How to Feed a Black Hole

What are AGN?

We now believe that in the centers of most, if not all, massive galaxies, there resides a supermassive black hole. In some cases, weighing over a billion times the mass of the sun. In most galaxies, these black holes lie dormant and can only be found through their gravitational influence. However, in a small fraction of galaxies, the supermassive black holes are seen to be 'active', astronomers call these 'Active Galactic Nuclei' or AGN. During these short phases, gas is accreting onto the black hole in an accretion disk. Gas in an accretion disk is heated to high temperatures and emits radiation through a wide range of wavelengths, most prominently in the X-ray, UV and optical part of the spectrum. AGN can often outshine the entire galaxy they reside in, but they span a very wide range in luminosities. 

How are black holes fed?

One question has puzzled astronomers since we first learned of the nature of AGN: how does a dormant supermassive black hole turn into an AGN? How is it triggered? Feeding even a very bright quasar requires a surprisingly sparse supply of gas: about the mass of the sun per year is required for bright AGN, while fainter AGN require considerably less than that. This might not seem like a lot, but AGN are known to be active for ten or even a hundred million years. If they are to be fed during that entire time, even a very small amount per year adds up to an impressive total mass. A bright AGN can swallow the entire gas supply of the galaxy it resides in during a single active phase.

There is another problem with feeding AGN: it is actually surprisingly difficult to funnel gas that is available in galaxies into their central black holes. The gas in galaxies is generally settled in a disk-like structure. Moving gas towards the center - where the black hole is located - requires stripping the gas of an overwhelming part of its angular momentum.  This requires some kind of a disturbance. There are different ways to achieve feeding the gas into the black hole, and in particular one process has become very popular amongst astronomers: mergers of galaxies. We will not touch on other possibilities in this post, but look at how mergers might trigger AGN and what the data tell us.

Galaxy mergers and black hole feeding

Simulation showing how gas in a merger is moved
towards the  supermassive black hole.
Image Credit: Phil Hopkins

When galaxies collide, both of them can carry considerable amounts of gas, and during the collision, the normal motions of the gas are disturbed and it can move to the center of the newly formed system to feed the black hole and start an active phase. In mergers, we therefore find good conditions to trigger luminous AGN since large amounts of gas become available and are funneled to the black hole within a short amount of time. Therefore, mergers are believed to be closely connected to AGN.

When AGN were first studied, it was also found that many were located in galaxies that looked very disturbed. In fact, many of the AGN in the vicinity of the Milky Way are located in galaxies that appear to have undergone mergers very recently. However, just looking at the incidence of merger features in galaxies is not sufficient, we must take into account what percentage of non-active galaxies show signs of merging. We need a so-called control sample. And while many galaxies showing AGN activity do show merger features, CANDELS researchers have shown that this just reflects the fact that galaxies in general often undergo interaction. So, what is happening? What is the real connection between mergers and AGN?

Does the luminosity of the AGN matter?

HST images of nearby luminous AGN showing clear signs of
interaction in their host galaxies
Image Credit: HubbleSite

One possibility I am interested in studying is that mergers are only responsible for some AGN. As mentioned earlier, depending on the luminosity of the AGN, the amount of gas required to feed it changes dramatically. Faint AGN can rather easily be triggered by smaller events, so they do not necessarily need to be connected to mergers. Also, when looking at the whole population of AGN, the brightest AGN form a minority in the overall AGN population. Could it be that mergers only trigger the most luminous of AGN?

To answer this question, we look at AGN at a low redshift (z=0.5-0.8) over a wide range of luminosities -- the brightest AGN in our sample are about a thousand times more luminous than the faintest ones. This also means that the brightest ones require about a thousand times more gas to shine as bright as they do. For all these AGN, we then look at a sample of control galaxies that are about equally massive and compare how asymmetric they appear. When galaxies undergo interaction, they appear asymmetric and disturbed, the more they settle, the more symmetric they will become. Comparing the levels of asymmetry in AGN hosting galaxies and normal control galaxies therefore lets us compare how likely they are to be connected to a recent galaxy interaction. 

It turns out that similar to previous studies, we find that host galaxies of AGN look no more disturbed than normal galaxies. However, because we choose AGN that are more nearby, we can also study these differences as a function of luminosity. This has not been studied previously. Dividing the AGN into different bins according to their luminosities, we can also determine if there are differences between AGN host galaxies and control galaxies only for certain AGN luminosities. We do not find any differences, even for the more luminous AGN where we would expect a stronger connection to mergers.

If the host galaxies of even luminous AGN are no more disturbed than normal galaxies, what does this mean? One possibility is that there is a very long delay between the collision of galaxies and the phase during which the AGN gets triggered. While this is possible, the delay would have to be very long for all merger features to fade. The other possibility is that the AGN we study are still not quite bright enough to see merger triggering in effect. The most luminous AGN are extremely rare, and even large fields cover only a few of them, the very brightest AGN are so rare that they are not found in CANDELS fields. Studying more extreme AGN might therefore lead us to understand how mergers and AGN are connected.