THE RADICAL CONSEQUENCES OF REALISTIC SATELLITE ORBITS FOR THE HEATING AND IMPLIED MERGER HISTORIES OF GALACTIC DISKS

Previous models of galactic disk heating in interactions invoke restrictive assumptions not necessarily valid in modern Delta CDM contexts: that satellites are rigid and orbits are circular, with slow decay over many orbital periods from dynamical friction. This leads to a linear scaling of disk heating with satellite mass: disk heights and velocity dispersions scale proportional to M-sat/M-disk. In turn, observed disk thicknesses present strong constraints on merger histories: the implication for the Milky Way is that <5% of its mass could come from mergers since z similar to 2, in conflict with cosmological predictions. More realistically, satellites merge on nearly radial orbits, and once near the disk, resonant interactions efficiently remove angular momentum while tidal stripping removes mass, leading to rapid merger/destruction in a couple of free-fall plunges. Under these conditions the proper heating efficiency is nonlinear in mass ratio, proportional to(M-sat/M-disk)(2). We derive the scaling of disk scale heights and velocity dispersions as a function of mass ratio and disk gas content in this regime, and show that this accurately describes the results of simulations with appropriate "live'' halos and disks. Under realistic circumstances, we show that disk heating in minor mergers is suppressed by an order of magnitude relative to the expectations of previous analyses. We show that the Milky Way disk could have experienced similar to 5-10 independent 1 : 10 mass ratio mergers since z similar to 2, in agreement with cosmological models. Because the realistic heating rates are nonlinear in mass, the predicted heating is dominated by the more stochastic, rare low mass ratio mergers, and the existence of populations with little or no thick disk does not require fundamental modifications to the cosmology. This also leads to important differences in the predicted isophotal shapes of bulge-disk systems along the Hubble sequence.