Missing matter found in the cosmic web

The location of nearly half of the ordinary matter in the Universe is unknown. X-ray observations suggest that this elusive ‘baryonic’ matter is hidden in the filamentary structure of the cosmic web.
Taotao Fang is in the Department of Astronomy and the Jiujiang Research Institute, Xiamen University, Xiamen, Fujian 361005, China.

Search for this author in:

We live in a dark Universe: just 5% of it consists of ordinary matter such as that found in atoms, whereas the rest is ‘dark’ matter and energy that cannot currently be detected directly1. However, observations of the nearby Universe suggest that up to 40% of this ordinary matter — which is made up primarily of particles known as baryons — is missing25. Baryonic matter is thought to be distributed through the Universe like a cosmic web, and the missing baryons are predicted to be located in the filamentary structures that connect the web, and in intergalactic space4. In a paper in Nature, Nicastro et al.6 report the detection of the X-ray absorption signatures of baryons in the spectra of a bright background object. The findings might finally reveal a major reservoir for baryonic matter.

Why have the missing baryons been so difficult to detect? One reason is that the density of the baryonic matter in the filaments is extremely low. The other reason is that the high temperature in the filaments causes the most abundant element (hydrogen) to be almost completely ionized — which means that it has no electrons to produce spectral features that could be used to detect it. However, there might be trace amounts of heavier elements such as oxygen, in which a few electrons are bound. These ions can produce detectable (but extremely weak) spectral features, typically in the X-ray and/or ultraviolet regions of the electromagnetic spectrum.

Nicastro et al. observed the X-rays emitted by a special type of astronomical object known as a BL Lacertae (BL Lac) object. These are typically extremely bright, and have no (or very few) intrinsic spectral features — which makes it easy to detect any absorption of their emissions by other objects between them and Earth, such as filaments in the cosmic web.

The BL Lac object studied by the authors is called 1ES 1553+113, and is more than 2,200 megaparsecs away. Nicastro and colleagues observed this target with the European Space Agency’s XMM-Newton X-ray Space Telescope over several periods, for a total observation time of about 1.75 million seconds (about 20 days). They thus obtained a spectrum with an extremely high signal-to-noise ratio, which allowed them to perform high-resolution spectroscopy of very weak spectral features (Fig. 1).

Emission spectrum of bright astronomical object called 1ES 1553+113

Figure 1 | The search for baryonic matter. Nicastro et al.6 used the XMM-Newton X-ray Space Telescope to detect the emission spectrum of a bright astronomical object called 1ES 1553+113. They observed lines superimposed on the spectrum, which they attribute to X-ray absorption by helium-like oxygen (oxygen ions that have just two bound electrons, not shown) in two filaments of the cosmic web located between the telescope and the emitter. The cosmic web is a massive structure composed of ordinary (baryonic) matter, such as that found in all atoms. If the authors’ attribution is correct, then the finding reveals the location of a major reservoir of baryonic matter. Distances and sizes of objects not shown to scale.

The authors discovered two highly statistically significant systems of absorption lines produced by helium-like oxygen (oxygen ions that have only two bound electrons) at redshifts of 0.43 and 0.36. Redshift measures the change in wavelength that occurs when light travels over astronomical distances, and is approximately proportional to the distance of the light-emitting object from Earth. The researchers also performed an optical survey of galaxies along the sight line towards 1ES 1553+113, and observed a high density of galaxies at the two redshifts associated with the absorption signals. Such densities are characteristic of the filamentary structures of the cosmic web. By combining the X-ray data with measurements of the ultraviolet emissions from 1ES 1553+113, Nicastro et al. estimated the density of the baryons associated with the X-ray absorbing features, and found that they account for 9–40% of the cosmic baryon density — suggesting that these features are a substantial reservoir of the missing baryons.

Weak X-ray absorption lines produced by baryons have been reported a few times before7,8, but most of the results were marginal, and in some cases debatable. What is remarkable about the current work is that it represents the first time both of the expected absorption lines for helium-like oxygen have been detected together (for the absorption system at redshift 0.43, although the statistical significance of one of the lines is marginal). The observation of two absorption lines from the same ion species is typically a good indication that the target ion species has been detected.

One concern is whether the observed X-ray-absorbing systems are truly located between Earth and 1ES 1553+113. The exact redshift of the BL Lac object is unknown; the best available estimate6 suggests that it is at least 0.41. This value is less than the redshift of one of the X-ray absorbers, which implies that this absorber is either part of the BL Lac object, or a misidentification of something else. Nicastro et al. argue that both scenarios are unlikely, but an accurate measurement of the redshift of 1ES 1553+113 is needed to resolve this issue.

It is also possible that the X-ray-absorbing systems are in galaxies, rather than in filamentary structures of the cosmic web — similar absorption systems have previously been detected in the Milky Way4. Nicastro and co-workers argue that this explanation is unlikely, partly because they did not find large galaxies similar to the Milky Way at the redshifts associated with the absorptions, but also because they did not detect any additional absorption lines from cold ions, which are typically found in galactic disks. These arguments are reasonable, but better observations are needed to rule out this scenario.

This type of observation, requiring more than a million seconds of exposure time, truly pushes the limits of the available instruments. Proposed space missions such as the Hot Universe Baryon Surveyor ( and the Advanced Telescope for High-Energy Astrophysics ( will have much more sensitive X-ray spectrometers, and might eventually provide a complete map of the missing baryons in the cosmic web.

An alternative approach for detecting the missing baryons is to use a phenomenon known as the Sunyaev–Zel’dovich effect, in which high-energy electrons scatter off photons in the cosmic microwave background (CMB; electromagnetic radiation left over as a remnant of the Big Bang), thereby slightly distorting the CMB spectrum. High-energy electrons outside galaxies, and probably also in the filaments of the cosmic web, could produce such a distortion9, yielding a signal that indicates the presence of baryons. In the meantime, Nicastro and colleagues’ findings offer a tantalizing glimpse of where the elusive missing baryons have been hiding.

Nature 558, 375-376 (2018)

doi: 10.1038/d41586-018-05432-2
Nature Briefing

Sign up for the daily Nature Briefing email newsletter

Stay up to date with what matters in science and why, handpicked from Nature and other publications worldwide.

Sign Up


  1. 1.

    Planck Collaboration. Astron. Astrophys. 594, 13 (2016).

  2. 2.

    Fukugita, M., Hogan, C. J. & Peebles, P. J. E. Astrophys. J. 503, 518 (1998).

  3. 3.

    Cen, R. & Ostriker, J. P. Astrophys. J. 514, 1 (1999).

  4. 4.

    Bregman, J. N. Annu. Rev. Astron. Astrophys. 45, 221–259 (2007).

  5. 5.

    Shull, J. M., Smith, B. D. & Danforth, C. W. Astrophys. J. 759, 23 (2012).

  6. 6.

    Nicastro, F. et al. Nature 558, 406–409 (2018).

  7. 7.

    Fang, T., Marshall, H. L., Lee, J. C., Davis, D. S. & Canizares, C. R. Astrophys. J. 572, L127 (2002).

  8. 8.

    Nicastro, F. et al. Nature 433, 495–498 (2005).

  9. 9.

    Hernández-Monteagudo, C. et al. Phys. Rev. Lett. 115, 191301 (2015).

Download references