Sunday, October 22, 2017

New gravitational wave detection with optical counterpart rules out some dark matter alternatives

The recently reported gravitational wave detection, GW170817, was accompanied by electromagnetic radiation. Both signals arrived on Earth almost simultaneously, within a time-window of a few seconds. This is a big problem for some alternatives to dark matter as this new paper lays out:

The observation is difficult to explain with some variants of modified gravity because in these models electromagnetic and gravitational radiation travel differently.

In modified gravity, dark matter is not made of particles. Instead, the gravitational pull felt by normal matter comes from a gravitational potential that is not the one predicted by general relativity. In general relativity and its modifications likewise, the gravitational potential is described by the curvature of space-time and encoded in what is called the “metric.” In the versions of modified gravity studied in the new paper, the metric has additional terms which effectively act on normal matter as if there was dark matter, even though there is no dark matter.

However, the metric in general relativity is also what gives rise to gravitational waves, which are small, periodic disturbances of that metric. If dark matter is made of particles, then the gravitational waves themselves travel through the gravitational potential of normal plus dark matter. If dark matter, however, is due to a modification of the gravitational potential, then gravitational waves themselves do not feel the dark matter potential.

This can be probed if you send both types of signals, electromagnetic and gravitational, through a gravitational potential, for example that of the Milky Way. The presence of the gravitational potential increases the run-time of the signal, and the deeper the potential, the longer the run-time. This is known as “Shapiro-delay” and is one of the ways, for example, to probe general relativity in the solar system.

The authors of the paper put in the numbers and find that the difference between the potential with dark matter for electromagnetic radiation and the potential without dark matter for gravitational radiation adds up to about a year for the Milky Way alone. On top come some hundred days more delay if you also take into account galaxies that the signals passed by on the way from the source to Earth. If correct, this means that the almost simultaneous arrival of both signals rules out the modifications of gravity which lead to differences in the travel-time by many orders of magnitude.

The logic of the argument is this. We know that galaxies cause gravitational lensing as if they contain dark matter. This means even if dark matter can be ascribed to modified gravity, its effect on light must be like that of dark matter. The Shapiro-delay isn’t exactly the same as gravitational lensing, but the origin of both effects is mathematically similar. This makes it plausible that the Shapiro-delay for electromagnetic radiation scales with the dark matter mass, regardless of its origin. The authors assume that the delay for the gravitational waves in modified gravity is just due to normal matter. This means that gravitational waves should arrive much sooner than their electromagnetic company because the potential the gravitational waves feel is much shallower.

The Shapiro-delay on the Sun is about 10-4 seconds. If you scale this up to the Milky Way, with a mass of about 1012 times that of the Sun, this gives 108 seconds, which is indeed about a year or so. You gain a little since the dark matter mass is somewhat higher and lose a little because the Milky Way isn’t spherically symmetric. But by order of magnitude this simple estimate explains the constraint.

The paper hence rules out all modified gravity theories that predict gravitational waves which pass differently through the gravitational potential of galaxies than electromagenetic waves do. This does not affect all types of modified gravity, but it does affect, according to the paper, Bekenstein’s TeVeS and Moffat’s Scalar-Vector-Tensor theory.

A word of caution, however, is that the paper does not contain, and I have not seen, an actual calculation for the delay of gravitational waves in the respective modified gravity models. Though the estimate seems good, it’s sketchy on the math.

I think the paper is a big step forward. I am not sold on either modified gravity or particle dark matter and think both have their pros and cons. To me, particle dark matter seems plausible and it works well on all scales, while modified gravity doesn’t work so well on cosmological (super-galactic) scales. On the other hand, we haven’t directly measured any dark matter particles, and some of the observed regularities in galaxies are not well explained by the particle-hypothesis.

But as wonderful as it is to cross some models off the list, ruling out certain types of modified gravity doesn’t make particle dark matter any better. The reason you never hear anyone claim that particle dark matter has been ruled out is that it’s not possible to rule it out. The idea is so flexible and the galactic simulations have so many parameters you can explain everything.

This is why I have lately been intrigued by the idea that dark matter is a kind of superfluid which, in certain approximations, behaves like modified gravity. This can explain the observed regularities while maintaining the benefits of particle dark matter. For all I can tell, the new constraint doesn’t apply to this type of superfluid (one of the authors of the new paper confirmed this to me).

In summary, let me emphasize that this new observation doesn’t rule out modified gravity any more than the no-detection of Weakly Interacting Massive Particles rules out particle dark matter. So please don’t jump to conclusions. It rules out certain types of modified gravity, no more and no less. But this paper gives me hope that a resolution of the dark matter mystery might happen in my lifetime.


