Thursday, April 08, 2010

OJ 287

OJ 287 is a quasar at a distance of approximately 3.5 Giga light-years with a quite unremarkable Wikipedia entry. I recently heard a talk by Mauri Valtonen from Tuorla Observatory about OJ 287 and was totally enchanted! That quasar deserves better than this little Wikipedia entry, so here's to OJ.

The figure below, from arXiv:0809.1280, shows the apparent magnitude of OJ 287 as recorded over the course of time:

[Click to enlarge]

There is several things one should note about this figure. First, a star's magnitude is a logarithmic measure (similar to the magnitude of an earthquake). That means the variability in the brightness ranges over many orders of magnitude. Second, the first measurements date back more than a century! In his talk, Valtonen explained that to obtain this data they have to sort through the old archives in observatories and find the photographic plates that might have captured the system back then. (Or rather, some student has to do this.)

To look at the physics, even by eye you can see some patterns in the data. There seems to be a long-term variation on about 60 years scale, with one peak at about 1910 and one at about 1970. Then there's a shorter variation with peaks about every 12 years. If you look closely, you can in the newer data also see that the peak is actually a double-peak. The more sophisticated way to extract these regularities is to do a Fourier-analysis of the data, which Valtonen and his collaborators did to quantify these periodicities. (Chad Orzel did a good job explaining Fourier transformation here.)

Sitting on this (and more) data, the physicist of course starts to wonder what might cause this particular variability. So Valtonen and his coworkers set up to develop a model to fit the data, and make predictions to test it. To make a long story short, here's what they came up with: OJ 287 is a binary system. It's a supermassive black hole with an accretion disk, and a smaller black hole in orbit around it. The masses of the objects can be inferred from the data, rspt they are parameters in the model. The slow accretion of the gas in the disk onto the large black hole and the outgoing jet is what makes the object visible. But in addition to this, on its orbit the small black hole will pass through the accretion disk, causing the gas to dramatically heat up. The heated gas flows out on both sides of the disk and radiates strongly for a time of several weeks.

The smaller black hole passing through the accretion disk is what causes the peaks in the brightness. It's a double peak because of the eccentricity of the orbit, and the long-term variation is due to the perihelion shift which for this system is quite sizeable. Below is my sketch of the situation. The red thing is supposed to be the accretion disk, and the black dots are the two black holes.

You find an actual fit to the data in Predicting the Next Outbursts of OJ 287 in 2006-2010, M. J. Valtonen et al 2006 ApJ 646 36. So now that one has a model that fits the available data, the good physicist goes and makes a prediction. That was done in the paper I just mentioned. It should be added that at this time there were other models for the system's variability, such as variations in the accretion disk or a wobble in the jet axis (the jet-axis of that system is almost exactly directed at us, which makes our observation very sensitive to changes in that axis.) So they made a prediction with the binary-system-accretion-disk model that the next outburst would be on September 13 2007, plus or minus 2 days uncertainty. And here's what happened:

[Figure 1 from arXiv:0809.1280]

The black squares is the data, the dashed line is the theoretical prediction. I was so impressed by this! I mean, look, that thing is Giga-lightyears away! There's all sorts of plasma physics and turbulences and weird things going on there. And they manage to predict the day the next outburst will take place.

But it's even better than that because the timing of the burst contains a big chunk of general relativity too: The system loses energy due to the emission of gravitational waves. This causes the orbit of the small black hole to shrink and the rotation to speed up. If the system would not be losing this energy, the September 2007 outburst would have been 20 days later.



  1. Dear Bee,

    thanks for sharing :-). I have never heard of OJ 287 before - a pity, as it is a fascinating object indeed!

    Cheers, Stefan

  2. If Sept 2007 was the emergent flare, the next entering flare is in 2016. Patience - and a space telescope or two taking a long hard look across the spectrum!

    Any gravitation theory competing with GR requires a surpassingly perfect strong field ansatz. Either that, or a GR founding postulate must be selectively falsified under heretofore unobserved circumstances.
    To criticize is to vounteer.

