Destruction of the Galactic Globular Cluster System

We investigate the dynamical evolution of the Galactic globular cluster system in considerably greater detail than has been done hitherto, finding that destruction rates are significantly larger than given by previous estimates. The general scheme (but not the detailed implementation) follows Aguilar, Hut, & Ostriker.

For the evolution of individual clusters, we use a Fokker-Planck code including the most important physical processes governing the evolution: two-body relaxation, tidal truncation of clusters, compressive gravitational shocks while clusters pass through the Galactic disk, and tidal shocks due to passage close to the bulge. Gravitational shocks are treated comprehensively, using a recent result by Kundić & Ostriker that the <ΔE²> shock-induced relaxation term, driving an additional dispersion of energies, is generally more important than the usual energy shift term <ΔE>. Various functional forms of the correction factor are adopted to allow for the adiabatic conservation of stellar actions in a presence of transient gravitational perturbation.

We use a recent compilation of the globular cluster positional and structural parameters, and a collection of radial velocity measurements. Two transverse to the line-of-sight velocity components were assigned randomly according to the two kinematic models for the cluster system (following the method of Aguilar, Hut, & Ostriker): one with an isotropic peculiar velocity distribution, corresponding to the present-day cluster population, and the other with the radially preferred peculiar velocities, similar to those of the stellar halo. We use the Ostriker & Caldwell and the Bahcall, Schmidt, & Soneira models for our Galaxy.

For each cluster in our sample, we calculated its orbits over a Hubble time, starting from the present observed positions and assumed velocities. Medians of the resulting set of peri- and apogalactic distances and velocities are used then as an input for the Fokker-Planck code. Evolution of the cluster is followed up to its total dissolution due to a coherent action of all of the destruction mechanisms. The rate of destruction is then obtained as a median over all the cluster sample, in accord with Aguilar, Hut, & Ostriker.

We find that the total destruction rate is much larger than that given by Aguilar, Hut, & Ostriker with more than half of the present clusters (52%-58% for the Ostriker & Caldwell model, and 75%-86% for the Bahcall, Schmidt, & Soneira model) destroyed in the next Hubble time. Alternatively put, the typical time to destruction is comparable to the typical age, a result that would follow from (but is not required by) an initially power law distribution of destruction times. We discuss some implications for a past history of the globular cluster system and the initial distribution of the destruction times, raising the possibility that the current population is but a very small fraction of the initial population with the remnants of the destroyed clusters constituting presently a large fraction of the spheroid (bulge + halo) stellar population.