By the time the universe was three minutes old, all the baryonic building blocks of normal matter had formed. As the young universe aged, a minority of those primordial baryons—a few percent—joined clumps of the far more abundant dark matter to condense and form the first galaxies.
Those galaxies grew by merging with each other. They grouped together in clusters. New generations of stars sprang from the gas left by their predecessors’ explosive demise. Throughout those processes, which are still going on, the overall distribution of baryons more or less persisted: Most baryons remain outside galaxy clusters.
Astronomers have confidence in their predictions of how much baryonic matter formed in the Big Bang. And their observations of luminous matter indicate how much baryonic matter lies in stars, galaxies, and galaxy clusters. What’s been harder to determine is the baryonic content of the intergalactic medium.
In principle, accounting observationally for all the IGM baryons is straightforward. Both the density fluctuations that led to the first galaxies and the mergers that formed their successors have squeezed dark and baryonic matter into a foamy network of widely spaced nodes, the galaxy clusters, connected by wispy filaments, the IGM. See figure 1.
Figure 1. Computer simulations like this one show that galaxies and clusters of galaxies form at the nodes of a foamy, filamentary web. The area of sky in the image is about 100000 light-years across. Stars appear in yellow. The colors from violet through blue and green to white correspond to gas of increasing density.
Figure 1. Computer simulations like this one show that galaxies and clusters of galaxies form at the nodes of a foamy, filamentary web. The area of sky in the image is about 100000 light-years across. Stars appear in yellow. The colors from violet through blue and green to white correspond to gas of increasing density.
Almost inevitably, a line of sight to a distant quasar passes through a succession of filaments at different redshifts. The potentially observable result is a quasar spectrum imprinted with absorption lines at helpfully discrete wavelengths. From the lines’ depths, one can infer the amount and distribution of the absorbing gas.
Where those lines show up in the electromagnetic spectrum depends not only on redshift but also on the temperature of the absorbing gas and the atomic species. At redshifts of 2 to 6, cold, neutral hydrogen absorbs at the Lyman-α transition in the near-UV and optical.
Astronomers discovered those lines, dubbed the Lyman-α forest, in the 1970s when 4-m telescopes equipped with photon-counting spectrometers came on line. With plausible assumptions about the ionization state of the clouds, they balanced the baryon budget: What was produced in the Big Bang matched the total seen in stars, galaxies, and the Lyman-α forest.
But a problem appeared. At lower redshifts, the Lyman-α forest thins out. The cold, neutral clouds that balance the baryon budget at higher redshifts are less abundant in the universe’s old, close-in neighborhoods. An apparent deficit had opened.
Hints of a solution emerged in the late 1990s when Renyue Cen and Jeremiah Ostriker of Princeton University tracked the fate of the IGM with a hydrodynamic computer simulation. 1 Their simulation, which spanned redshifts from 3 to 0, showed that the large-scale collapsing and squeezing that forms the universe’s foamy structure also shock-heats the IGM’s baryons. As the universe ages, the IGM’s temperature climbs.
At the lowest redshifts, shock-heated hydrogen and other elements are ionized. Their absorption lines pass out of the optical and into the far-UV and x ray. The baryons weren’t missing from the nearby IGM; they were just too hot to see—at least with the instruments and telescopes of the past century.
Now, Charles Danforth and Michael Shull of the University of Colorado at Boulder have used the UV spectrometers aboard the Far Ultraviolet Spectroscopic Explorer ( FUSE ) and the Hubble Space Telescope ( HST ) to survey quasar absorption lines in the IGM. Their survey has turned up 40% of the missing baryons in the nearby universe. 2 The rest, they and others say, lie in even hotter x-ray-emitting gas. New missions, such as NASA’s Constellation-X and ESA’s X-ray Evolving Universe Spectroscopy , are expected to find some of them.
