Other extragalactic objects
Although the TeV blazars were not initially predicted as TeV emitters, now it is clear that the discovery of large TeV fluxes from these distant objects was possible because of the relativistic bulk motion of Y-ray production regions towards the observer, leading to the Doppler boosting of fluxes by several orders of magnitude. The detection of isotropically radiating sources of similar intrinsic power can hardly be realized in the foreseeable future.
In order to produce TeV Y-ray flux at the level of 0.1 Crab (approximately several time 10-12 erg/cm2s), the Y-ray luminosity of an "isotropic" point source at a distance d should exceed L > 3 x 1042 (d/100 Mpc)2 erg/c. Thus, one may hope to detect isotropically emitted Y-rays only from the most powerful extragalactic objects, like radiogalaxies or clusters of galaxies. On the other hand, the requirement for the Y-ray luminosity of sources within 10 Mpc is quite modest, so one may expect, on pure phenomenolog-ical grounds, detectable Y-ray fluxes from some nearby prominent galaxies. The radiogalaxies Centaurus A and M 87, as well the starburst galaxies M 82 and NGC 253 are obvious TeV source candidates, basically because of their overall power, a significant part of which is released in nonthermal forms. The first evidence of TeV radiation from an extragalactic object, the radiogalaxy Centaurus A, was obtained already in the 1970s (Grindlay et al., 1977). Recently, the CANGAROO and HEGRA collaborations reported TeV signals from two other nearby galaxies – NGC 253 (Itoh et al., 2002) and M 87.
Although the flux of Y-rays above 730 GeV from M 87 has been obtained using the reliable technique of shower reconstruction with the HEGRA stereoscopic system of telescopes, the statistical significance of about 4a is marginal, and therefore the result needs an independent confirmation. Unfortunately, the tiny signal detected at « 0.03 Crab level is not achievable by other current instruments. But the next generation of IACT arrays in both hemispheres should be able to verify this discovery rather quickly, and, in the case of confirmation, to provide detailed studies of the energy spectrum over two decades, from 100 GeV to 10 TeV, localise the position and measure the possible extension of the TeV source with an accuracy of about 1-2 arcminutes. For the distance to M 87 of about 16 Mpc, this angular scale corresponds to 5-10 kpc. If the extension of the TeV source exceeds several arcminutes, this would be an indication that the TeV emission is produced within the giant elliptical galaxy M 87. Both the inverse Compton scattering of relativistic electrons on 2.7 K CMBR and interactions of cosmic ray protons with the ambient interstellar gas can be responsible for the observed TeV emission. In particular, approximately 1057(n/0.1 cm-3)-1 erg total energy in > 10 TeV protons is required to be accumulated in the elliptical galaxy (n is the average number density of the interstellar gas) in order to explain the observed y-ray flux by p-p interactions. Another possible site of TeV emission is the famous kpc scale jet with several bright knots detected at radio, optical and X-rays. The latter is believed to have a synchrotron origin and be produced by electrons with energies up to 100 TeV. If so, within reasonable model parameters, detectable fluxes of inverse Compton TeV Y-rays also can be expected.
The detection of a TeV signal from the starburst galaxy NGC 253 has been reported by the CANGAROO group at the 11a level (Itoh et al., 2002). Despite the very high statistical significance, this detection perhaps should be considered yet as tentative, because of possible systematic effects which are less controllable in the case of a single telescope, especially for the large angular size of the excess region ~ 0.25° claimed by the authors. For the distance to the source 2.5 Mpc, this corresponds to the liner size of about 10 kpc, just the size of the radio halo discovered earlier around NGC 253. This prompted the authors to propose that the TeV radiation originates from the same radio halo region due to the inverse Compton scattering of electrons (Itoh et al., 2003a), which proceeds with very high efficiency in this region. Because of severe radiative losses, it is unlikely that the TeV electrons are produced in the galactic disk and then transported to the halo region. Thus, one has to assume that the electron accelerator(s) are located immediately in the radio halo. If so, the particle acceleration by strong termination shocks of the galactic wind (see e.g. Jokipii and Morfil, 1985) could be a possible source of TeV electrons in the halo.
