6.5 Dust in Spirals
Although it is usually assumed that the disks of spiral galaxies are optically thin, there have been some suggestions that dust in spirals might obscure much of the stellar light (e.g. Disney, Davies and Phillipps 1989; Davies 1990; Valentijn 1990). This would have major implications for the deduced DM fraction in spirals. Valentijn (1990) in particular has claimed that spiral disks are opaque, so that their stellar masses are higher than usual estimates. He further argues that the exponential scale-length of the absorbing material is greater than that of the stellar disk, so that the mass-to-light ratio of spirals increases with radius.
Gonzalez-Serrano and Valentijn (1991) studied the dwarf spiral UGC 2259 and found that the rotation curve could be fit with a purely disk-like mass distribution, provided absorbing material was obscuring much of the starlight. They also showed that a similar model could reproduce the rotation curve of the well-studied galaxy NGC 3198, so that a dark halo was unnecessary.
While such models can in principle explain the rotation curves of spirals, they contain several problems. As van der Kruit (1992) has emphasized, galaxies are often seen through the disks of foreground spirals, and reddened globular clusters are observed through the disk of M31. In the Milky Way, our ability to observe both North and South of the plane argues against an opaque disk. Burstein, Haynes and Faber (1991) and Freeman (DMW) have studied spiral galaxy samples for dust obscuration, but find that the effect is smaller than claimed by Valentijn (1990). Moreover, Burstein et al. (1991) have pointed out that dusty spirals have similar (M/LV) to dust-free ones. This is because the presence of dust both dims and reddens the underlying galaxy. Correcting for the dimming pushes the galaxy luminosity up so that the stellar mass is greater, but correcting for the reddening implies the true stellar population has a lower mass-to-light ratio. These two effects roughly cancel each other out, implying that the usual estimates of the disk mass are reasonable and that DM is required to explain most rotation curves.
It is worth pointing out that there are various arguments that suggest the dominant mass component in spirals is considerably rounder than an exponential disk (see Section 9 below). Thus while Gonzalez-Serrano and Valentijn's (1991) model can reproduce rotation curves, other considerations probably rule out the idea. The presence of dark halos around ellipticals also contradicts the opaque disk model, if the popular notion that such galaxies form by the merger of spirals is correct.