Identification of dark matter

Zwicky (1933) was the first to infer the existence of large quantities of unseen mass, in studies of the dynamics of clusters of galaxies, but it was many years before the ``missing mass'', or dark matter, problem was widely recognised. Van den Bergh (1999) reviews the early history of this subject, while Ashman (1992) gives a thorough account of the problem in the context of galaxies. It is, in fact, HI galaxy rotation curves which provide the best evidence we have for the existence of dark matter; here the dynamical interpretation, and its implication, is entirely unambiguous. Equally interesting, though less clear-cut, are the indications that the Universe as a whole is composed principally of dark matter -- this evidence is discussed extensively by Peebles (1993).

Because stars usually contribute the bulk of the visible mass, at least in galaxies, much attention has been given to the possibility that dark matter is composed of stars which are of low luminosity. This category includes: low-mass main-sequence stars; brown dwarfs; old white dwarfs; neutron stars; and black holes -- a diverse group which has required a variety of techniques to constrain their total mass contribution. It is beyond the scope of this paper to summarise these efforts; good reviews are given by Trimble (1987), Ashman (1992) and Carr (1994). However, a unifying feature of this group is that they are all rather dense objects and so constitute strong gravitational lenses even when they are in the halo of our Galaxy. As pointed out by Paczynski (1986), this enables indirect searches for such objects via photometric monitoring of millions of Magellanic cloud stars, and such gravitational lensing events have now clearly been detected (Alcock et al 1997). It is not currently known whether these lensing events are caused by dark matter, or by known stellar populations associated with the Galaxy and the Magellanic Clouds (Sahu 1994).

If it turns out that stars don't fit the bill, what might the dark matter be composed of? The lack of success of very deep searches for previously uncatalogued material, conducted with the most modern instrumentation, across the whole range of the observable spectrum, promoted the suspicion that dark matter has no electromagnetic interaction. In this case one imagines the dark matter to be composed of massive neutrinos, for example, or perhaps an elementary particle which has not yet been detected? This idea was strengthened by detailed computations of the abundances of light elements which result from primordial nucleosynthesis in a hot Big Bang cosmology. These calculations, when compared to the observed abundances, suggest that only a small fraction of the closure density is in the form of baryonic (ordinary) material -- see the review by Schramm & Turner (1998). This picture of weakly interacting dark matter readily lends itself to numerical simulations of the growth of structure (galaxies, clusters, superclusters), from an initially near-homogeneous universe. These simulations demonstrate that dynamically `cold' dark matter can reproduce the observed structural characteristics of the Universe, at least approximately, while simultaneously remaining consistent with measurements of the Cosmic Microwave Background (CMB) anisotropies. `Hot' dark matter gives a much poorer representation of the observed structure (Davis et al 1985), suggesting that the dark matter is not composed of neutrinos, but is a type of particle which has yet to be detected. These results have given impetus to a variety of experiments designed to detect individual dark matter particles, or their decay products -- see the many contributions to Spooner (1997).

Contrary to these general trends in the field, Pfenniger, Combes & Martinet (1994: PCM94 hereafter) advanced the view that the dark matter associated with galaxies might be composed of cold gas clouds. PCM94 motivated this suggestion with arguments related to galaxy dynamics and evolution, emphasising the astrophysical appeal of dark matter in this form, and noting that if the temperature of the cold gas were close to that of the CMB then the clouds would be very difficult to detect (see also Combes & Pfenniger 1997). PCM94 additionally proposed that the cold gas should have a fractal structure - an idea which was developed in Pfenniger & Combes (1994) - and that it should be distributed in a thin disk. However, neither of these are essential features for a dark matter model based on cold gas, and one can equally well imagine a spherical halo of individual or clustered clouds (de Paolis et al 1995; Gerhard & Silk 1996). These quasi-spherical distributions of dark matter are more palatable to most dynamicists than highly flattened distributions, but even so there remain plenty of contentious issues relating to the physics of the putative cold gas clouds; these are addressed in §§4,5; first we turn to the main observational evidence for their existence.

© Copyright Astronomical Society of Australia 1997