The measured TeV spectrum is very steep (see Fig. 2.22), with a slop described by a photon index 3.74 ± 0.27 (Itoh et al., 2003b). Such a steep spectrum can be explained, most naturally, assuming an exponential cutoff in the acceleration spectrum at energies around several TeV. On the other hand in order to avoid the conflict with the flux upper limits set by EGRET at lower energies, it is necessary to assume a rather flat electron spectrum with spectral index p close to 2. Note that these values characterise the cooled (steady-state) electron spectrum, and since the radiative (synchrotron and Compton) cooling time of > 100 GeV electrons, which are responsible for > 100 MeV Y-rays, does not exceed 106 years, one has to assume unusually hard acceleration spectrum with differential spectral index p — 1 ~ 1. Also, the magnetic field in the halo required by this model is somewhat lower (by a factor of 2 or 3) than the field B ~ 6 ^G derived from the polarization measurements (Beck, 1994).
For different assumptions about the steady-state spectrum of electrons, Itoh et al. (2003b) estimated the total energy of electrons between 1055 and 1056 erg (Itoh et al., 2003). It is interesting to compare this estimate with the energy in cosmic ray protons, assuming that the bulk of TeV Y-rays is produced in interactions of cosmic rays with the ambient gas. A proton differential spectrum with a power-law index slightly less than 2 and with a break or exponential cutoff around 10 TeV can readily describe the TeV spectrum observed by CANGAROO, as well as accommodate the EGRET upper limits at GeV energies. For such a spectrum, the energy required in TeV protons is estimated Wp ~ LY tno ~ 5 x 1057(n/0.01)-1 erg, where LY « 1040 erg/s is the detected TeV Y-ray luminosity, tno is the characteristic time of n0-production, and n is the number density of the ambient gas. Even for a very low gas density in the halo, the total energy in protons seems not unreasonably high. Moreover, in the galactic disk, where the average gas density can be as large as n — 10 cm-3, the required energy in cosmic rays can reduced to a rather modest level, < 1055 erg. The model of Y-ray production in the galactic disk does not agree, however, with the angular size of the TeV source reported by the CANGAROO collaboration. On the other hand, any strong assumption concerning the origin of the TeV emission seems a bit premature. Both the energy spectrum and the angular size of the source (and, perhaps, even the very fact of the TeV signal) should be verified by independent measurements. Such measurements with the CANGAROO-III and H.E.S.S. telescope arrays will be available within the next 1 or 2 years.
Fig. 2.22 Differential fluxes of Y-rays from the starburst galaxy NGC 253 based on the data obtained by the CANGAROO group in 2000 and 2001.By the dotted line the Crab flux is shown for a reference.
Next generation of IACT arrays
The success of ground-based gamma-ray astronomy in the 1990s not only led to remarkable astrophysical discoveries, but also elucidated the promising detection techniques to be developed in the context of next generation instruments. Among a variety of competing designs, the stereoscopic arrays consisting of 10m diameter or larger aperture telescopes is identified as the most convincing approach that can facilitate a qualitative improvement in performance at an affordable cost, and, at the same time, promises a fast scientific return.
Atmospheric Cherenkov radiation
Observations of cosmic Y-rays from the ground is possible by detecting the secondary products of interactions of the primary Y-ray photon with the atmosphere, either directly or through their electromagnetic radiation. The atmospheric Cherenkov telescopes detect the Cherenkov light emitted in the atmosphere by secondary electrons which are produced in the cascades initiated by primary cosmic rays – protons, nuclei, electrons, Y-rays. Generally, the ground-based technique based on the direct detection of secondary cascade products – electrons, photons, muons, hadrons – effectively works in the > 1 TeV primary energy region, even for particle detectors installed at very high mountain altitudes, closer to the shower maximum. At smaller primary energies the cascades die out in the upper atmosphere. But, fortunately, the atmospheric Cherenkov radiation potentially makes accessible primary Y-rays down to energies of about several GeV.
In the Cherenkov light, a Y-ray induced shower looks like an object of (approximately) elliptical shape that starts at an altitude of some 20-25 km (the point of the first interaction) and extends down to several km. The relativistic electrons move along the shower axis without significant lateral displacement, thus the light is emitted predominantly at the characteristic Cherenkov angle, which at the the maximum of the shower development of about 10 km above sea level (a.s.l.) is close to 1° (see e.g. Hillas, 1996). This results in a pool of Cherenkov light on the ground with a radius « 120 m at average mountain altitude of about 2 km a.s.l. (see Fig. 2.23).
