designed “analog sum” hardware trigger, provided the first ground-based measurement of gamma-ray emission from the Crab pulsar Aliu and et al. (2008). The initial MAGIC flux measurement above was, like the Fermi-LAT result, consistent with an exponential cut-off. The existence of an exponential cut-off is a natural consequence of emission due to curvature radiation, as favored by various models (e.g. Tang et al. (2008)). Both VERITAS and MAGIC recently presented new results, which challenge this paradigm Aliu and et al. (2011); Aleksić and et al. (2011); Saito and for the MAGIC collaboration (2011). Pulsed emission is observed to extend up to well beyond , and the combined LAT-IACT spectrum can best be fit with a broken power law. The explanation for this high energy component is an open question, at present. Gamma-ray opacity arguments require that the emission zone of the highest energy photons must be at least 10 stellar radii from the surface of the neutron star - much further than previously assumed. The results require either a substantial revision of existing models of high energy pulsar emission, or the addition of a new component, not directly related to the MeV-GeV emission.

Figure 11: Left: The spectral energy distribution of the Crab Nebula from Abdo et al. Abdo and et al. (2010). Their fit to the synchrotron component is shown (blue dashed line), as well as inverse Compton spectra from Atoyan and Aharonian Atoyan and Aharonian (1996) for assumed magnetic field strengths of (solid red line), (dashed green line) and (dotted blue line). Right: High Energy spectrum of the Crab Pulsar. The black dashed line shows a fit of a power law with exponential cut-off to the Fermi-LAT data alone; the solid line shows a broken power law fit to the LAT and IACT data. The red dashed line and bowtie shows a power law fit to the VERITAS points alone. Figure courtesy of N. Otte (priv. comm.).

As mentioned above, the Crab has been used as a standard candle in TeV astronomy, on the assumption that its emission was steady. This is now demonstrably false, at least at energies below , with the detection of multiple day-scale flaring events Striani and et al. (2011); Abdo and et al. (2011), and long term variation in the hard X-ray/ soft gamma-ray regime Wilson-Hodge and et al. (2011). The GeV spectrum during flares indicates that the emission is confined to the synchrotron component of the SED, a conclusion which is supported by the rapid timescale of the events (since the inverse Compton or Bremsstrahlung cooling time of the emitting electrons is much greater than the observed flare duration). At higher energies, some evidence for an enhanced flux during HE flare states has been presented by ARGO-YBJ Di Sciascio and for the ARGO-YBJ Collaboration (2011). IACT measurements do not support these results, but are not necessarily in conflict, given the differing duty cycles. Detailed measurements with IACTs during future flare states are required to resolve this question.

5.3 Pulsar Wind Nebulae

Pulsar wind nebulae are the most abundant class of known VHE emitters in the Galaxy, with firm examples, and numerous other sources where the PWN association is more tentative (for reviews see e.g. de Oña Wilhelmi (2011); Kargaltsev and Pavlov (2010); Slane (????)). The essential emission mechanisms - shock accelerated leptons producing synchrotron and inverse Compton radiation - have already been described for the case of the Crab PWN, but the Crab is far from the typical object. Understanding of the structure and evolution of PWN has advanced significantly over the past few years, in particular thanks to the high resolution X-ray imaging provided by Chandra (see Gaensler and Slane Gaensler and Slane (2006) for a detailed review). Initially, the PWN expands uniformly from the central pulsar, while at later stages the nebula may be confined and distorted by the reverse shock from the expanding supernova remnant (SNR).

At TeV energies, young PWN are usually still embedded within their parent SNR and are point-like, within the angular resolution of IACTs. They are positionally coincident with a bright X-ray synchrotron nebula powered by a pulsar with very high spin-down luminosity (e.g. G0.9+0.1 Aharonian and et al. (2005a), HESS J1813-178 Aharonian and et al. (2005b), G54.1+0.3 Acciari and et al. (2010)). More evolved PWN, with ages years, are usually much larger, and their TeV emission can be spatially resolved and mapped. The pulsar powering the nebula is often offset from the center of the TeV emission, probably for reasons related to density gradients in the medium surrounding the SNR Blondin et al. (2001). Remarkably, the TeV nebulae are often two or three orders of magnitude larger than the corresponding X-ray PWN, and the TeV PWN sizes tend to increase with age, while the X-ray PWN sizes show the opposite trend. This can be understood as a result of the fact that the electron population which is responsible for the TeV inverse Compton flux has lower energies than the electrons which produce the X-ray synchrotron emission. They therefore cool more slowly, and survive for longer so, while the X-ray nebula is dominated by freshly accelerated particles, the TeV nebula can record the entire history of particle propagation away from the pulsar. A natural result of this is that the observed TeV spectrum should vary with distance from the pulsar. One of the best examples of this is HESS J1825-137, associated with the PWN of the pulsar PSRJ1826-1334 Aharonian and et al. (2006a). Figure 12 shows the spatially dependent spectra for this source, which soften with increasing distance from the pulsar. This is interpreted as the natural effect of both inverse Compton and synchrotron cooling of the electron population during propagation. A counter-example is the case of Vela-X Aharonian and et al. (2006b); Hinton et al. (2011), in which no spectral variability is seen over the extended nebula, suggesting that cooling does not play an important role.

Figure 12: Inset: H.E.S.S. gamma-ray excess map for HESS J1825-137. The wedges show the radial regions with radii in steps of in which the energy spectra were determined. The main figure shows the differential energy spectra for the regions illustrated in the inset, scaled by powers of 10 for the purpose of viewing. The spectrum for the analysis at the pulsar position is shown as a reference along with the other spectra as dashed line. Figure from Aharonian and et al. (2006a): see that paper for full details.

