Boxy/Peanut Galaxies Formation Scenarios

Various scenarios have been proposed to explain the nature and formation of boxy/peanut-bulge spiral galaxies. Among other possibilities, May, van Albada, & Norman (1985) showed that the action of an external torque on a spheroidal system can produce a box-shaped bulge. We will concentrate here on two other mechanisms which have their own problems when faced with observations, but appear promising at the moment. We will describe those two mechanisms in the rest of this section.

Accretion Mechanism

The first scenario to form boxy/peanut bulges is through the accretion of material onto a host spiral galaxy. Binney & Petrou (1985) have shown that it is possible to generate cylindrically rotating boxy/peanut bulges using a distribution function (specifying the phase-space density) approach. In addition to the dependence of the distribution function on the two integrals of motion E (energy) and (angular momentum around the symmetry axis), orbits reaching a given height above the plane are favoured through a third integral, giving rise to the boxy/peanut shape when the bulge is seen in projection. Cylindrical rotation follows naturally. Rowley (1986, 1988) later showed that for a two-integral distribution function, a truncation depending on both E and instead of a conventional high energy cut-off also leads to boxy/peanut cylindrically rotating systems.

Binney & Petrou (1985) argued that the best way to form the required distribution function is through slow accretion of material. If the orbit of a satellite galaxy with velocity dispersion much lower than its orbital speed decays toward a host galaxy with a decay time-scale much longer than the orbital time, the material shed by the satellite will naturally give rise to the required boxy/peanut bulge shape.

Clearly, a whole range of such scenarios is possible, from the accretion of one or more small satellites galaxies to the merger of two galaxies of similar size. However, accretion scenarios face several problems. Firstly, in order to form a boxy/peanut-shaped bulge, the accreted material must have a particular distribution of orbital energy and angular momentum. It seems unlikely that several small accreted satellite galaxies would all satisfy this criterion. The remaining satellites should still be visible, but they are not (Shaw 1987). Secondly, in the case of a large companion, the increased velocity dispersion and smaller decay time preclude the necessary clustering of the accreted material in phase-space, and the end result of the accretion won't be a boxy/peanut-shaped bulge (Binney & Petrou 1985). Thirdly, we believe the merger of two spiral galaxies (or one spiral and one small elliptical) can not account for the large number of boxy/peanut-bulge galaxies seen: such mergers would require fairly precise alignment of the spin and orbital angular momenta of the two galaxies, and this seems an unlikely route for the production of about 1/3 of spiral bulges.

It seems that the only viable accretion scenario left to form boxy/peanut bulges is through the accretion of a small number of moderate size companions. In fact, while boxy/peanut-bulge spirals are not found preferentially in clusters (Shaw 1987), the best examples of this class of object are seen in small groups (e.g. NGC 128). Therefore, while accretion is unlikely to be the primary formation mechanism of boxy/peanut galaxies, it is likely that it plays a role in some cases. This view is supported by the fact that the related X-shaped galaxies can be formed through the accretion of a satellite galaxy (Whitmore & Bell 1988; Mihos et al. 1995). Consequently, observers should look for evidence of accretion in boxy/peanut-bulge galaxies, and part of our observational program does that (see §3). On the other hand, we should note that Shaw (1993) did not detect any arcs, shells, or filaments optically around boxy/peanut bulge spirals. This argues against any kind of recent accretion or merger models.

Bar Buckling Mechanism

The second and currently fashionable scenario proposed to form peanut/boxy bulges is through the buckling and thickening of a bar in a strongly barred spiral galaxy. The buckling or fire-hose instability was first considered by Toomre (1966) in an idealised model: basically, if the vertical velocity dispersion in a disk is less than about a third of the velocity dispersion in the plane, buckling modes will develop. Toomre's (1966) results were confirmed by many authors (e.g. Fridman & Polyachenko 1984; Araki 1985).