    Oh to be around in 10,000 years when the 10^8 solar mass BH merges with its 1.80x10^10 solar mass primary.

  3. Nice post, but I have some reservations. First there is really no way to tell from your post or the paper how impressive the prediction really is as we cannot see the Fourier transform. If there is a clear periodicity on FT then predicting next event may be easier then it looks based on time-domain data.

    The actual graph in the arxiv paper is quite different from the one embedded in this post and it's much less spectacular in that there are clearly additional peaks trailing the first one which the model fails to predict.

    It is far from obvious whether the model can be fixed to do it. Perhaps if one changed the disk into rings and made it coplanar with orbiting BH?

    Finally I'm also not so sure about the GR comment at the end of the post, the information whether there is energy loss should already be encoded in the data, so how come they can make an additional "no-GR waves" prediction? If the system were losing energy in the past it will continue to do so. The next event should be determined by the kinematics alone.

  4. Hi Tspin:

    The results of the Fourier trafo are in the ApJ paper (it's open access), see table 1, Fig 2 & 3, Section 2.2 & 2.3 (I left out most details in the post for obvious reasons). The point is that the events are not strictly periodic (otherwise the prediction of the time of the outburst would have been easy, and pretty much model independent). If you look at my sketch (second figure in my post), you see why it's not strictly periodic. The prediction was (as I wrote) Sept 13 +/- 2 days, which was an exact hit. The competing models were not able to make that prediction. Note that I carefully wrote I was impressed they were able to predict the timing, as I'm not sure how well the magnitude was predicted in advance.

    For what the GR comment is concerned, that is content of the Nature paper. I don't understand your objection. You have a model for the system. It's a binary system, there's some parameters for the orbit and the masses. There's some more parameters for the disk and the model describing the outburst caused by passage through the disk and so on. You have some equations describing that orbit. These differ whether or not there's an energy loss through gravitational radiation. You take the best fit parameters to either orbit. Then the prediction for the next outburst differs for whether you have an energy loss or haven't. Note that the data previously available was just barely enough to determine the parameters of the model and make the prediction. Thus you couldn't have inferred from the previous data whether or not there was an energy loss. Best,


  5. Hi Bee,

    Thanks for this truly interesting piece about the observations and progress in understanding celestial objects like OJ 287. I do wish however they would offer up a more inspirational name for something that is so immensely attractive in nature. That is I can just see myself trying to strike up a conversation today with someone beginning with saying ‘what do you think about all this stuff about QJ 287’;-). You must admit for being the name for what stand as as being the mother of all black holes its somewhat uninspiring.

    However it is amazing really just how far in the last while we’ve gone in understanding what not long ago Einstein said just couldn’t exist. Then still it is possible, although the mass of such objects are undeniably what they are, what being truer of their natures is still much in question. I also find it interesting that with examining things from the opposite end of the size scale, with the LHC (something having another non inspiring name) that this may have us better understand whether we have things right or not. Not that I claim to know better, yet I’m thinking not.



  6. I won't be able to drink another glass of orange juice now without remembering binary blackhole systems. How cool is that?!

  7. Regarding competing GR theories ....

    MOND has allegedly been recently falsified.

    What about John Moffat's MOG, is that still in play? (This is a response to Uncle Al's reply, which got me thinking ...)

  8. Phil Warnell said... "However it is amazing really just how far in the last while we’ve gone in understanding what not long ago Einstein said just couldn’t exist."

    Perhaps it was the 0-D, singularity, of no size at all but enormous mass which Einstein objected to?

    And isn't it true that a zero-dimensional object, aka singularity, can never be proved to exist? So in that case, can't say Einstein was wrong.

    Perhaps matter can collapse to an extremely small and dense 3-D object, which can *mimic* a hypothetical singularity, but which is not a 0-D singularity.