5.4 Supernova Remnants

The search for the origin of the cosmic rays triggered the development of gamma-ray astronomy, and continues to motivate many gamma-ray observations. Chief among these is the study of supernova remnants, which are believed to efficiently accelerate particles at the shock front where the expanding SNR encounters the surrounding medium (e.g. Drury et al. (1994)). This likely occurs through diffusive shock acceleration (first order Fermi acceleration), in which charged particles are reflected from magnetic inhomogeneities and repeatedly cross the shock front, gaining energy with each crossing (see e.g. Malkov and O’C Drury (2001); Blandford and Eichler (1987)). As well as plausibly providing enough energy to explain the observed Galactic cosmic ray population, this process naturally produces a power law distribution of particle energies with an index of , which matches the cosmic ray spectrum (after accounting for diffusion and escape). In recent years, the importance of magnetic field amplification by the accelerated particles themselves has been increasingly recognised, and plays a particular role in explaining the existence of the highest energy Galactic cosmic rays, around the cosmic ray knee region (at ).

The evidence for efficient leptonic acceleration in SNRs is now clearly established (e.g. Uchiyama and et al. (2007)); however, the question of whether SNR are efficient hadron accelerators is more difficult to answer. A definitive measurement would be the detection of high energy neutrinos from an SNR, but the expected fluxes are likely below the sensitivity thresholds of current neutrino observatories. Gamma-ray observations may provide the key, since the interactions of high energy nuclei with target material produce neutral pions, which decay immediately into gamma-rays. Disentangling the spectral signature of this process from other sources of gamma-ray emission (i.e. leptonic inverse Compton and bremsstrahlung processes) is difficult, but not impossible. Two classes of gamma-ray source are of interest for these studies: those which can be clearly associated with SNR shells, based on the gamma-ray morphology, and sources which are coincident with a massive volume of target material, such as molecular clouds.

The definitive association of gamma-ray emission with an SNR shell is often difficult to make, due to the presence of other potential counterparts, particularly PWN. A handful of shell-type SNRs have been unequivocally identified as gamma-ray sources by IACTs. This identification can be made on the basis of the observed shell morphology, (RXJ 1713.7-3946 Aharonian and et al. (2006), RXJ 0852.0-4622 (Vela Jr) Aharonian and et al. (2007), HESS J1731-3467 Abramowski and et al. (2011), SN1006 Acero and et al. (2010) and, possibly, RCW 86 Aharonian and et al. (2009)), or, in the case of Tycho’s SNR Acciari and et al. (2011), on the positional coincidence, coupled with the fact that the progenitor was a known Type Ia exposion, and so no compact object is present. Figure 13 shows the gamma-ray map for the first SNR shell to be resolved, RXJ 1713.7-3946. The recent addition of Fermi-LAT observations to the broadband spectrum of RXJ 1713.7-3946 Abdo and et al. (2011) are consistent with a leptonic origin as the dominant mechanism for the gamma-ray emission. A counter-example is illustrated by the spectrum in Figure 13, which corresponds to Tycho’s SNR. In this case, the fact that the broadband gamma-ray spectrum can be fit with a single hard power law from to favours a hadronic origin Morlino and Caprioli (2011); Giordano and et al. (2012). This interpretation is not completely compelling, however, given the large statistical errors in the measurements, and the impact of various unknown parameters such as the SNR distance, and possible enhancements of the gamma-ray flux due to a nearby molecular cloud. Additionally, Atoyan and Dermer Atoyan and Dermer (2011) describe a two-zone leptonic model which provides an acceptable spectral fit. Future measurements of the spectrum below , and deeper exposure at TeV energies, will further test the differing interpretations.

Figure 13: Left: H.E.S.S. map of gamma-ray excess events for RXJ 1713.7-3946 - the first SNR shell to be resolved at TeV energies. Figure from Aharonian and et al. (2006). Right: The broadband SED of Tycho’s SNR from Giordano and et al. (2012), together with models for the various emission components (dominated by hadronic processes in the gamma-ray band). See paper for details.

The intensity of gamma-ray emission due to hadronic interactions depends upon the flux of high energy nuclei, and also upon the density of target material. Regions of high matter density (e.g. molecular clouds with densities ), situated close to sites of particle acceleration (such as SNRs), can therefore be expected to produce a large gamma-ray flux due to hadronic interactions. At TeV energies, likely candidates for this process have been identified. The task of identification is complicated, both for the usual reasons of source confusion, and also because the evidence for a molecular cloud / SNR interaction is only definitive in those cases where the cloud morphology is visibly deformed by the expanding SNR, and/or where sites of hydroxyl (OH) maser emission indicate the presence of shocked molecular material. One of the best examples of this source class is the old remnant W28 Aharonian and et al. (2008). H.E.S.S. observations of this region show four distinct sites of emission, with three of the four showing a resolvable angular extent (). Each of the TeV sources is positionally coincident with a molecular cloud. Assuming that the TeV emission is due to hadronic cosmic rays interacting with the cloud material, the cosmic ray density is inferred to be a factor of 10 to 30 times greater than in the solar neighbourhood.

5.5 Star forming regions

The process of diffusive shock acceleration is not limited to supernova remnant shells. An alternative scenario invokes particle acceleration at the shock formed by the collision between the supersonic stellar winds of massive stars in close binary systems. Stellar winds may also become collectively important in large assemblies of massive stars. The combined effect of the stellar winds, coupled with the effect of multiple SNRs, results in an overall wind from the cluster which forms a giant superbubble in the interstellar medium. Particle acceleration can occur where the cluster wind interacts with the surrounding medium (e.g. Butt (2009)).

Massive star associations are naturally likely to host other potential source counterparts for TeV emission, such as compact object binary systems, individual supernova remnants and pulsar wind nebulae. A case in point is the young, open cluster Westerlund 2, containing the Wolf-Rayet binary system WR 20a. This was originally suggested as a plausible counterpart to the unidentified source HESS J1023-575, with the emission presumed to be connected to either the Wolf-Rayet binary, or the combined cluster wind Aharonian and et al. (2007). A re-assesment of this region, informed by a deeper H.E.S.S. exposure and results from Fermi-LAT, alters the picture H.E.S.S. Collaboration (2011). The LAT detects an energetic pulsar, PSR J1022-5746, which drives a PWN which is bright in GeV gamma-rays Ackermann and et al. (2011). Given the ubiquity of bright TeV PWN, this now seems the most likely explanation for the TeV source. A similar conclusion may arise for the first TeV source to be linked with a massive star association, TeV 2032+4130 (coincident with the Cyg OB2 association). In this case, the LAT pulsar (PSR J2032+4127) is sufficiently energetic to explain the TeV emission, although no PWN has been detected as yet.

