X-ray astronomy

When appearances are deceptive

The sharpest X-ray image ever obtained of a portion of the Milky Way resolves a seemingly diffuse X-ray emission into discrete sources. These sources are likely to be stars of the 'garden variety' in the Sun's vicinity.

With blurry vision, one can imagine strange and wonderful things that later turn out to be erroneous. Astronomers have learned this lesson before, from the hypothetical canals on Mars1 to claims of rotating spiral arms in the Andromeda galaxy2. Even the imaginary 'man in the moon' figure may be the result of poor vision.

In its early years, X-ray astronomy had similar difficulties interpreting the diffuse, extragalactic X-ray background (XRB) emission from the sky3. Was this X-ray glow produced by hot gas spread between the galaxies or a blur of unresolved point sources? Only with the development of modern X-ray detectors and telescopes, particularly the Chandra X-ray Observatory4, was most of this cosmic XRB radiation resolved into discrete sources. The 'background' turned out to be a myriad of distant quasars and other galaxies whose bright cores are powered by supermassive black holes5.

On page 1142 of this issue, Revnivtsev et al.6 address a similar problem — that of deciphering the hitherto unresolved X-ray emission from the Milky Way's disk of gas and stars, the Galactic X-ray ridge emission7,8. Their analysis resolves most of this apparently diffuse X-ray emission into discrete sources. The results indicate that many of these sources are stars, one of the less exotic early suggestions9,10.

As its name suggests, the Galactic X-ray ridge is a ridge of X-ray emission in the Galaxy: at energies above several kiloelectronvolts (keV), the diffuse XRB emission exhibits an excess near the Galactic disk that extends a few hundred parsecs (about 800 light years) in height above the disk.

Its spectrum9 is characteristic of a hot plasma at 100 million kelvin, including a prominent emission line of iron at 6.7 keV. Such emission is similar to that from hot intergalactic gas confined to clusters of galaxies. The problem with this hot-plasma description of the Galactic ridge emission is that the radiative energy losses from hot electrons in the plasma would be enormous11, perhaps 1043 erg s−1, exceeding all likely sources of heat in the Galaxy, even supernovae. Furthermore, the gravity in the Milky Way disk is much too weak to confine this hot plasma, which should instead be flowing away at large velocities in a Galactic wind. This phenomenon has been observed in other galaxies, such as M82 (Fig. 1).

Figure 1: Galactic wind.


A horizontal disk of stars (white) and a perpendicular galactic wind of hot plasma (purple) are visible in this image of the starburst galaxy M82. Revnivtsev and colleagues6 show that the apparently diffuse X-ray emission from the Milky Way's disk is not due to a Galactic wind but instead to discrete sources such as stars. This colour-coded, visible-light image combines observations taken by the WIYN 3.5-metre telescope and the Hubble Space Telescope.

The sharp optics of the Chandra Observatory allowed Revnivtsev et al.6 to detect point sources in an ultra-deep image obtained from a million-second (equivalent to about 12 days) exposure of a small (16 × 16 arcminute square) region of sky near the Galactic Centre (see Fig. 1 on page 1143). In total, they detected 473 X-ray sources with a reliable degree of confidence. These sources account for a substantial fraction of the X-ray emission: in the energy range 6.5–7.1 keV, and particularly near the prominent 6.7-keV iron emission line, the authors' analysis resolves more than 80% of the emission into discrete sources, most of which are probably stars.

Revnivtsev and colleagues' work demonstrates the power of combining ultra-long X-ray sky exposures with exquisite telescope optics. In many ways, the Chandra Observatory marks a high point of X-ray imaging, with mirrors capable of resolving objects to an accuracy of 0.5 arcsecond — the size of a small coin viewed from a distance of 10 kilometres. Most of the X-ray sources found in the Chandra image are probably around the same distance away as the Galactic Centre is from Earth, which is approximately 8.5 kiloparsecs (27,700 light years). At this distance, the brightest X-ray sources have luminosities of 1032 erg s−1, just 3% of the Sun's total luminosity at all wavelengths.

The minimum luminosities detectable, with the sensitivity of the mirrors and cameras aboard Chandra, are 100 times fainter, around 1030 erg s−1. These X-ray sources are likely to be white dwarfs with luminosities of 1031–1032 erg s−1 accreting matter from companion stars, and binary stars with strong coronal activity — akin to flares on the Sun, but with X-ray luminosities of 1031 erg s−1 or less. Such stellar X-ray sources are of the common 'garden variety' in the Sun's neighbourhood12. However, at the distance of the Galactic ridge from Earth, their combined light becomes a diffuse blur, the X-ray equivalent of the many stars that make up the Milky Way, as Galileo first saw with his telescope in visible light.

Regarding the value of high-resolution imaging, it is worth reflecting on the greater significance of Revnivtsev and colleagues' observations. The resolution of the seemingly diffuse X-ray emission into discrete sources eliminates a physically implausible explanation of that emission that required enormous energy sources of hot plasma. The discrete-source model offers the chance to apply techniques of optical (visible-light) astronomy, such as stellar-population synthesis and surface-brightness fluctuations — both based on the assumption that galaxies are made up of a finite number of stars — to study galaxies other than our own in the X-ray waveband. For example, the apparently diffuse X-ray emission of distant galaxies may also be dominated by unresolved faint stellar sources and supernova remnants7,9,10.

As with the extragalactic X-ray background, the sources that produce the Galactic ridge X-ray emission proved to be less exotic than some early suggestions. Perhaps there are lessons here, and astronomers should bear that in mind as they explore sources of faint backgrounds at many wavelengths, from X-ray to infrared. As Revnivtsev and colleagues' work6 demonstrates, sometimes the exotic explanation can be set aside by more accurate imaging and spectroscopy.


  1. 1

    Lowell, P. Astrophys. J. 26, 131–140 (1907).

  2. 2

    van Maanen, A. Astrophys. J. 81, 336–337 (1935).

  3. 3

    Giacconi, R., Gursky, H., Paolini, F. & Rossi, B. Phys. Rev. Lett. 9, 439–443 (1962).

  4. 4

    Weisskopf, M. C. et al. Astron. Soc. Pacif. 114, 1–24 (2002).

  5. 5

    Hasinger, G. et al. Astron. Astrophys. 365, L45–L50 (2001).

  6. 6

    Revnivtsev, M. et al. Nature 458, 1142–1144 (2009).

  7. 7

    Worrall, D. M., Marshall, F. E., Boldt, E. A. & Swank, J. H. Astrophys. J. 255, 111–121 (1982).

  8. 8

    Warwick, R. S., Pye, J. P. & Fabian, A. C. Mon. Not. R. Astron. Soc. 190, 243–260 (1980).

  9. 9

    Koyama, K., Makishima, K., Tanaka, Y. & Tsunemi, H. Publ. Astron. Soc. Jpn 38, 121–131 (1986).

  10. 10

    Worrall, D. M. & Marshall, F. E. Astrophys. J. 267, 691–697 (1983).

  11. 11

    Tanaka, Y. Astron. Astrophys. 382, 1052–1060 (2002).

  12. 12

    Sazonov, S., Revnivtsev, M., Gilfanov, M., Churazov, E. & Sunyaev, R. Astron. Astrophys. 450, 117–128 (2006).

Download references

Author information

Rights and permissions

Reprints and Permissions

About this article

Cite this article

Shull, M. When appearances are deceptive. Nature 458, 1119–1121 (2009). https://doi.org/10.1038/4581119a

Download citation


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.