Identity parade clears cosmic collisions of the suspicion of promoting black hole growth

January 5, 2011
Selected galaxies from the COSMOS survey. Credit: NASA, ESA, M. Cisternas (Max-Planck Institute for Astronomy)

( -- What happens when galaxies crash together? For years, these cosmic collisions have been blamed for triggering violent outbursts at the hearts of galaxies. Now, a remarkable piece of detective work has given a verdict: galactic mergers do not usually whet the appetite of the black holes that power these active galactic nuclei, meaning other, less dramatic phenomena are responsible.

Most galaxies, including our own, have a huge but well-behaved black hole at their heart, while some have messy eaters that suck in vast amounts of matter which then shines brightly as it falls towards oblivion. This causes the telltale bright spots at the centre of galaxies known as (AGN). Why are the two types so different? Until now, the leading theory has been that mergers between galaxies are instrumental in driving matter into the , making them grow.

In a new study, the largest of its kind so far, astronomers set up an identity parade of galaxies to test this theory. Comparing 140 active galaxies with a control group of over 1200 comparable inactive galaxies, they found that there has been no significant link between AGN activity and for at least the past eight billion years. Therefore, other phenomena such as instabilities within galaxies, collisions of molecular clouds or tidal disruption by other galaxies flying by must instead be to blame.

The results will be published in the Astrophysical Journal on 10 January.

The video will load shortly.
This video offers a zoom on the largest sky survey ever carried out by Hubble: COSMOS. The COSMOS survey is a detailed map of a patch of sky roughly 10 times the size of the area covered by the full moon, which has been studied by Hubble and other telescopes at different wavelengths. Data from the COSMOS survey has been used for much important research, including creating a 3-dimensional map of the Universe’s dark matter distribution, and a study into the relationship between major mergers of galaxies and active galactic nuclei. Credit: ESA/Hubble (M. Kornmesser & L. L. Christensen)

The emission of radiation from active galactic nuclei is driven by the behaviour of matter such as gas clouds and even stars as it heats up and falls into the galaxy's supermassive central black hole. But an open question in the physics of active galaxies is precisely how matter crosses the final few hundreds of light-years to reach the immediate neighbourhood of the black hole before being swallowed.

Team leader Mauricio Cisternas from Germany's Max Planck Institute for Astronomy explains: "A study of this scope has become possible only recently, as the large surveys undertaken using the Hubble Space Telescope have become available. These have given us a huge sample of galaxies, both active and inactive, meaning that we can now study many distant galaxies in exquisite detail. Before these surveys, we hadn't examined many active galaxies at large cosmic distances in sufficient detail."

Cisternas and his team chose 140 active galaxies from the COSMOS survey. The COSMOS field is an area of sky roughly 10 times the area covered by the Moon, in the constellation of Sextans (the Sextant), which has been comprehensively mapped by Hubble and other telescopes at different wavelengths. It contains several hundred thousand distant galaxies of all types. The team was able to identify active galaxies from among these using X-ray observations from ESA's XMM-Newton space telescope, and they then studied the more detailed optical images of them taken by the NASA/ESA Hubble Space Telescope.

For each of the in the study, they selected nine non-active galaxies at roughly the same distances, and thus roughly in the same stage of cosmic evolution, from the same Hubble images. This gives a grand total of just over 1400 galaxies that the team could then test for the telltale signs of mergers.

"You can usually tell when galaxies have been involved in a merger," explains Knud Jahnke, co-author of the study. "Instead of the neat, geometric spiral or smooth elliptical shapes you usually see in Hubble images, colliding galaxies typically look distorted and warped. We planned to find out whether these misshapen galaxies were more likely than regular ones to host active nuclei."

Identifying whether or not a galaxy is distorted is a matter of judgement for which the expert eye of a trained astronomer is far better than any computerised assessment. To harness this human expertise without introducing the risk of unwitting bias, Cisternas set up a kind of identity parade of galaxies, in which he had modelled and removed the bright spot that reveals the AGN. Ten galaxy experts, based at eight different institutions, independently assessed whether each of the was distorted or not, without being told which had an AGN.

None of the experts found a significant correlation between a galaxy's activity and its distortion, that is, between its black hole being well-fed and its involvement in a major merger.

While mergers are a common phenomenon, and are thought to play a role at least for some AGN, the study shows that they provide neither a universal nor a dominant mechanism for feeding black holes. By the study's statistics, at least 75%, and possibly all, of AGN activity over the last eight billion years must have a different explanation. Possible ways of transporting matter towards a central black hole include instabilities of structures like a spiral galaxy's bar, the collisions of giant within the galaxy, or the fly-by of another galaxy that does not lead to a merger (known as galactic harassment).