Other unidentified sources which have been linked with massive star clusters and associations include HESS J1646-458 (Westerlund 1) H.E.S.S. Collaboration (2011), HESS J1614-518 (Pismis 22) Ohm and et al. (2010), HESS J1848-018 (W43, which hosts Wolf-Rayet star WR121a) Chaves et al. (????) and W49A Brun and for the H. E. S. S. Collaboration (2011), a massive star forming region. For all of these, however, the evidence that particle acceleration in stellar winds is the driving force behind the gamma-ray emission is not definitive (e.g. Lemoine-Goumard and et al. (2011)).

5.6 Compact Object Binary Systems

Despite many early unconfirmed claims, the first definitive detection of a TeV gamma-ray binary system was not published until 2005. The population has grown slowly, and now consists of four clearly identified systems, plus marginal evidence for transient emission associated with Cyg X-1 Albert and et al. (2007). The gamma-ray emission from binaries is believed to be powered either by accretion (most likely onto a black hole), or by a pulsar wind. In the case of accretion, particle acceleration takes place in relativistic jets (e.g. Bosch-Ramon and Khangulyan (2009)). In the pulsar wind scenario, the acceleration occurs either in shocks where the pulsar wind encounters the circumstellar environment (e.g. Dubus (2006)), or possibly within the pulsar wind zone itself Sierpowska-Bartosik and Torres (2008). The detection of Cyg X-1, if confirmed, would be extremely important, since there is no doubt that this system hosts a black hole. This is in contrast to all of the other TeV binaries, in which the compact object is either known to be, or may be, a neutron star. Here we briefly summarize the results for each of the four well-studied objects.

PSR B1259-63/LS 2883: This was the first gamma-ray binary system to be firmly detected at TeV energies, and the first known variable VHE source in our Galaxy Aharonian and et al. (2005). The system comprises a pulsar orbiting a massive B2Ve companion. The orbit is highly eccentric (), with a period of 3.4 years. The TeV emission exhibits two peaks, approximately 15 days before and after periastron. Various authors have attempted to explain the double bumped VHE lightcurve within a ’hadronic disk scenario’, in which a circumstellar disk provides target material for accelerated hadrons, leading to production and subsequent TeV gamma-ray emission Kawachi and et al. (2004); Chernyakova and et al. (2006). The 2007 H.E.S.S. observations disfavour this, since the onset of TeV emission occurs prior to periastron, well before interactions with the disk could be expected to play a significant role Aharonian and et al. (2009). Leptonic scenarios have also been discussed in e.g. Kangulyan et al. Khangulyan et al. (2007).

The recent discovery of an extended and variable radio structure in PSR B1259-63/LS 2883 at phases far from periastron provides definitive evidence that non-accreting pulsars orbiting massive stars can produce variable and extended radio emission at AU scales Moldón et al. (2011). This is important, since similar structures in LS 5039 and LS I +61303, where the nature of the compact object is not certain, have been used to argue for the existence of jets driven by accretion onto a black hole.

LS 5039: LS 5039 consists of a compact object, either neutron star or black hole, orbiting a massive O6.5V (23 M) star in a 3.9 day orbit. Observations by H.E.S.S. in 2004 revealed that LS 5039 is a bright source of VHE gamma-rays Aharonian and et al. (2006). Unlike PSR B1259-63 (and, to a lesser extent, LS I +61303) LS 5039 is almost perfectly suited to TeV observations, with a short orbital period and a convenient declination angle, allowing sensitive observations at all phases over numerous orbits. The VHE emission measured by H.E.S.S. is modulated at the orbital period, peaking around inferior conjunction, when the compact object is closest to us and co-aligned with our line-of-sight (Figure 14). The spectrum is also orbitally modulated, appearing significantly harder around inferior conjuction (), but with an exponential cut-off at . At GeV energies, the source is detected by Fermi-LAT at all orbital phases, with the emission peaking close to superior conjuction, in apparent anti-phase with the VHE results Abdo and et al. (2009). A sharp spectral cut-off at is observed in the LAT data near superior conjunction, indicating that the VHE spectra cannot be simply a smooth extrapolation of the lower energy emission.

Figure 14: The H.E.S.S. flux (bottom) and photon index (top) for LS 5039 as a function of orbital phase. Figure from Aharonian et al. Aharonian and et al. (2006).

LS I +61303: Similar to LS 5039, LS I +61303 consists of a compact object, either neutron star or black hole, in this case orbiting a B0Ve star with a circumstellar disk (12.5 M) in a 26.5 day orbit. The detection of a variable VHE source at the location of LS I +61303 with MAGIC Albert and et al. (2009), later confirmed by VERITAS Acciari (2008), established this source as a gamma-ray binary. The object is now one of the most heavily observed locations in the VHE sky, with deep exposures by the two observatories spread over half a decade. Despite this, the VHE source is much less well-characterized than LS 5039, owing to its relatively weak VHE flux, and an inconvenient orbital period which closely matches the lunar cycle, making observations over all orbital phases almost impossible within a single observing season. VHE emission was originally detected close to apastron, between phases . In contrast to this, VHE observations between 2008 and 2010 showed that, at least during the orbits when the source was observed, the apastron flux was much lower than during the previous detections Acciari and et al. (2011); MAGIC Collaboration (2011). The detection of TeV emission by VERITAS during a single episode close to superior conjunction complicates the picture even further. As with LS 5039, the Fermi-LAT GeV emission peaks closer to periastron, and the spectrum displays sharp cut-off, at . Long term variability in the GeV band has also been observed Hadasch et al. (2012).

HESS J0632+057: This TeV source was serendipitously detected during H.E.S.S. observations of the Monoceros Loop SNR region, and noted as a potential binary primarily because of its small angular extent Aharonian and et al. (2007). Subsequent observations revealed it to be a variable TeV source Acciari and et al. (2009), and co-located with a variable radio and X-ray source, at the position of a massive Be star, MWC 148 Hinton and et al. (2009). X-ray observations with Swift recently provided definitive proof of its binary nature, with the measurement of a periodicity in the lightcurve Bongiorno and et al. (2011). The TeV light curve displays a broad gamma-ray flare close to the X-ray maximum, with a duration of Maier et al. (2011). No GeV source has been detected.