Because of absorption and scattering in the atmosphere, the Cherenkov light arrives with a spectrum that peaks at wavelengths around 300-350 nm. This light is very faint, with a density ranging from 100 to several 100 photons/m2 (depending on the altitude) for a 1 TeV Y-ray photon, and very brief – it lasts only several nanoseconds. This implies that the Cherenkov telescopes must have large (1 m2) optical reflectors to image the Cherenkov light onto a multi-pixel camera. The latter should be sensitive to the visible (closer to blue) light and sufficiently fast with a typical "time gate" of about 10 nanoseconds.
Fig. 2.23 Illustration of detection of the Cherenkov light from a shower initiated by a primary very high energy 7-ray photon.
A characteristic image of a Y-ray induced shower obtained with such a camera is shown in Fig. 2.23. The total intensity of the image is a measure of the primary energy, the orientation of the image in the camera correlates with the arrival direction of the primary particle, and the shape of the image contains information about the the origin of the air shower (induced by a cosmic ray proton/nucleus or by a Y-ray photon). These three features comprise the basis of the Imaging Atmospheric Cherenkov Telescope (IACT) technique.
Stereoscopic detection of Cherenkov images
The huge detection area of showers determined effectively by the radius of the Cherenkov light pool, the effective separation of electromagnetic and hadronic showers and the good accuracy of reconstruction of shower parameters are three remarkable attributes of the IACT technique. The critical advantage of air-shower imaging is the discrimination it provides against a potentially overwhelming rate of the cosmic-ray induced shower events (Weekes and Turver, 1977, Stepanian et al., 1983, Plyasheshnikov and Bignami, 1985, Hillas, 1985). Two independent image criteria are roughly equal in CR rejection power by a single telescope – the intersection of the shower track with the source location and the apparent shower width. The first criterion is valid for point sources, but does not depend on the origin (hadronic or electromagnetic) of showers. The second one is based on the inherently larger transfer momentum of secondaries produced in hadronic interactions. With single imaging telescopes, the Whipple, CAT and HEGRA groups have achieved approximately 300:1 rejection of cosmic ray showers with a factor of 2 loss in gamma-ray events.
Viewing the shower from several vantage points provides additional degrees of freedom, and thus allows further significant improvement of the Y-ray registration technique. The concept of stereo imaging is based on the simultaneous detection of a single air shower in different projections by at least two telescopes separated at a distance comparable with the "effective radius" of the Cherenkov light pool. The stereoscopic approach allows (i) unambiguous and precise reconstruction of shower parameters on an event-by-event basis, (ii) superior rejection of hadronic showers, and (iii) effective suppression of the background light from different sources – the night sky background (NSB), local muons, etc..
Fig. 2.24 Performance of the HEGRA stereoscopic system of telescopes for reconstruction of the arrival direction of primary Y-rays. Distribution of reconstructed directions of showers detected in the 2-telescope coincidence mode for 12 h ON-source observations of the Crab Nebula.
Compared with single ("stand alone") telescopes, which can adequately measure the shower inclination only in the direction perpendicular to the plane containing the telescope axis, the stereoscopic approach allows full reconstruction of the arrival direction of an individual primary photon with accuracy of about 0.1 ° or less. Apart from the good directional information, the stereoscopic systems make use of the fact that the Cherenkov images of a shower detected in different projections are only partially correlated. Therefore, the stereoscopic measurements increase the efficiency of rejection of hadronic (background) showers at both the hardware (trigger) and software levels, and consequently improve significantly the flux sensitivity compared to the sensitivity of single telescopes. Finally, the precise detection of shower parameters, in particular the shower maximum and the position of the shower core, allows energy determination of primary Y-rays with resolution as good as 10 per cent.
The only disadvantage of the stereoscopic approach is a non-negligible loss in the detection rate because of the overlap of the shower detection areas of individual telescopes. However, this loss of statistics is largely compensated for by a significant reduction of the energy threshold when the telescopes operate in the coincidence mode.
