The GRIFFIN Project—Formation of Star Clusters with Individual Massive Stars in a Simulated Dwarf Galaxy Starburst

Abstract

We describe a population of young star clusters (SCs) formed in a hydrodynamical simulation of a gasrich dwarf galaxy merger resolved with individual massive stars at sub-parsec spatial resolution. The simulation is part of the griffin (Galaxy Realizations Including Feedback From INdividual massive stars) project. The star formation environment during the simulation spans seven orders of magnitude in gas surface density and thermal pressure, and the global star formation rate surface density (Σ SFR ) varies by more than three orders of magnitude during the simulation. Young SCs more massive than M * ,cl ∼ 10 2.5 M form along a mass function with a power-law index α ∼ −1.7 (α ∼ −2 for M * ,cl 10 3 M ) at all merger phases, while the normalization and the highest SC masses (up to ∼ 10 6 M ) correlate with Σ SFR . The cluster formation efficiency varies from Γ ∼ 20% in early merger phases to Γ ∼ 80% at the peak of the starburst and is compared to observations and model predictions. The massive SCs ( 10 4 M ) have sizes and mean surface densities similar to observed young massive SCs. Simulated lower mass clusters appear slightly more concentrated than observed. All SCs form on timescales of a few Myr and lose their gas rapidly resulting in typical stellar age spreads between σ ∼ 0.1 − 2 Myr (1σ), consistent with observations. The age spreads increase with cluster mass, with the most massive cluster (∼ 10 6 M ) reaching a spread of 5 Myr once its hierarchical formation finishes. Our study shows that it is now feasible to investigate the SC population of entire galaxies with novel high-resolution numerical simulations.

Star Clusters Near and Far

“…On the other hand, CFE can change from a few to very high rates in galaxies with larger stellar masses, indicating that the positions of galaxies in the Γ vs. Σ SFR relation will change as a function of the evolutionary phase that the galaxy experiences. This was indeed recently outlined also in the simulations analysed by Lahén et al (2019).…”

“…In recent years, the increasing computational power combined with improved numerical recipes that account for sub-galactic scale physics such as stellar feedback and self-consistently evolving multi-phase ISM, has made possible to follow cluster formation and evolution in combination with galaxy evolution (e.g., Kruijssen et al 2011Renaud et al 2015;Li et al 2017;Choksi et al 2018;Lahén et al 2019;Li and Gnedin 2019). However, computational resources are not unlimited and the initial generation of works modelling YSC populations focus on isolated or merging galaxies in idealized, non-cosmological simulations (e.g., Kruijssen et al 2011;Renaud et al 2015;Lahén et al 2019). These setups were designed for developing methodologies and numerical methods, but they lack the cosmological context that determines and drives galaxy assembly history.…”

Where did the globular clusters of the Milky Way form? Insights from the E-MOSAICS simulations

“…They find that mergers can increase the GC formation rate by a factor of ∼3 over 5 Gyr, but without any disruption mechanism it is difficult to say what fraction of these proto-GCs would survive for a significant time post-merger. Recently, the approach of simulating isolated dwarf galaxies has been pushed to mass resolutions of 4 M by Lahén et al (2020), which allow them to study the formation of individual, resolved massive stars, and look at the formation of clusters from first principles in a galactic environment.…”

On the early evolution of massive star clusters: the case of cloud D1 and its embedded cluster in NGC 5253

“…In particular, we adopted a static distribution of stars and gas in the parental cloud and thus considered a snapshot of a more complicated dynamical process (e.g. Lahén et al 2019Lahén et al , 2020. We also did not consider a possible mass segregation and thus a possible difference in massive and low mass star distributions.…”