Thursday, November 12, 2015

Mysteriously quiet space baffles researchers

The Parkes Telescope. [Image Source]

Astrophysicists have concluded the yet most precise search for the gravitational wave background created by supermassive black hole mergers. But the expected signal isn’t there.

Last month, Lawrence Krauss rumored that the newly updated gravitational wave detector LIGO had seen its first signal. The news spread quickly – and was shot down almost as quickly. The new detector still had to be calibrated, a member of the collaboration explained, and a week later it emerged that the signal was probably a test run.

While this rumor caught everybody’s attention, a surprise find from another gravitational wave experiment almost drowned in the noise. The Parkes Pulsar Timing Array Project just published results from analyzing 11 years’ worth of data in which they expected to find evidence for gravitational waves created by mergers of supermassive black holes. The sensitivity of their experiment is well within the regime where the signal was predicted to be present. But the researchers didn’t find anything. Spacetime, it seems, is eerily quiet.

The Pulsar Timing Array project uses the 64 m Parkes radio telescope in Australia to monitor regularly flashing light sources in our galaxy. Known as pulsars, such objects are thought to be created in some binary systems, where two stars orbit around a common center. When a neutron star succeeds in accreting mass from the companion star, an accretion disk forms and starts to emit large amounts of particles. Due to the rapid rotation of the system, this emission goes into one particular direction. Since we can only observe the signal when it is aimed at our telescopes, the source seems to turn on and off in regular intervals: A pulsar has been created.

The astrophysicists on the lookout for gravitational waves use the fastest-spinning pulsars as enormously precise galactic clocks. These millisecond pulsars rotate so reliably that their pulses get measurably distorted already by minuscule disturbances in spacetime. Much like buoys move with waves on the water, pulsars move with the gravitational waves when space and time is stretched. In this way, the precise arrival times of the pulsars’ signals on Earth gets distorted. The millisecond pulsars in our galaxy are thus nothing but a huge gravitational wave detector that nature has given us for free.

Take the pulsar with the catchy name PSR J1909-3744. It flashes us every 2.95 milliseconds, a hundred times in the blink of an eye. And, as the new experiment reveals, it does so to a precision within a few microseconds, year after year after year. This tells the researchers that the the noise they expected from supermassive black hole mergers is not there.

The reason for this missing signal is a great puzzle right now. Most known galaxies, including our own, seem to host huge black holes with masses of more than a million times that of our Sun. And in the vastness of space and on cosmological times, galaxies bump into each other every once and then. If that happens, they most often combine to a larger galaxy and, after some period of turmoil, the new galaxy will have a supermassive binary black hole at its center. These binary systems emit gravitational waves which should be found throughout the entire universe.

The prevalence of gravitational waves from supermassive binary black holes can be estimated from the probability of a galaxy to host a black hole, and the frequency in which galaxies merge. The emission of gravitational waves in these systems is a consequence of Einstein’s theory of General Relativity. Combine the existing observations with the calculation for the emission, and you get an estimate for the background noise from gravitational waves. The pulsar timing should be sensitive to this noise. But the new measurement is inconsistent with all existing models for the gravitational wave background in this frequency range.

Gravitational waves are one of the key predictions of General Relativity, Einstein’s masterwork which celebrates its 100th anniversary this year. They have never been detected directly, but the energy loss that gravitational waves must cause has been observationally confirmed in stellar binary systems. A binary system acts much like a gravitational antenna: it constantly emits a radiation, just that instead of electromagnetic waves it is gravitational waves that the system sends into space. As a consequence of the constant loss of energy, the stars move closer together and the rotation frequency of binary systems increases. In 1993 the Physics Nobel Prize went to Hulse and Taylor for pioneering this remarkable confirmation of General Relativity.

Ever since, researchers have tried to find other ways to measure the elusive gravitational waves. The amount of gravitational waves they expect depends on their wavelength – roughly speaking, the longer the wavelength, the more of them should be around. The LIGO experiment is sensitive to wavelengths of the order of some thousand km. The network of pulsars however is sensitive to wavelengths of a several lightyears, corresponding to 1016 meters or even more. At these wavelengths astrophysicists expected a much larger background signal. But this is now excluded by the recent measurement.

Estimated gravitational wave spectrum. [Image Source]

Why the discrepancy with the models? In their paper the researchers offer various possible explanations. To begin with, the estimates for the number of galaxy mergers or supermassive binary black holes could be wrong. Or the supermassive black holes might not be able to form close-enough binary systems in the mergers. Or it could be that the black holes experience an environment full with interstellar gas, which would reduce the time during which they emit gravitational waves. There are many astrophysical scenarios that might explain the observation. An absolutely last resort is to reconsider what General Relativity tells us about gravitational-wave emission.