The first attempt of stereoscopic observations of air showers using the so-called "double-beam" technique, which also contained certain elements of the modern imaging Cherenkov technique, led to the tentative detection of a Y-ray signal from the radiogalaxy Centaurus A (Grindlay et al., 1975). Both the energy threshold and the flux sensitivity 10-11 ph/cm2s above 300 GeV) achieved by the two 7-m aperture optical reflectors were quite impressive even in the standards of the current ground-based instruments. However, the full potential of the stereoscopic approach has been convincingly demonstrated two decades later by the HEGRA system of 5 relatively small, 3 m diameter telescopes equipped with medium resolution (0.25° pixel size) cameras. The observations of the standard TeV candle, the Crab Nebula for the performance of this instrument. Fig. 2.24 shows the potential of the HEGRA IACT array for the reconstruction of the arrival direction of Y-rays. The angular resolution of about 0.1 ° coupled with the effective rejection of hadronic showers allowed observations of point Y-ray sources with a sensitivity characterised by a robust 10a detection of a TeV signal from a Crab-like source for just 1 hour observation time. This is a factor of 1.5-2 better than the sensitivity of the best single telescopes, CAT and Whipple, despite the fact that they operate at lower energy thresholds, and therefore have significantly higher detection rates. The ability of this instrument to detect Y-ray sources at an energy flux level around 10-12 erg/cm2s after approximately 100 h observation time (see Fig.1.2) was demonstrated by the discovery of Y-rays from 3 objects (Cas A, M 87, and TeV J2132+4131). This instrument was able to localise point-like TeV sources with an accuracy of better than 1 arcmin, and to perform detailed spectral studies with « 12 per cent energy resolution. Despite the small area of reflectors and the use of medium resolution cameras, the effective energy threshold of the HEGRA system of about 500 GeV was only a factor of two higher than the energy threshold of the Whip-ple telescope with an order of magnitude larger mirror area. Moreover, it has been demonstrated that with an optimised "topological" trigger, the energy threshold of the HEGRA telescope system could be reduced to 350 GeV (Lucarelli et al., 2003).
One of the principal issues in the planning of next generation detectors for ground-based gamma-ray astronomy is the choice of the energy domain. If one limits the energy region to a relatively modest threshold around 100 GeV, the performance of the telescope arrays and their implementation can be predicted with confidence.In practice, an energy threshold of 100 GeV can be achieved with a stereoscopic system consisting of 10 m diameter optical reflectors equipped with conventional PMT-based high resolution cameras. The technology for building large telescopes with high quality mirrors and multi-channel cameras was also rather well understood. These circumstances soon resulted in a logical outcome – 3 quite similar proposals for 10m diameter class imaging telescope arrays, CANGAROO-III (Collaboration of Australia and Nippon for a Gamma Ray Observatory in the Outback) in Australia, H.E.S.S. (High Energy Stereoscopic System) in Namibia and VERITAS (Very Energetic Imaging Telescope Array System) in southern Arizona, USA. With their superior angular resolution and background rejection, that should provide a TeV flux sensitivity in the 10 mCrab range, and an energy threshold of about 100 GeV, these instruments, together with a single 17m diameter imaging telescope MAGIC, will dominate the field for the next 5 to 10 years.
Both the arrangement (four or more telescopes separated from each other at distances of about 100m) and the design of individual telescopes of these three projects are quite similar, although the H.E.S.S. collaboration has taken perhaps the most ambitious steps.
Fig. 2.25 The H.E.S.S. Phase-I four telescope system.
The Phase-I system (see Fig.2.25) to be completed in 2004, consists of 4 telescopes. The reflector of a H.E.S.S. telescope is composed of 380 round 60 cm glass, aluminised and quartz coated mirrors. Arranged in a Davis-Cotton design, they form a 15 m focal length, 107 m2 area dish with a reflectivity of about 85%. Due to the superior quality of the mirrors and the alignment system, an excellent mirror point spread function has been achieved out to 2 degrees from the optical axis. Over almost the entire 5° FoV of 960-pixel camera, the spot is essentially contained in a single 0.16° pixel. The performance of the first telescope (Fig.2.26), which started taking data in mid 2002, is consistent with expectations.
