Formation of Gaseous Shells

HI observations have revealed in several shell galaxies the presence of gaseous shells slightly displaced from the stellar shells radially, in the outward direction. We propose a mechanism to form this gaseous shells, based on the well-known phase-wrapping process of the companion matter in a merger, with nearly radial orbits. The mechanism relies on the existence of a clumpy interstellar matter, and on dynamical friction experienced by the companion core. Shells are sharp-edged features, formed during interactions and mergers, through phase-wrapping of debris (Quinn 1984, Dupraz & Combes 1986, Hernquist & Quinn 1989). Recent HI observations have revealed the existence of associated gaseous shells, slightly displaced from the stellar ones, questioning the validity of the phase-wrapping mechanism result is that the HI shells follow the curvature of the stellar shells, but are shifted about 1kpc outside. There are two ways shells can be formed:-(1)-in minor mergers, shells correspond to phase-wrapping of the stars liberated from the small companion (e.g. Quinn 1984);-(2)-in major mergers, shells correspond to phase-wrapping of the debris falling back into the merged-object potential (Hernquist & Spergel 1992, Hibbard & Mihos 1995). But what is the fate of gas? due to dissipation, it falls towards the center, as in the simulations of Weil & Hernquist (1993), and there should not exist gaseous shells. There exists a population of small and dense gas clouds, that have very low dissipation. This gas has a behaviour intermediate between stars and diffuse gas, and remains available to form shells. Already Kojima & Noguchi (1997) have simulated the sinking of a disk satellite into an elliptical, with a sticky particle code, instead of SPH, and found no segregation between gas and stars. We have also simulated the phenomenon, with a cloud-cloud collision code, to be able to control the dissipation rate. With strong dissipation, the gas component, after a few oscillations back and forth in the primary's potential, settles in the center, as previously. But with small dissipation, only a small fraction of the gas falls into the potential well, most of it form shells (cf figure 1).