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

10. CONCLUSIONS AND FUTURE PROSPECTS

10.1. Reappraisal of Baryonic Dark Matter Candidates

By way of summarizing the key points of this review, and also because it provides an opportunity to mention some candidates that have not yet been covered, we conclude with a reappraisal of the various baryonic dark matter candidates (cf Carr 1990, Dalcanton et al 1994).

SNOWBALLS     Condensations of cold hydrogen can be excluded in most mass ranges. In order to avoid being disrupted by collisions within the age of the Universe, they must have a mass of at least 1 g (Hegyi & Olive 1983, Wollman 1992). Constraints in the mass range above this have been discussed by Hills (1986): Snowballs are excluded by the upper limit on the frequency of encounters with interstellar meteors between 10^-3g and 10⁷g, by the number of impact craters on the Moon between 10⁷g and 10¹⁶g, and by the fact that no interstellar comet has crossed the Earth's orbit in the last 400 years between 10¹⁵g and 10²²g. The limits are marginally stronger for halo objects, because of their larger velocities, and are shown in Figure 3. Hegyi & Olive (1983) have argued that snowballs would be evaporated by the microwave background, but Phinney (1985) has pointed out that this only happens below a mass of 10²²g. De Rujula et al (1992) have claimed an even stronger limit on the grounds that snowballs smaller than 10^-7 M ~ 10²⁶g would be evaporated within the age of the Universe by their own heat; this is also indicated in Figure 3. Another argument against snowballs is that, since one would expect only hydrogen to condense, the cosmic helium abundance would be increased to an unacceptably high value if the fraction of the Universe going into them were more than (1 - Y_min / Y_max), where Y_min is the minimum primordial abundance ( 0.2) and Y_max is the maximum presolar helium abundance ( 0.3). This suggests that the fraction must be less than 30%.

BROWN DWARFS Fragmentation could in principle lead to objects smaller than 0.08 M and there may be evidence that such brown dwarfs form prolifically in cooling flows (Section 9.1). Such objects might be detectable as infrared sources; it is not surprising that IRAS has not found them but ISO or SIRTF could be expected to detect brown dwarfs with masses down to 0.01 M (Section 9.5). Another important signature of brown dwarfs, in either our own or other galactic halos, is the intensity fluctuations in stars or quasars induced by their microlensing effects. This effect would be observable for objects over the entire brown dwarf mass range and may have already been found (Sections 7.2, 7.3, and 7.7). Observations of microlensing on different timescales could also give information about the mass spectrum of the brown dwarfs (De Rujula et al 1991). The brown dwarf scenario currently appears to be the most plausible. In any case, the combination of infrared and microlensing searches should soon either confirm or eliminate it.

M-DWARFS     Stars in the range 0.3-0.8 M are excluded from solving any of the dark matter problems by background light limits (Section 5.1). Lower mass hydrogen-burning stars would also seem to be excluded by source count constraints and infrared measurements of other galaxy halos (Section 9.3).

WHITE DWARFS     These would be the natural end-state of stars with initial mass in the range of 0.8-8 M and they could certainly fade below detectability if they formed sufficiently early in the history of the Galaxy. The fraction of the original star that is left in the white dwarf remnant is low but one could still produce a lot of dark matter if there were many generations of stars (Larson 1986). In some sense white dwarfs are the most conservative candidates, since we know that they form prolifically today. The problem is that one needs a very contrived mass spectrum if they are presumed to make up galactic halos: The IMF must be restricted to between 2 and 8 M to avoid producing too much light or too many metals (Ryu et al 1990) and even then one must worry about excessive helium production (Section 5.3). However, this scenario would have many interesting observational consequences, such as an abundance of cool white dwarfs (Tamanaha et al 1990) and a large number of X-ray sources formed from white dwarf binaries which have coalesced into neutron stars (Silk 1993). A potential problem is that the fraction of white dwarfs in binaries might produce too many type 1a supernovae (Smecker & Wyse 1991), although this might actually be required to explain the high-velocity pulsars moving towards the disk in our own Galaxy (Eichler & Silk 1992). Even if white dwarfs do not have a high enough density to explain the halo dark matter, they could still explain the dark matter in the Galactic disk (if this exists).

NEUTRON STARS     Although neutron stars would be the natural end-state of stars in some mass range above 8 M, the fact that the poorest Population I stars have metallicity of order 10^-3 places an upper limit on the fraction of the Universe's mass that can have been processed through the stellar precursors - this probably precludes their explaining any of the dark matter problems (Section 5.2). The only way out is to adopt the proposal of Wasserman & Salpeter (1993) in which the neutron stars arc in clusters, so that their nucleosynthetic products are trapped within the cluster potentials. Even in this scenario, the neutron stars contain only 1% of the halo dark matter; most of the mass is in asteroids. Nevertheless, the small admixture of neutron stars has an intriguing consequence since collisions between the neutron stars and asteroids are supposed to explain gamma-ray bursts.

STELLAR BLACK HOLES     Stars larger than some critical mass M_BH 25-50 M may leave black hole rather than neutron star remnants, with most of their nucleosynthetic products being swallowed. However, they will still return a substantial amount of heavy elements through winds prior to collapsing (Maeder 1992), so normal stellar black holes are probably excluded. In any case, stellar black holes could not provide the disk dark matter because the survival of binaries in the disk requires that the local dark objects are smaller than 2 M (Section 6.5). Stellar black holes could also be detected by their lensing effects on the line-to-continuum ratio of quasars; this already excludes black holes from having a critical density below 300 M or a tenth critical density (required for halos) below 20 M (Section 7.5).

VMO BLACK HOLES     Since stars larger than some critical mass M_c 200 M undergo complete collapse, they may be better candidates for the dark matter than ordinary stars. However, VMOs are radiation-dominated and therefore unstable to pulsations; these pulsations are unlikely to be completely disruptive, but they could lead to considerable mass loss and possible overproduction of helium (Section 5.3). Another important constraint on the number of stellar black hole remnants is provided by background light limits. Although these can be obviated if the stars burn at a sufficiently high redshift, the scenario is becoming increasingly squeezed by the FIRAS data (Section 8.2). However, VMO black holes are relatively unconstrained by lensing effects, since the line-to-continuum constraint only applies below 300 M (Section 7.5). Laser interferometry might just detect the gravitational wave background generated by a large population of VMO black holes, especially if they form in binary systems (Section 8.4).

SUPERMASSIVE BLACK HOLES     We have seen that SMOs larger than 10⁵ M would collapse directly to black holes without any nuclear burning due to relativistic instabilities. However, halo black holes would heat up the disk stars more than is observed unless they were smaller than about 10⁶ M (Section 6.1), so they would have to lie in the narrow mass range 10⁵-10⁶ M, and the survival of globular clusters (Section 6.2) and dynamical friction effects (Section 6.3) probably exclude even this range. If the dark matter in clusters comprises black holes, then the absence of unexplained tidal distortions in the visible galaxies implies that they must be smaller than 10⁹ M (Section 6.5). The number of SMO black holes is also constrained by macrolensing searches: Their density parameter must be less than 0.4 between 10⁷ and 10⁹ M and less than 0.02 between 10¹¹ and 10¹³ M (Section 7.2). The background gravitational waves generated by the formation of SMO black holes could in principle be detected by space interferometers or the Doppler tracking of interplanetary spacecraft (Section 8.4).