Lines of sight
To look for the missing baryons, Danforth and Shull combed through 36 quasar spectra taken in the UV by FUSE and HST . The sample is fixed for now: The FUSE mission, which began in 1999, ended last year when one of its gyroscopes failed and, although HST is still in orbit, its UV spectrometer fell victim in 2004 to an electrical fault.
Not all the spectra yielded usable data. Some were too noisy. Others came from quasars that were too close: Their lines of sight didn’t pass through enough absorbing filaments to yield a significant sample. And for a handful of more distant quasars, the lines of sight passed through so many absorbers that reliably assigning redshifts proved impossible.
The final sample included 28 quasars and hundreds of absorption lines from various atomic species and from redshifts up to 0.4. Figure 2 shows an example. Finding several absorbing filaments at different redshifts in such a spectrum might seem daunting. Atomic physics helps. Corrected for redshift, lines from a given species will always appear at the same wavelengths.
Figure 2. The UV spectrum from the quasar PG 1211+143 as observed by the Far Ultraviolet Spectroscopic Explorer (FUSE) and the Space Telescope Imaging Spectrograph (STIS) on the Hubble Space Telescope (HST). The broad peaks originate from the quasar itself, whereas the narrow absorption lines originate from filaments or clouds of gas that lie at various redshifts between Earth and the quasar.
Figure 2. The UV spectrum from the quasar PG 1211+143 as observed by the Far Ultraviolet Spectroscopic Explorer (FUSE) and the Space Telescope Imaging Spectrograph (STIS) on the Hubble Space Telescope (HST). The broad peaks originate from the quasar itself, whereas the narrow absorption lines originate from filaments or clouds of gas that lie at various redshifts between Earth and the quasar.
To identify an absorption line, Danforth and Shull would proceed iteratively. First they’d assume a line belonged to Lyman-α and then look for Lyman-β at the same redshift. If they found it, they’d look for lines from other species at the same redshift.
In addition to Lyman-α, Danforth and Shull found lines from six different ionic species: O5+, N4+, C2+, Si3+, Si2+, and Fe2+.
The lines appear to arise from two distinct thermal phases. The majority phase, traced by Lyman-α, is in the form of photoionized gas at 104 K; it closes 30% of the baryon deficit. The minority phase, traced by O5+ and the other ions, is in the form of shock-heated gas at 105 to 106 K; it closes 10% of the baryon deficit.
The abundances of oxygen, carbon, and other elements made in stars are surprisingly high: about 0.1 of the solar values. Evidently, stellar material travels millions of light-years from its original galactic host.
Shocked—shocked
It would have been a surprise if all the missing baryons turned up in the warm, UV-absorbing gas that Danforth and Shull found. Simulations by Cen and Ostriker and by others predict that some gas, especially the gas closest to the sites of newly merged galaxies, should be at tens of millions of degrees. At that temperature, oxygen and carbon become more highly ionized and their absorption lines pass into the x-ray band.
Chandra and XMM-Newton can barely detect those x-ray absorption lines. The two observatories have yielded weak and contradictory evidence. On the other hand, an XMM-Newton image recently analyzed by Norbert Werner of SRON Netherlands Institute for Space Research and his collaborators revealed faint hot gas between two clusters of galaxies, Abell 222 and Abell 223. 3 One observation constitutes consistent, but not compelling, evidence of the whereabouts of the last slice of missing baryons.
Beyond the satisfaction of balancing the baryon budget, what would observations of the warm and hot IGM reveal about the universe? The number of absorbers that Danforth and Shull have found agrees with the number of filaments that simulations predict. Those simulations, and their underlying assumptions, will be tested further as the sensitivity of observations and the power of computers improve.
Observations of the nearby IGM could also help astronomers tackle one of their most pressing and difficult problems: the role of feedback in galaxy formation. One expects the formation of large-scale structure to heat the IGM as galaxies form, but not necessarily to enrich it. The presence of stellar material deep in the IGM suggests that galaxies could influence the formation of other galaxies by expelling hot, energetic gas into the IGM.
Fourteen billion years after the Big Bang, the universe remains a violent place.
