The expanding X-ray halo of GRB 031203

The story in a nutshell

The European Space Agency's X-ray observatory, XMM-Newton, has imaged a spectacular set of rings which appear to expand, with a speed a thousand times faster than that of light, around the point in the sky where a powerful gamma-ray explosion took place in early December. This is the first time that such a fascinating event, called an `echo', is seen in X-ray wavelengths. This echo forms when the powerful radiation of a gamma-ray burst, coming from far away, crosses a slab of dust in our Galaxy and scattered by it, like the beam of a lighthouse in clouds. Using the expanding rings to precisely pin-point the location of this dust, astronomers can identify places where new stars and planets are likely to form.

Images of GRB 031203

X-ray image of GRB 031203 and its
dust halo X-ray simulation of GRB
031203 and its dust halo

The image to the left shows how the X-ray afterglow looked at different times through the XMM-Newton observation. The ring ("halo") structure around the afterglow (in the centre) is something that has never been seen before from a gamma-ray burst. The image to the right is a computer simulation of the X-ray halos produced as X-rays from the gamma-ray burst scatter off two screens of dust in our Galaxy at distances of 1390 and 880 parsecs (4530 and 2870 light years) from us. The outermost ring is produced by the closer of the two dust screens. For higher resolution images click here (observation, simulation), or for a high resolution .tiff file of the observation click here.

Background info

For a basic introduction to Gamma-Ray Bursts (GRBs) click here, or for a more detailed but very readable account, try this article by Peter Meszaros here. A brief history of gamma-ray bursts can be found here.

Press releases

The official press release by ESA can be found here (html or text only) and the release by PPARC can be found here (html and text only). The local (University of Leicester) announcement is here. The story has been reported by,, Space Flight Now, Atlas Aerospace, PhysicsWeb,,,, The images also made Astronomy Picture of the Day on 30th Jan 2004 and HEASARC picture of the week on 2nd Feb 2004.

When did GRB 031203 happen?

On 2003 December 03 (Wednesday) at 22:01 (GMT) the Integral satellite recorded a burst of gamma-rays in its IBIS detector (from the Imager on Board the Integral Satellite). The burst lasted for about 30 seconds and then faded. An automatic detection system (IBAS - Integral Burst Alert System) sent out a message over the internet less than 1.5 hours later. Then, about 6 hours and 10 minutes after the burst (at 04:10 GMT on 2003 December 04), the orbiting X-ray observatory XMM-Newton started observing in this direction and saw in X-rays the fading "afterglow" of the gamma-ray burst. The XMM-Newton observation lasted for just longer than 16 hours.

Where is GRB 031203?

Integral pinpointed the direction of the burst pretty accurately, and then the quick XMM-Newton observation improved upon this. The burst happened in the southern hemisphere sky in the direction of the constellation "Puppis" (aka. "The Ship's Stern"). The exact position given by XMM-Newton was RA = 08h 02m 30.2s, Dec = -39d 51m 04s (J2000).

In this direction lies a small galaxy, no more than a spec with all but the largest telescopes (see some images here). A spectrum taken with the 6.5m Baade telescope in Chile (info) measured a redshift of z=0.105. This corresponds to a distance of 450 Mpc (450 million parsecs) or about 1.46 billion light years away. Almost certainly the gamma-ray burst occurred in this galaxy.

How bright was GRB 031203?

Integral detected 4 × 10-13 Joules worth of gamma-rays (equivalent to 1.3 × 10-14 Watts of power during the 30s the burst lasted). This tiny amount was enough to trigger Integral's sensitive instruments.

Of course, this is the energy (power) as recorded from Earth. As the gamma-ray burst (almost certainly) occurred at a distance of 1.46 billion light years, this means that in order for us to see it as bright as this it must have shone with a power 8 × 1014 times that of the Sun (800 thousand billion Solar luminosities!). This is equivalent to a power of 3 × 1041 W during the 30 seconds of burst. In the 30 seconds this Gamma-Ray Burst released almost as much energy - in the form of electromagnetic radiation - as the Sun will over 1 billion years, 1/10 of its lifetime. And this was not particularly bright as GRBs go!

What makes the halo?

XMM-Newton saw the fading "afterglow" of the GRB, a much dimmer X-ray source decaying away after the initial burst had finished. Around the X-ray afterglow was a "halo" caused by X-rays from the initial burst (which was much brighter than the afterglow) scattered off layers of dust in our Galaxy, much in the same way as fog scatters the light from a car's headlights.

