Research > Background Material

The curves show the percentage of all refereed publications on SAO/NASA ADS that have abstracts containing the combination of keywords shown in the legend each year. Abstracts discussing AGN or quasar outflows were growing relatively steadily over time since ~1990; however, since ~2000 the number of abstracts mentioning AGN feedback has been growing rapidly. We relate this to: (1) three papers that characterise M_(BH)-host galaxy relationships that are in the top-100 of all-time astronomy publications (i.e., top 0.009\% citation counts based on SAO/ADS; Magorrian+98; Ferrarese+00 and Gebhardt+00) and (2) a series of galaxy formation analytical models (e.g., Silk & Rees 98), semi-analytical models (e.g., Benson+03) and hydrodynamical simulations (e.g., DiMatteo+05) where AGN-driven outflows are required to explain these and other observables of galaxy populations. It has been become increasingly popular to attempt to derive observed outflow properties to test AGN feedback models.

The figure shows the ratio of stellar mass to halo mass as a function of halo mass for three different runs of a simulation (Somerville+08) and for the semi-empirical relationship (Moster+13). The shaded region shows the 16th and 84th percentiles of the fiducial model that includes energy injection from AGN and star formation (SF). The right y-axis shows the efficiency for turning baryons into stars (M_stellar/fb*M_halo; where the factor of fb=0.17 is the cosmological baryon fraction). The impact of including star formation feedback in the model is to reduce the efficiency of converting baryons into stars in low mass haloes. For massive haloes, energy injection from AGN is required in order to reduce these efficiencies. Such effects are required in the model in order to reproduce many observable properties of the massive galaxy population.

A schematic diagram to illustrate the relationships between fuel supply, galaxy growth and black hole growth. Both AGN and star formation are fuelled by cold gas that originates from a shared (potentially hot) gas reservoir inside the galaxy halo. This gas reservoir can be fed by gas-rich mergers, by recycled material from internal galactic processes and by accretion of gas from intergalactic material. The amount of gas and the ability for this gas to cool determines the amount of usable fuel that can be used for feeding black hole growth and star formation. In the case of providing the fuel for black hole growth the material has the additional challenge of losing sufficient angular momentum to reach the inner sub-parsec region of the galaxy. Both processes are known to inject energy and momentum (via radiation, winds and jets) that can reduce the availability of usable fuel through ionising, heating, shocking or expelling material, and hence provide self-regulatory feedback mechanisms. A key component of most galaxy formation models is that these two processes can also have a positive or negative impact on the usable fuel supply for the other process (black and grey arrows).

Plots to highlight typical optical emission-line diagnostics used to identify AGN. All data shown are taken from the SDSS DR7 database for extragalactic sources with z<0.4. Left: Emission-line flux ratios of [O III]5007/Hbeta versus [N II]6584/Halpha for sources with FWHM(Halpha)<1000 km/s (to remove the majority of Type 1 AGN). The dashed lines are taken from Ho et al. 1997 and Kauffmann et al. 2003 to identify different source classifications (i.e., Type 2 AGN [shown in red]; LINERS and H II galaxies [shown in black]). Right:[O III]5007/Hbeta versus FWHM(Halpha). Broad Halpha emission-line profiles (i.e., FWHM >~1000 km/s) may indicate the presence of an AGN broad line region. It is possible for reasonably broad emission lines (i.e., 1000 km/s ~< FWHM ~< 2000 km/s) to be produced in the NLR due to outflows therefore care needs to be taken when classifying AGN with these ``intermediate'' line widths.

Schematic diagram of a data cube. A data cube provides information in three dimensions: two spatial dimensions (i.e., an image in [x,y]) and a third dimension of wavelength (equivalently velocity or frequency). It is therefore possible to obtain an image of the target at a single wavelength or collapsed over wavelength slices. Furthermore, at every spatial pixel of the datacube a spectrum can be extracted.