A number of competing processes likely contribute to the variability observed in the gamma-ray binaries. In particular, the efficiency of inverse Compton gamma-ray production, as well that of VHE gamma-ray absorption (through pair production), changes as a function of orbital phase. This can go some way towards explaining the apparent phase shift between the GeV and TeV lightcurves, for example in the case of LS 5039. There are clearly other effects which contribute, however, as demonstrated by the long term instability of LS I +61303. The sharp GeV cut-offs in these systems are also difficult to explain, and may indicate that the GeV and TeV emission components do not have the same origin.

A final comment should be reserved for the Fermi-LAT source 1 FGL J1018.6-5856, which was recently identified as a GeV binary system, with a period of Corbet and et al. (2011). This source resides in a complex region, coincident with the center of the SNR G284.3-1.8, and close to a LAT pulsar (at a distance of 35’). The extended, unidentified HESS source, HESS J1018-589, overlaps with the GeV binary location, and appears to consist of a point-like source overlaid on a diffuse structure de Ona Wilhelmi and et al. (????). While this is indicative of a new TeV binary there is, as yet, no evidence for variability in the point-like emission, making the identification still uncertain.

5.7 Globular Clusters

A single globular cluster, Terzan 5, has been suggested as the probable counterpart of a TeV source (HESS J1747-248) Domainko and et al. (2011). If the association is correct, the emission is likely related to the large population of millisecond pulsars in this cluster, which provide an injection source of relativistic leptons Venter et al. (2009). Inverse Compton gamma-ray photons result when these electrons upscatter low energy photons of the intense stellar radiation field. The H.E.S.S. source is extended, and slightly offset from the cluster core (although there is significant overlap). The probability of this being simply a chance positional coincidence is .

Globular clusters have also been favored targets in searches for dark matter particle annihilation signatures, since they may have been generated in dark matter mini-haloes before the formation of galaxies took place, and thus retain a significant dark matter component Komatsu and et al. (2009); Wood and et al. (2008). Limits have been placed on NGC 6388, M15, Omega Centauri, 47 Tuc, M13, and M5 (Abramowski and et al. (2011), and references therein).

5.8 Unidentified objects

IACTs are able to locate point sources with reasonably good accuracy (typically for a moderately strong source). Extragalactic TeV sources can therefore usually be firmly identified with a single counterpart at other wavelengths. For Galactic sources, identification poses more of a problem, and around one third of the Galactic sources lack a firm identification. While the diffuse Galactic background emission, which dominates at GeV energies, is not significant, the TeV sources themselves are mostly extended, and can often be plausibly associated with multiple counterparts. PWN, in particular, often have their brightest TeV emission offset from the location of the parent pulsar or X-ray PWN, which may not yet have been detected (a number of new energetic pulsars have been located in follow-up observations of unidentified TeV sources). In other cases, despite deep X-ray and radio follow-up observations, no reasonable counterpart has yet been found.

The unidentified Milagro sources also pose some interesting questions. While IACTs have identified sources associated with these objects, the TeV sources are typically much smaller in angular extent, and cannot account for the entire Milagro flux. One possible explanation for this is that there is a diffuse high energy component to the emission, which is difficult to resolve with IACTs, given their relatively small fields of view.

6 The Future

The field of ground-based gamma-ray astronomy has expanded dramatically over the past 10 years, but it is worth noting that the observatories currently operating are far from reaching the physical limits of the detection techniques. We therefore conclude this review with a brief discussion of some of the instrumental developments which can be expected over the next decade.

The IACTs currently operating have all made significant efforts to maintain, and improve, sensitivity since they were first commissioned. H.E.S.S. recoated the telescope mirrors, and successfully developed sophisticated analysis tools with which to exploit the data. VERITAS relocated their original prototype telescope to provide a more favorable array layout, halving the time required to detect a weak source. Most significantly, MAGIC added a second telescope of similar design to the first, providing a stereo pair with a baseline of . All of these observatories have further upgrade plans. Both MAGIC and VERITAS are implementing camera upgrades - in the case of VERITAS this involves the replacement of all of the camera PMTs with more sensitive, super-bialkali devices in summer 2012. H.E.S.S. are constructing H.E.S.S. II - the addition of a single large telescope, with mirror area, to the centre of the array. Figure 15 shows the telescope structure in November 2011. The mirrors and camera will be installed in the first half of 2012.

Figure 15: The steel structure for the H.E.S.S. II telescope. Mirrors and camera will be installed in the first half of 2012. (Note that the H.E.S.S. II telescope is located in between the two H.E.S.S. I telescopes shown in the image, not in the foreground). Figure courtesy of the H.E.S.S. collaboration http://www.mpi-hd.mpg.de/hfm/HESS/

Novel approaches are also being pursued by many smaller projects. A particularly nice example of this is FACT (the First G-APD Cherenkov Telescope), which has recently demonstrated the application of Geiger-mode Avalanche Photodiodes for Cherenkov astronomy. The FACT telescope consists of a 1440-pixel G-APD camera at the focus of one of the original HEGRA telescopes. Geiger-APDs hold great promise as a potential replacement for PMTs, since they are robust, and offer much better photon conversion efficiency. Figure 16 shows some “first light” images from FACT. Other projects in development include GAW (Gamma Air Watch), MACE (Major Atmospheric Cerenkov Telescope Experiment) and LHASSO (Large High Altitude Air Shower Observatory). LHASSO is an ambitious project which will be located near the site of the ARGO-YBJ experiment in Tibet. It is planned to consist of four water Cherenkov detectors, two IACTs, three fluorescence telescopes and a large scintillator array.