Unknown said...
This comment has been removed by the author.
Sabine Hossenfelder said...


How do you scoop a paper when you cite it? I don't understand what you mean.

Shantanu said...

Unknown, that paper does not say anything about Dark matter emulators. It however does do an independent calculation of Shapiro delay and we have cited that paper and mentioned that our calculation is complementary to that analysis.

Uncle Al said...

Solve a problem? No, remove the problem.

"gravitational potential is described by the curvature of space-time" and/or by Einstein-Cartan spacetime torsion, and mixed [1,2]. Torsion is chiral (a left foot) validated with opposite shoes (excluded from physics by Green's function).

10^(-9) chiral anisotropic space-time torsion softens exact Noetherian angular momentum conservation - Milgrom acceleration for the Tulley-Fisher relation. It enables net matter baryogenesis - Sakharov conditions. Torsion interactions are selective to fermionic matter - hadrons not massless boson photons. Look – geometric Eötvös experiments, enantiomorphic space groups P3(1)21 versus P3(2)21 single crystal α-quartz test masses.

[1] arXiv:1609.09275, arXiv:1705.06317
[2] DOI:10.1090/tran/6964, teleparallel spacetimes re Weitzenböck canonical metrics on four-manifolds

JimV said...

With all the new data resulting from LIGO detections, the Nobel committee's choice is looking better and better, in retrospect.

Bruce Rout said...

This is a very good analysis Sabine. However there is also a theory that dark matter doesn't exist in the first place and it is all a hoax and a scam. Nevertheless, a very good analysis.

David Brown said...

"In modified gravity, dark matter is not made of particles." In Milgrom's MOND, there is not necessarily an exclusion of all hypothetical dark matter particles. Also, MOND might be empirically valid even though Einstein's field equations are, after quantum averaging, 100% correct. How might this happen? There might be MOND-chameleon particles that have variable effective mass depending upon nearby gravitational acceleration. In other words, the empirical successes of MOND might be explained as follows: In the standard form of Einstein's field equations, replace the -1/2 by -1/2 + MOND-chameleon-fake-function, where this function is approximately a constant over the range of gravitational accelerations for which MOND works. The MOND-chameleon-fake-function would be an apparent, but not real, failure of Newtonian-Einsteinian gravitational theory.

dlb said...

I'm a long time reader of this blog, but almost never post. I would like to thank you again for taking the time to summarize papers like this for us.

I realize your blog is about quantum gravity phenomenology, and not dark matter or general relativity (if you know anyone who blogs at this level of expertise, I'd be delighted to know about them), but I have a question (hopefully on topic).

You write "But this paper gives me hope that a resolution of the dark matter mystery might happen in my lifetime". What kind of experiments or observations can we expect to settle this issue?

For instance, I once hoped that the Gaia satellite would give us the answer.
. By probing the gravitational potential above the galatic plane with enough precision to be able to see the inprint of dark matter.
. By observing the relative motion of stars in binary systems when they are very far from each other.

Theophanes Raptis said...

@Rout: A "scam" is indeed always possible since, however much we learn, there is still a possibility that our ignorance is still infinite in as much as removing a portion of the integers cannot exaust infinity - not even a countable one at that! Yet, for our world, equipped with such dreadful instrumentation what are the chances that we do not see what trully is there? Unless of course, the "scam" is so ellaborate, taking place in our own minds without us even noticing, a slight slippery path in the tree of choices in how to put up all the evidence together. And then, there still are professional magicians able to steal your wallet while your mind is distracted, how would the poor average Joe know?

Frank Hubeny said...

I would be interested in hearing follow ups on what those supporting Moffat's modified gravity have to say about this data. Thanks for presenting it.

Sabine Hossenfelder said...


This is an interesting question. I was thinking about this yesterday. If it's a particle, then detecting the particle will convince almost everyone that's it. If it's not a particle, I'm not sure what it would take.

Rob van Son (Not a physicist, just an amateur) said...

"However there is also a theory that dark matter doesn't exist in the first place and it is all a hoax and a scam."

The theory that "X is a scam and does not exist" is very popular for many values of "X". However, supporting evidence that this theory is valid has historically been limited to only a handful of values of X. Also, values of X have been very sensitive to fashion and new X's have been substituted whenever old ones fell out of fashion. This has made proving or disproving the theory challenging. Therefore, I find it very difficult to evaluate your assertion that there are theories that tells us that dark matter is a scam. I find it especially difficult to imagine a motivation for thousands of astronomers over decades to collude in a scam to "invent" dark matter.

