Nature: Dark Matter Revealed
Dark Matter Revealed by the First Stars

Rennan Barkana (Tel Aviv University) [link]

Nature, advance online publication on Wednesday, Feb. 28'th 2018 (in print Thursday, Mar. 1'st).

This webpage contains a short non-technical outline, selected links to audio/video and other websites, and a technical note about the EDGES measurement.



Brief outline

DM radio sky
An example of the predicted radio wave pattern on the sky due to the effect of dark matter in the early Universe.

Dark Matter

Physicists wish to answer some big questions: What is the Universe made of? What laws do its constituents follow? What is the history of the Universe from the distant past to the present, and how will it evolve in the future?

In recent decades, many of the biggest discoveries in physics have come from astronomy (dark matter, dark energy, gravitational waves, and black holes). This includes major progress on what makes up the Universe. For millenia, people speculated that stars are made of an exotic, cosmic substance. Then, in the 19'th century physicists discovered that light carries fingerprints of the atoms that emitted the light. Thus, it was discovered that the Sun and other stars are composed of the same kinds of chemical elements as are known to us on Earth. Starting in the 1970's, astronomers increasingly came to realize that most of the matter in the Universe is actually different from this "ordinary" matter that we know. The rotation of galaxies including our own Milky Way is so rapid that, according to Newton's theory of gravity (as generalized by Einstein), the strong gravity that causes this motion must come from a mass that is much larger than what we see in stars and gas. This mass is invisible (neither emits nor blocks light), so it is known as "dark matter". That there is five times more dark matter than ordinary matter is confirmed by observations of the remant glow of the early Universe.

Today the Universe is expanding, and it cools as it expands. If we look back in time, before the Universe had expanded so much, it was much hotter and denser. At an early time, conditions throughout the Universe were similar to the physical conditions in the Sun today, and the whole Universe glowed with light. This light still fills the Universe, but due to the cosmic expansion, it has been transformed into radio waves known as the Cosmic Microwave Background (CMB).

The First Stars

Observations of the CMB show that the early Universe was very different from today. It was highly uniform (with the density and temperature nearly the same everywhere), filled with dark matter, ordinary matter (mostly hydrogen), and light. Regions that were slightly denser than average experienced strong gravity, which eventually pulled together enough material to form galaxies of stars. This gradually led to the complex Universe we see today, with stars and planets separated by near-empty darkness.

Now, astronomers explore our distant past, billions of years back in time. Unlike archaeologists on Earth, however, who can only study remnants of the past, astronomers can see the past directly. It takes the light from distant objects a long time to reach us, and we see these objects as they were back when they emitted their light. For example, we see our own Sun not as it is now, but as it was 8 minutes ago. Other stars in the sky we see from many years ago. If astronomers look out far enough, they can see the first stars as they actually were in the early Universe.

For over a decade, astronomers have been searching for a radio-wave signal that had been predicted to be a sign of the formation of the first stars. The idea was that when the first stars formed, their radiation changed the internal state of the hydrogen atoms that filled the Universe. This should cause the hydrogen to absorb some of the radio waves of the CMB. Radio astronomers should thus be able to detect when the first stars formed by detecting this absorption. The point is that this occurred so early, and the corresponding signal is so distant, that we cannot see those first stars in any other way. Note that this field is called "21-cm cosmology" since the hydrogen atoms absorb radio waves at a wavelength of 21 centimeter.

March 2018 Discovery

Bowman and colleagues reported the first such detection, obtained with their EDGES radio telescope. The detection at a frequency of 78 MegaHertz corresponds to a signal that dates back to 180 million years after the Big Bang (today the Universe is 13.7 billion years old). The signal, though, had a much larger amplitude (corresponding to deeper absorption) than expected, presenting a major puzzle. I realized that this surprising signal can be explained by combining two factors: the first stars, and dark matter. The first stars in the universe turned on the radio signal as expected, while the dark matter collided with the ordinary matter and cooled it down. Extra-cold material naturally explains the strong radio signal.

The EDGES observation and my dramatic interpretation were announced in March 2018, and reported in two separate papers in Nature. If true, this exciting discovery will have enormous implications, and represent another example of deriving fundamental physics from astronomical observations. Indeed, this is the first direct clue about the nature of the mysterious dark matter that makes up most of the matter in the Universe. For example, physicists expected that dark matter would be made up of heavy particles, but the discovery indicates low-mass particles. This insight alone has the potential to reorient the search for dark matter.