Figure 16: First light cosmic ray images from FACT (First G-APD Cherenkov Telescope). Four different events are shown; the upper right event shows the characteristic ring image produced by a local muon. The camera contains 1440 Geiger-mode avalanche photodiodes, installed on one of the original HEGRA telescopes at the Roque de los Muchachos on La Palma. Figure courtesy of the FACT collaboration http://fact.ethz.ch/first

For the particle detectors, the next stage in instrumentation is HAWC (the High Altitude Water Cherenkov Observatory). HAWC will consist of 300 individual water Cherenkov tanks, at an altitude of in Mexico. The final array is expected to be 15 times more sensitive than Milagro, and will be a powerful tool for surveying, and for observations of transient phenomena. A prototype system is already in operation, and science operations will start in spring 2012, with completion of the full array expected in 2014.

The most ambitious future project is CTA (the Cherenkov Telescope Array). This is described in detail in Actis and et al. (2011). Briefly, it comprises an array of imaging atmospheric Cherenkov telescopes covering , providing a factor of 5-10 improvement in sensitivity in the - range, and extending the energy range both above and below these values. Both a northern and a southern site are envisaged, and the array will be operated as an open observatory. Multiple telescope designs are planned, including small (few), medium () and large () diameter reflectors, as well as two-mirror telescope designs, such as the Schwarzchild-Couder Vassiliev and Fegan (????).

In conclusion, TeV gamma-ray astronomy now describes a broad astronomical discipline which addresses a wide, and expanding, range of astrophysical topics. With planned instrumental developments, it is not unreasonable to expect the source catalogue to exceed 1000 objects within the next decade. Much of the TeV sky remains relatively unexplored. Less than of the sky has been observed with sensitivity at , and the sensitive exposure to transient events and widely extended sources at these energies is much lower. The likelihood of continued exciting results is certain, both for the known sources and source classes, and for new discoveries.


