10.1 Non-Baryonic Dark Matter

Much of the interest in non-baryonic DM candidates was sparked by theoretical arguments that favor a universe with the critical density (e.g. Guth 1981; Steinhardt 1982; Peebles 1988; Schramm 1991). Despite attempts to modify standard nucleosynthesis calculations by including inhomogeneities and diffusion processes in the early universe (Applegate and Hogan 1985; Alcock, Fuller and Mathews 1987; Malaney and Fowler 1988), it has proven difficult to obtain _b = 1 and still satisfy constraints from light element abundances (e.g. Hogan 1990; Malaney and Mathews 1992). Thus a universe with = 1 requires non-baryonic DM, which raises the possibility that dark halos are made of such material.

Non-baryonic candidates are usually classified as either hot or cold (cf. Bond and Szalay 1983). In a universe dominated by hot DM, clusters of galaxies form first, whereas cold DM universes form galaxies by hierarchical clustering of smaller units. (``Seeded'' hot DM models, however, produce galaxies before clusters; see, e.g., Turok [1991] and Cen at al. [1991]). Pure hot DM has fallen from favor for a number of reasons, such as evidence that galaxies form before clusters and apparent disagreements with observations of large-scale structure. Cold DM also has problems, but its popularity has proved more resilient.

Constraints on non-baryonic DM candidates are obtained in two ways: direct limits from laboratory experiments, and indirect astrophysical constraints. Information is also obtained as a spin-off from some accelerator experiments. Laboratory experiments are often aimed at finding specific candidates, but in practice they tend to constrain the parameter space of hypothetical DM particles. Recent discussions on the particle physics motivations for non-baryonic DM candiates, as well as the experimental techniques and results in this field, have been given by Krauss (1990; 1991), Primack, Seckel and Sadoulet (1989) and Turner (1992) amongst others (see also papers in the proceedings edited by Audouze and Tran Thanh Van [1988] and Holt, Bennet and Trimble [1991]).

Many of the cold DM candidates fall under the general category of WIMPs (Weakly Interacting Massive Particles). Results on the Z boson resonance place the most stringent limits on the properties of WIMPs. Combined with direct search experiments, they rule out the heavy neutrino as a viable cold DM candidate (Krauss 1991). Apparently, these limits also rule out the sneutrino and make life difficult but not impossible for the neutralino. In other words, certain WIMPs have been excluded, whereas others remain possible candidates. Importantly, Krauss (1991) notes that these constraints demand that WIMPs have properties that will make then considerably harder to detect than earlier ideas suggested.

One astrophysical technique of detecting WIMPs is by detecting -rays or other photons that may be produced when WIMPs annihilate. The flux from a smooth halo is very low, but Lake (1990c) pointed out that if the DM has a lumpy distribution, as is probable in many cosmological models, then the signal may be detectable. Annihilations in the Earth also produce detectable signals and may prove a better way of constraining WIMP properties (Kamionkowski 1991).

One of the most popular cold DM candidates is the axion. The upper bound on its mass is about 10^-3 eV, based on observations of neutrinos from 1987A (e.g. Raffelt and Seckel 1988). The lower bound used to be around 10^-5 eV, but there has been some recent discussion that this may have to be raised (Davis 1986; Wilczek 1992). It is therefore possible that axions may be ruled out, but there is little consensus at the moment on the lower mass limit.

Theorists are clearly having their choices of cold DM candidates trimmed by such experiments. Hot DM is having even more trouble, primarily because of astrophysical constraints. The high DM densities inferred in some dwarf spheroidal galaxies such as Draco and Ursa Minor are particularly useful in this regard. I noted in Section 4 that the central density in these objects is not well-determined, but that there are firm lower limits based on the virial theorem. Tremaine and Gunn (1979) showed that considerations of the phase-space density of dark halos provides a lower limit to the mass of a hot DM particle such as the neutrino. (This is a ``light'' neutrino rather than the cold DM heavy neutrino mentioned above.)

For the neutrino or similar particle to contribute a significant DM density, it must have a mass around 30 eV. A greater mass leads to a density in excess of the critical density required to close the Universe. If such particles constitute the DM in an isothermal dark halo, the neutrino mass must satisfy

(10.3)

(e.g. Spergel 1991) where r_c is the core radius of the DM halo, ₀ is the central DM density, and g is a parameter of order unity.

If the halos around Draco and Ursa Minor are made of neutrinos, equation (10.3) and current limits on ₀ require a core radius 100 times that of the stellar distribution or an enormous dark mass within the core (Gerhard and Spergel 1992). By considering the time scale for dynamical friction to drag such objects into the Milky Way, Gerhard and Spergel (1992) conclude that, if the halos of Draco and Ursa Minor are made of neutrinos, then m 80 eV. Such a mass is well above the value that produces closure density.

There have been attempts to circumvent these limits using highly anisotropic distribution functions for the neutrinos. Gerhard and Spergel (1992) argue that such distributions are unstable, so that the limit of Tremaine and Gunn (1979) is still valid. This appears to rule out neutrinos as viable candidates for the DM in the halos of Draco and Ursa Minor. However, the high central densities of these objects also require that the halos virialized at very early epochs, at least if the DM is dissipationless. Indeed, their formation would have occurred at higher redshifts than expected for universes dominated by cold DM (Lake 1990b). If one wants to overcome the early formation epoch, a solution is to suppose that DM in these two dwarfs dissipated, which would require it to be baryonic (Kormendy 1990; Lake 1990b). Alternatively, as mentioned in Section 4, there have been suggestions that there is no DM in dwarf spheroidals.

A very specific non-baryonic DM candidate is a neutrino-like particle that decays to produce hydrogen-ionizing photons. Sciama (1988) showed that such photons could explain the ionization of Ly- clouds and the IGM at high redshifts. Sciama (1990a) also showed that if such photons were responsible for the observed ionization of HI clouds in the Milky Way, then the characteristics of the decay were tightly constrained. In particular, the particle would decay into a photon with an energy close to 14 eV. This picture also required a fairly flattened halo for the Galaxy (Sciama 1990b). Salucci and Sciama (1990) showed that such a mass model agreed well with observations of the Milky Way rotation curve.

Davidsen et al. (1991) used the Hopkins Ultraviolet Telescope to search for the predicted spectral signature in the cluster A655. If the DM in this cluster was comprised of the decaying DM particles, then the 14 eV line was well within the sensitivity of the instrument. However, the line was not detected. A non-detection was also reported by Fabian, Naylor and Sciama (1991) from IUE observations of the quasar 3C 263 which lies at the center of a moderately rich cluster. However, Fabian et al. (1991) suggested that the spectral line could be absorbed by cold gas clouds such as those found in some nearby clusters (White et al. 1991). Another argument to rehabilitate the model is that the DM in the center of clusters is partly baryonic (as suggested by results described below), so that the strength of the 14 eV line is less than predictions which assume all the cluster DM is comprised of decaying particles.