An Old Friend: Supernova Remnant RX J1713.7–3946

September 2016

Previous | Index | Next

Fig. 1: Shown on the top is an optical image of the constellation Scorpius with the stars that form the constellation highlighte. The approximate position of the supernova remnant RX J1713.7-3946 is marked with a red star. The white rectangle is the area pictured in the ancient image drawn on the bottom ([1] taken from [2]). This image was produced by a Chinese astronomer, who noted: A GUEST STAR APPEARED WITHIN THE ASTERISM WEI ("tail of a dragon") DURING THE SECOND LUNAR MONTH OF THE EIGHTEENTH YEAR OF THE TAI-YUAN REIGN PERIOD (February 27 to March 28, AD393), AND DISAPPEARED DURING THE NINTH LUNAR MONTH (October 22 to November 19, AD393). The Chinese character in the circle means "wei" and corresponds more or less to the position of RX J1713.7-3946.

The constellation Scorpius hosts one of the most important sources of very-high-energy gamma rays: the supernova remnant RX J1713.7-3946. Possibly associated with a star explosion seen by Chinese astronomers in the year AD393 (see Fig. 1), this object is one of the brightest sources of gamma rays in the H.E.S.S. energy range in the sky and its detailed studies (see SOM January 2005 and references [3]-[5]) earmark the breakthrough of ground-based gamma-ray astronomy with Cherenkov telescopes some twelve years ago. This was the first resolved gamma-ray source in the sky, it allowed us to study a Galactic particle accelerator in detail, and by comparisons with measurements at other wavelengths (X-ray and radio data) it allowed us to probe the supernova remnant (SNR) paradigm as an explanation for the origin of Galactic cosmic rays. We are now about to release an updated H.E.S.S. measurement of this supernova remnant, 9 years after our previous publication [5], using a larger dataset and much improved analysis tools more than doubling the previous sensitivity. As seen in Fig. 2, the morphology measurement and angular resolution have improved dramatically over the years, gamma-ray astronomy is turning into precision science!

Fig. 2: Animation of the H.E.S.S. images of RX J1713.7-3946 produced at three different phases. In 2004, we had barely commissioned the telescope array when one of our first datasets obtained with only two out of the four telescopes revealed this impressive source, which is twice the size of the full moon in the sky! The 2006 image is a deep exposure measurement of the remnant obtained with the full four-telescope array. By then we had developed background modelling and subtraction techniques to produce a real gamma-ray excess image with a good angular resolution of 3.6 arc minutes. With more data, and a much better understanding of the system performance and hence improved sensitivity, the 2016 image is probably in many respects demonstrating the ultimate performance of the current-generation instruments like H.E.S.S., MAGIC, or VERITAS. With an angular resolution below 3 arc minutes and superb event statistics from 160 hours of deadtime-corrected observation time, morphological details at parsec scale can now be investigated and compared in detail to X-ray images. Further improved measurements will likely only become available once CTA [6], our next-generation facility, comes online in the early 2020's.

Speaking about the SNR paradigm - how can we prove that this beast accelerates protons and not only electrons? If we knew precisely what the ambient magnetic field strength as well as the ambient matter and photon field densities were, we could simply calculate the spectral energy distributions and compare these to data. Unfortunately, since none of the above parameters is really precisely known, we can only make assumptions and derive parameters from model curves that match the data. This is illustrated in the plot of the H.E.S.S. and the Fermi-LAT data, compared to electron and proton models, shown in Fig. 3.

Fig. 3: An electron and a proton model fit to the gamma-ray data of H.E.S.S. and the Fermi-LAT are shown in this figure on the top. In the electron scenario, the X-ray data at 1-10 keV energies are also relevant, and while they are not shown in this plot, they have been used to fit the model curves. On the bottom, the corresponding particle energy spectra are shown.

So what's the point? As you can see from the figure, both models can in principle fit the data, and these fits tell us the corresponding physical parameters of the source. This is where the work of an experimental collaboration like H.E.S.S. stops and the work of theorists will start. Questions like: is 10 micro-Gauss a reasonable magnetic field strength, or a density of 1 particle per cm3, can anyone understand why there is a break in the particle spectrum at 2-3 TeV, are the energetics reasonable, are we looking at gamma rays from only electrons or only protons, or a mix of both? Time and lots of theory papers will hopefully tell.

And here is a final interesting bit of the H.E.S.S. result that is really new: Fig. 4 illustrates that when comparing the radial size of the supernova shell of RX J1713.7-3946 it clearly extends further when seen in gamma rays than in X-rays!

Such effects from accelerated particles leaving the main shock region have long been predicted in Diffusive Shock Acceleration theory [7], but have so far never been seen. Well, that has changed now! We believe that the gamma rays extending further out than the X-rays are exactly this: the X-rays mark the end of the shock region, the gamma rays are either from completely detached (escaped) particles or else from particles in the forward shock (shock precursor) region. Even if it is too early to tell with certainty which one of the two scenarios we are seeing, this is an important achievement: we are ticking off an item on the long list of cosmic-particle-acceleration theorists by showing them for the first time gamma-ray images of particles in the process of leaving the main accelerating shock!

Fig. 4: The H.E.S.S. gamma-ray image with XMM X-ray contours overlaid is shown on the left-hand side. We have divided the supernova remnant in 5 wedges to compare radial profiles of the gamma-ray and X-ray data. For wedge number 3, this comparison is shown on the right, clearly demonstrating that the gamma-ray data extend further out than the X-ray data.

[1] Shen, Y. 500, Sung Shu (History of the Sung Dynasty), 25.
[2] Wang, Z. R., Qu, Q.-Y., Chen, Y., A&A, 318, L59 (1997).
[3] H.E.S.S. Collaboration, Aharonian et al., Nature, 432, 75 (2004).
[4] H.E.S.S. Collaboration, Aharonian et al., A&A, 449, 223 (2006).
[5] H.E.S.S. Collaboration, Aharonian et al., A&A, 464, 235 (2007).
[6] CTA web page at
[7] See e.g. the review by Hillas, A. M., Journal of Physics G: Nuclear and Particle Physics, 31, R95 (2005).