Frederika Phipps
Gather.town id
CSF05
Poster Title
Infant Globular Clusters: Extremely Dense Star Forming Factories at High Redshift
Institution
University of Edinburgh
Abstract (short summary)
We present the results of a study of the demographics of low-mass stellar systems at high redshift in a suite of cosmological simulations from the First Billion Years Project (FiBY). We have identified a population of objects at z=6 which we believe show characteristics to be early ancestors of local Universe globular clusters (GCs), which we dub “infant GC candidates”. These candidates are compact in size, have little dark matter and high gas content, thus they are likely to host further episodes of star formation. From their evolution until z=6, the process of star formation in these objects appears to be bursty in nature, which could be an early manifestation of the ‘multiple population phenomenon’ we currently observe in low-redshift GCs. In our simulations, these high-density GC ancestors tend to form in groups located within the disks of their host galaxies. We also propose a new link between present-day globulars and their high-redshift counterparts, which we have identified by studying the global properties of these infant GC candidates and the relationship with their large-scale environment. We find that the relation between the specific star formation rate of a host galaxy and the mass of the most massive GC holds across a large range of redshifts, thus providing a powerful way to probe both the early evolution of this class of stellar systems as well star formation across cosmic time.
Plain text (extended) Summary
We explore a suite of high-resolution cosmological simulations at high redshift to investigate theoretical scenarios concerning the formation of old, low-mass stellar systems with a particular focus on globular clusters (GCs). The First Billion Years (FiBY) project is a set of high-resolution, physics rich, cosmological SPH simulations. They track metal pollution for 9 elements and include prescriptions for supernova feedback and formation of both Population II and III stars. In this work, a simulated box of volume (4Mpc)^3 with 684^3 particles per type was used. The mass resolution is 1250 and 6160 Msol for SPH and dark matter particles, respectively.
From our ‘agnostic exploration’, two distinct groups of objects are naturally identified (see Figure 1) in the plane of stellar fraction (fstar = mass of baryons in stars / total mass of baryons) versus baryon fraction (fb = mass of baryons / total mass of baryons and dark matter). The first group of objects has fb = 1 and we hypothesise that these objects could be infant GCs. The second group lies along the line of fstar = 1 and they could be akin to proto ultra-faint dwarfs.
We compare the half-mass radii, stellar density, metallicity and stellar velocity dispersion of these two groups against other systems from the simulation that have a similar stellar mass. We find that the fb = 1 group were denser than the other systems at all masses and were far more compact. When comparing to the Milky Way GC properties, we find that the fb = 1 group have similar metallicities and stellar masses. This further supports our hypotheses that these objects are infant GC candidates.
In Figure 2, we plot size versus total cluster mass for the fb =1 and fstar = 1 groups along with local Universe data for GCs, young massive clusters and ultra-compact dwarfs. We find that our infant GC candidates have slightly higher total masses than the local Universe GCs due to the simulated objects having copious amounts of gas. This indicates the potential for further star formation and mass growth before feedback clears out all the gas, pushing our candidates into the regime observed for local Universe GCs.
By fitting redshift-zero relations, we can determine whether formation or evolutionary processes are more important in developing GC systems. In Figure 3 we plot the GC system mass vs the halo mass of the host galaxy. We overlay two different redshift-zero fits to local Universe data. The good agreement between the redshift-zero relations and the distribution of infant GC candidates implies that this relation could be set at formation.
We also investigate the relation between the most massive GC in the system and the specific star formation rate of the host galaxies (Figure 4). We compare our findings with observational data at low redshift. As we also fit a simple analytical expression to the data, we note that this relation can be further investigated observationally, at different redshifts.
Finally, we study the galactic environments of the fb = 1 objects and find that they are preferentially found in the spiral arms of their host galaxies at locations of high gas surface density. These regions of a galaxy are rich in giant molecular clouds, high-density gas and clustered star formation - further evidence that these fb = 1 objects can be associated with the progenitors of GCs.
Further details can be found in an article accepted for publication and selected as a highlight in Astronomy & Astrophysics (arXiv link provided on the poster).
From our ‘agnostic exploration’, two distinct groups of objects are naturally identified (see Figure 1) in the plane of stellar fraction (fstar = mass of baryons in stars / total mass of baryons) versus baryon fraction (fb = mass of baryons / total mass of baryons and dark matter). The first group of objects has fb = 1 and we hypothesise that these objects could be infant GCs. The second group lies along the line of fstar = 1 and they could be akin to proto ultra-faint dwarfs.
We compare the half-mass radii, stellar density, metallicity and stellar velocity dispersion of these two groups against other systems from the simulation that have a similar stellar mass. We find that the fb = 1 group were denser than the other systems at all masses and were far more compact. When comparing to the Milky Way GC properties, we find that the fb = 1 group have similar metallicities and stellar masses. This further supports our hypotheses that these objects are infant GC candidates.
In Figure 2, we plot size versus total cluster mass for the fb =1 and fstar = 1 groups along with local Universe data for GCs, young massive clusters and ultra-compact dwarfs. We find that our infant GC candidates have slightly higher total masses than the local Universe GCs due to the simulated objects having copious amounts of gas. This indicates the potential for further star formation and mass growth before feedback clears out all the gas, pushing our candidates into the regime observed for local Universe GCs.
By fitting redshift-zero relations, we can determine whether formation or evolutionary processes are more important in developing GC systems. In Figure 3 we plot the GC system mass vs the halo mass of the host galaxy. We overlay two different redshift-zero fits to local Universe data. The good agreement between the redshift-zero relations and the distribution of infant GC candidates implies that this relation could be set at formation.
We also investigate the relation between the most massive GC in the system and the specific star formation rate of the host galaxies (Figure 4). We compare our findings with observational data at low redshift. As we also fit a simple analytical expression to the data, we note that this relation can be further investigated observationally, at different redshifts.
Finally, we study the galactic environments of the fb = 1 objects and find that they are preferentially found in the spiral arms of their host galaxies at locations of high gas surface density. These regions of a galaxy are rich in giant molecular clouds, high-density gas and clustered star formation - further evidence that these fb = 1 objects can be associated with the progenitors of GCs.
Further details can be found in an article accepted for publication and selected as a highlight in Astronomy & Astrophysics (arXiv link provided on the poster).
URL
phipps@roe.ac.uk
Poster file