Paper detail

The Star-Formation Properties of the Observed and Simulated AGN Universe: BAT vs EAGLE

In this paper we present data from 72 low redshift, hard X-ray selected AGN taken from the {\it Swift}-BAT 58 month catalogue. We utilise spectral energy distribution fitting to the optical to IR photometry in order to estimate host galaxy properties. We compare this observational sample to a volume and flux matched sample of AGN from the EAGLE hydrodynamical simulations in order to verify how accurately the simulations can reproduce observed AGN host galaxy properties. After correcting for the known +0.2 dex offset in the SFRs between EAGLE and previous observations, we find agreement in the SFR and X-ray luminosity distributions; however we find that the stellar masses in EAGLE are $0.2 - 0.4$ dex greater than the observational sample, which consequently leads to lower sSFRs. We compare these results to our previous study at high redshift, finding agreement in both the observations and simulations, whereby the widths of sSFR distributions are similar ($\sim0.4-0.6$ dex) and the median of the SFR distributions lie below the star forming main sequence by $\sim0.3-0.5$ dex across all samples. We also use EAGLE to select a sample of AGN host galaxies at high and low redshift and follow their characteristic evolution from $z=8$ to $z=0$. We find similar behaviour between these two samples, whereby star formation is quenched when the black hole goes through its phase of most rapid growth. Utilising EAGLE we find that 23\% of AGN selected at $z\sim0$ are also AGN at high redshift, and that their host galaxies are among the most massive objects in the simulation. Overall we find EAGLE reproduces the observations well, with some minor inconsistencies ($\sim$ 0.2 dex in stellar masses and $\sim$ 0.4 dex in sSFRs).