Swift progress

The agile, choreographed response of the Swift satellite to γ-ray bursts tests models to an unprecedented degree. Results from two recent long bursts suggest that the models are good, but require some tweaking.

The Swift satellite1 is NASA's latest tool for investigating the mysterious phenomenon of γ-ray bursts, or GRBs. These intense bursts of high-frequency γ-ray and X-ray radiation were discovered four decades ago2; since Swift's launch in November 2004, one such burst has activated its detectors every few days. On page 985 of this issue, Tagliaferri et al.3 present observations from Swift that suggest strongly that the ‘prompt’ emission of the burst and its so-called afterglow are separate radiation components. This is consistent with models in which the prompt phase of γ-radiation, lasting a few to a few hundred seconds, results from internal shocks within the medium that produces them, whereas the longer-lasting afterglow, at lower frequencies, is the outcome of interactions between ejected and surrounding material.

The durations of GRBs, together with their spectral properties, suggest a classification into short and long bursts4. How far away both classes originate was unknown until a distance scale was established, at first on a statistical basis5 by comparing the directions of GRBs with models of their possible distribution in the Galaxy. This was followed by direct determination through the discovery of X-ray afterglows6 and the measurement of redshifts through optical line spectroscopy7. Separate models for the origins of short and long GRBs also emerged. Short GRBs are believed to be associated with the merger of compact binaries: a pair of neutron stars, for example. This explanation is not yet certain, and the Swift mission is designed to clarify the nature of the short bursts.

For long GRBs, the favoured ‘collapsar’ model8 starts from the idea of a rapidly rotating, massive star that has undergone extreme gravitational collapse and formed a central black hole. Complex processes extract some of the gravitational binding energy from the disk of material that forms around this black hole; directed along the axis of rotation, this energy breaks through the surface of the star at almost the speed of light. This high-energy radiation shockwave is sent out along a narrow jet and can be spotted by detectors — if the jet is pointing the right way.

According to the collapsar model, the prompt energetic structure of a long GRB is the result of dissipation by internal, relativistic shocks, which may last seconds or minutes, at a radius of about 1014 cm from the centre of the collapsed star9. At this point, much of the energy remains in storage as kinetic energy and is tapped later10, at a radius of about 1016 cm, when the outflow collides with the interstellar medium or the particle wind from the progenitor star. This two-stage model, known as the internal–external model11, predicts a transition from an erratic, prompt γ-ray outburst to a smoother X-ray afterglow (Fig. 1). Late bumps in the light curves that are observed from long GRBs, as well as directly observed spectral features, are reminiscent of features from stel- lar explosions — supernovae — and provide evidence in support of the collapsar model.

Figure 1: Staged burst.

The flux F of a γ-ray burst (GRB) over time t, as recorded by the Swift satellite and reported by Tagliaferri et al.1. The prompt γ-ray emission usually lasts for a only few seconds, and is believed to originate from shocks within the relativistic jet. The longer-lasting afterglow, in the X-ray to optical and infrared bands and typically outshining the GRB's host galaxy, decays following a power law, with the flux falling as t−1 during the first 24 hours, and faster (t−2) thereafter. In the transition region between prompt emission and afterglow, the gradient is even steeper (t−3). A bump in the light curves sometimes observed a few weeks after the trigger is predicted by the collapsar model for long GRBs, and is believed to be light from an associated supernova.

Tagliaferri and colleagues observed3 two long bursts, GRB050126 and GRB050219a, which lasted about 20–30 seconds and were detected initially by Swift's Burst Alert Telescope (BAT) at γ-ray frequencies. True to its name, Swift rapidly — in 129 seconds in the first case and 87 seconds in the second — slewed its X-Ray Telescope (XRT) into the line of the burst to allow unprecedented monitoring, at low X-ray frequencies, of the afterglow. The Swift data show that, as the prompt emission fades, an initial afterglow appears that declines very rapidly over a few hundred seconds. This early afterglow is, however, distinct from the later afterglow familiar from pre-Swift data — with the high sensitivity and excellent time coverage of Swift, we are now witnessing the X-ray light curves in the transition period from prompt emission to afterglow.