  • Weekes and et al. (1989) T. C. Weekes, et al., ApJ 342 (1989) 379–395.
  • Hinton and Hofmann (2009) J. A. Hinton, W. Hofmann, ARA&A 47 (2009) 523–565.
  • Aharonian et al. (2008) F. Aharonian, J. Buckley, T. Kifune, G. Sinnis, Reports on Progress in Physics 71 (2008) 096901.
  • Chadwick et al. (2008) P. M. Chadwick, I. J. Latham, S. J. Nolan, Journal of Physics G Nuclear Physics 35 (2008) 033201.
  • Aharonian (2004) F. A. Aharonian, Very high energy cosmic gamma radiation: a crucial window on the extreme Universe, River Edge, NJ: World Scientific Publishing, 2004.
  • Morrison (1958) P. Morrison, Nuovo Cimento 7 (1958) 858.
  • Cocconi (????) G. Cocconi, in: Proc. International Cosmic Ray Conference, Moscow, 2 (1959) 309.
  • Galbraith and Jelley (1953) W. Galbraith, J. V. Jelley, Nature 171 (1953) 349–350.
  • Weekes and Turver (????) T. C. Weekes, K. E. Turver, in: R. D. Wills & B. Battrick (Ed.), Recent Advances in Gamma-Ray Astronomy, 124 (1977) 279.
  • Hillas (1985) A. M. Hillas, in: F. C. Jones (Ed.), Proc. 19th International Cosmic Ray Conference, NASA Goddard Space Flight Center, 7 (1985) 231.
  • Punch and et al. (1992) M. Punch, et al., Nature 358 (1992) 477.
  • Quinn and et al. (1996) J. Quinn, et al., ApJL 456 (1996) L83.
  • Daum and et al. (1997) A. Daum, et al., Astroparticle Physics 8 (1997) 1–11.
  • Wakely and Horan (????) S. P. Wakely, D. Horan, in: Rogelio Caballero, Juan Carlos D’Olivo, Gustavo Medina-Tanco, Lukas Nellen, Federico A. Sánchez, José F. Valdés-Galicia (Ed.), Proc. 30th International Cosmic Ray Conference, Mérida, Yucatán, Mexico, 3 (2008) 1341.
  • McKay and et al. (1993) T. A. McKay, et al., ApJ 417 (1993) 742.
  • Aharonian and et al. (2002) F. Aharonian, et al., A&A 390 (2002) 39–46.
  • Abdo and et al. (2009) A. A. Abdo, et al., ApJL 700 (2009) L127–L131.
  • Abdo and et al. (2007) A. A. Abdo, et al., ApJL 664 (2007) L91–L94.
  • Cao et al. (2011) Z. Cao, S. Z. Chen, for the ARGO-YBJ collaboration, ArXiv e-prints, 1110.1809 (2011).
  • Amenomori and et al. (2009) M. Amenomori, et al., ApJ 692 (2009) 61–72.
  • Costamante and Ghisellini (2002) L. Costamante, G. Ghisellini, A&A 384 (2002) 56–71.
  • Abdo and et al. (2011) A. A. Abdo, et al., ApJ 727 (2011) 129.
  • Krawczynski and et al. (2004) H. Krawczynski, et al., ApJ 601 (2004) 151–164.
  • Aharonian (2000) F. A. Aharonian, New Ast. Rev. 5 (2000) 377–395.
  • H.E.S.S. Collaboration (2010) H.E.S.S. Collaboration, A&A 520 (2010) A83.
  • Gaidos and et al. (1996) J. A. Gaidos, et al., Nature 383 (1996) 319–320.
  • Biller and et al. (1999) S. D. Biller, et al., Physical Review Letters 83 (1999) 2108–2111.
  • H.E.S.S. Collaboration (2011) H.E.S.S. Collaboration, Astroparticle Physics 34 (2011) 738–747.
  • Aharonian and et al. (2007) F. Aharonian, et al., ApJL 664 (2007) L71–L74.
  • Aleksić and et al. (2011) J. Aleksić, et al., ApJL 730 (2011) L8.
  • Wagner and H.E.S.S. collaboration (2010) S. J. Wagner, H.E.S.S. collaboration, in: AAS/High Energy Astrophysics Division, Meeting #11 (2010) #27.06.
  • Aleksić and et al. (2011) J. Aleksić, et al., A&A 530 (2011) A4.
  • Rector et al. (1999) T. A. Rector, J. T. Stocke, E. S. Perlman, ApJ 516 (1999) 145–162.
  • Neronov et al. (2010) A. Neronov, D. Semikoz, I. Vovk, A&A 519 (2010) L6.
  • Aleksić and et al. (2010) J. Aleksić, et al., ApJL 723 (2010) L207–L212.
  • Stecker et al. (1992) F. W. Stecker, O. C. de Jager, M. H. Salamon, ApJL 390 (1992) L49–L52.
  • Aharonian and et al. (2006) F. Aharonian, et al., Nature 440 (2006) 1018–1021.
  • MAGIC Collaboration (2008) MAGIC Collaboration, Science 320 (2008) 1752.
  • Orr et al. (2011) M. R. Orr, F. Krennrich, E. Dwek, ApJ 733 (2011) 77.
  • Aharonian et al. (1994) F. A. Aharonian, P. S. Coppi, H. J. Voelk, ApJL 423 (1994) L5–L8.
  • Gabici and Aharonian (2007) S. Gabici, F. A. Aharonian, Ap&SS 309 (2007) 465–469.
  • Hooper and Serpico (2007) D. Hooper, P. D. Serpico, Physical Review Letters 99 (2007) 231102.
  • Domínguez et al. (2011) A. Domínguez, M. A. Sánchez-Conde, F. Prada, JCAP 11 (2011) 20.
  • Tavecchio et al. (2011) F. Tavecchio, G. Ghisellini, G. Bonnoli, L. Foschini, MNRAS 414 (2011) 3566–3576.
  • Dermer and et al. (2011) C. D. Dermer, et al., ApJL 733 (2011) L21.
  • Aharonian and et al. (2003) F. Aharonian, et al., A&A 403 (2003) L1–L5.
  • Aharonian and et al. (2006) F. Aharonian, et al., Science 314 (2006) 1424–1427.
  • Acciari and et al. (2008) V. A. Acciari, et al., ApJ 679 (2008) 397–403.
  • Albert and et al. (2008) J. Albert, et al., ApJL 685 (2008) L23–L26.
  • Abramowski and et al. (2012) A. Abramowski, et al., ApJ 746 (2012) 151.
  • Acciari and et al. (2009) V. A. Acciari, et al., Science 325 (2009) 444.
  • Aharonian and et al. (2009) F. Aharonian, et al., ApJL 695 (2009) L40–L44.
  • Stawarz et al. (2006) Ł. Stawarz, F. Aharonian, S. Wagner, M. Ostrowski, MNRAS 371 (2006) 1705–1716.
  • Lombardi and et al. (2011) S. Lombardi, et al., ArXiv e-prints, 1111.0143 (2011).
  • Hildebrand and et al. (2011) D. Hildebrand, et al., ArXiv e-prints, 1110.5358 (2011).
  • Brown and Adams (2011) A. M. Brown, J. Adams, MNRAS 413 (2011) 2785–2790.
  • VERITAS Collaboration (2009) VERITAS Collaboration, Nature 462 (2009) 770–772.
  • Acero and et al. (2009) F. Acero, et al., Science 326 (2009) 1080.
  • Lacki et al. (2011) B. C. Lacki, T. A. Thompson, E. Quataert, A. Loeb, E. Waxman, ApJ 734 (2011) 107.
  • Mannheim et al. (2010) K. Mannheim, D. Elsässer, O. Tibolla, ArXiv e-prints, 1010.2185 (2010).
