Discovering new Supernova remnants with H.E.S.S.

March 2018

Previous | Index | Next

Fig 1: Positions of the new TeV supernova remnant shell candidates (shown in the bottom panels) on the H.E.S.S. Galactic plane survey flux map (HGPS, shown at the center of the top panel). The HGPS flux map is overlaid on the CO(1-0) all-sky map from the Planck satellite.

In the Galactic Plane survey that has been conducted with H.E.S.S. in the decade between 2004 and 2014, a significant number of new astrophysical sources have been discovered through their TeV radiation. Many of the sources have counterparts in other wavebands, permitting an association and a physical interpretation of the TeV sources, e.g. in pulsar wind nebula scenarios. The astrophysical objects behind many of the new TeV sources, however, remain currently unknown because of lacking or unclear counterparts in other wavebands.

Some TeV sources without any counterparts at all (dubbed "dark accelerators") have been suggested to be clouds of relativistic protons, that have formed as leftovers of old supernova remnants. A supernova remnant develops after the explosion of a massive star, sending shock waves into the interstellar medium where matter is heated and particles are accelerated to relativistic speeds. An old supernova remnant would have accelerated the relativistic particles at earlier phases of its lifetime. Today, the remnant itself would not be detectable in any waveband any more, the only visible trace would be the relativistic particle cloud which would still remain near the original position of the supernova remnant (e.g. [1]).

It is evidently very difficult to prove the above scenario for an arbitrary unidentified TeV source. But it should be possible to discover hitherto unknown supernova remnants that shine in TeV gamma rays because of their relativistic particles and that still look morphologically very much like supernova remnants (presumably therefore being not "old" but perhaps "middle-aged" supernova remnants). Indeed, the H.E.S.S. collaboration has now conducted such a search for new supernova remnants in the H.E.S.S. Galactic Plane survey data. The final results of this search have recently been accepted for publication in A&A [2]. The search was solely based on the expected morphological appearance of a supernova remnant in a TeV image, basically simply a ring-like or shell-like object. The final source sample consists of three sources, namely HESS J1534-571, HESS J1912+101, and HESS J1614-518, see Fig. 1.

All three sources share some extreme commonalities. Their apparent angular sizes are between 0.4 deg and 0.5 deg, i.e. they are as large as the full moon on the sky. At their likely distances between one and a few kiloparsec (1 kiloparsec = 3300 light years), this translates into physical sizes in a likely range between 15 and over 50 parsec, i.e. they are amongst the largest known supernova remnants.

The question however is: are we really sure that those sources are supernova remnants, and not something else? We face of course a dilemma. If the sources are first discovered in the TeV band, that probably means that they are faint or invisible in the other wavebands in which supernova remnants are traditionally discovered; most of the supernova remnant discoveries happen using radio telescopes, and in some cases also using X-ray satellite data. But we need the information from the other wavebands to make sure that the TeV shell-like sources are really supernova remnants; without this information, they remain formally only supernova remnant candidates. Luckily, for HESS J1534-571, a faint supernova remnant candidate counterpart has in fact been discovered also in the radio band [3]. Therefore, for this source, the supernova remnant nature is confirmed.

The confirmation that HESS J1534-571 is really a supernova remnant makes us confident that our methods and assumptions for identifying new supernova remnants using the TeV band are correct. So, let us for the moment assume that all three sources are supernova remnants, and that they are indeed intrinsically very faint in the other wavebands (which is actually just an assumption, because e.g. HESS J1912+101 is very difficult to observe with current X-ray satellites due to observational issues, and in general radio and X-ray observations of the three sources are technically challenging because of their large angular sizes). Let us in addition speculate that they share commonalities regarding their physical processes which lead to similar appearances.

Their large sizes then tell us that the supernova explosions have happened in very tenuous areas, probably wiped empty from interstellar material by the massive progenitor stellar winds. The possibility (again, we are not sure but speculate) that the remnants are faint in X-ray emission tells us that the supernova remnant shocks have slowed down considerably, already since quite a long time (thousands of years). The fact that the shocks nevertheless still shine brightly in TeV gamma-rays likely tells us that the shocks have hit the "walls" of material blown away by the progenitor winds and have slowed down there. This scenario pretty much matches our expectations as outlined at the beginning, that we are seeing "middle-aged" supernova remnants before they release the bulk of relativistic protons into the surroundings and then "die" (i.e. become invisible).

Fig 2: (Top:) TeV surface brightness map of HESS J1614-518 derived with H.E.S.S. The green circle denotes the position and extension of the GeV counterpart detected with Fermi-LAT. The inset on the bottom left denotes the point spread function of the TeV map. (Bottom:) Spectral energy distribution of HESS J1614-518 in the GeV (Fermi-LAT) and TeV (H.E.S.S.) bands, strongly supporting the identifications of the two sources being due to the same astrophysical object.

Do we have other means to test whether this scenario could be correct? It would in fact be very useful to show that the relativistic particles shining in TeV gamma-rays are really protons (and not electrons which can also very efficiently emit TeV gamma rays). Indeed, the adjacent GeV waveband in principle offers such a test. Unfortunately, the objects behind both HESS J1534-571 and HESS J1912+101 are not bright enough to produce counterparts in the GeV band that would permit such a test there (a likely GeV counterpart to HESS J1534-571 has recently been found by Araya [4], but the information is sparse). But HESS J1614-518 is bright enough to have a clearly identified counterpart in the GeV band (see Fig. 2, the identification is based both on the matching morphology and on the matching spectrum). The combined TeV-GeV spectrum shows that the two sources (the GeV source and the TeV source) are just incarnations of the same particle population, visible in the two adjacent wavebands. The GeV counterpart itself does not help with the question whether the source is indeed a supernova remnant or not. But do the GeV data at least show the hoped-for signature for protons (the so-called pion bump)? Unfortunately not (yet). But perhaps with more data and improved analysis methods, the GeV band may tell us in the future that the TeV- (and GeV-) emitting particles are indeed protons, which would add an important clue towards the interpretation of the TeV sources.


[1] R. Yamazaki et al., MNRAS 371 (2006) 1975
[2] H. Abdallah et al. (H.E.S.S. collaboration), A&A in press, arXiv:1801.06020
[3] A.J. Green et al., PASA 31 (2014) 42
[4] M. Araya et al. ApJ 843 (2017) 12