Annu. Rev. Astron. Astrophys. 1994. 32: 531-590 Copyright © 1994 by Annual Reviews. All rights reserved

10.2. Best Bet Candidates

The various constraints on the form of baryonic dark matter discussed in this review are brought together in Table 2; the shaded regions are excluded by either dynamical, nucleosynthetic, lensing, or light constraints. This assumes that the objects all have the same mass, so that one does not have extra constraints associated with assumptions about the IMF. The dotted regions may also be excluded but this is less certain. Table 2 does not include the dynamical constraints on the nonbaryonic candidates, but it should be noted that only cold inos could explain the presence of dark matter in galaxies. However, hot inos could explain the cluster and background dark matter; large-scale structure and microwave anisotropy observations may even require a mixture of hot and cold inos (Taylor & Rowan Robinson 1992).

Whether the dotted region in Table 2 is excluded depends on whether one believes that the primordial nucleosynthesis constraint permits the cluster dark matter to be baryonic. This is possible only if one invokes inhomogeneous cosmological nucleosynthesis, but this scenario should still be taken seriously. It would be remarkable if the Universe came through the quark hadron phase transition with no fluctuations at all; the surprise is that the resulting light element abundance are relatively insensitive to these. On the other hand, it seems clear that even inhomogeneous nucleosynthesis will not permit baryons to have the critical density. A critical baryon density is excluded in most mass ranges anyway. The prospects for the "maximal BDM" scenario therefore seem bleak.

Table 2 Constraints on baryonic dark matter candidates ^a

a The shaded regions are excluded by at least one of the limits discussed in the text and the dotted regions are improbable. SMO, VMO, and BH refer to the black hole remnants of Supermassive Objects, Very Massive Objects, and ordinary stars respectively; WD = white dwarf; MD = M-dwarf; BD = brown dwarf.

The prime message of Table 2 is that one could not expect any single candidate to explain all four dark matter problems. On the other hand, the table does constrain the possible solutions:

1. The local dark matter (if it exists) could be white dwarfs or brown dwarfs but observations of the Population I IMF gives no reason for expecting this; presumably it could not be inos since these are nondissipative and so would not settle into a disk.

2. The halo dark matter could be brown dwarfs, white dwarfs, or VMO black holes; all of these possibilities require a departure from the standard Population II IMF, but the first probably requires the least radical departure since observations may already indicate a preponderance of low mass stars at small metallicity (white dwarfs - although in a sense the most conservative candidate - require cutting the IMF off at both ends).

3. The cluster dark matter may be partly baryonic, especially if galactic halos are baryonic, but we have seen that it could only be dominated by baryonic dark matter if one invokes inhomogeneous cosmological nucleosynthesis.

4. The background dark matter (if it exists) would have to be inos; if the inos are cold, one would expect both the halo and cluster dark matter to be a mixture of WIMPs and MACHOs.

Finally, we should comment on the possibility of Primordial Black Holes (PBHs). Although these are not baryonic (since they form mainly from radiation rather than from gas), they share many of the features of their baryonic counterparts. These could certainly contribute to the dark matter in principle, and those smaller than 10⁵ M (which form before the time of cosmological nucleosynthesis) could have the critical density. However, whether they form from initial inhomogeneities (Hawking 1971, Carr 1975), from phase transitions (Hawking et al 1982, Crawford & Schramm 1982, Dolgov & Silk 1993), or from the collapse of cosmic loops (Polnarev & Zemboricz 1988, Hawking 1989), fine-tuning is required to get an interesting cosmological density because the fraction of the Universe going into PBHs at time t must be ~ 10^-6t^1/2, where t is in seconds. Therefore, despite the invocation of PBHs to explain gravitational lensing effects (Section 7.3), this does not seem very likely.