Galactic Mergers Fail to Feed Black Holes


The large black holes that reside at the center of galaxies can be hungry beasts. As dust and gas are forced into the vicinity around the black holes, it crowds up and jostles together, emitting lots of heat and light. But what forces that gas and dust the last few light years into the maw of these supermassive black holes?

It has been theorized that mergers between galaxies disturbs the gas and dust in a galaxy, and forces the matter into the immediate neighborhood of the black hole. That is, until a recent study of 140 galaxies hosting Active Galactic Nuclei (AGN) – another name for active black holes at the center of galaxies – provided strong evidence that many of the galaxies containing these AGN show no signs of past mergers.

The study was performed by an international team of astronomers. Mauricio Cisternas of the Max Planck Institute for Astronomy and his team used data from 140 galaxies that were imaged by the XMM-Newton X-ray observatory. The galaxies they sampled had a redshift between z= 0.3 – 1, which means that they are between about 4 and 8 billion light-years away (and thus, the light we see from them is about 4-8 billion years old).

They didn’t just look at the images of the galaxies in question, though; a bias towards classifying those galaxies that show active nuclei to be more distorted from mergers might creep in. Rather, they created a “control group” of galaxies, using images of inactive galaxies from the same redshift as the AGN host galaxies. They took the images from the Cosmic Evolution Survey (COSMOS), a survey of a large region of the sky in multiple wavelengths of light. Since these galaxies were from the same redshift as the ones they wanted to study, they show the same stage in galactic evolution. In all, they had 1264 galaxies in their comparison sample.

The way they designed the study involved a tenet of science that is not normally used in the field of astronomy: the blind study. Cisternas and his team had 9 comparison galaxies – which didn’t contain AGN – of the same redshift for each of their 140 galaxies that showed signs of having an active nucleus.

What they did next was remove any sign of the bright active nucleus in the image. This means that the galaxies in their sample of 140 galaxies with AGN would essentially appear to even a trained eye as a galaxy without the telltale signs of an AGN. They then submitted the control galaxies and the altered AGN images to ten different astronomers, and asked them to classify them all as “distorted”, “moderately distorted”, or “not distorted”.

Since their sample size was pretty manageable, and the distortion in many of the galaxies would be too subtle for a computer to recognize, the pattern-seeking human brain was their image analysis tool of choice. This may sound familiar – something similar is being done with enormous success with people who are amateur galaxy classifiers at Galaxy Zoo.

When a galaxy merges with another galaxy, the merger distorts its shape in ways that are identifiable – it will warp a normally smooth elliptical galaxy out of shape, and if the galaxy is a spiral the arms seem to be a bit “unwound”. If it were the case that galactic mergers are the most likely cause of AGN, then those galaxies with an active nucleus would be more probable to show distortion from this past merger.

The team went through this process of blinding the study to eliminate any bias that those looking at the images would have towards classifying AGN as more distorted. By both having a reasonably large sample size of galaxies and removing any bias when analyzing the images, they hoped to definitively show whether the correlation between AGN and mergers exists.

The result? Those galaxies with an Active Galactic Nucleus did not show any more distortion on the whole than those galaxies in the comparison sample. As the authors state in the paper, “Mergers and interactions involving AGN hosts are not dominant, and occur no more frequently than for inactive galaxies.”

This means that astronomers can’t point towards galactic mergers as the main reason for AGN. The study showed that at least 75% of AGN creation – at least between the last 4-8 billion years – must be from sources other than galactic mergers. Likely candidates for these sources include: “galactic harrassment”, those galaxies that don’t collide, but come close enough to gravitationally influence each other; the instability of the central bar in a galaxy; or the collision of giant molecular clouds within the galaxy.

Knowing that AGN aren’t caused in large part by galactic mergers will help astronomers to better understand the formation and evolution of galaxies. The active nuclei in galaxies that host them greatly influence galactic formation. This process is called ‘AGN feedback’, and the mechanisms and effects that result from the interplay between the energy streaming out of the AGN and the surrounding material in the center of a galaxy is still a hot topic of study in astronomy.

