1994A&A...290..709S

The accretion of a small satellite by a big elliptical galaxy is acknowledged as a key phenomenon in the dynamics of galaxies. It is held responsible for the formation of most fine structures which are observed in ellipticals. Circular orbits for the satellite have been studied in most work so far. However, they may not be representative of the general phenomenon, as their geometry prevents the process from being time-dependent in a genuine way. Here is the first of two articles devoted to the dynamical study of orbital decay during a head-on encounter. Analytical calculations give access to the mechanism of dynamical friction during the event. They consist in the coupled solution of the linearised Poisson and collisionless Boltzmann equations. Our results confirm that time now plays an active role: the drag force undergone by the satellite derives from the transient response of the galaxy and depends on the full past history of the event. As a consequence, dynamical friction is not strictly proportional to the local density as classically expected from Chandrasekhar's formula. Our study is refined by a harmonic analysis of the process. We extend these results by means of restricted three-body simulations. They allow us to deal with a compact satellite and to follow the accretion until merging with the galaxy. We adopt the Multiple Three-Body Algorithm, which naturally accounts for dynamical friction and proves equivalent to the analytical method inasmuch as self-gravity of the galactic wake is unimportant. A satellite along a radial trajectory is braked in a step-like way: its orbital energy is essentially dissipated when it crosses the galactic core. We discuss the dependence of this process on the parameters of the collision: initial energy, mass and radius of the intruder. The size is most important, because it determines both the nature of the galactic response and the ultimate fate of the object. However, we do not properly describe the strong distortion of the inner galaxy when hit by the intruder, nor do we include the disruption of the satellite in the tidal field of the elliptical. These issues will be discussed in a forthcoming article (Seguin & Dupraz 1994: Paper II) in the light of self-consistent N-body simulations and within a broader study of eccentric collisions between the galaxy and the satellite.