Timing the starburst AGN connection

The existence of a correlation between the mass of stars in the bulge of local galaxies, and the masses of the supermassive black holes at the centres of these galaxies ( M-sigma relation ), suggests a joint mechanism by which both objects grow together. The difference in scales is enormous however, and finding a physical process which can robustly link the masses of a stellar bulge 1000 times more massive than the central supermassive black hole, has been keeping theorists busy for many years.

Here we start by measuring the recent star formation history of galaxies with bulges (and thus black holes) in the Sloan Digital Sky Survey (Figure 1 below). We then "peel-off" the strip on the outer edge of the distribution - these galaxies have recently undergone the strongest starbursts (coloured dots in Figure 1). The shape of the galaxy spectrum allows us to measure precisely the age of the starburst. We can then use the nebula emission lines to measure the accretion onto the black hole.

In the local Universe, most black hole growth occurs in a quiescent mode, inside bulges with some ongoing star formation, but nothing spectacular ( Wild et al. 2007 ). At high redshift the situation is expected to be vastly different however. The higher gas mass fractions, more frequent collisions between gas rich galaxies leading to more frequent strong bursts of star-formation ( starbursts ) makes the Universe a much more active place. In order to better understand how black holes and galaxy bulges grew together, it therefore makes sense to look at the most extreme cases in the local Universe.

The results are presented below in Figure 2 and 3. Averaging over all starburst galaxies in our sample, we find that more black hole growth occurs as the starburst ages. An apparent offset between the beginning of the starburst and the growth of the black hole has been noted previously by Davies et al. 2007 for a small, inhomogeneous sample of very nearby, mostly unobscured, AGN. Our results are the first to observed this offset in a statistically complete sample of starburst galaxies, necessarily including black holes in both their "on" (AGN) and "off" (no observational signatures) state.

In Figure 3, we observe that the cause of this decrease in global accretion onto the black hole is caused by a deficit in the number of AGN in young starbursts, compared to older starbursts. However, we note that there is an excess of AGN in both young and old starbursts, compared to bulges which are undergoing more ordinary levels of star formation - this proves the existence of a starburst-AGN connection for the first time, based on a large homogenous sample of starburst galaxies with a well defined control sample.

The implications of this result are that (i) black hole accretion does not track mass loss from fast stellar winds and supernovae; (ii) the probability of an accretion event increases only after the mass loss rate from supernovae and fast stellar winds has decreased, possibly suggesting a supresion of accretion by supernovae at early times; (iii) extrapolating a simple model to 10Gyr, in which the black hole eats ~0.5% (none) of the mass of winds ejected by low mass stars after (before) 250Myr, leads to a black hole mass-bulge mass relation intriguingly close to that observed in the local Universe today. If stars are the primary source of fuel for growing the black hole and also play a major role in the feedback processes that limit this growth, then it starts to be less puzzling that black hole mass is linked to the stellar mass of the bulge.


Figure 1: The selection of a complete sample of the strongest starbursts in the local Universe.


Figure 2: The growth of the central supermassive black hole in a bulge undergoing a strong burst of star-formation, as the starburst ages.


Figure 3: The distribution of black hole growth and accretion rates, for young and old starbursts, alongside a well-defined control sample of bulge galaxies undergoing more ordinary levels of star formation (matched in stellar mass, black hole mass, stellar mass density and redshift to the starburst samples)

Abstract: The mass of super massive black holes at the centre of galaxies is tightly correlated with the mass of the galaxy bulges which host them. This observed correlation implies a mechanism of joint growth, but the precise physical processes responsible are a matter of some debate. Here we report on the growth of black holes in 400 local galactic bulges which have experienced a strong burst of star formation in the past 600 Myr. We combine stellar continuum indices with H-alpha luminosities to measure a decay timescale of 300 Myr for the decline in star formation after a starburst. During the first 600 Myr after a starburst, the black holes in our sample increase their mass by on-average 5% and the total mass of stars formed is about 10^3 times the total mass accreted onto the black hole (similar to the ratio of stellar to black hole mass observed in present-day bulges). We find that the average rate of accretion of matter onto the black hole rises steeply at late times (roughly 250 Myr after the onset of the starburst). We show that this time dependence is consistent with a simple model in which roughly 0.5% of the mass lost by intermediate mass stars in the bulge is accreted by the black hole, but with a suppression in the efficiency of black hole growth at early times plausibly caused by supernova feedback, which is stronger at earlier times. We suggest this picture may be more generally applicable to black hole growth, and could help explain the strong correlation between bulge and black hole mass.