Combes & Sanders (1981) were the first to associate the formation of boxy/peanut-shaped bulges in spiral galaxies with the thickening of a bar in 3D N-body simulations. Many groups have subsequently reproduced and developed those results (e.g. Combes et al. 1990; Raha et al.\ 1991). Basically, after the bar develops, it thickens, and when looking at the system edge-on, the bar appears boxy when seen end-on and peanut-shaped when seen side-on. Although the thickening of a bar can be caused by instabilities associated with resonances between the bar motion and the vertical oscillations of the stars (e.g. Combes & Sanders 1981; Combes et al. 1990), this effect is probably not dominant, and the primary cause of the thickening is more likely to be the buckling instability itself. Bar formation makes the orbits in the bar more eccentric and aligns their principal axes without greatly affecting the motion perpendicular to the plane. The bar then becomes unstable to buckling modes, buckles, and settles with an increased vertical velocity dispersion and thickness (e.g. Raha et al. 1991). Notwithstanding how the bar thickened, the final boxy/peanut shape is probably due to ``orbit-trapping'' around the 2:2:1 (banana and anti-banana) periodic orbit family (Pfenniger & Friedli 1991; Raha 1992; for the notation and an excellent review on the subject, see Sellwood & Wilkinson 1993).

The bar buckling scenario to form boxy/peanut-bulge spirals is attractive for many reasons. Firstly, the fraction of boxy/peanut-bulge systems among edge-on spirals is similar to the fraction of strongly barred spiral galaxies observed in more face-on systems, at least for early-type spirals (30%, Shaw 1987). The statistics for later types are not as conclusive, probably because of difficulties in classifying their generally smaller bulges. Secondly, N-body simulations of buckling bars reproduce the cylindrical rotation seen in boxy/peanut bulges. Thirdly, with the buckling instability, even isolated galaxies can develop a boxy/peanut-shaped bulge. This is in agreement with the fact mentioned that boxy/peanut-bulge spirals are not found preferentially in clusters (Shaw 1987).

The mechanism discussed here also has its drawbacks. The main problem faced by the bar buckling mechanism is that the observed boxy/peanut bulges are usually shorter with respect to the scale of the host galaxy than are the strong bars seen in simulations, although no proper statistics have been compiled. Also, interactions between the bar, the disk (through spiral arm activity), and the bulge and dark halo components (through dynamical friction) might affect the evolution of the bar. Particularly, transfer of angular momentum from the bar to a spheroidal component can be very efficient (Sellwood 1980; Weinberg 1985). Recent results by Debattista & Sellwood (1997) have shown that massive halos with a high central density can slow down bars very effectively, although a number of effects that could decrease this efficiency have been suggested (Sellwood & Debattista 1997). It is interesting to note though that the bar keeps growing during the process, and the boxy/peanut appearance isn't affected. In addition, many ideas have emerged recently about how a bar can be destroyed (e.g. very violent buckling instability, interaction with a massive companion), but the presence or growth (due to bar-driven material) of a central mass concentration is the most promising mechanism. A relatively small mass at the center of the bar can rapidly destroy the bar by affecting the orbit families associated with it (see e.g. Hasan & Norman 1990; Norman, Sellwood, & Hasan 1996).

Associated with those questions is the issue of the stability of spiral disks to bar formation. Most simulations start with an equilibrium disk unstable to bar formation. But, that instability is so quick to act that it is uncertain whether such a disk would have formed in the first place. Recent simulations by Mihos et al. (1995) suggest interactions and mergers as a possible way to excite strong bar modes in a dynamically stable disk. The bar can then buckle and give rise to a boxy/peanut-shaped bulge in the usual manner. This scenario couples the accretion and bar buckling mechanisms.

Despite the problems mentioned, the formation of boxy/peanut-shaped bulge spiral galaxies through the buckling of a strong bar in a disk is the mechanism favoured at the moment. It offers a natural and efficient way to form boxy/peanut bulges, and the qualitative predictions of the models agree with what is known at the moment about this class of object. In addition, this scenario offers numerous opportunities for observers to test its validity. A large part of the observational programs described in §3 is related to this hypothesis, and the new spectroscopy results presented in §4 are aimed directly at testing it.

© Copyright Astronomical Society of Australia 1997