Mergers in the more distant past than 8 billion years might yet correlate with AGN – this study only rules out a certain population of these galaxies – and this is a question that the team plans to take on next, pending surveys by the Hubble Space Telescope and the James Webb Space Telescope. Their study will be published in the January 10 issue of the Astrophysical Journal, and a pre-print version is available on Arxiv.

Source: HST news release, Max Planck Institute for Astronomy, Arxiv paper

20 Replies to “Galactic Mergers Fail to Feed Black Holes”

  1. Active Galactic Nuclei, I like that much better then Black Hole. Black holes just suck. 🙂

    I still dont buy the Status Quo hypothesis of what these things actually are. Why should these Galactic Nuclei have to (devour or feed on) anything ? I’m a lay-person, so is there some flaw in the cosmic maths that needs this interpretation?

    I’m curious, UT had an article (last year) pointing out that galactic accretion discs can form via interactions with the interstellar medium (dust) ‘without needing mergers’. This study could form the basis of validation of such a hypothesis. Where is the unity of ideas in scientific circles?

  2. Why all the complicated hypotheses? The simplest answer to me as to why an AGN is active could just be that there is more material near the black hole to get sucked in!

  3. Presumably AGN are active because the supermassive black hole within is sucking down an accretion disk of material, which is the source of the radiation. This should eventually run out of steam if there is no fresh supply of gas and dust to populate the accretion disk. However, the link between galactic mergers and AGN might be more correlation than causation?

  4. Fozzie January 5, 2011 at 2:34 am

    The thing is, 1) there just can’t be enough material in the central regions to continuously power most AGN for any period of time without significant material inflow. 2) observations generally show that most galaxies are very gas-poor in their nuclear regions, and gas rich in the outer regions, and 3) gas in the outer regions of galaxies should be in stable orbits around the galactic centre in the absence of any perturbation.

    Putting all of this together means that disturbances to galactic systems are generally required in most models to feed material to the central regions of AGN to account for both their observed prevalence and luminosity distribution. The generally emerging consensus is that the lowest luminosity AGN can be fed through internal stochastic processes such as bar instabilities and other exotica, whereas the most powerful AGN are thought to require major galactic collisions/interactions to feed in the required amount of material. More middle-of-the-range AGN should be able to be triggered by more minor interactions or weak gravitational encounters.

    More recently, minor mergers (mergers between the AGN host and a small dwarf companion) have been proposed as a viable AGN triggering mechanism because they can cause substantial amounts of material to be fed to the AGN host while intrinsically disturbing the morphology of the system less obviously and for less time. I haven’t read their paper, but perhaps their findings support this emerging viewpoint indirectly…

  5. “observations generally show that most galaxies are very gas-poor in their nuclear regions, and gas rich in the outer regions, ”

    > should have just left that as gas poor in the nuclear regions – not all galaxies are gas rich in the outer regions, particularly ellipticals of course…

  6. “Active Galactic Nuclei, I like that much better then Black Hole. Black holes just suck.”

    Nice pun 🙂 , however, the terms aren’t really interchangeable. Black holes power AGN; the term AGN referring more to the observational fact that some galaxies appear to emit radiation from their core regions in an amount and with characteristics that cannot be explained merely by the presence of stars or associated activity… I guess in a nutshell the AGN is the observable energetic activity powered by the BH engine.

    Anyway – AGN triggering is one of the most persistent problems in astronomy today. There is no reason why the current paradigm absolutely has to be correct, just that this is where a vast amount of complex and often apparently conflicting observational evidence has led us. There are many aspects of the current understanding that are very well supported by observation, while there is still a great deal that remains uncertain. One thing is for sure – there are some very exciting new projects in the offing that will address these questions decisively. In 5 years time we will have a great deal more clarity on the issue.