    Only in the past year have I begun to have appreciation for the idea that, in general, mathematicians who also "do physics" may have very different mental processing and view of reality compared to physicists who also use math in their work.

    But I notice that other people have that conception too ... including about black hole theory:

    "The fanatics of the religion of Mathematical Physics routinely talk about zero-dimensional 'entities.' One of these is the infamous singularity, usually defined as the number of beers it takes for a mathematician to say that the laws of Physics break down or are undefined or something stupid like that.

    In Physics, there are no objects that are 0-D. We're done!

    But assuming that a singularity is 0-D, we must use this proposal consistently. Relativists are famous for glossing over definitions or tailoring them retroactively to suit their theories. The 0-D singularity suddenly acquires dimensions out of the blue:

    "Calculations using Kerr geometry describe the singularity as ring-shaped." [28]

    Perhaps the mathematicians who came up with this nonsense reasoned that if they jiggle the non-dimensional singularity of a static black hole they would end up with a 3-D singularity in a dynamic black hole. This is again a case of too much Math and a total disregard for the basics of Physics. In the real world, as opposed to the fantasy world of Mathematics, you can shake your 0-D love as much as you want. You will never end up with a 3-D heart!"

    LOL, there is more, from where that was quoted, here:
    Although blatantly extreme, I found it very humorous, and thought provoking.

  9. Hi Bee, thanks for clarification.

    I just thought it unlikely they had enough data to predict next flare but not enough to detect energy loss due to GR waves. But if that's the case then ok.

  10. William:
    Actually, Kerr singularities are two-dimensional.

    The three-dimensional version would be the toroidal singularity at the heart of a (purely mathematical) toroidal black hole. A tbh has major-axis rotation keeping the ring open, and minor-axis rotation keeping the ring "fat".

    There've been some research papers on these beasties, but they're a bit obscure - partly because some mathematicians use a different definition of "toroidal" to the rest of us, so papers tended to end up with the word "torus-like" instead. Which kinda messes up database searches if you're using "toroidal black hole" as a keyphrase.

  11. Hi tspin,

    I read through the paper "Predicting the Next Outbursts of OJ 287 in 2006-2010" today and was a bit disappointed that I couldn't find a very precise prediction - but this seems not to have been the point of that paper. Instead, it discusses in detail the complicated processes responsible for the actual light curve, which has quite a complicated "fine structure". In particular, there are many possible reasons for outbursts, and good data are necessary to pinpoint the typical signature of the impact outbursts.

    The prediction which was tested in September 2007 was published in April of that year: Mauri J. Valtonen: "New Orbit Solutions for the Precessing Binary Black Hole Model of OJ 287" (ApJ 659 1074 - it's also open access, so you can have a look).

    In this paper, the previous impact outburst of November 2005 was used to refine the details of the orbit of the secondary black hole, so that a prediction for September could be made. Now, the orbital parameters were known well enough to discriminated between the cases with and without energy loss by gravitational waves.

    From the abstract:

    The parameters of the system to be determined are the orbital period, fixed by the separation of the 1947.30 and 1983.00 outbursts; the orientation of the major axis of the orbit at a given time, fixed by the 1972.99 outburst; the time delay factor, which is a function of the disk thickness, fixed by the 2005.82 outburst; and the precession rate of the binary, fixed by the 1913.02 outburst. A unique solution is found for both the case of gravitational radiation and that of no gravitational radiation. The 2007 September outburst begins 2007 September 9-16 in the former models and 2007 October 8 in the latter model.

    Cheers, Stefan

  12. Phil, Bee, etal, have you seen these? ...

    Black holes don't exist:

    A primer in important math concepts:

    Besides making me laugh frequently, I think there are some very good and valid points made there. Yes? No?


  13. William: Stopped watching at 1:46. The defining property of a black hole is not the singularity but the horizon. You might find this interesting. Apologies if that came later in the video. Best,


  14. This comment has been removed by the author.

  15. I found that interesting, but one sentence in the first long paragraph confused me based on my understanding, possibly misunderstanding, of black holes:

    If the gravitational force on the surface gets so strong not even light can escape, we call this surface an event horizon.

