2.3. The Matter Composition of the Universe
While we know more about the other one-third of the universe - the matter part - important questions remain. According to the current best census, the visible part of ordinary matter - that associated with stars - contributes only about 1% of the total. What we can see with telescopes is literally the tip of an enormous iceberg.
The rest of the matter in the universe is dark, and its existence is inferred from its gravitational effects. While the case for dark matter holding together galaxies (as well as clusters of galaxies) has been around for a long time (Zwicky, 1933, Rubin et al., 1980), the nature of the dark matter in the universe is still unknown. In fact, we still speak with more certainty about what dark matter is known not to be. Based upon simple accounting, we have all but eliminated the possibility of dark matter being made of neutrons and protons, and established a strong case for a new form of matter.
The accounting of ordinary matter involves three different methods, all of which arrive at the same answer. The most precise of these methods comes from consideration of the formation of light elements during big-bang nucleosynthesis (BBN). Hydrogen, helium, deuterium, and lithium are produced in the first few minutes of the big bang. However, only if the density of ordinary baryons is within a narrow range is the predicted production consistent with what we actually measure (see Figure 5). The production of deuterium is the most sensitive indicator of the baryon density. Measurements made with the 10-meter Keck Telescopes of the amount of deuterium in high-redshift clouds of gas (seen by their absorption of light from even more distant quasars in the Lyman series of lines), together with the theory of big-bang nucleosynthesis yield a density of ordinary matter of 3.8 ± 0.2 × 10-31 g cm-3, or only about 4% of the critical density (Burles et al., 2001).
Figure 5. The predicted abundance of the light elements vs. baryon density. The vertical band indicates the narrow range of baryon densities consistent with the deuterium measurements; the boxes (and open box for 3He) indicate the range in baryon density (horizontal extent of box) that is consistent with the measured light-element abundance (vertical extent of box). The overlap of the boxes with the deuterium band indicates the general consistency of the observed abundances of the other light elements with their predicted abundances for this baryon density. (Note, for the B scale at the top, h2 = 0.5 is assumed.)
Two other determinations are consistent with the nucleosynthesis argument: First, the net absorption of light emitted from very distant quasars by intervening gas (which exists in clouds of gas known as the Lyman-alpha forest after the multiplicity of redshifted absorption features produced by the individual clouds) indicates a similar value for the baryon density. This probes ordinary matter at a time and place when the bulk of the baryons are still expected to still be in gaseous form (z ~ 3 - 4). The second constraint comes from measurements of the CMB, which yield an independent baryon density consistent with that determined from nucleosynthesis. Our best accounting of ordinary matter comes from this early, simpler time, before many stars had yet formed.
Our accounting of baryons at the present epoch, in the local universe, is not as complete. Baryons in stars account for only about one-quarter of all the baryons; the rest are optically dark. While a number of possibilities for the baryonic dark matter (from planets to black holes) have been considered, it now appears that the most plausible reservoir for most of the unseen baryons is warm and hot ionized gas surrounding galaxies within groups and clusters. In fact, in rich clusters the amount of matter in hot intercluster gas exceeds that in stars by a large factor. But since only a few percent of galaxies are found in these unusually rich clusters, the bulk of the dark baryons are still unaccounted for.
While not all of the dark baryons are accounted for, baryonic dark matter itself only accounts for about one-tenth of all dark matter. The evidence that the total of amount of dark matter is much greater - about one-third of the critical density - has gradually become firm, as several, independent (and increasingly higher precision) measures have yielded concordant results (Sadoulet, 1999 Griest and Kamionkowski, 2000).
Clusters of galaxies provide a laboratory for studying and measuring dark matter in a variety of ways. Perhaps most graphically, dark matter can be seen in its effect on more distant background galaxies whose images can be distorted and multiplied by dark-matter gravitational lensing effects. This and other techniques (applied in the x-ray, radio, and optical) have determined the ratio of the total cluster mass to ordinary matter (predominantly in the hot x-ray emitting intracluster gas): averaged over more than fifty clusters the ratio is about 8 (Mohr et al., 1999, Grego et al., 2001). Assuming that clusters provide a "representative sample" of matter in the Universe, the total amount of matter can be inferred from the baryon density. That number is about one third of the critical density.
What then is this nonbaryonic dark matter? The working hypothesis is weakly interacting elementary particles produced in the early universe. Before discussing specific particle candidates, let's review the constraints from astrophysical observations. First, because dark matter is diffusely distributed in extended halos around individual galaxies or in a sea through which cluster galaxies move, dark-matter particles must not interact with ordinary matter very much, if at all. Otherwise, dark matter would by now have dissipated energy and relaxed to the more concentrated structures where only baryons are found. At the very least, we can be confident that the constituents of nonbaryonic dark matter are uncharged, and have only very weak interactions.
In addition, the formation of structure in the universe tells us that early on dark-matter particles must have been cold (i.e., moving at non-relativistic speeds) rather than hot (i.e., moving relativistically). If the dark matter had been hot, then these fast-moving particles would have smoothed out the smaller density irregularities, which seed the formation of galaxies and clusters, by streaming from high-density regions to low-density regions. The first objects to form would have been the largest structures (the superclusters) and smaller objects (galaxies) would have only formed later by fragmentation. However, this is inconsistent with observations.
The deep image of the sky obtained by HST in 1995 (the Hubble Deep Field; see Figure 6), along with other observations by ground-based telescopes, identified the epoch when most galaxies formed as a few billion years after the big bang (at redshifts of order 1 to 3). The Sloan Digital Sky Survey, as well as x-ray observations from space and other ground-based telescopes, have shown that clusters form later (redshifts less than about 1). Finally, superclusters, which are loosely bound collections of a few clusters, are forming just today. This sequence is inconsistent with hot dark matter.
Figure 6. The deepest image of the sky in visible light obtained by the Hubble Space Telescope in 1995 (the Hubble Deep Field). This image revealed the time when typical galaxies (like our own Milky Way) were forming (redshifts z ~ 1 - 3). In this image of about one-forty-millionth of the sky, there is one star and more than 1500 galaxies (image courtesy of NASA).
Nonetheless, there is at least one hot dark-matter particle that we do know exists - the neutrino. Two experiments, one undertaken in Canada, the other in Japan, now provide evidence that neutrinos have mass (Fukuda et al., 2002, Ahmad et al., 2001, Ahmad et al., 2002). The experiments, which are studying solar neutrinos and atmospheric neutrinos, have placed a lower limit on the mass of the heaviest neutrino at about 0.05eV. This implies that neutrinos contribute at least 0.1% of the mass-energy budget of the universe. However, the cosmological considerations just discussed cap the contribution of neutrinos - or any hot dark matter candidate - to be less than about 5%. This leaves the bulk of the dark matter still to be identified. We will return to the other particle candidates for dark matter later.