However reality may be not quite so simple. A more detailed examination suggests that negative feedback in momentum-driven winds by supermassive black holes falls short of explaining the observed MBH - correlation by a factor of a few (Silk & Nusser 2010). Moreover comparison of baryonic fractions with bulge-to-disk ratios in nearby galaxies demonstrates that AGN alone do not eject significant amounts of baryons (Anderson & Bregman 2010). If negative feedback in momentum-driven winds by supermassive black holes cannot explain the MBH - correlation, something else is needed.
3.1. Its not SN, its not AGN: maybe its both!
A plausible addition to the physics is inclusion of star formation, induced, enhanced and quenched by the SMBH outflows. There is extensive evidence, recently compiled by Netzer (Netzer et al. 2010), that demonstrates the intimate connection of AGN luminosity and star formation rate over a wide dynamic range. If AGN-driven outflows trigger star formation, the star formation rate is boosted by a factor tdyn / tjet, and the outflow momentum is amplified by supernovae (Norman and Silk 2009). The star formation rate boost factor amounts to vcocoon / ~ 10-100. The outflow momentum is amplified by supernovae. Consequently, the momentum supplied to the gas is boosted by the combination of AGN and star formation. Of course the causal direction is uncertain, and indeed the phenomena could be mutually self-regulating. To go beyond phenomenology, many details need to be refined, the most pressing perhaps being the nature of the black hole growth.
3.2. Triggering is observed
There are examples of jet-induced global star formation, as seen locally in Minkowski's object (Croft et al. 2006), and jet-induced CO formation (and excitation) at high redshift. CO is a prerequisite for star formation, and has been detected in large amounts in the host galaxies of high redshift quasars. The conversion ratio to H2 is uncertain however, and renders any conclusions uncertain. The SFE in early type galaxies, possible sites of AGN, containing CO is elevated (Wei et al. 2010). An actively accreting massive black hole in the dwarf starburst galaxy Henize 2-10 (Reines et al. 2010) is suggestive of current epoch triggering. At high redshift (Wang et al. 2010), the presence of a SMBH favours the higher SFE seen in ULIRGs.
3.3. The role of AGN at high z
Are active galactic nuclei aftermaths or precursors to star formation? Most data points to a relative increase of black hole mass with redshift. Contrary to earlier claims, even SMGs, typically the most massive galaxies at z~ 2 in which the star formation rate is high, reveal black hole mass to bulge mass ratios that agree with the local value (Hainline et al. 2010). The most extreme case is that SMBH in z ~ 6 quasars lie high (Riechers et al. 2010). The initiation of SMBH growth remains a mystery. The observed accretion rate is ~ 10-3 of the star formation rate in most AGN. The accretion rate tracks the star formation rate at z 2 but any conclusion about the ratio at higher z is confused by the fact that up to 80% of ultraluminous infrared starbursts (ULIRGs) have buried AGN Imanishi et al. 2010). AGN are generally thought to be responsible for quenching of star formation, especially at high z, rendering massive ellipticals red. However the preceding discussion suggests that there may have been a prior, short-lived, phase of triggering. The various correlations between AGN content via emission line diagnostics and stellar population content and age are inconclusive in elucidating this issue (Schawinski et al. 2007).