5. WORKING MODEL

The most likely model for the origin of clumps in clumpy galaxies is that they form by gravitational instabilities in rapidly assembled disks. The clumps are confined to within 100 pc of the mean disk, they are young star-forming regions (not diverse merged galaxies), the clump masses are 10⁷ M - 10⁸ M, sometimes 10⁹ M, and these masses appear to be the ISM Jeans masses with the measured turbulent speeds and gas column densities. For example, M_Jeans ~ ⁴ / G² ~ 10⁸ M if ~ 30-50 km s^-1 and _gas ~ 100 M pc^-2. These dispersions are consistent with observed HII dispersions (Förster-Schreiber et al. 2006, Förster Schreiber et al. 2009, Weiner et al. 2006, Genzel et al. 2006, Genzel et al. 2008, Puech et al. 2007, Law et al. 2009) and this column density is typical for the inner disk regions of spiral galaxies today. It is also comparable to what Tacconi et al. (2010) observed directly using CO emission.

There are many consequences of having such large clumps in a galaxy disk (Noguchi 1999, Immeli et al. 2004a, Immeli et al. 2004b, Bournaud, Elmegreen & Elmegreen 2007). They contribute strongly to the total disk potential, so they interact gravitationally, experience strong dynamical friction, and lose angular momentum to the outer disk. This all causes them to migrate rather quickly to the disk center where they contribute to a growing bulge (Elmegreen et al. 2008a). Star formation in the center can get triggered by their merger too, and this adds to the bulge. At the same time, their disruption in the disk causes it to smooth out, and this, combined with their angular momentum transfer, gives the disk an exponential radial profile (Bournaud, Elmegreen & Elmegreen 2007). All of this disk evolution can happen within 0.5-1 Gyr.

Stirring from the clumps also thickens the disk and this probably produces the thick disk component of today's spiral galaxies (Bournaud, Elmegreen & Martig 2009). Thick disks can also form by minor mergers, both through the stirring of existing disk stars and the dispersal of the merger remnants (Quinn et al. 1993, Walker et al. 1996). However, thick disks formed in this way flare out at the edge, and real thick disks do not seem to do this (Yoachim & Dalcanton 2006, Bournaud & Elmegreen 2009). Stirring by internal processes automatically makes a thick disk with an approximately constant scale height, because the stirring force from clump gravity is proportional to the disk restoring force from gravity. In the case of a merger, the stirring force is proportional to the companion galaxy mass and independent of disk restoring force, so the disk is dispersed much further in the outer regions than the inner regions (Bournaud & Elmegreen 2009).

There is also a possible connection with nuclear black holes if the dense clusters that are likely to be present in the cores of the individual disk clumps form intermediate mass black holes by stellar coalescence, as proposed for dense clusters by Ebisuzaki et al. (2001), Portegies-Zwart & McMillan (2002), and others. If the clumps form black holes in this way, then these black holes will migrate into the disk center along with the clumps, and possibly merge to make a massive nuclear black hole. Simulations of this process obtain a correlation between the black hole mass and the bulge velocity dispersion that is similar to what is observed (Elmegreen et al. 2008b). If the clumps make globular clusters too (Shapiro et al. 2010), then the correlation between globular cluster number and central black hole mass (Burkert & Tremaine 2010) might be explained in the same way.