Could there still be a causal connection between mergers and activity in the more distant past? That is the next question the group is gearing up to address. Suitable data is bound to come from two ongoing observational programmes (Multi-Cycle Treasury Programs) with the Hubble Space Telescope, as well as from observations by its successor, the James Webb Space Telescope, which is scheduled for launch after 2014.

Explore further: Rain of giant gas clouds create active galactic nuclei

Related Stories

Rain of giant gas clouds create active galactic nuclei

July 8, 2010

Galaxies like our own were built billions of years ago from a deluge of giant clouds of gas, some of which continue to rain down. Now new calculations tie the rain of giant clouds of gas to active galactic nuclei (AGN), the ...

How often do giant black holes become hyperactive?

December 20, 2010

( -- A new study from NASA's Chandra X-ray Observatory tells scientists how often the biggest black holes have been active over the last few billion years. This discovery clarifies how supermassive black holes ...

Active galaxies are different near and far

January 6, 2009

( -- An ongoing X-ray survey undertaken by NASA's Swift spacecraft is revealing differences between nearby active galaxies and those located about halfway across the universe. Understanding these differences will ...

The Universe: It's not as violent as we think

September 17, 2004

The Universe has experienced far fewer collisions among galaxies than previously thought, according to a new analysis of Hubble Space Telescope data by an ANU researcher. Astronomer Dr Alister Graham, from the Research School ...

Recommended for you

Solar minimum surprisingly constant

November 17, 2017

Using more than a half-century of observations, Japanese astronomers have discovered that the microwaves coming from the sun at the minimums of the past five solar cycles have been the same each time, despite large differences ...


Adjust slider to filter visible comments by rank

Display comments: newest first

not rated yet Jan 05, 2011
You know, this video made me re-think of something that I was curious about in high school - back 30+ years ago: How is it that we and presumably an infinite number of others in a shell around the event can see the same point of light in space 14 billion light years (82,175,990,400,000,000,000,000 miles) ago? I know we are seeing very old photons, but how many photons would have to be generated from a single light source so that an infinite number of observers in a shell 14 billion light years from the source - all see the same thing? That is a long ways, and the spread over that distance would seem to indicate an incalculable photon density at the source. This just seems a curiosity to me.
4.5 / 5 (2) Jan 05, 2011
That is a long ways, and the spread over that distance would seem to indicate an incalculable photon density at the source. This just seems a curiosity to me.

You can calculate the amount of power of a distant EM source through the usage of the inverse squared law, in meters. The farthest object I know of was allegedly 13.1 billion ly distant.

Essentially, use the formula for surface area of a sphere.

So convert ly to meters and then square this number gives: 1.536 *10^52 meters squared, and multiply by 4pi.

Then multiply by the amount of energy per second you obtain across 1 square meter of detector surface, this takes into account all the photons moving in directions that don't intersect the individual viewer. This will give the spherical power output of a point light source at the object's location.

The Sun is approximately: 3.82*10^26 watts.

Based on distance and apparant magnitude, Sirius should be: 1.014*10^28 watts.
not rated yet Jan 05, 2011
For a ridiculously distant object with apparant magnitude 30, the Sun would APEAR to be 1,927,823 times brighter. However, these objects are 13 billion light years distant. By the time you reverse the formula for Newtons' law, you find that the most distant quasars are some ~8*10^45 times more powerful than the sun in terms of the entire spectrum of radiation they produce. They produce about 145 times as much energy in a second or so as the Sun will in it's entire life span.
not rated yet Jan 05, 2011
Yet, however minutely in distance we move a detector around the circumference we still get photons from the same source. It seems that the limitation is our ability to sense, as we use larger lenses to grab more photons or larger arrays to grab more EM to increase signal gain. But the signal is still there, everywhere. Or, do we use larger lenses (just considering light here) because there are larger gaps between the photons as the distance from the source increases?

not rated yet Jan 05, 2011
yes there would be larger gaps, thats why sometimes a multiple setup of telescopes in a large baseline are used to get our head around it
4.7 / 5 (3) Jan 05, 2011

You also have to realize that many exposures of these very distant objects can take hours, days, or weeks for the telescopes to gather enough photons to present a clear picture.

I used to kind of wonder along the same lines as you. I had a thought that perhaps instead of a photon being a single little wavelet, it is more of a 3 dimensional spherical wave emanating in all directions from it's point of origin, thus allowing an infinite number observers to detect it.

However, given the research conducted on the quantum nature of light, I have not seen any evidence or suggestion by any of the scientific community for this being the case.

The most likely explanation is that these light sources just pump out truly immense quantities of photons.

