The missing link

Article metrics

Some stars might be 'magnetars', powered by magnetism instead of fusion. The discovery of an X-ray burst from an anomalous X-ray pulsar suggests that this type of star can be added to the list.

About 60 years ago, physicists realized that ordinary stars are powered by nuclear fusion. Since then, astronomers have found other types of star that are powered by gravitational binding energy, by radioactivity or by rotational energy. Over the past decade, there has been speculation that stars powered entirely by magnetism — 'magnetars' — could also exist. Magnetism is quite common in the Universe. The Earth has a magnetic field of 1 gauss (G), and the Sun of 10 G. For comparison, a typical toy magnet on a refrigerator door has a strength of about 100 G, and the strongest sustained magnetic fields produced in laboratories are around 106 G. But magnetars are thought to have magnetic fields a billion times stronger still. On page 142 of this issue, Gavriil, Kaspi and Woods1 present evidence that an anomalous star 15,000 light years from Earth is a magnetar.

There is already one class of astronomical objects that astronomers think are magnetars2,3. These are the 'soft-gamma-ray repeaters' (SGRs), which emit intense bursts of low-energy gamma-rays. One such burst was so powerful it caused detectable changes in the Earth's ionosphere. Initially confused with extragalactic gamma-ray bursts, SGRs are believed to be a subclass of the young neutron stars within our Galaxy4. The bursts are so intense that they can only be explained by the existence of strong magnetic fields confining the electrons that are radiating gamma-rays (similar to the magnetic confinement of plasma in a fusion reactor). The discovery of regular pulsations with long periods5 added further support for the existence of intense magnetic fields in SGRs.

Whether another subclass of neutron stars are magnetars has been more controversial. These are a group of X-ray pulsars with similarly long periods — the so-called 'anomalous X-ray pulsars' (AXPs). Pulsars are thought to be spinning neutron stars, and ordinary X-ray pulsars are powered by the matter that they accrete from a companion star. But, despite careful searches, no plausible companions have been identified around AXPs. Nonetheless, AXPs show persistent and strong X-ray emission, well above the level that can be supported by their store of rotational energy. The source of energy in these objects is mysterious, which is why these X-ray pulsars are called anomalous. The long rotation period of AXPs also suggests a family resemblance to SGRs.

According to the magnetar model devised by Thompson and Duncan6, the decay of the intense magnetic field in a magnetar provides a source of energy and powers the radiation emission. By analogy with the Sun (Fig. 1), loops of magnetic field could occasionally reconnect, producing a flare. But in the case of magnetars, the field strength is 1012 times greater than that in the loops on the Sun, and the resulting flares are proportionally more intense. Furthermore, the strong magnetic field might move portions of the outer layer, or crust, of the neutron star — as though moving continental plates. Like earthquakes, a major flare could be followed by minor ones. Once the stress has built up once more, the reconnection process happens all over again, which explains the repetitive appearance of the flares.

Figure 1: The surface of the Sun photographed by the TRACE spacecraft.


Hot ionized gas forms coronal loops, tracing out the magnetic-field structure at the star's surface.

There are, however, several issues still outstanding for the magnetar model. To start with, there is the problem of the evidence for strong magnetic fields in SGRs. From the spectra of emitted light, astronomers have been able to measure field strengths directly in a number of astrophysical objects, including the Sun and ordinary neutron stars. Yet there had been no such direct spectral evidence for SGRs: evidence for their strong magnetic fields came only indirectly, through their intense bursts of radiation. Second, the magnetar model postulates that both AXPs and SGRs are magnetars, and yet the two types of star seem quite different in that no bursts of radiation had ever been detected from AXPs. So although the case for SGRs as magnetars is plausible, for AXPs it has up to now been considerably weaker.

But during a routine monitoring programme using the Rossi X-ray Timing Explorer, Gavriil et al.1 found bursts from an AXP. They argue quite persuasively that these bursts arise from the AXP and not from an unrelated source. This discovery at last establishes a strong link between AXPs and SGRs.

There is a plausible explanation for why the bursts from AXRs are fainter than those from SGRs, and thus harder to detect. AXPs tend to be located at the centre of supernova remnants, and so are undoubtedly young neutron stars. In contrast, SGRs are not associated with supernova remnants, implying that SGRs must be older than AXPs7. It is possible that the crust of young neutron stars is more malleable or plastic, and thus AXPs may be unable to support superstrong magnetic loops that reconnect to generate the intense bursts. But as AXPs age, the crust could become less plastic and consequently be capable of supporting strong magnetic loops — up to a point.

Gavriil et al.1 also made spectroscopic measurements. They found that the bursts have a strong feature in their spectrum that can reasonably be interpreted as being due to protons gyrating in the intense fields of a magnetar. A similar spectral feature has recently been reported for an SGR8. Both detections await firm confirmation — and will motivate many further observations — but if they are real then we have direct evidence for strong field strengths.

It seems that magnetars do exist. Astronomers can feel quite satisfied to have postulated, discovered and confirmed a new class of cosmic object. And physicists will be excited by the possibilities of magnetars: where magnetic field strengths exceed about 1014 G, all sorts of strange effects arising from the quantum theory of electrodynamics are potentially detectable (for example, the intense fields of magnetars could polarize the vacuum). Sensitive experiments could look for such effects and make good use of one of nature's greatest laboratories.


  1. 1

    Gavriil, F. P., Kaspi, V. M. & Woods, P. M. Nature 419, 142–144 (2002).

  2. 2

    Thompson, C. & Duncan, R. C. Astrophys. J. 408, 194–217 (1993).

  3. 3

    Pacynski, B. Acta Astron. 42, 145–153 (1992).

  4. 4

    Kulkarni, S. R. & Frail, D. A. Nature 365, 33–35 (1993).

  5. 5

    Kouveliotou, C. et al. Nature 393, 235–237 (1998).

  6. 6

    Thompson, C. & Duncan, R. C. Mon. Not. Astron. Soc. 275, 255–300 (1995).

  7. 7

    Gaensler, B. M., Slane, P. O., Gotthelf, E. V. & Vasisht, G. Astrophys. J. 559, 963–972 (2001).

  8. 8

    Ibrahim, A. I. et al. Astrophys. J. 574, L51–L55 (2002).

Download references

Author information

Correspondence to Shri Kulkarni.

Rights and permissions

Reprints and Permissions

About this article

Cite this article

Kulkarni, S. The missing link. Nature 419, 121–123 (2002) doi:10.1038/419121a

Download citation

Further reading

  • Astrophysics in 2002

    • Virginia Trimble
    •  & Markus J. Aschwanden

    Publications of the Astronomical Society of the Pacific (2003)


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.