Annu. Rev. Astron. Astrophys. 2004. 42: 603-683 Copyright © 2004 by Annual Reviews. All rights reserved

7. COMPLICATIONS

This section covers issues that complicate the identification of pseudobulges. They do not threaten the conclusion that pseudobulges form by secular evolution of disks; on the contrary, they make it likely that such evolution has operated even in situations where this is not obvious. They highlight areas that need further work.

7.1. Pseudobulges Do Not Have To Be Flat

Several dynamical heating processes are expected to puff pseudobulges up in the axial direction. We are fortunate that some pseudobulges are disky enough so that we can detect the smoking gun that points to a secular origin. It is plausible that others are so similar to classical bulges that they cannot easily be recognized.

One heating mechanism that we have already discussed is bar buckling (Section 4.5). If the density profile along the ridge line of the bar is not too steep, then buckled bars produce box-shaped structures that can easily be recognized when they are seen edge-on.

Resonant vertical heating by the bar may also be important (Pfenniger 1984, 1985; Pfenniger & Norman 1990; Friedli 1999). It is not limited to the relatively few stars that are in resonance at any one time, because the radii of all resonances change as the central concentration increases.

Another heating mechanism involves in-the-plane instabilities that result if the nuclear disk gets too dense for its velocity dispersion. Secular gas inflow is slow compared with the disk's rotation velocity and even the disk's radial velocity dispersion _r. If, as a result, star formation builds up the disk surface density µ without much changing _r, then the disk gets less stable. Toomre (1964) showed that violent instability sets in when Q _r / _crit 1, where _crit = 3.36 Gµ / . Gas inflow increases both the density and the epicyclic frequency, but density wins and _crit increases as the evolution proceeds. As Q drops toward 1, instabilities should form and heat the growing pseudobulge in the disk plane. Toomre (1966) showed further that buckling instabilities heat the disk vertically if _r gets bigger than about 3.3 times the vertical velocity dispersion even if there is no bar. So heating in the plane results in heating perpendicular to the plane. Because the density increases rapidly toward the center, it is unlikely that the result will look box-shaped. Rather, the thickness of the pseudobulge is likely to be larger at smaller radii, much like in a classical bulge.

Finally, at radii that are smaller than the disk thickness, it is no longer relevant that the incoming gas comes from a disk. There is no reason why the innermost parts of a pseudobulge should be flattened at all.

7.2. Pseudobulges Do Not Have To Be Young

Most of this review emphasizes evolution in progress, because this is the easiest way to see that evolution is happening at all. However, we do not mean to create the mistaken impression that pseudobulges must be young or that they must be made of young stars.

Secular evolution is by definition slower than nonequilibrium processes such as mergers, but it can have timescales that are much shorter than a Hubble time. Rates are uncertain, but evolution is thought to be possible on timescales as short as ~ 5 galactic rotations. Bars started to be reasonably abundant at least 5 Gyr ago (Abraham et al. 1999, van den Bergh 2002, van den Bergh et al. 2002). Therefore it is possible that secular evolution built some pseudobulges quickly 5 Gyr ago and then stopped.

Also, heating by bars can elevate disk stars to scale heights characteristic of pseudobulges. These stars can be as old as the oldest disks.

So we expect that pseudobulges have a range of stellar population ages from nearly zero to at least 5 Gyr (e.g., Bouwens, Cayón & Silk 1999). Much older stellar populations are not out of the question; how much older is plausible is not known.

If building a central mass concentration of 5-10% of the disk mass destroys a bar, does secular evolution then stop? Do we already know the maximum bulge-to-disk ratio B / D that secular evolution can produce?

The answer is probably "no": (a) Many SB0 and SBa galaxies have bulge-to-disk ratios of ~ 1. So bars can coexist with surprisingly large central mass concentrations. (b) Once bars grow nonlinear, simple heuristic arguments about how to make a successful bar lose force. In particular, it may no longer be necessary that the radius of ILR be small. (c) The simulations that demonstrate bar suicide do not take into account enough physics. Many do not include gas. Almost none allow the bar to interact with all of the components that we see in galaxies. The competition between angular momentum sinks that help to strengthen bars and the damaging effects of pseudobulge building may be weighted more in favor of the angular momentum sinks than current simulations suggest. (d) Gravitational tickling of an unbarred galaxy with an encounter (but not a merger) may re-excite a bar (Noguchi 1988; Gerin, Combes & Athanassoula 1990; Barnes & Hernquist 1991).