A standard stereoscopic system of telescopes should consist of several telescopes, because the quality of reconstructed shower parameters continues to be improved up to three or four triggered telescopes. After reaching the maximum possible suppression of the cosmic ray background by simultaneous detection of air showers in different projections – limited basically by intrinsic fluctuations in cascade development – the further improvement of the flux sensitivities for a given energy threshold can be achieved by an increase of the shower collection area, i.e. construction of an array consisting of several stereoscopic systems – cells. A possible design of a homogeneous multi-cell array of 100 GeV-threshold class telescopes is shown in Fig. 2.27a. The detection area of a telescope array consisting of a large number of 100mx100m 4-telescope cells, n0 ^ 10, is almost energy-independent, A « n0A0 where A0 ~ 104 m2 is the area of a single cell. Thus, approximately n0 = 100 cells are required in order to approach to a highly desired detection area of about 1 km2. This is, however, an expensive and hardly affordable approach. Also, it has another major disadvantage – significant loss in the collection area per cell above the energy threshold 100 GeV, and, therefore, an inadequate (for the total number of telescopes) sensitivity at TeV energies.
Fig. 2.26 The first H.E.S.S. telescope equipped with 960-pixel camera.
From this point of view, an array consisting of several cells located at large distances from each other (significantly larger than the Cherenkov pool radius), seems a more effective, and economically better justified, approach. A possible design for such an array is shown in Fig.2.27b. At energies close to the energy threshold, the cells operate independently, thus we should expect the same sensitivity as in the case of a "homogeneous" array consisting of the same number of cells. However, at energies significantly above the energy threshold there should a significant gain in the collection area. At very high energies, a "stereoscopic trigger" applied to some of the telescopes from different cells should provide also a significant gain of shower detection in the so-called large-zenith-angle mode.
It is important to note that the so-called 100 GeV energy threshold arrays provide, in fact, the best energy flux sensitivity at TeV energies. For optimisation of the Y-ray detection at energies around 100 GeV, we must reduce the energy threshold of individual cells to 10 GeV or so, by using larger, > 20 m aperture class reflectors and/or using higher quantum efficiency optical detectors. On the other hand, reduction of the detection threshold to such low energies is a big issue itself. It will help not only to improve significantly the flux sensitivities at 100 GeV, but also will open an exciting scientific research area in the intermediate for the ground- and space-based gamma-ray astronomies interval between 10 and 100 GeV.
Sub-10 GeV ground based detectors?
The forthcoming stereoscopic systems of large imaging telescopes, with their superior energy-flux sensitivity between 10-13 — 10-12 erg/cm2s, perfectly suit the energy range from approximately 10 TeV down to 100 GeV (or perhaps even 30 GeV) – a spectral domain in its own right from the point of view of both the main scientific motivations and the specific astrophysical source populations that emit most effectively in this energy band.
It is expected that the next generation major satellite Y-ray mission GLAST with a large pair-conversion telescope (Fig. 2.28) will have similar energy-flux sensitivity in the 30 MeV to 10 GeV energy range, and extent the exploration of the gamma-ray sky with still reasonable sensitivity to 100 GeV, or even beyond (see Fig. 1.2). Thus the gap between space-based and ground-based instruments will finally disappear. This is an important condition for cross-calibration of two different detection techniques.
At the same time, the astrophysical significance of the overlap of detection domains of the satellite-borne and ground-based instruments should not be overestimated.
Fig. 2.27 Two possible designs of future IACT arrays.
Although at GeV energies GLAST will improve the EGRET sensitivity by almost two orders of magnitude, the capability of GLAST and, in fact, of any post-GLAST space project at energies well beyond 10 GeV will be quite limited because of the limited detection area, even if the Moon would be used in (far) future as a possible platform for installation of very large, e.g. 100 m2 area pair-conversion detectors.
The impressive sensitivity of GLAST shown in Fig. 1.2 can be achieved from an approximately one year all-sky survey. For the persistent Y-ray sources this is an adequate estimate, taking into account that a huge number of sources will be simultaneously covered by the large, almost ~ 2n steradian homogeneous FoV of GLAST.