  7. Galactic mergers do not necessarily mean their black holes merge or even enter into a tight orbit around each other. I suspect it might be common for galaxies to merge and for their large black holes to continue on an orbit which leaves the merged galaxy. It requires a large amount of drag or friction to force billions of solar masses to congeal. This is easy to understand with gas clouds, but less so with multi-million solar mass black holes. It also requires a near dead on impact trajectory for the two black holes to actually merge.


  8. I’m far from the grade level most of you are at but, given that we have a certain confidence in the inflation model and that we also believe that that laws of physics/conservation are universal and unchanging, at least in our bubble; smile, would it not be reasonable to presume that mass formation and interaction dynamics would differ dramatically at different stages of the expansion?
    LC, wouldn’t the probability for BH mergers be higher, gravitationally, in the early confined and denser environment? This study is for the 4-8 BY old galaxies. What about the 12+ BY old ones? As Astrofiend said, hopefully well have more clarity in the next 5 years or so.

    1. given that we have a certain confidence in the inflation model … would it not be reasonable to presume that mass formation and interaction dynamics would differ dramatically at different stages of the expansion?

      Close. What you refer to as “the inflation model” is actually the standard cosmology, where the inflation mechanism is but a part. It is also only one mechanism for universe expansion.

      Nancy’s recent post has a great figure of the observable universe in the standard model. Notice how “inflation” points only to the early exponential expansion where the universe really ‘blows up’?

      The later, much more sloven pace is AFAIU just the universe freewheeling, near enough, from the first stage. (Mostly because there’s no reason to collapse if it is still expanding.) The inflation mechanism has stopped working.

      This freewheeling continues until we come to the last part of the universe timeline, where dark energy has grown enough in relative importance to start another “accelerated expansion”. But again no inflation.

      So inflation is a mechanism, and as such it can be modeled by inflation models. This is actually ongoing and so is the concurrent testing for inflation as an observable fact. Because while the standard cosmology with all its mechanisms and its 5 free parameters is decently tested, inflation on its lonesome isn’t.

      (Technically the confidence for the later is just shy of the conventional 3 sigma for physics hypotheses. It is hoped, at least from my layman side, that the current Planck mission will nail this among other deeds.)

      Btw, beware, the great figure starts off not so great. I will explain why in the parenthesis, jump over to the conclusion if it is too technical:

      [Notice how the figure says that the volume that becomes our observable universe comes out of “quantum fluctuations”? This is poor allusions to “a single big bang, from a point” and to mathematically inspired “quantum gravity” both.

      In reality, if we take inflation seriously, there is likely an ongoing inflation process. In any case, our universe arose from a “Planck volume” among others in some volume that afterwards appeared as a local end of inflation. (I.e. as seen from after the crazy exponential expansion.)

      And what ties the putative inflation process together with our local volume of universe is what is called “semiclassical worldlines”, a much more nicely behaved physical process than the imagined “quantum gravity fluctuations”.

      Semiclassical worldlines is a common and useful approximation in these parts of physics; quantum gravity and its unobserved phenomena of “fluctuations” is not. And incidentally, while it is controversial, the only Planck volume probe observed so far (in the form of photon timing from early universe events) tests the prediction of smooth worldlines, and so absence of “fluctuations”, nicely. It is “smooth sailing” as special relativity tells us about spacetime on all scales, whatever that means exactly.]

      In short: The large later part of the figure, which shows the standard cosmology, is great. The first small part of the figure, which shows speculation, and old and troubled one to boot, not so much.

      1. More explanatory: “our universe arose from a “Planck volume”” – our universe arose from a minute “Planck volume”. Those suckers are really small, which is why we have trouble (need tremendously high energies and/or long times) to see into such scales.