Some caution is needed, since the EDGES measurement is the first of its kind. Independent confirmation is now being attempted, and much more information is expected within a few years. I predict that the dark matter produced a specific pattern of radio waves on the sky (see the image), which can be detected with a large array of radio antennas. One such array is the SKA (Square Kilometre Array), the largest radio telescope in the world, now under construction. Such an observation with the SKA would confirm that the first stars have indeed revealed the dark matter. While the SKA may directly measure the predicted radio images on the sky, other radio arrays such as HERA may first achieve a statistical detection of the expected pattern.


Selected Links

Links in Hebrew

  1. Short animated video explaining the discovery (Hebrew)
  2. Interview on Israel radio (Tel Aviv radio) program (with Gidi Gov), Mar. 4'th: audio , link
  3. Interview on Israel radio (Galei Zahal) news magazine (with Efi Triger), morning of Mar. 1'st: audio
  4. Interview on Israel radio (Kan Tarbut) news magazine, morning of Mar. 14'th, about Stephen Hawking's death: audio
  5. Hour-long public talk (TAU astronomy club, May 1'st 2018): link
  6. LONG in-depth video interview (with Dror Foyer) for Halalit TV link
  7. Ynet
  8. Tel Aviv University
  9. HaAretz
  10. IsraelHaYom
  11. Mako
  12. Maariv
  13. Israeli National Geographic (July 2018)
  14. Haaretz article on dark matter (July 2018)

Links in English

  1. Short animated video explaining the discovery (English)
  2. Chosen by Physics World as one of the top ten Breakthroughs of the Year for 2018: link
  3. FQXi podcast (15-min interview): audio, link
  4. BBC World Service Weekend (May 13, 2018): audio
  5. NPR report: audio , link
  6. Nature podcast: audio
  7. The (technical) paper: Nature website, Free read-only link
  8. New York Times
  9. Tel Aviv University
  10. The Guardian
  11. Jerusalem Post
  12. Qanta Magazine
  13. Science News
  14. Cnet
  15. Inside Science
  16. Chicago Tribune
  17. Digital Journal
  18. Space.com
  19. Science Magazine


  20. Technical note about the EDGES measurement

    EDGES check
    The best-fit amplitude of the 21-cm absorption signal versus the intensity of the sky (i.e., the foreground). The blue and green points are based on the same data, but with different binning in time and using different foreground models. Red curves show the best-fit amplitude (long-dashed curve) for the best data (i.e., the data taken when the sky temperature was lowest), along with the one-sigma error range (short-dashed curves). The magenta solid curve (plus dotted curves for the one-sigma range) shows the expected signal amplitude if it were proportional to the foreground level. Points taken from Extended Data Table 2 in the EDGES paper [Nature 555, 67 (2018)].

    One worry about the EDGES measurement is that the apparent 21-cm signal could be caused by frequency structure in the foreground (i.e., not coming from cosmic dawn) radio emission, or by an instrumental distortion of the foreground that could remain in the data if the instrumental modeling and calibration are inaccurate. This possibility can be tested for, since the foreground is dominated by emission from our Galaxy; as the Galaxy changes position in the sky (especially, going above or below the horizon), the foreground becomes brighter by up to a factor of 3 from the best data. A residual due directly to the foreground, or to an instrumental distortion of the foreground, would be expected to grow approximately in proportion to the foreground level.

    The figure (above or to the left) shows how the signal amplitude varies with the foreground level. The measurements all lie close to the amplitude derived from the best data (i.e., within the range indicated by the horizontal red curves). These measurements are completely inconsistent with being proportional to the foreground level (in which case they would all lie within the band indicated by the rising magenta curves).

    One possibility to evade this test is to have the apparent 21-cm signal produced by an extragalactic foreground that does not vary strongly with position on the sky. However, radio emission at such low frequencies is thought to come from synchrotron emission, which is spectrally smooth even when coming from a single electron. The spectrum should be further smoothed by the superposition of the large number of synchrotron sources that contribute to the global (sky averaged) 21-cm signal. I am not aware of a published claim that synchrotron sources can produce as sharp a frequency structure as that measured by EDGES.