Fig. 2.28 The GLAST 7-ray telescope.The main element of the telescope is the tracker consisting of four-by-four array of tower modules. Each module consists of layers of silicon-strip detectors and thin large-Z material sheets converting 7-rays into electron positron pairs. The silicon-strip detectors track with high precision the secondary electrons. The segmented CsI(Tl) calorimeter measures the energy of the absorbed electron-positron pair, and thus gives information about the the energy of the primary 7-ray photon. The active anti-coincidence shield consisting of segmented plastic scintillator tiles, provides effective rejection of the charged particle background from primary cosmic rays and from the Earth albedo secondary products.
However, the small, approximately 1 m2 , detection area limits the potential of this instrument for detailed studies of the temporal and spectral characteristics of highly variable sources like blazars or solitary events like GRBs at multi-GeV energies. In this regard, GLAST can hardly match the performance of current X-ray detectors that have similar detection areas but operate in a regime of photon fluxes that exceeds the fluxes of MeV/GeV Y-rays by many orders of magnitude.
The need in a powerful instrument to study transient phenomena with adequate high energy Y-ray photon statistics, has motivated the idea/concept of extension of the domain of the imaging atmospheric Cherenkov technique, with its huge shower collection area > 104 m2, down to energies of about 5 GeV. This requires high elevations of the order of 5 km. That is why the concept is called 5@5.
Very high altitudes allow significant gain in the density of the Cherenkov light. Two other key requirements to achieve sub-10 GeV energy threshold are the stereoscopic approach and very large optical reflectors. Monte Carlo studies show that the operation of a stereoscopic system consisting of > 3 telescopes, each of which is a 25 to 30 m diameter optical reflector equipped with a conventional PMT-based multichannel camera with pixel size less than 0.1 ° , at an altitude of about 5 km above sea level, should indeed allow reduction of the effective detection threshold down to several GeV.
The successful realization of a several GeV energy threshold telescope array would largely depend on the availability of exceptional sites with a dry and transparent atmosphere at an altitude as high as 5 km. Nature does provide us with such an extraordinary site – the Liano de Chajnantor in the Atacama desert in Northern Chile. This site with its very arid atmosphere has been chosen for the installation of one of the most powerful future astronomical instruments – the Atacama Large Millimeter Array (ALMA), a US-European project which will consists of sixty four 12m aperture radio antennas. The very large flat area of this unique site could certainly accommodate a Cherenkov telescope array as well. Another attractive feature of this site seems is an adequate infrastructure which will be built up for ALMA. A potential site for installation of a similar array in the Northern Hemisphere would be a site close to the Yangbajing Laboratory (Tibet, China) at 4.3 km a.s.l. where the low-threshold air-shower array ARGO-YBJ is under construction, and other detectors have operated for a number of years.
An array of such large telescopes located at a relatively modest, but more comfortable altitude around 2 km a.s.l. would have somewhat larger, by a factor of 2 or 3, energy threshold. The lower Cherenkov light density at these elevation can be compensated for by more sensitive cameras based on novel (unfortunately, yet not available) fast (nanosecond) detectors of optical radiation with quantum efficiency exceeding 50%. Because of the unavoidable increase of the night sky background. such cameras cannot,however, provide proportional reduction of the energy threshold. Finally we note that combination of several such (sub)arrays with spacing of approximately 200 to 400 m, as shown in Fig. 2.27b, can provide an excellent flux sensitivity around 100 GeV.
Reduction of the energy threshold down to several GeV is a critical issue for a number of astrophysical and cosmological problems, e.g. for study of Y-radiation from pulsars and cosmologically distant objects like quasars and gamma-ray bursts. Therefore, preference should be given to very high elevations. Moreover, with sub-10 GeV threshold arrays one may significantly gain in sensitivity because of the effect related to the so-called rigidity cutoff. At such low energies the cosmic ray background is dominated by electrons. However, for the geomagnetic latitudes of both the ALMA and Yangbajing sites, the geomagnetic field effectively prevents the electrons and protons with energies less than 10 to 15 GeV from entering the Earth’s atmosphere (see e.g. Lipari, 2002). This may have an impact on the background caused by electrons, and consequently would improve the flux sensitivity at energies below the rigidity cutoff.