      2. Thank you for taking your time to reply. Although I sounded quite naive I actually have a few years of trying to absorb the cosmology controversy. Also I have been a member of the U.T. forum since close to it’s beginning and it has been part of my inspiration to seek a deeper insight into the hidden mysteries. I even suffered through the heated “iron/electric sun” b.s. I’m now retired and can now spend more time getting more confused.
        I have seen that depiction in Nancy’s “Shedding new light…” article more than once,
        and I do understand that “inflation theory” is but a part of the “currently” most accepted, however controversial, cosmological model. Btw, I do try to stick with the everyday worldliness and try to avoid Plank’s and LC’s world whenever possible.
        Just kidding, I do look forward to your math explanations of the “un-worldliness” and because of your posts, I have forced myself to visit the likes of Susskind, Write, Baez etc. Now I find myself looking out the sides and back of my head attempting to avoid the past.
        Anyway, while I said “more confined @12+by”, I was relating to the smaller space at the point of continued “expansion” following the last scattering (1st stars per the graphic and discussion in the article) and progressing to now. Now I realize that what happened in between is under the microscope, err, telescope/spectroscope and largely unknown and that is the point of this article and from what I get from it, it sounds like we’re moving forward and hopefully I’ll be around a few more years to share the excitement of these and many other discoveries that are on the horizon. Any word from Switzerland lately?

    2. Galactic coalescence occurs for galaxies in a cluster. These are gravitationally bound to each other and are in mutual orbits. The dynamics is no different from the motion of a binary star system; it is just bigger, more massive and the orbital period much longer. If the orbit takes the galaxies close to each other they then interact with each other. Most of a galaxy is dark matter which does not interact by means other than gravity — for all practical purposes. So the DM halos will continue merrily along their dynamic path. The luminous stuff is mostly gas, which does interact and friction slows that down so it sticks at a collision spot. Stars are somewhat slowed down, but they also largely continue onwards. So about 20% of galactic galactic mass is stopped, the 70% DM continues onwards, and stars continue with some drag. So this is a sort of splitting up — a crash.

      What about the huge black holes (SMBH)? They are concentrated masses, a bit like bullets, and they continue onwards. Gravity is a conservative force, so there is nothing, other than a direct BH collision, about gravity which will somehow cause the SMBHs to They will largely continue along their orbit with the DM halos around them. Depending on the relative velocities of the two galaxies the luminous matter that sticks together can gravitationally change the dynamics of the DM halo and SMBHs so as to put them in a tighter orbit. Eventually the whole gemish might settle into a larger elliptical galaxy.

      The break up of structure and the separation of the SMBHs from luminous matter that undoubtedly occurs would turn AGN activity off. These results are not that surprising.


  9. “Black holes power AGN; the term AGN referring more to the observational fact that some galaxies appear to emit radiation from their core regions in an amount and with characteristics that cannot be explained merely by the presence of stars or associated activity…”

    The activity in the centre of the AGN galaxies is caused probably by the accelerating ionised gases whizzing madly around their galactic centres, therefore emitting the radiation. Variability in the black hole in the centre is consuming stars and gas at a rapid rate, in which the matter “screams” before reaching the event horizon.

    If galaxies absorb little ones to make the ‘mother’ galaxy bigger, the material mostly avoids the core, but is torn apart by the gravitational forces – gas and stars are either spewed in all direction and/or are taken by the host galaxy. Assuming the little galaxies have black holes, the gravity is exactly the same as its acts on the stars and the gas. These injected black holes just orbit within the hot galaxy and it is highly improbable that they will encounter the central black hole engine.

    This explains this paper, where the central engine driving the AGN is independent from galaxy mergers.

    Fozzie view is probably correct here.

    Steve also said;

    “This should eventually run out of steam if there is no fresh supply of gas and dust to populate the accretion disk. However, the link between galactic mergers and AGN might be more correlation than causation?”