  • Komin and et al. (2011) N. Komin, et al., in: Proc. 32nd International Cosmic Ray Conference, Beijing, (2011), Arxiv e-prints, 1201.0639.
  • Abdo and et al. (2010) A. A. Abdo, et al., A&A 512 (2010) A7.
  • Williams and et al. (2009) D. A. Williams, et al., in: astro2010: The Astronomy and Astrophysics Decadal Survey, volume 2010 of Astronomy, p. 316.
  • Abdo and et al. (2009a) A. A. Abdo, et al., Science 323 (2009a) 1688.
  • Abdo and et al. (2009b) A. A. Abdo, et al., ApJ 707 (2009b) 580–592.
  • Acciari and et al. (2010) V. A. Acciari, et al., ApJ 720 (2010) 1174–1180.
  • Aleksić and et al. (2010) J. Aleksić, et al., A&A 517 (2010) A5.
  • Aharonian and et al. (2009) F. Aharonian, et al., A&A 495 (2009) 505–512.
  • Acciari and et al. (2011) V. A. Acciari, et al., ArXiv e-prints, 1109.0050 (2011).
  • Aune and for the VERITAS Collaboration (2011) T. Aune, for the VERITAS Collaboration, ArXiv e-prints, 1111.1326 (2011).
  • Porter et al. (2011) T. A. Porter, R. P. Johnson, P. W. Graham, ARA&A 49 (2011) 155–194.
  • Abramowski and et al. (2011) A. Abramowski, et al., Physical Review Letters 106 (2011) 161301.
  • Vivier and for the VERITAS collaboration (2011) M. Vivier, for the VERITAS collaboration, ArXiv e-prints, 1110.6615 (2011).
  • Aleksić and et al. (2011) J. Aleksić, et al., JCAP 6 (2011) 35.
  • H. E. S. S. Collaboration (2011) H. E. S. S. Collaboration, Astroparticle Physics 34 (2011) 608–616.
  • Aharonian and et al. (2009) F. Aharonian, et al., ApJ 691 (2009) 175–181.
  • Aharonian and et al. (2008) F. Aharonian, et al., Astroparticle Physics 29 (2008) 55–62.
  • Aharonian and et al. (2006) F. Aharonian, et al., ApJ 636 (2006) 777–797.
  • Gast and et al. (2011) H. Gast, et al., in: Proc. 32nd International Cosmic Ray Conference, Beijing, (2011).
  • Tsuchiya and et al. (2004) K. Tsuchiya, et al., ApJL 606 (2004) L115–L118.
  • Kosack and et al. (2004) K. Kosack, et al., ApJL 608 (2004) L97–L100.
  • Albert and et al. (2006) J. Albert, et al., ApJL 638 (2006) L101–L104.
  • Aharonian and et al. (2006a) F. Aharonian, et al., Physical Review Letters 97 (2006a) 221102.
  • Aharonian and et al. (2006b) F. Aharonian, et al., Nature 439 (2006b) 695–698.
  • Aharonian and et al. (2009) F. Aharonian, et al., A&A 503 (2009) 817–825.
  • Acero and et al. (2010) F. Acero, et al., MNRAS 402 (2010) 1877–1882.
  • Kennel and Coroniti (1984) C. F. Kennel, F. V. Coroniti, ApJ 283 (1984) 710–730.
  • de Jager and et al. (1996) O. C. de Jager, et al., ApJ 457 (1996) 253.
  • Abdo and et al. (2011) A. A. Abdo, et al., Science 331 (2011) 739.
  • Abdo and et al. (2010) A. A. Abdo, et al., ApJ 708 (2010) 1254–1267.
  • Kuiper and et al. (2001) L. Kuiper, et al., A&A 378 (2001) 918–935.
  • Aliu and et al. (2008) E. Aliu, et al., Science 322 (2008) 1221.
  • Tang et al. (2008) A. P. S. Tang, J. Takata, J. J. Jia, K. S. Cheng, ApJ 676 (2008) 562–572.
  • Aliu and et al. (2011) E. Aliu, et al., Science 334 (2011) 69.
  • Aleksić and et al. (2011) J. Aleksić, et al., ApJ 742 (2011) 43.
  • Saito and for the MAGIC collaboration (2011) T. Y. Saito, for the MAGIC collaboration, ArXiv e-prints, 1109.6100 (2011).
  • Atoyan and Aharonian (1996) A. M. Atoyan, F. A. Aharonian, MNRAS 278 (1996) 525–541.
  • Striani and et al. (2011) E. Striani, et al., ApJL 741 (2011) L5.
  • Wilson-Hodge and et al. (2011) C. A. Wilson-Hodge, et al., ApJL 727 (2011) L40.
  • Di Sciascio and for the ARGO-YBJ Collaboration (2011) G. Di Sciascio, for the ARGO-YBJ Collaboration, ArXiv e-prints, 1107.3404 (2011).
  • de Oña Wilhelmi (2011) E. de Oña Wilhelmi, in: M. Burgay, N. D’Amico, P. Esposito, A. Pellizzoni, & A. Possenti (Ed.), American Institute of Physics Conference Series, 1357 (2011) 213-220, pp. 213–220.
  • Kargaltsev and Pavlov (2010) O. Kargaltsev, G. G. Pavlov, X-ray Astronomy 2009; Present Status, Multi-Wavelength Approach and Future Perspectives 1248 (2010) 25–28.
  • Slane (????) P. Slane, in: D. F. Torres & N. Rea (Ed.), High-Energy Emission from Pulsars and their Systems, Astrophysics and Space Science Proceedings (2011) 373.
  • Gaensler and Slane (2006) B. M. Gaensler, P. O. Slane, ARA&A 44 (2006) 17–47.
  • Aharonian and et al. (2005a) F. Aharonian, et al., A&A 432 (2005a) L25–L29.
  • Aharonian and et al. (2005b) F. Aharonian, et al., Science 307 (2005b) 1938–1942.
  • Acciari and et al. (2010) V. A. Acciari, et al., ApJL 719 (2010) L69–L73.
  • Blondin et al. (2001) J. M. Blondin, R. A. Chevalier, D. M. Frierson, ApJ 563 (2001) 806–815.
  • Aharonian and et al. (2006a) F. Aharonian, et al., A&A 460 (2006a) 365–374.
  • Aharonian and et al. (2006b) F. Aharonian, et al., A&A 448 (2006b) L43–L47.
  • Hinton et al. (2011) J. A. Hinton, S. Funk, R. D. Parsons, S. Ohm, ApJL 743 (2011) L7.
  • Drury et al. (1994) L. O. Drury, F. A. Aharonian, H. J. Voelk, A&A 287 (1994) 959–971.
  • Malkov and O’C Drury (2001) M. A. Malkov, L. O’C Drury, Reports on Progress in Physics 64 (2001) 429–481.
  • Blandford and Eichler (1987) R. Blandford, D. Eichler, Phys. Rep. 154 (1987) 1–75.
  • Uchiyama and et al. (2007) Y. Uchiyama, et al., Nature 449 (2007) 576–578.
  • Aharonian and et al. (2006) F. Aharonian, et al., A&A 449 (2006) 223–242.
  • Aharonian and et al. (2007) F. Aharonian, et al., ApJ 661 (2007) 236–249.
  • Abramowski and et al. (2011) A. Abramowski, et al., A&A 531 (2011) A81.
  • Acero and et al. (2010) F. Acero, et al., A&A 516 (2010) A62.
  • Aharonian and et al. (2009) F. Aharonian, et al., ApJ 692 (2009) 1500–1505.
  • Acciari and et al. (2011) V. A. Acciari, et al., ApJL 730 (2011) L20.
  • Abdo and et al. (2011) A. A. Abdo, et al., ApJ 734 (2011) 28.