The range of sensitivities that can be achieved by a 5@5 type instrument after approximately 100 hours observation of a point source is shown in Fig. 1.2. Comparing the sensitivity curves in Fig. 1.2 one may conclude that 5@5 and GLAST are competitors. However, they are, in fact, highly complementary. Note that the sensitivity curves for GLAST and 5@5 shown in Fig. 1.2 correspond to very different times (100 h for 5@5 versus 1 year for GLAST). While GLAST with its almost 2n FoV can provide very effective simultaneous monitoring of a large number (hundreds or even thousands) of sources, and also enables studies of the galactic and extragalactic components of the diffuse Y-ray background, 5@5 has an obvious advantage for the search and study of highly variable or transient Y-ray sources. On the other hand, 5@5 is a detector with a small field of view, therefore it requires special strategies for the search and study of multi-GeV emitters. Because of overlap of the energy intervals covered by these two instruments, GLAST may serve as a perfect "guide" for 5@5. Generally, all sources that will be detected by GLAST can be potential targets for observations with 5@5.
The concept of 5@5 is not only motivated by the possibility of coverage of the as yet unexplored region of multi-GeV Y-rays. In fact, 5@5 combines two advantages of the current ground-based and satellite-borne Y-ray domains – large photon fluxes at GeV energies (typically, 10-8 to 10-6 ph/cm2s above 1 GeV versus 10-12 to 10-10 ph/cm2s above 1 TeV), and enormous detection areas in the TeV domain (up to 105 m2 at TeV energies to « 1 m2 at GeV energies). This makes 5@5 a unique "Gamma-ray timing explorer" (with a sensitivity to detect EGRET sources with hard spectra extending to 10 GeV, in exposure times of a few seconds to several minutes) for the study of transient non-thermal Y-ray phenomena like rapid variability of Blazars, synchrotron flares in Microquasars, the high energy (GeV) counterparts of Gamma Ray Bursts, etc.
The capability of 5@5 will not be limited to highly variable source studies. This instrument, in fact, has significantly broader objectives related to detailed Y-ray spectrometry of emission in the energy interval from 10 to 100 GeV from Y-ray sources like SNRs, Plerions, Pulsars, the Galactic Disk as a whole, large (kpc to Mpc) scale jets in radiogalaxies, rich galaxy clusters, etc.. Therefore, 5@5 will be complementary in its capabilities to the forthcoming 10m class telescope arrays (spectrometry typically above 100 GeV) and GLAST (spectrometry below 10 GeV). The good energy resolution in the energy interval between 10 and 100 GeV, supported by an adequate Y-ray photon statistics, is crucial also for cosmological studies through (i) probing the cosmic optical and near infrared background radiation, (ii) detecting Y-rays from large scale cosmological structures, and (iii) searching for characteristic line emission from the non-baryonic Dark Matter Halos.
The successful realization of the 5@5 concept in the form of an operating high-altitude Cherenkov telescope array during the lifetime of GLAST was identified as a major observational goal and classified as one of the highest priority objectives in the field of gamma-ray astronomy (Buckley et al., 2002).
Large field-of-view detectors
The imaging atmospheric Cherenkov telescopes are designed for observations of Y-rays from objects with well determined positions. However, the high sensitivity of stereoscopic arrays coupled with relatively large (4 degree or more) field-of-view homogeneous imaging cameras, may allow quite effective sky surveys as well, as has been demonstrated by the HEGRA CT-system. The next generation telescope arrays with an energy threshold around 100 GeV and with significantly improved sensitivity at TeV energies, will provide deeper surveys. In particular, it is expected that all point TeV sources with fluxes at the 0.1 Crab level can be be revealed within one steradian of the sky during a one-year survey. A 5@5 type instrument will have a similar potential at GeV energies.