 You have just witnessed the birth of a new mystery in physics.

[This post previously appeared at Starts with a Bang.]


Phillip Helbig said...

"An absolutely last resort is to reconsider what General Relativity tells us about gravitational wave emission."

I don't see how this is even a last resort. The observations you mention of the binary pulsar demonstrate energy loss exactly as predicted by GR when gravitational waves are emitted from such a system.

By the way, one of my pet peeves is the lack of a hyphen in two-word adjectives. So, above, it should be "gravitational-wave emission", since it is not the emission which is gravitational. Think of it this way: a high-energy physicist is a physicist who works with high-energy phenomena. The next time you meet a high energy physicist, ask him what he has been smoking. :-)

Phillip Helbig said...

Known pulsars ---> Known as pulsars

Sabine Hossenfelder said...


Thanks, I've fixed that. Yes, it would require a lot of fudging and fumbling to get rid off the binary slowdown by other means than GR. I'm sure it can be done somehow though, just introduce a few new parameters and you can fit anything ;) Best,


marten said...

Space is not quiet. Take for instance CMBR.

Noa Drake said...

Suppose we trust in the correctness of the GR prediction, could it be that these gravitational waves get absorbed after emission ? (Absorbed how ? Suggestions ? By the vacuum , curved spacetime itselfaround the system ?) Causing the waves to die out before reaching any areas measurable by us. Just a spur of the moment, I can take it if I get corrected , and learn something in the process.

radii said...

or there is no space-time and no gravitational waves

Thierry Periat said...

(citation) Gravitational waves are one of the key predictions of General Relativity, Einstein’s masterwork which celebrates its 100th anniversary this year.(end citation)

Yes but this prediction results in some way from a mathematical trick: a linearization. I don't contest the exactitude of the demonstration but I ask if it is meaningful to apply such simplification to a essentially non-linear phenomen (gravitation)? Why do we trust our mathematics in that case?

Would it not be intelligent to consider that mathematical rules applying in the vicinity of black holes should be modified? Thus modifying our predictions too?

(citation) Suppose we trust in the correctness of the GR prediction, could it be that these gravitational waves get absorbed after emission ? (Absorbed how ? Suggestions ? By the vacuum , curved spacetime itselfaround the system ?) Causing the waves to die out before reaching any areas measurable by us.(end citation) Although it is actually science-fiction, one could be seducted by the idea - vacuum as a metastable superfluid?

Soup-a-Chicken said...

For the unworthy, the paper is freely available on the archive.

עמיר ליבנה בר-און said...

What can cause SMBHs to merge less frequently than expected? (that is, either early in cosmic evolution or almost not at all)

Is the configuration of matter near galaxy centers relevant? If so, can this be a sign of an uneven distribution of dark matter?

Barry said...

It seems that this result can be easily explained in terms of adding large masses of gas to supermassive black hole mergers. In this way, the SMBH can trade angular momentum with the gas and cross the PTA frequency window faster than by GW emission alone.

Uncle Al said...

Inspiraling degenerate bodies have a well-characterized vast radiance of gravitational radiation. Propagation is not well-characterized, for space is filled with real (matter and radiation; phenomena of dark matter and dark energy) and virtual (Casimir effect to the Dirac sea) stuff. Consider pumping, damping, resonance, Q overall; vacuum Reynolds number and wave propagation. Spacetime as an inhomogeneously filled medium for gravitational wave propagation may exhibit refraction, dispersion, dissipation, dichroism, gyrotropy, scattering.... Effectively shielding EM radiation is high art, or drive under a bridge with your AM radio on.

Tom Andersen said...

The emission scenario feels pretty robust. Galaxies are quite large relative to the space they inhabit and interact often.

Could high frequency gravitational waves wash out the signal? If they can it sounds like a paper to come up with some spectra and intensities that would do it.

Sabine Hossenfelder said...


What you say is incorrect. A linearization is not a "trick" it's an approximation. It is perfectly valid up to some precision and the gravitational field of large black holes is weak. Besides this, black hole mergers can now be treated numerically, so one doesn't have to rely on what's analytically feasible. Best,


Google Made Me Make This Account I Dont Want It said...

It seems to me you don't need GR to predict a signal from the GW detectors. You can make the prediction with much simpler ingredients:

1. The observation of binary pulsar energy loss.

2. Conservation of energy (that lost energy must go somewhere...)

3. The geometry and scaling of quadropole interactions (which is required unless you introduce new fundamental constants with dimensions of length or time?) - to translate the binary pulsar observations to other potential sources and detectors.