Although the Swift data broadly support the notion of two components to GRB emission, the decay of the burst's prompt brightness within the first few minutes is unexpectedly steep, falling with the inverse cube of time, t−3. How should we interpret this? The authors cite several possibilities3, but most of these are undermined by the ‘non-thermal’ spectrum of the radiation that follows a power law, rather than the ‘blackbody’ form assumed in the models. The most intriguing option that would provide such a rapid decay is ‘inverse Compton’ scattering of low-energy photons by high-energy electrons in the reverse shock of the burst. (Blast waves lead to two shocks, one propagating ahead into the external medium, and another propagating back through the ejected material of the jet.) If the inverse Compton mechanism dominates — which would be true if the outflow contained a very large abundance of electron–positron pairs — prompt ‘optical’ flashes that are also not observed might be suppressed12. Robotic telescopes on the ground with swift responses should soon constrain these speculations.

The standard model used to explain GRB afterglows is known as the fireball model and is based on a few main ingredients. First, the extreme brightness of the GRB, in conjunction with the large distances involved, implies a huge energy release. Second, the brevity of the bursts indicates a small volume in which this energy is released. Taken together, these seem to imply the existence of a dense, opaque fireball driving the burst to relativistic speeds.

If there really were a fireball, this would imply emission in all directions, and the observed GRB energies would be difficult to explain within the confines of conventional astrophysics. But commonly observed breaks in the afterglow suggest that the outflow is indeed jetted and is thereby limited to a small solid angle, reducing the energy required. After correcting for this effect, the bursts seem to behave approximately as ‘standard candles’13 — objects whose observed brightness depends only on how far away they are — making them useful measures of cosmological distance14.

Continued monitoring of the sky with Swift, with other detectors such as HETE-2 and INTEGRAL, and with the future missions GLAST-GBM, Agile and EXIST, will further expand our understanding of GRBs. They will remain a source of amazement and wonder for some time to come.


  1. 1

    Gehrels, N. et al. Astrophys. J. 611, 1005–1020 (2004).

    ADS  CAS  Article  Google Scholar 

  2. 2

    Klebesadel, R. W., Strong, I. B. & Olson, R. A. Astrophys. J. 182, L85–L88 (1973).

    ADS  Article  Google Scholar 

  3. 3

    Tagliaferri, G. et al. Nature 436, 985–988 (2005).

    ADS  CAS  Article  Google Scholar 

  4. 4

    Kouveliotou, C. et al. Astrophys. J. 413, L101–L104 (1993).

    ADS  CAS  Article  Google Scholar 

  5. 5

    Meegan, C. et al. Nature 355, 143–145 (1992).

    ADS  Article  Google Scholar 

  6. 6

    Costa, E. et al. Nature 387, 783–785 (1997).

    ADS  CAS  Article  Google Scholar 

  7. 7

    Metzger, M. Nature 387, 878–880 (1997).

    ADS  CAS  Article  Google Scholar 

  8. 8

    MacFadyen, A. I. & Woosley, S. E. Astrophys. J. 524, 262–289 (1999).

    ADS  CAS  Article  Google Scholar 

  9. 9

    Rees, M. & Meszaros, P. Astrophys. J. 430, L93–L96 (1994).

    ADS  Article  Google Scholar 

  10. 10

    Rees, M. & Meszaros, P. Mon. Not. R. Astron. Soc. 258, 41p–43p (1992).

    ADS  CAS  Article  Google Scholar 

  11. 11

    Piran, T. Rev. Mod. Phys. 76, 1143–1210 (2004).

    ADS  Article  Google Scholar 

  12. 12

    Kobayashi, S. et al. Astrophys. J. Lett. (submitted); preprint at (2005).

  13. 13

    Bloom, J. S., Frail, D. & Kulkarni, S. R. Astrophys. J. 594, 674–683 (2003).

    ADS  Article  Google Scholar 

  14. 14

    Ghirlanda, G. et al. Astrophys. J. 613, L13–L16 (2004).

    ADS  CAS  Article  Google Scholar 

Download references

Author information



Rights and permissions

Reprints and Permissions

About this article

Cite this article

Hartmann, D. Swift progress. Nature 436, 923–924 (2005).

Download citation

Further reading


By submitting a comment you agree to abide by our Terms and Community Guidelines. If you find something abusive or that does not comply with our terms or guidelines please flag it as inappropriate.