    I disagree. The food supply of the central core is being supplied by the gross disturbance near and away from the core, where the gas and stars are not orbiting in nice near circular orbits but in highly elliptical ones. Stars and gas plunge sometimes into the inner bulge then move back beyond it. Each close approach mixes up the core, causing some material or stars towards then into the black holes grasp, thus forming an AGN that can last for a very long time. (We could also assume in galaxy formation that the absorption of lots of merging galaxies just again stir up the mix. If there is no big central black hole, then some galaxies will only form normal galaxies, and this process does little more than stir up the gas or force gases to collide inducing star formation. Joseph Silk and a few others in the 1980s had very similar ideas in this kind of process.)
    As Lawrence said; “It also requires a near dead on impact trajectory for the two black holes to actually merge.”; but the question is what is the probability for that to happen? In actuality it is so infinitely small, that it is not worth considering. I mean. The area of a one million solar mass central black hole might be about a single parsec or so, The galaxies in which these monsters exist are +30,000 parsecs across; in compared areas the changes of close encounters with another monster hole is incredibly small.
    Also the inference here is clear too. The monster black holes place in many, if not the majority, of galaxies, were stared in the earliest periods of the big bang and formed very quickly when the mass was close enough together to create them.
    In a nutshell, as this paper says; AGN or ‘normal’ galaxies were born that way/ They were likely not created by evolutionary events such as mergers.

    Note: The crazy notons by our mate / jimhenson / aka Muppet in Jon’s recent story “First Observational Evidence Other Universes?”. Here he talks about colliding black holes in the earliest stages of the universe; he is alluding to the earliest formations of these monster black holes for the precursors of galaxy formation. Its a familiar concept quite difficult to observationally confirm. He must of read i somewhere. (By the way, his quite ludicrous conclusion of; “the largest collision of this stream triggered what we see as the big-bang” is utter nonsense.)

    Anyway, I can add sources to many of these points, but I’ve prattled on long enough…


    1. Oops! I said “The area of a one million solar mass central black hole might be about a single parsec or so,”

      By this I mean the gravitational influence dragging the material and accelerating the gas. I don’t mean the accession disk or the size +million solar mass event horizon. Sorry.

    2. The coalescence takes a long time to happen. In the immediate aftermath of the collision the DM halo and black holes continue to orbit as if almost nothing happened. The luminous matter will however clump up fast and the complicated gravitational interplay will see that material chewed up like a piece of gum. This does rob the kinetic + gravitational energy of the DM halo and black holes (eg friction) so the DM halo will eventually settle into a single clump and the BHs will settle in as well. The resulting galaxy will then have two SMBHs that will over time have their orbital energy decrease, due to heating up of luminous matter, and they may after many billions of years coalesce.

      A dead on collision at the start is extremely unlikely. A black hole has about 1.5km Schwarzschild radius per solar mass. A one million solar mass BH has a radius only somewhat larger than the radius of the sun. A billion solar mass BH has its horizon out to about where Saturn is. Compared to the scale of an entire galaxy that is very small.


      1. Great explanation and I appreciate it. I’ll add orbital mechanics on my to do list. lol

      2. The orbit of two galaxies is a classic two body problem in Newtonian mechanics. Where things diverge is due to the size of these galaxies, so that as they get close to each other tidal forces and other processes start to kick in. However, since DM only interacts by gravity (FAPP), and black holes (even huge ones) are little particles or bullets they orbit on as if virtually nothing is going on.

        The system does eventually settle down. The interactions between luminous matter, in particular gas, converts a lot of this orbital energy into heat and is a source of “friction.” So given some number of billions of years the two galaxies do eventually coalesce into a single elliptical galaxy.


  10. AN IMPORTANT NOTE: Also read Nancy Atkinson’s later story on JANUARY 5, 2011; “Star Birth and Death in the Andromeda Galaxy”
    Here the X-ray image supplied in the story shows the real violence of the colliding ionised gases in the core. The material (and stars) are being mixed up the variable gravitational field likely being destabilised by galaxies mergers.
    If you look at the XMM Newton’s view in X-Ray. at the inner core of the central bulge the area is quite minuscule compared to the bulk of the galaxy!

    1. I don’t mean to be a pest but this is intriguing and I thank you as well for your help. I don’t know the density but doesn’t this Alpha H filtered image show a substantial amount of far reaching gas surrounding the core as well as what looks like a collision remnant somewhat perpendicular to the galaxy plane? I did get permission to include the link.

Comments are closed.