Nevertheless, the development of a ground-based technique allowing simultaneous coverage of a significant (1 steradian or so) fraction of the sky is recognised as a high priority issue. This would be important, for example, for monitoring of Y-ray activity of a large number of highly variable blazars and, for independent detection of solitary VHE Y-ray events, e.g. TeV counterparts of GRBs. The general requirements to the performance of the future wide FoV Y-ray ground-based detectors is dictated by the parameters of the upcoming imaging Cherenkov telescope arrays. Namely, the energy threshold of these instruments should be close to 100 GeV and their sensitivity for a one year survey should be at least 0.1 Crab in order to be compatible to the sensitivity of the Cherenkov telescope arrays achieved for several hour (1 night) observation time. Two possible approaches have been proposed in this regard – (i) very large FoV imaging air Cherenkov telescope technique based on refractive optics (Kifune, 2001) and (ii) dense air shower particle arrays or large water Cherenkov detectors installed at very high, 4 km or higher altitudes (for a review see Hoffman et al., 1999).
The first technique requires several technological innovations -very large UV transparent Fresnel lenses, mega-pixel/nanosecond/high-quantum-efficiency detectors, a huge data handling capability, etc. This approach may receive a significant support from the cosmic ray community which recently proposed to build a similar, but space-based downward looking atmospheric fluorescence detectors for registration of extremely high energy cosmic rays (the OWL and EUSO projects).
The second approach does not face serious technological challenges. The feasibility of both the high altitude air shower array and water Cherenkov techniques have been convincingly demonstrated by the Tibet and Milagro collaborations (for a review, see e.g. Buckley et al., 2002). Moreover, ARGO, a new air shower detector under construction by the Chinese-Italian collaboration in Tibet, is expected to have performance rather close to the above requirements (Bacci et al., 2002). An ambitious idea of constructing a Milagoro type instrument with 10 times more physical area and located at 4 km a.s.l. (G. Sinnis, private communication), promises further significant improvement of sensitivity of the 100 GeV threshold, large field-of-view ground-based technique.
IACT arrays for probing PeV Y-rays
The current trend to reduce the energy threshold of ground-based Y-ray detection technique concerns both the atmospheric Cherenkov telescopes and the air-shower detectors. As a result, presently there is only little instrumental activity in the energy domain above 10 TeV. On the other hand, this energy region is of high astrophysical interest. For example, the detection of 100 TeV Y-rays from shell type SNRs would be a key proof that the shocks in SNRs accelerate protons up to energies 103 TeV. Another interesting issue is related to 10-100 TeV Y-rays from nearby extragalactic sources, for example from radiogalaxies M 87 or Cen A – potential sites of acceleration of the observed highest energy cosmic rays.
The fluxes of Y-rays typically decrease very rapidly with energy. This limits the capability of the traditional air-shower arrays because of both limited proton/gamma separation power and limited detection area, and consequently low Y-ray photon statistics. Any meaningful study of cosmic Y-rays beyond 10 TeV requires a detection area of about 1 km2.
At large zenith angles of about 60°, the collection area of the atmospheric Cherenkov detectors increases rapidly. Thus the use of IACT systems at such large angles can improve the Y-ray statistics in the multi-TeV region. For many astrophysical objects, on the other hand, the observation time within a single night is very short in this mode, because the Y-ray source sets rapidly below the horizon. Besides, even small variations of the atmospheric transparency add non-negligible uncertainties in the derivation of the shower parameters obtained at large zenith angles. Also, an exploitation of the large-zenith angle technique requires very small pixel size, and therefore quite expensive multi-channel cameras.
An effective and straightforward approach seems the use of an array of imaging telescopes optimised for detection of Y-rays in the 10 to 100 TeV region. Such an array can consists of telescopes of rather modest, approximately 10 m2 mirror size, and separated from each other by a distance essentially larger than the conventional 100 m. Depending on the tasks, and the configuration of the imagers (multi-pixel cameras), the optimum distance varies between 300 to 500 m. The requirement to the pixel size of the imagers is also quite modest, from 0.25° to 0.5°, however they have to cover very large field-of-view, 6 to 8 degree, in order to collect showers from distances up to 500 m or more. Monte Carlo simulations show (Plyashesh-nikov et al., 2000) that an array consisting of 9 or more such telescopes can provide an extraordinary collection area exceeding 1 km2 , a reasonable efficiency for suppression of hadronic showers, good angular resolution of several arcmin and reasonable energy resolution. Such an array can serve also as an effective tool for the study of the energy spectrum and mass composition of cosmic rays up to the so-called "knee" around 103 TeV.