The Dynamics of Star Cluster Formation

Doktorsavhandling, 2020

How do star clusters form? We care, since these are the birth sites of most stars, perhaps including our own Sun. There are a great variety of different theoretical models of cluster formation and our main goal in this thesis is to examine the implications of these for the dynamical evolution of a cluster's stellar population, including the ejected stars. In contrast to the majority of previous cluster formation studies, the focus of this work is on detailed modeling, using the Nbody6 code, of stellar dynamics, including binaries, with the structure, kinematics and star formation of the natal gas cloud explored with simple analytic prescriptions. In particular, we adopt the Turbulent Clump Model of pressure-truncated singular polytropic spheres, which sets global and local initial conditions of the newly formed stars. In a first paper, exploring a fiducial 3,000 solar mass clump, we investigated the effects of overall clump density, global star formation efficiency, degree of primordial mass segregation, degree of primordial binarity and binary population properties on the dynamical evolution of the cluster. Here, like most previous works, we assumed stars are formed very quickly, i.e., approximated as instantaneously, compared to the free-fall time of the clump. In our next work, after implementing a major code development to Nbody6 that allows modeling of gradual formation of stars, we investigated how the timescale of cluster formation, parameterized via the star formation efficiency per free-fall time, affects its early dynamical evolution. This is the first time that such a study, including a realistic binary population, has been carried out. We showed that star clusters that form rapidly expand more quickly after they emerge from the gas, while slowly-formed clusters evolve into a much more stable configuration during the gas rich phase. We also showed how the stellar population is affected by the timescale of formation, including the frequency of runaway/walkaway stars, stellar age gradients and primordial binary processing. We have then carried out preliminary explorations of a broad range of star-forming clump parameters, i.e., with masses from 300 to 30,000 solar masses and background cloud mass surface densities from 0.1 to 1 g cm². For the largest clusters simulated, we make use of a GPU-enabled version of the code. Further improvements to the modeling that have been implemented include global elongation of the clump so that nonspherical, including very filamentary, initial conditions can be studied. Models with internal spatial and kinematic substructure for the birth locations of the stars, based on hydrodynamic simulations of supersonic turbulence, have also been studied. In parallel, we have also carried out two projects that focus on observed systems related to dynamical ejections within the Orion Nebula Cluster (ONC). First, we examined a particular set of runaway stars associated with the Orion KL massive star forming region and carried out a systematic exploration of N-body simulations to understand the properties of the dynamical ejection that produced them. Second, we have performed a census of runaway stars from the ONC using Gaia data, estimating the total unbound population from the cluster. We have compared these results with our cluster formation simulations leading to new constraints on the star formation rate and dynamical age of the system.