4.3.2. Binary Galaxies

Binary galaxies, whether they are spiral-spiral or elliptical-spiral pairs, potentially offer the best means for measuring the total masses of galaxies. Clearly, flat rotation curves can never provide a measure of the total mass of a galaxy. Binary galaxies, however, can be thought of as test particles that can be used to extend the rotation curve to large distances. This technique was realized as early as 1937 by Holmberg who first applied it to a sample of binary galaxies to derive rather ambiguous and confusing results. Although the technique has much promise, there are several technical difficulties:

Because of small scale clustering, it is difficult to actually identify from a radial velocity catalog, a sample of binary pairs that are physically bound into a mutual orbit. Many binaries will simply be unbound projections. Treating those cases as if they were true binaries will cause a clear bias.

Small scale clustering of galaxies also means that an individual galaxy may feel the gravitational tug of more than just one nearby galaxy.

Only one projection of the relative velocity between the two galaxies can be measured. Furthermore, the orbits may be highly radial instead of circular and this seriously affects the relation between the measured mean velocity dispersion for binary pairs in some sample and the derived Mass.

Dynamical friction effects are potentially operative in binary galaxies and hence the observed relative velocities may reflect this frictional drag process instead of the total mass.

To date, the literature contains a rather large dispersion of results for binary galaxies. These results are clearly dependent on choice of sample and assumptions about orbits. However, in general, the data are inconsistent with the point mass representation of a galaxy and therefore argue that galaxies must have extended mass distributions.

A significant improvement on the binary galaxy approach has been made by Zaritsky et al. (1996) who have used a well-identified sample of small, satellite galaxies which are in orbit about one isolated large galaxy. The isolation criteria simplifies the dynamical analysis and allows for an ensemble average of all satellites. For a sample of 115 satellites located around 69 spiral host galaxies, Zaritsky et al. derive a characteristic mass of 2 x 10¹² M and halo radius of > 200 kpc for a luminous spiral. The data show no evidence for a decrease in velocity dispersion out to galactocentric radii as large as 400 kpc. To date, this remains the best evidence that spirals are surrounded by very large dark halos. The interesting aspect of the Zaritsky et al. result is the very large halo size which is inferred means that the actual space density of the dark matter in the galactic potential is low, much lower than the stellar density. For instance, a typical spiral galaxy has 2 x 10¹¹ M of stars confined to a disk (cylinder) of radius 15 kpc and thickness 1 kpc. The corresponding density is 0.3 M pc^-3. For a spherical dark matter halo of mass 2 x 10¹² M and radius 200 kpc, the corresponding density is 6 x 10^-5 M pc^-3.