I do find it very interesting to hear that there is now a new way to test theories of dark matter and modified gravity.

Thanks Dr B.

Andy Miller said...

Sabine, let me clarify/correct my comments about TeVeS. It was not Moffat who originally pointed out that TeVeS results in stellar instability but rather Seifert (2007). And its violation of local Lorentz invariance is mentioned on page 353 of the 2015 book Gravity: Where Do We Stand? Peron et al.. (Which doesn't seem to mention STVG at all. Sigh…)

Tom Andersen said...

In addition to a dark matter particle detection, a way to see dark matter 'better' will be with better telescopes like the new Mauna Loa and James Webb scopes. Astronomy has been so successful in the past 100 years - every major step forward in capability solves and creates lots of new problems.

Things one might see about dark matter - better more detailed maps of dm density (or whatever the effect is), or associated glow, etc.

Axil said...

Prof. Pierre Sikivie: "It has long been known that axions produced by vacuum realignment during the QCD phase transition in the early universe form a cold degenerate Bose gas and are a candidate for the dark matter. More recently it was found that dark matter axions thermalize through their gravitational self-interactions and form a Bose-Einstein condensate (BEC). On time scales long compared to their rethermalization time scale, almost all the axions go to the lowest energy state available to them. In this behaviour they differ from the other dark matter candidates. Axions accreting onto a galactic halo fall in with net overall rotation because almost all go to the lowest energy available state for given angular momentum. In contrast, the other proposed forms of dark matter accrete onto galactic halos with an irrotational velocity field. The inner caustics are different in the two cases. I'll argue that the dark matter is axions because there is observational evidence for the type of inner caustic produced by, and only by, an axion BEC."

There is dark matter theory that shows evidence of BEC formation on the galactic scale. If you need to get to superfluidity, the axion BEC is what is required.

Also related to Scalar field dark matter

"The dark matter can be modeled as a scalar field using two fitted parameters, mass and self-interaction. In this picture the dark matter consists of an ultralight particle with a mass of O(10e−22) eV when there is no self-interaction. If there is a self-interaction a wider mass range is allowed. The uncertainty in position of a particle is larger than its Compton wavelength, and for some reasonable estimates of particle mass and density of dark matter there is no point talking about the individual particle's position and momentum. The dark matter is more like a wave than a particle, and the galactic halos are giant systems of condensed bose liquid, possibly superfluid. The dark matter can be described as a Bose–Einstein condensate of the ultralight quanta of the field and as boson stars. The enormous Compton wavelength of these particles prevents structure formation on small subgalactic scales, which is a major problem in traditional cold dark matter models.

This dark matter model is also known as BEC dark matter or wave dark matter. Fuzzy dark matter and ultra-light axion are examples of scalar field dark matter."

An axion BEC on the galactic scale meets the need for the dark matter particle to produce superfluid effects, and in certain approximations, behaves like modified gravity.

Unknown said...

@Axil, yes fuzzy/BEC dark matter still seems to be in good shape regarding observation.

Andy Miller said...

On October 30, Green, Moffat, and Toth published a new preprint, demonstrating that the preprint of Boran et al. grossly misrepresented Moffat's STVG alternative theory of gravity, and that STVG is consistent with the neutron-start merger GW170817/GRB170817A. (One the other hand, theories attempting to reproduce the behaviour of MOND are indeed ruled out by the merger.) To me, it seems difficult to reconcile Boran et al.'s preprint with their having ever actually read Moffat's publications. As Sabine has pointed out, such ignorance of the scientific literature is all too common in the field of interpretive cosmology.

I think that a well-known psychological phenomenon is impeding progress in cosmology — we humans tend to prefer beliefs that we find personally comfortable and convenient. It is far more convenient to publish results from refined and extended Dark Matter simulations that have been developed over years than to learn how to create and study solutions of a new gravitational theory. For many, the intellectual barrier is significant, possibly leading them to seek comfort in the demise of STVG. After Einstein's General Relativity appeared in 2015/2016, and Eddington confirmed in 1919 its prediction of the precession of the orbit of Mercury, Eddington was accused of being one of only three people in the world who understood GR. Eddington was puzzled, and wondered aloud who might be the third! Although GR is now familiar to many, STVG is more complex. Nevertheless, this last year has seen the appearance of about a dozen preprints from those independently studying the predictions of STVG for particular features in cosmological observations. Now with MOG apparently the only gravitational theory left that is capable of seriously challenging the dark matter paradigm, we can hope for accelerated progress.