Challenging the unified model of AGN
This plot shows the halo occupation number of AGN in bins of the halo mass for satellite AGN (left panel) and central AGN (right panel) at z=0.25 for our super-large 909Mpc/hr simulation run. The mean number of all satellite and central AGN is shown as blue lines. For comparison, we also show the mean number of all black holes (BHs) without any luminosity threshold), which is – above our resolution threshold – equivalent to the number of galaxies. Thus, it is very promising that the slope for satellite BHs is about 1 (dashed line), as expected for galaxies, meaning that all galaxies contain central SMBHs.
Interestingly, the halo occupation distribution (HOD) of different AGN types looks very different (green and red lines). In particular, the satellite HOD slope for quasar-mode AGN (=AGN with high Eddington ratios, green lines) is smaller than for AGN with very low Eddington ratios. The HOD of central AGN consists of two different regimes: Low mass haloes (=isolated galaxies) are dominated by quasar-mode AGN (green line), whereas more massive haloes (=galaxy groups and galaxy clusters) are dominated by inefficiently accreting AGN (red line).
This clearly demonstrates that different AGN types prefer different environments, or in other words: AGN are no random events! This strongly contradicts the unified model of AGN, which assumes that all AGN are intrinsically similar (see also ‘Science’).