The Magellanic Clouds are a natural laboratory for investigating the evolution of stellar populations in dynamically interacting systems. Their populations are well studied and provide a basis for standard candle and population evolution studies. In addition, a variety of dynamical studies and simulations exploit the Milky Way-Clouds-Magellanic Stream interaction to both infer its complex history and constrain Milky Way mass models (Fujimoto & Sofue 1976, Lin & Lynden-Bell 1977, Davies & Wright 1977, Lin & Lynden-Bell 1982, Murai & Fujimoto 1986, Lin et al. 1995; see Lin et al. for a thorough historical discussion). The link between chemical evolution and externally driven structural evolution is a hard problem, and to date, there has been little work devoted to a self-consistent dynamical picture of Cloud evolution. To this end, this paper focuses on one specific aspect--the dynamical interaction between the Milky Way and the Large Magellanic Cloud--and ignores the likelihood of a significant interaction with the Small Magellanic Cloud (SMC) in the past or the possibility that the SMC originated in a tidal disruption event. This simplified scenario by itself admits a rich set of interacting mechanisms.
Recent work by myself and others points out that time-dependent tidal forcing can have significant evolutionary consequences for globular clusters and dwarf or cannibalized galaxies (Chernoff et al. 1986, Aguilar et al. 1988, Weinberg 1994c, Gnedin & Ostriker 1997, Murali & Weinberg 1997ab, Vesperini 1997, Weinberg 1997). The same physics applies to non-isotropic distributions such as disks or disks embedded in halos. For example, Sellwood et al. (1998) explored the importance of these resonant mechanisms to thickening host disks by dwarfs and excitation of bending waves. Weinberg (1998, Paper I) found that the resonant interaction between the Milky Way and LMC is sufficient to excite a warp and cause lopsided asymmetries, depending on the Galactic halo potential and LMC mass. In that work, the LMC was structurally fixed. Here, we turn the tables by structurally fixing the Milky Way and applying the same physics to LMC evolution.
It is straightforward to see that the magnitude of such a disturbance to the LMC is large. Either the subtended size of the LMC on the sky or the rotation curve combined with an estimate of the Milky Way mass enclosed in the LMC orbit leads to a tidal radius of approximately 11 kpc. At 5 kpc from the center, the ratio of the tidal force to the self force has only dropped to approximately 20% assuming a flat rotation curve, a significant perturbation. Although this ratio drops quickly further inward, the effect of the tidal force is amplified by a spectrum of resonances between the LMC-Milky Way orbital frequencies and internal LMC orbital frequencies. Simultaneously, the LMC disk axis precesses due to the coupling with its orbit about the Galaxy. This induces an additional interaction between the LMC disk and halo. Altogether, these mechanisms result in enhanced angular momentum and energy transfer between the orbit and internal motions. They thicken the disk, populate the spheroid and drive mass loss. The latter mechanism has direct analogy to globular cluster evolution (see refs. cited above).
This paper explores this basic picture as follows. First, §2 summarizes the inference of the LMC orbit and mass needed to estimate the time-dependent tidal force. We will then explore the underlying dynamical interaction in several steps. Analytically, effectively irreversible changes in energy and angular momentum occur at resonances between the frequencies of the applied perturbation and the stellar orbital frequencies. First, I will describe the results of a restricted computation which sums the effect of all the resonances directly. This idealized model treats the evolution of a disk without self-gravity in a fixed halo potential (§3.1). We find that the disk is notably heated and thickened in several gigayears. Although the omission of self-gravity surely leads to an overestimate of the heating, the simple model serves to illustrate the potential importance of the underlying physical mechanisms.
This example is followed up in §3.2 with a full n-body
simulation using the force algorithm described in Weinberg
(1999). This code uses a basis expansion
tailored to the density profile and is well-suited to following a
slowly changing system. The collisionless evolution is
gravitationally self-consistent with the caveat that the Galactic mass
model and LMC orbit remains fixed. The n-body results substantiate
the simple restricted example in §3.1, although the rate
of disk thickening is smaller due to the self-gravity of the disk.
Specifically, the simulations predict a thickening rate of 70 pc/Gyr
at a roughly constant rate over the duration (about 4 Gyr). The tail
of the torqued distribution populates the LMC halo region. At the
same time, the energy input does work on the potential and causes
overall expansion of the disk. This offsets the increase in velocity
dispersion that might be observed from heating in a fixed potential;
in fact, expansion wins and the velocity dispersion observed at a
inclination at one disk scale length is very slowly
dropping.
These results lead to a number of interesting predictions and implications. First, the stellar component should be as extended as the halo. Moreover, this is done without isotropizing the distribution since a modest change in direction of the orbital plane (and therefore its angular momentum vector) is relatively easy. In fact, Olszewski et al. (1996) outline the evidence that nearly all components of the LMC have disk-like kinematics regardless of their extent. Second, the dark halo and the kinematically evolved and extended stellar component are preferentially stripped. The stripped material continues to orbit with the LMC and slowly spread in phase. Because the stripped stars are not part of the disk, they do not lie along the HI-defined Magellanic Stream but rather in a much more diffuse distribution. In other words, this scenario does not suggest looking in the gas stream for the stars. Finally, a thickened bound stellar component in the LMC and an extended unbound cloud surrounding the LMC will increase the rate of self microlensing and we will estimate the effect in §4. A final discussion with implications for the LMC and satellite systems in general is presented in §5 followed by a summary of results in §6.