Multiple populations in globular clusters: the distinct kinematic imprints of different formation scenarios

Several scenarios have been proposed to explain the presence of multiple stellar populations in globular clusters. Many of them invoke multiple generations of stars to explain the observed chemical abundance anomalies, but it has also been suggested that self-enrichment could occur via accretion of ejecta from massive stars on to the circumstellar disc of low-mass pre-main sequence stars. These scenarios imply different initial conditions for the kinematics of the various stellar populations. Given some net angular momentum initially, models for which a second generation forms from gas that collects in a cooling flow into the core of the cluster predict an initially larger rotational amplitude for the polluted stars compared to the pristine stars. This is opposite to what is expected from the accretion model, where the polluted stars are the ones crossing the core and are on preferentially radial (low-angular momentum) orbits, such that their rotational amplitude is lower. Here we present the results of a suite of N-body simulations with initial conditions chosen to capture the distinct kinematic properties of these pollution scenarios. We show that initial differences in the kinematics of polluted and pristine stars can survive to the present epoch in the outer parts of a large fraction of Galactic globular clusters. The differential rotation of pristine and polluted stars is identified as a unique kinematic signature that could allow us to distinguish between various scenarios, while other kinematic imprints are generally very similar from one scenario to the other.