The DIANOGA simulations of galaxy clusters: characterising star formation in protoclusters

Abstract

Aims. We studied the star formation rate (SFR) in cosmological hydrodynamical simulations of galaxy (proto-)clusters in the redshift range 0 < z < 4, comparing them to recent observational studies; we also investigated the effect of varying the parameters of the star formation model on galaxy properties such as SFR, star-formation efficiency, and gas fraction. Methods. We analyse a set of zoom-in cosmological hydrodynamical simulations centred on 12 clusters. The simulations are carried out with the GADGET-3 Tree-PM smoothed-particle hydro-dynamics code which includes various subgrid models to treat unresolved baryonic physics, including AGN feedback. Results. Simulations do not reproduce the high values of SFR observed within protocluster cores, where the values of SFR are underpredicted by a factor ≳4 both at z ∼ 2 and z ∼ 4. The difference arises as simulations are unable to reproduce the observed starburst population and is greater at z ∼ 2 because simulations underpredict the normalisation of the main sequence (MS) of star forming galaxies (i.e. the correlation between stellar mass and SFR) by a factor of ∼3. As the low normalisation of the MS seems to be driven by an underestimated gas fraction, it remains unclear whether numerical simulations miss starburst galaxies due to overly underpredicted gas fractions or overly low star formation efficiencies. Our results are stable against varying several parameters of the star formation subgrid model and do not depend on the details of AGN feedback. Conclusions. The subgrid model for star formation, introduced to reproduce the self-regulated evolution of quiescent galaxies, is not suitable to describe violent events like high-redshift starbursts. We find that this conclusion holds, independently of the parameter choice for the star formation and AGN models. The increasing number of multi-wavelength high-redshift observations will help to improve the current star formation model, which is needed to fully recover the observed star formation history of galaxy clusters.

Is there enough star formation in simulated protoclusters?

“…Notably, the strong dependence on numerical resolution found in our analysis is in conflict with some findings from Bassini et al (2020). Bassini et al (2020) performed zoomed-in simulations of 12 galaxy clusters, selected from a parent dark-matter-only simulation of 1 h −1 cGpc on a side, all with M 200 > 8 × 10 14 h −1 M at z 0, with a resolution that is similar to TNG300-1 and 10 times better than that of the simulations used by Granato et al (2015). They reported a similar level of discrepancy to our analysis in SFR with observations for both protoclusters and individual galaxies.…”

“…The discrepancy with observation by more than a factor of 2-4 shown here is consistent with findings of some previous studies (e.g. Granato et al 2015;Bassini et al 2020) that used completely different sets of simulations and subgrid models, indicating that the issue lies not with the particular simulations and models, but is more general. The predictions of the empirical models are even lower than the simulations, typically by a factor of 2-3.…”

“…However, they found that the SFR calculated within a 2 Mpc aperture is not any higher than that from Granato et al (2015), who used lower resolution simulations. This may indicate that details of the numerical convergence of the SFR are subject to particular subgrid prescriptions and numerical implementations used in the simulations, as well as redshift (the comparison with Granato et al 2015in Bassini et al 2020 was made at a lower redshift of z 2). Moreover, the resolution is not the only difference between the simulations used in Granato et al (2015) and Bassini et al (2020).…”

“…Comparison between observed protoclusters and theoretical predictions have been attempted before, but the findings in most (if not all) cases is that the SFRs in simulated protoclusters at high z fall significantly short (typically by a factor of 2-5) of matching the observations, while simulations predict much higher SFRs and stellar masses for clusters at low redshifts than obserations (e.g. Granato et al 2015;Bassini et al 2020). Lovell et al (2020) found about a factor of 2 deficit in their simulations relative to observations of submm source number counts in single-dish surveys (although their simulations only had a side length of 150 Mpc, which is too small to contain even a single protocluster with SFRs as extreme as reported in the literature, as we discuss later).…”

Radio galaxies in galaxy groups: kinematics, scaling relations, and AGN feedback

“…At higher radii the curves of radio and nonradio BGGs decrease in a similar fashion, while the probability density of simulated BGGs already steepens at ∼ 0.1 Rvir. Nevertheless, in simulations BGGs tend to be more massive than their observational counterparts (e.g., Bahé et al, 2017;Henden et al, 2019;Bassini et al, 2020). For this reason, they are closer to the centre and exhibit less spread in velocity.…”