4. Time reversal symmetry (because the detector is the time-reversal of the source).

Putting together 1 to 4, it is hard to escape the conclusion that whatever the cause of the binary-pulsar observations, *something* should show up at the GW detectors (even if GR is bogus and GWs are fiction).

If we really do fail to detect GWs, then the revision to our understanding of the universe is going to require much more than just a clever modification of GR to hide GWs under-the-carpet somehow.

Kevin Van Horn said...

I'm interested in hearing whether you think Carver Mead's proposed alternative to general relativity, G4v (gravity with 4-vector potentials), is worth taking seriously. I'm wondering whether its predictions match or differ from GR's in this case.

Arun said...

Hi Bee, is there some way that dark matter could act as an attenuating transmission medium for gravitational waves?


Leo Vuyk said...

No black hole merging signal could be another indication that our description of black hole qualities is wrong. So we need a new black hole theory without dual merger black holes.

Sabine Hossenfelder said...


Interesting idea. You would have to come up with a dark matter model that has a frequency-dependent response, then possibly you could do it. Makes me wonder now if the superfluid dark matter that we discussed the other day could do this. Best,


Leo Vuyk said...

If dark matter is the same as black holes , which also do not merge but remain distant from other black holes by an assumed but unknown process, then we can explain why the dark matter halos around galaxies remain in a halo shape as the remnants of super nova black hole families through the years.

Arun said...

Hi Bee, if anyone can work it out, it is you :)

Sabine Hossenfelder said...


Science doesn't work that way. Why not just say, the graviational wave background is suppressed "by an assumed but unknown process, then we can explain" the observations. True, but nothing will be learned from it. There isn't any known process that will prevent black holes from merging, period. Best,


Sabine Hossenfelder said...


Yes - there are also 300 other things that I can do and that I'd rather do...

Stuart said...

The graviton is not a messenger.

akidbelle said...

An absolutely last resort would rather be that GRT is not the end of the story. Something absolutely impossible in most physicists minds - I think. But on the other hand each time GRT is tested out of the solar system, it needs a patch: dark matter, dark energy, no gravitational waves (for those I know).
So I think that the community has no doubt, see no uncertainty in theories.

Emmette Davidson said...

עמיר ליבנה בר-או asks "What can cause SMBHs to merge less frequently than expected?"

If I’m reading this correctly, colliding galaxies will intermingle as the two (solar system sized) SMBH event horizons “merge” into a single galactic nucleus; however, the actual merger of the two singularities - each of point dimension with infinite density – should be strictly controlled by general relativity theory, which apparently predicts energy dissipation via gravity wave emissions as the two points orbit closer and closer (below the solar system sized event horizon). Thus, such binaries should produce gravity waves indefinitely and we could be detecting them from all directions throughout time; yet with detectors on for eleven years we find none.

So the question is rather, what prompts massive singularities to merge quickly and without issue, such that the GR predicted binaries either cannot form or don’t persist?

Peter Thieberger said...

General Relativity (GR) has been experimentally confirmed where gravity is relatively strong. But the outskirts of galaxies and groups of galaxies, where gravity is weak, don’t move at all as prescribed by GR. That problem has been “fixed” by inventing dark matter and by placing as much of it as necessary, wherever necessary to make things work as expected. No candidate dark matter particles have been found, and there is no other independent evidence for its existence. The alternative explanation by Milgrom, called MOND, uses a single, new, universal constant a0 =~ 1.2E-10 m/s^2 such that GR works in regions where the gravitational acceleration is larger than a0, while a gravitational acceleration proportional to 1/r instead of 1/r^2 prevails in regions where the acceleration is smaller than a0. This simple prescription is surprisingly successful, often more successful than dark matter. And this is achieved using a single constant rather than the multidimensional ad-hoc adjustments required by dark matter explanations.

The non-observation of gravitational waves is an independent indication that GR may not be as universally valid as we thought. Gravitational waves would be produced as expected from GR, which has been indirectly but very convincingly confirmed. They can’t however be expected to propagate as predicted in regions of space where the field equations are different. What these different equations may be hasn’t been worked out so far in a satisfactory way. The validity of this scenario, where the failure to observe gravitational waves is conjectured as being due to their failure to propagate through intergalactic space, will be supported or falsified by the future results of LIGO, by the development of a satisfactory MOND theory and/or by the discovery of dark matter particles.

Peter Thieberger