Dynamical Constraints on the Origin of Multiple Stellar Populations in Globular Clusters

TL;DR

This study uses N-body simulations to explore how primordial gas expulsion influences the formation of multiple stellar populations in globular clusters, suggesting rapid gas loss is crucial for current cluster properties.

Contribution

It provides new constraints on initial cluster conditions and gas expulsion timescales necessary to produce observed multiple stellar populations.

Findings

Short gas expulsion timescales (<10^5 yr) are essential.

A substantial amount of residual gas was present after second-generation star formation.

Predicted anti-correlation between second-generation star fraction and cluster mass.

Abstract

We have carried out a large grid of N-body simulations in order to investigate if mass-loss as a result of primordial gas expulsion can be responsible for the large fraction of second generation stars in globular clusters (GCs) with multiple stellar populations (MSPs). Our clusters start with two stellar populations in which $10\%$ of all stars are second generation stars. We simulate clusters with different initial masses, different ratios of the half-mass radius of first to second generation stars, different primordial gas fractions and Galactic tidal fields with varying strength. We then let our clusters undergo primordial gas-loss and obtain their final properties such as mass, half-mass radius and the fraction of second generation stars. Using our N-body grid we then perform a Monte Carlo analysis to constrain the initial masses, radii and required gas expulsion time-scales of GCs with MSPs. Our results can explain the present-day properties of GCs only if (1) a substantial amount of gas was present in the clusters after the formation of second generation stars and (2) gas expulsion time-scales were extremely short ($\lesssim 10^5$ yr). Such short gas expulsion time-scales are in agreement with recent predictions that dark remnants have ejected the primordial gas from globular clusters, and pose a potential problem for the AGB scenario. In addition, our results predict a strong anti-correlation between the number ratio of second-generation stars in GCs and the present-day mass of GCs. So far, the observational data show only a significantly weaker anti-correlation, if any at all.