X-ray and Observational Astronomy

News September 2002: More X-ray emission lines from XMM-Newton gamma ray bursts

Left: The XMM-Newton EPIC PN X-ray image of GRB001025a. The green box shows the region, 9 arcminutes across, in which the gamma ray burst could have originated. The brighter of the two sources faded during the X-ray observation, and is assumed to be the afterglow of the GRB. Right: The X-ray spectrum of the GRB afterglow is not a power law (a, b), but is well fit by line-dominated spectral models (d, e).
	Left: The XMM-Newton EPIC PN X-ray image of GRB010220. The green circle shows the BeppoSAX 4 arcminute radius circle locating the burst. The brightest X-ray source, in the centre, was assumed to be the GRB afterglow. Right:The X-ray spectrum of this GRB afterglow. The solid line shows a power law fit, the residuals to this fit indicate the presence of an emission line at 3.9 keV

Gamma ray bursts are the most energetic explosions in the universe, yet they usually last much less than 100 seconds, making them difficult to study. We have learnt much more about them from the relatively long-lived afterglows discovered a few years ago by the BeppoSAX satellite.

Earlier this year, in April, the first XMM-Newton X-ray spectrum of a gamma ray burst afterglow was published in Nature. This spectrum showed emission lines from the elements magnesium, silicon, sulfur, argon and calcium. The Nature article argued that the link between supernovae and gamma ray bursts was strengthened by the detection of 50 million degree gas, which naturally emits these these spectral lines. All previous reports of X-ray emission lines had been identified with iron, as would be expected if the X-rays were due to illumination of the stellar remnant by the gamma rays.

Thus the Nature result has been controversial, with some workers suggesting that the analysis was flawed, and others pointing out the new implications of the result. What has been missing so far is observations of a similar quality which might confirm or refute the thermal emission model used to explain the data published in Nature.

Darrach Watson, and others at the University of Leicester and the ESA XMM-Newton operations centre have now analysed two more gamma ray burst afterglows observed by XMM-Newton. While these were observed later after the burst than the original Nature burst, they were still bright enough to see whether similar emission lines could be seen.

GRB001025a (shown at the top of this page) was observed 45 hours after the burst on Oct 25, 2000. The precise position of this burst is not known, but has been approximately determined from an interplanetary network of satellites which recorded the burst arrival time. The region of origin is a box on the sky which included two X-ray sources, the brighter of these was seen to fade during the XMM-Newton observation, making it most likely that this was the GRB afterglow.

The X-ray spectrum of the afterglow of this burst is shown on the right. The top plot compares the data (crosses) with a simple power law model (line) for the EPIC PN and MOS cameras. Below that the residuals from this and other fits are shown. It is clear that the power law is not a good fit to the data, while the fits to models including emission lines from Mg, Si, S, Ar, and Ca (thermal plasma model in d), or 5 emission lines with a power law (e) are good descriptions of the data. Monte Carlo simulations of this dataset suggest that the chance of the spectrum actually being a pure power law is not more than 0.13%.

GRB010220 (shown in the second image) was located by the BeppoSAX satellite on Feb 20, 2001. It was observed by XMM-Newton 14.8 hours after the burst. There are 4 X-ray sources in the positional error circle, the brightest is in the centre and has been assumed to be the GRB afterglow. None of the sources was seen to fade.

The spectrum from the EPIC PN camera can be seen to the right of the X-ray image. It has been fit with a power law with absorption due to our Galaxy. Below the spectrum the difference between the model and the data can be seen. It is clear that there is a feature at around 3.9 keV, the statistical significance of this is >99%. This spectrum is also well fit by a optically thin thermal plasma emission model, like GRB001025a and the GRB published in Nature.

There are no optical spectra of these two GRB afterglows, in the case of GRB001025a the thermal plasma emission line fit implies a redshift for the GRB of z = 0.53. For the other burst, the identification of the line cannot be certain, and so a redshift cannot be measured.

These results show that thermal X-ray emission, such as was reported earlier this year in Nature, is unlikely to be unique to that gamma ray burst. We have to admit that the while the statistical quality of these results is good, we have not yet seen a completely overwhelming demonstration that the X-ray afterglows of gamma ray bursts include a thermal component. However, if the X-ray spectra of GRB afterglows are generally thermal rather than power laws, then the intrinsic curvature of a thermal spectrum may account for some of the odd results previously reported on the basis of poorer quality spectra. These include different X-ray and optical absorption and so-called 'dark GRBs' which have little optical emission, possibly the use of a power law X-ray emission model is causing an over-estimate of the optical flux in some cases.

This work was recently accepted for publication in Astronomy & Astrophysics.

The XMM-Newton EPIC camera used to obtain this result was built by a team led by Dr Martin Turner of the University of Leicester. A similar camera will be carried by the NASA satellite Swift, due to be launched in the autumn of 2003. Swift will detect 2 or 3 new gamma ray bursts each week, rapidly maneuvering to point its X-ray and optical telescopes at the burst and afterglow within 20 seconds. This satellite has the capability to revolutionise gamma ray burst research.

Other news items are available here.

Back to group page.

Last updated: 2002 September 13 by Julian Osborne