There is probably some(vast) distance from a lightsource where you can no longer detect it due to the improbability of collecting any of it's far flung photons, let alone accurately tracing their origin back to a source
not rated yet Jan 06, 2011
Have we witnessed actual stars eaten by black holes yet?
not rated yet Jan 06, 2011
So, there are points of reference where you can be focusing a sensor on a source of energy, but simply won't detect it - provided you have sufficient photon spread due to distance. There just won't be any photons to detect - or would it just be a matter of time before a photon comes along?

I'm not a scientist nor a physicist, I'm just a person of curiosity. But, I'd have to wonder at the mass represented by a sphere (cloud?) of photons 26 - 28 billion light years in diameter via the good 'ole E=MC^2. Yes, I know photons are supposed to be massless and come in different frequencies, but still - that a photon can be absorbed and by doing so change the energy level of an atom enough to release electrons tells me that something has to be there of value in the photon to add enough energy to to do this. It just seems natural that some substrate must exist in the photon to store its energy.
1 / 5 (2) Jan 06, 2011

You are so correct to question.

LIGHT travelling unimpeded is not unimpeded. It hits Suns, planets, moons, asteroids, etc.,etc., and is absorbed. Yet some LIGHT will continue on. Scientists believe LIGHT does not "die."

As for density, is there one? Is there a build-up or a singularity of composition? As for a souce, there is no one source; Suns are immemorial.
not rated yet Jan 06, 2011
For a ridiculously distant object with apparant magnitude 30, the Sun would APEAR to be 1,927,823 times brighter. However, these objects are 13 billion light years distant. By the time you reverse the formula for Newtons' law, you find that the most distant quasars are some ~8*10^45 times more powerful than the sun in terms of the entire spectrum of radiation they produce. They produce about 145 times as much energy in a second or so as the Sun will in it's entire life span.

So, is there a correlation where the younger universe had more energy available than our present one?
5 / 5 (5) Jan 06, 2011

Photons indeed have mass (they have momentum, which has the classical formula of p=mv). They just don't have a rest mass: you can't "stop" a photon, put it on a balance, and weight it that way -- a photon must by definition always travel at the speed of light for as long as it exists.

The energy-mass conversion, as you point out, has a proportionality constant equal to c^2. If you were to take just 1 kg of matter, and convert it completely to energy, how many visible photons is that?

Let's say we choose a wavelength of 475 nm (blue color). Photons at that wavelength carry 2.615 eV of energy each, which translates to 2.9x10^-17 eV/c^2 of mass. 1 kg of matter is equivalent to ~5.6x10^35 eV/c^2 of energy.

So to have the equivalent of 1 kg of matter, you'd need to collect ~1.9x10^52 blue photons (of wavelength 475 nm.)
5 / 5 (6) Jan 06, 2011

You can think about photons as spatially confined wave packets traveling in "straight lines" or more correctly along geodesics in spacetime (the "particle" view) or as perpetually expanding spherical emanations that instantaneously collapse to a single random point upon measurement (the "wave" view according to the Copenhagen Interpretation.)

To me, the corpuscular view is easier to visualize and think about on cosmic scales, though mathematically the two are equivalent (just different sides of the same wave-particle duality coin.)

From the corpuscular view, a light source spits out a swarm of photons in every direction, and as you get farther away from it, the distances between photons grow. Eventually, per square meter of detector area you start seeing a "trickle-in" effect where individual photons arrive sporadically, one-by-one, separated by long periods of no signal at all. That's when it takes days or weeks for enough photons to trickle in, to make a picture.
5 / 5 (1) Jan 07, 2011

So, can we determine the distance from an energetic object based on photon count over a known sensor size over time? Provided of course you can group a singular object or neighboring objects via frequency or other method to classify photon types from a specific source.
5 / 5 (5) Jan 07, 2011
can we determine the distance from an energetic object based on photon count
Not quite that easily. To do this, you will need additional information. Namely, the intrinsic brightness of the object in question (i.e. how many photons it puts out per unit time in the first place), as well as any optical barriers (e.g. interstellar gas and dust) between you and the object that might absorb the photons, or scatter them away.

For Type 1A supernovas, the intrinsic brightness happens to be confined to a narrow range, due to the physical mechanisms that drive these explosions. So, Type 1A's are useful as "standard candles" (also easily identifiable by a very particular spectrum of emission and luminosity vs. time curve), and you can indeed estimate their distance based on their apparent luminosity.

But for most other objects, there is no reliable way to know what their true luminosity is. For very far-away objects, people rely mostly on redshift to estimate their distance.

Please sign in to add a comment. Registration is free, and takes less than a minute. Read more

Click here to reset your password.
Sign in to get notified via email when new comments are made.