    Here is my confusion. Irrespective of what a black hole actually is (a singularity "point"? A really dense ball?), the single most important part of a BH is a sphere of space a certain distance from the black hole's center known as "the event horizon". I think we all agree on that.

    But the event horizon is not the "surface" of the black hole (for Russians: frozen star), mainly it's a "shell" of a certain radius (beyond the black hole "surface") inside of which, as you said, nothing can escape not even light.

    If I am wrong pls correct my mis-knowledge Bee, as you have taught us much and we thirst for more, thanks in advance.

  16. Hi Steven,

    What you say is correct. I don't know what confused you about my sentence. The word "surface" doesn't imply there actually "is" something at this surface. Every point in space is on some surface. It's just a subset of points. (Every point "is" on some "holographic screen" harharharharhar.) Best,


  17. What confused me Bee was the last bits: "we call this surface the event horizon" which relates earlier in the sentence to "on the surface." It's a silly semantic thing Bee, I'm sorry for being a weenie on this, and hoo-boy I'm glad I was right as Black Holes ain't my current thing (but they will be soon ... I'm just re-remembering/studying that which I learned/loved/forgot about Special Relativity years ago ... the Gen Rev stuff is next). :-)

    But again, with each comment you challenge my brain and wonderfully so. You don't like Holographic/ AdS/CFT?

    I like it. And Superstrings too. As interesting ... Geometrical Mathematics!

    Not necessarily Physics.

    And if I'm right and they're not (as I strongly am leaning they're not): Physics (better word: Reality), then the last GREAT achievement in Physics was 1973's Discovery of Asymptotic Freedom by Politzer, Wilczek, and Gross.

    Thank goodness for Astronomical Technology and Computer Science. They've joined Mathematics as Physicists' best friends. Not sure the Mathematicians are happy about sharing. Sorry, Witten.

  18. Hi Bee,

    If we look at things simplistically, just before a mass shrinks below its Schwartzchild Radius, the greatest gravitational force encountered would be what it is at the surface of the collapsing sphere, with the force diminishing to zero as one moves to the centre of the mass, since the pull of gravity (in the Newtonian perspective) would have the force increasingly offsetting itself the closer one gets. So isn’t it somewhat misleading to say all is being drawn to a singularity, rather than saying things are had to be drawn to pass through a single point and that point until total collapse would remain as having zero gravity. So even here there seems to be some contradiction as the surface during collapse will always be the region having the greatest level of force, while the centre remaining at zero until one reaches a point that is mathematically undefined as all reaching the singularity.

    Shades of the Poincare's Conjecture, for which our Russian Mathematical genius has now proven that no matter how small something is diminished to be in three dimensions, at its limit will still remain as being a sphere, as opposed to a point. The central contention of Schwarzschild’s conjecture being there is no known force that can resist such a eventual end in collapse, yet I wonder if it being a mathematical impossibility would serve just as well.

    Then of course there are many things in mathematics that appear as simply impossible, such as the Banack-Taski Paradox ,which allows that a sphere can be broken apart and then reassembled to be a sphere of any greater or lesser size you could have imagined, as to able in having a pea become the size of the sun. Which has one to question what is the actual meaning, when considering the sphere, as which can we find as being small and which being large;-)



  19. Hi Phil,

    Well, I wouldn't say stuff is "drawn" into the singularity because the singularity doesn't exist on all space-like hypersurfaces, thus you could plausibly say it doesn't exist yet, so how can stuff be drawn into something that doesn't exist. It would be more accurate to say that all world-lines end in the singularity (see posts on causal diagrams) or that matter density reaches infinite values. Thus, yes, your account is more accurate. Best,



COMMENTS ON THIS BLOG ARE PERMANENTLY CLOSED. You can join the discussion on Patreon.

Note: Only a member of this blog may post a comment.