The rings look "thin" for two reasons. First, the initial "flash" of X-rays (which accompanied the gamma-rays) was quite short (only 30 seconds or so). Secondly the dust layers causing the scattering are rather thin (in comparison to the distances involved they are quite thin) - we call them "dust sheets." There are two rings because there are two quite distinct dust sheets in this direction through our Galaxy.

On cold nights it is possible to see a "halo" around the Moon - see this image - although this is caused by refraction of visible light (not scattering of X-ray light). For more information about the Moon's halo, click here

How far away is the dust?

From the expansion of the rings on the sky the distance to the dust screens can be measured. The distances are 2,900 and 4,500 light years away from Earth (the inner ring is caused by the more distant dust screen). In more astronomer-friendly units the distances are 882 ± 20 parsec (outer) and 1388 ± 32 parsec (inner). Probably the nearest of the two dust sheets is associated with the Gum nebula, bubble of hot gas resulting from many supernova explosions.

How big is the halo?

The radius of the outer ring measured about 274 seconds of arc at the end of the XMM-Newton observation (about 21.4 hours after the burst). This is 0.075 of a degree (for comparison, the radius of the full moon is about 0.25 of a degree). At the distance of the nearest dust screen (2,900 light years) the halo size corresponds to a projected size of 3.8 light years (1.17 parsec) in radius. However, this is not a very meaningful number; the halo is an optical effect caused by dust scattering X-ray light, its not actually a tangible "thing" as such. Nevertheless it is fun to think that the halo images (as seen from Earth) appeared to expand at about 1,500 times the speed of light!

Why does the halo expand?

Nothing really moves faster than light, it is just an optical effect. What is happening is that the X-rays scattered by larger angles travel along a slightly longer path between the dust and the Earth. The speed of light is constant (2.998 × 108 m/s, or 134 million mph) and so the X-rays travelling the longer distance reach us slightly later. This means there is a delay - we see the X-rays scattered at smaller angles first. This is why the halo rings seem to expand - not because anything is actually "moving" but because there is a time delay between seeing the X-rays that are scattered by smaller angles (which we see as a smaller ring) and those scattered by larger angles (larger ring).

A schematic diagram of this is available as either html, PS or PDF formats.

What was special about GRB 031203?

GRB 031203 happened in a direction which points through the plane of the Galaxy (ie: the long way through) and so through a large amount of intervening dust. In this direction there happen to be two concentrations of Galactic dust. The other interesting quirk about GRB 031203 is that it showed a "soft spectrum" - this is X-ray astronomer speak for saying that, by comparison with other GRBs, it showed more low-energy X-rays. As lower energy X-rays are more efficiently scattered by dust this increased the chance of seeing a halo.

Was this previously predicted?

Indeed it was. As far back as 1965 there was theoretical work on what happens to X-rays scattered off interstellar dust grains (the 1965 paper by J. W. Overbeck). Through the years the ideas were refined and eventually halos around bright Galactic X-ray sources (X-ray binaries) were first seen back in 1983 (the 1983 Nature paper by D. P. Rolf). In a 1995 paper based on ROSAT observations, P. Predelh and J. Schmidtt measured dust scattered X-ray halos around 28 X-ray sources (24 X-ray binaries and 4 supernova remnants). The possibility of halos around GRBs was considered in various theoretical papers but until this observation, none had been observed.

What did we do?

Based on our first analysis of the data we issued a GCN reporting the first ever detection of an X-ray halo (or echo) around a gamma-ray burst (GCN 2489). Following from this a more thorough analysis was presented in a scientific paper to be published in the Astrophysical Journal. The paper can be viewed here. All the GCNs about GRB 031203 can be viewed from here.

The XMM-Newton data were made into an animation. The animation was constructed by co-adding the images obtained from the three EPIC instruments (MOS1, MOS2 and pn) and smoothing the result with a Gaussian of width 8 arcsec (sigma). The data were divided into consecutive 5 ksec (83 minute) intervals and the consecutive images made into a movie clearly showing the expansion of the halo rings. (The first three animations used the same data but different methods of plotting.)

Movies of the GRB 031203 halo

anim.gif 1-2.5 keV band.
Scale = 0.0 - 1.0.
5 ks frame time with 'in-between' frames (2.5 ks) produced by interpolation.
2.2 Mb file
anim_hi.gif As above with higher image resolution
5.6 Mb file
anim2.gif RGB colour scheme (R=1-2.5 keV; G=2.5-4 keV; B=4-10 keV).
Scales: 0.0-0.75 (R), 0.0-0.25 (G), 0.0-0.3 (B)
simulation Animation of the simulated data
(by R. Willingale)

Maintained by Simon Vaughan (sav2 at star. le. ac. uk)
Last updated: 02/02/2004
Back to homepage