École Nationale Supérieure de
Techniques Avancées

Abstract

In this work, we address the issue of abundance anomalies in globular clusters. First, we define what globular clusters are, and we explain that observational data highlight "abundance anomalies" : the chemical compositions measured within these clusters do not match those predicted by standard models of star formation. Several hypotheses have been studied in the literature to date, but one of the most popular, and the one we are going to focus on, is that of supermassive star pollution. The idea is simple : stars formed from gas whose chemical composition was altered by the presence of a supermassive star, which explains why they have such distinctive chemical compositions today. The question we are investigating is : can a supermassive star form in a globular cluster in formation ? We are exploring the idea of formation by runaway collision : a very rapid increase of the number of collisions between stars would occur in the cluster, leading to the rapid formation of a supermassive star. We develop a Monte-Carlo code to simulate these collisions in a forming globular cluster. We compare our results to an analytical model developed in [1], which only partially takes collisions into account. We show that our way of taking into account all possible collisions (for all the stars in the cluster) highly increases the chances of a supermassive star forming. We also study the impact of taking these collisions into account on the density required for a cluster to form a supermassive star. The phenomenon of segregation in globular clusters is also studied.