  • Morlino and Caprioli (2011) G. Morlino, D. Caprioli, ArXiv e-prints, 1105.6342 (2011).
  • Giordano and et al. (2012) F. Giordano, et al., ApJL 744 (2012) L2.
  • Atoyan and Dermer (2011) A. Atoyan, C. Dermer, ArXiv e-prints, 1111.4175 (2011).
  • Aharonian and et al. (2008) F. Aharonian, et al., A&A 481 (2008) 401–410.
  • Butt (2009) Y. Butt, Nature 460 (2009) 701–704.
  • Aharonian and et al. (2007) F. Aharonian, et al., A&A 467 (2007) 1075–1080.
  • H.E.S.S. Collaboration (2011) H.E.S.S. Collaboration, A&A 525 (2011) A46.
  • Ackermann and et al. (2011) M. Ackermann, et al., ApJ 726 (2011) 35.
  • H.E.S.S. Collaboration (2011) H.E.S.S. Collaboration, ArXiv e-prints, 1111.2043 (2011).
  • Ohm and et al. (2010) S. Ohm, et al., in: J. Martí, P. L. Luque-Escamilla, & J. A. Combi (Ed.), Proc. High Energy Phenomena in Massive Stars, Astronomical Society of the Pacific Conference Series, 422 (2010) 265.
  • Chaves et al. (????) R. C. G. Chaves, M. Renaud, M. Lemoine-Goumard, P. Goret, in: F. A. Aharonian, W. Hofmann, & F. Rieger (Ed.), American Institute of Physics Conference Series, 1085 (2008) 372-375.
  • Brun and for the H. E. S. S. Collaboration (2011) F. Brun, for the H. E. S. S. Collaboration, Proc. of the 25th TEXAS Symposium on Relativistic Astrophysics, ArXiv e-prints, 1104.5003 (2011).
  • Lemoine-Goumard and et al. (2011) M. Lemoine-Goumard, et al., ArXiv e-prints, 1109.4733 (2011).
  • Albert and et al. (2007) J. Albert, et al., ApJL 665 (2007) L51–L54.
  • Bosch-Ramon and Khangulyan (2009) V. Bosch-Ramon, D. Khangulyan, International Journal of Modern Physics D 18 (2009) 347–387.
  • Dubus (2006) G. Dubus, A&A 456 (2006) 801–817.
  • Sierpowska-Bartosik and Torres (2008) A. Sierpowska-Bartosik, D. F. Torres, Astroparticle Physics 30 (2008) 239–263.
  • Aharonian and et al. (2005) F. Aharonian, et al., A&A 442 (2005) 1–10.
  • Kawachi and et al. (2004) A. Kawachi, et al., ApJ 607 (2004) 949–958.
  • Chernyakova and et al. (2006) M. Chernyakova, et al., MNRAS 367 (2006) 1201–1208.
  • Aharonian and et al. (2009) F. Aharonian, et al., A&A 507 (2009) 389–396.
  • Khangulyan et al. (2007) D. Khangulyan, S. Hnatic, F. Aharonian, S. Bogovalov, MNRAS 380 (2007) 320–330.
  • Moldón et al. (2011) J. Moldón, S. Johnston, M. Ribó, J. M. Paredes, A. T. Deller, ApJL 732 (2011) L10.
  • Aharonian and et al. (2006) F. Aharonian, et al., A&A 460 (2006) 743–749.
  • Abdo and et al. (2009) A. A. Abdo, et al., ApJL 706 (2009) L56–L61.
  • Albert and et al. (2009) J. Albert, et al., ApJ 693 (2009) 303–310.
  • Acciari (2008) V. A. e. a. Acciari, ApJ 679 (2008) 1427–1432.
  • Acciari and et al. (2011) V. A. Acciari, et al., ApJ 738 (2011) 3.
  • MAGIC Collaboration (2011) MAGIC Collaboration, ArXiv e-prints, 1111.6572 (2011).
  • Hadasch et al. (2012) D. Hadasch, D. F. Torres, T. Tanaka, R. H. D. Corbet, A. B. Hill, R. Dubois, G. Dubus, T. Glanzman, S. Corbel, J. Li, Y. P. Chen, S. Zhang, G. A. Caliandro, M. Kerr, J. L. Richards, W. Max-Moerbeck, A. Readhead, G. Pooley, ArXiv e-prints, 1202.1866 (2012).
  • Aharonian and et al. (2007) F. A. Aharonian, et al., A&A 469 (2007) L1–L4.
  • Acciari and et al. (2009) V. A. Acciari, et al., ApJL 698 (2009) L94–L97.
  • Hinton and et al. (2009) J. A. Hinton, et al., ApJL 690 (2009) L101–L104.
  • Bongiorno and et al. (2011) S. D. Bongiorno, et al., ApJL 737 (2011) L11.
  • Maier et al. (2011) G. Maier, for the VERITAS Collaboration, J. Skilton, for the H.E.S.S. Collaboration, ArXiv e-prints, 1111.2155 (2011).
  • Corbet and et al. (2011) R. H. D. Corbet, et al., The Astronomer’s Telegram 3221 (2011) 1.
  • de Ona Wilhelmi and et al. (????) E. de Ona Wilhelmi, et al., in: Proc. 38th COSPAR Scientific Assembly, Bremen, Germany, 38, (2010) 2803.
  • Domainko and et al. (2011) W. Domainko, et al., in: Proc. 32nd International Cosmic Ray Conference, Beijing, (2011), Arxiv e-prints, 1112.4974.
  • Venter et al. (2009) C. Venter, O. C. De Jager, A.-C. Clapson, ApJL 696 (2009) L52–L55.
  • Komatsu and et al. (2009) E. Komatsu, et al., ApJS 180 (2009) 330–376.
  • Wood and et al. (2008) M. Wood, et al., ApJ 678 (2008) 594–605.
  • Abramowski and et al. (2011) A. Abramowski, et al., ApJ 735 (2011) 12.
  • Actis and et al. (2011) M. Actis, et al., Experimental Astronomy (2011) 121.
  • Vassiliev and Fegan (????) V. V. Vassiliev, S. J. Fegan, in: Rogelio Caballero, Juan Carlos D’Olivo, Gustavo Medina-Tanco, Lukas Nellen, Federico A. Sánchez, José F. Valdés-Galicia (Ed.), Proc. 30th International Cosmic Ray Conference, Mérida, Yucatán, Mexico, 3 (2008) 1445-1448.
Comments 0
Request Comment
You are adding the first comment!
How to quickly get a good reply:
  • Give credit where it’s due by listing out the positive aspects of a paper before getting into which changes should be made.
  • Be specific in your critique, and provide supporting evidence with appropriate references to substantiate general statements.
  • Your comment should inspire ideas to flow and help the author improves the paper.

The better we are at sharing our knowledge with each other, the faster we move forward.
The feedback must be of minimum 40 characters and the title a minimum of 5 characters
Add comment
Loading ...
This is a comment super asjknd jkasnjk adsnkj
The feedback must be of minumum 40 characters
The feedback must be of minumum 40 characters

You are asking your first question!
How to quickly get a good answer:
  • Keep your question short and to the point
  • Check for grammar or spelling errors.
  • Phrase it like a question
Test description