Monday, May 18, 2015

Galactic Cluster Collision Observations Disfavor Heavy Particle Dark Matter

El Gordo is the largest known example of two galaxies colliding.  Comparing the X-ray spectrum emissions of hot colliding gas with the visible light from the stars are the cores of the galaxies in the galactic clusters, provides an observational foundation from which the nature of dark matter phenomena in the clusters can be inferred.

The New Result

Scientists who have done this find that a heavy particle model of dark matter is a poor fit to this data.
This distinctive configuration has allowed the researchers to establish the relative speed of the collision, which is extreme (~2200km/second), as it puts it at the limit of what is allowed by current theory for dark matter.

These rare, extreme examples of clusters caught in the act of colliding seem to be challenging the accepted view that dark matter is made up of heavy particles, since no such particles have actually been detected yet, despite the efforts being made to find them by means of the LHC (Large Hadron Particle Collider) accelerator in Geneva and the LUX (Large Underground Xenon Experiment), an underground dark matter detector in the United States. In Tom Broadhurst's opinion, "it's all the more important to find a new model that will enable the mysterious dark matter to be understood better." Broadhurst is one of the authors of a wave-dark-matter model published in Nature Physics last year.

This new piece of research has entailed interpreting the gas observed and the dark matter of El Gordo "hydrodynamically" through the development of an in-house computational model that includes the dark matter, which comprises most of the mass, and which can be observed in the Xray region of the visible spectrum because of its extremely high temperature (100 million kelvin). Dr Broadhurst and Dr Molnar have managed to obtain a unique computational solution for this collision because of the comet-like shape of the hot gas, and the locations and the masses of the two dark matter cores that have passed through each other at an oblique angle at a relative speed of about 2200 km/s. This means that the total energy release is bigger than that of any other known phenomenon, with the exception of the Big Bang.
The underlying study to this interpretative description from a university PR office is here (preprint here). The abstract (with a few mathematical symbols converted to words) states that:
The distinctive cometary X-ray morphology of the recently discovered massive galaxy cluster "El Gordo" (ACT-CT J0102–4915; z = 0.87) indicates that an unusually high-speed collision is ongoing between two massive galaxy clusters. A bright X-ray "bullet" leads a "twin-tailed" wake, with the Sunyaev-Zel'dovich (SZ) centroid at the end of the northern tail. 
We show how the physical properties of this system can be determined using our FLASH-based, N-body/hydrodynamic model, constrained by detailed X-ray, SZ, and Hubble lensing and dynamical data. 
The X-ray morphology and the location of the two dark matter components and the SZ peak are accurately described by a simple binary collision viewed about 480 million years after the first core passage. We derive an impact parameter of about 300 kpc, and a relative initial infall velocity of about 2250 km s–1 when separated by the sum of the two virial radii assuming an initial total mass of 2.15 × 1015 M ☉ and a mass ratio of 1.9. 
Our model demonstrates that tidally stretched gas accounts for the northern X-ray tail along the collision axis between the mass peaks, and that the southern tail lies off axis, comprising compressed and shock heated gas generated as the less massive component plunges through the main cluster. 
The challenge for ΛCDM will be to find out if this physically extreme event can be plausibly accommodated when combined with the similarly massive, high-infall-velocity case of the Bullet cluster and other such cases being uncovered in new SZ based surveys.
In the pertinent part of the conclusion the paper states that:
This massive merging cluster with an infall velocity of 2250 km s−1 at the the time the Universe was half of its age (at the redshift of z = 0.87) is apparently very unusual in the standard ΛCDM models, for which extensive simulations have been performed to determine the expected probability distribution of relative velocities (Thompson & Nagamine 2012). At face value, such extreme cases present a challenge to our understanding of structure formation and may lead to a better understanding of dark matter and/or alternative theories of gravity.
Basically,  then, the example is notable because structure formation of such a huge scale tends to come much later in standard ΛCDM models than is observed.  Thompson, Dave & Nagamine 2014, however, finds that the deficit of merging clusters of this type may have simply been an artifact of an inadequate simulation program, and reproduce the observed number of large cluster combinations from  the standard ΛCDM models in a simulation using a different methodology.

Wave, Scalar Field and Boson Star Dark Matter

More background on wave dark matter is found here and here.  A key point is that "In these wave dark matter mechanisms, the substructure of the dark matter (the perturbations from spherical symmetry) is what drives the substructure in the regular matter (spiral patterns and shells)."  The observed texture of luminous matter distributions in elliptical galaxies (which often have "shells") and spiral galaxies (which often have spiral arms rather than being smooth disks) is a major motivation for this theory.

Another notable observation about wave dark matter theories in these links is this one:
[T]his is where our discussion intersects with the fascinating works of many others who have studied “scalar field dark matter” and “boson stars.” In the boson star case, the motivation is quantum mechanical, so the above scalar field f is supposed to represent the overall wave function for a very large number of very tiny bosons with masses on the order of 10−23 eV. We relate our constant Υ to the mass of the Klein-Gordon equation by noting that the Compton wavelength in both cases is λ = 2π Υ = h m ≈ 13 light years if we take m ≈ 10−23eV , or equivalently, Υ ≈ 1/(2 light years). These other motivations are interesting as well but should be distinguished from the purely geometric motivations provided in this paper.
The related scalar field dark matter theory is also described here and was originally conceived in 1994 by Ji and Sin.

A recent analysis of cosmological data in the context of the theory concludes that the right boson mass is actually about 10 orders of magnitude smaller than previously assumed.  My own intuition is to note that the bosons of scalar field dark matter are suspiciously close in mass-energy magnitude to the hypothetical graviton.  And, indeed, the mass predicted in this recent analysis is precisely the mass a graviton would need to have in order to reproduce the observed cosmology without the cosmological constant.  This mass also corresponds to a Compton wave length of approximately the same length of the size of the universe (i.e. about 13 billion light years).

I suspect that a lot of the discussion of the massless graviton confounds particle rest mass, which an infinite range force boson should not have, and mass-energy which is something which gravity couples to and is present even in massless bosons like the photon.  It is not at all obvious to me that "massive gravity" theories are in any way distinguishable from "mass-energyful gravity" theories, in which case "massive gravity" is not inconsistent with a graviton that has zero rest mass as expected (but I would seem to be wrong).  A hypothetical graviton, while it lacks rest mass, should have mass-energy of some small, finite amount.

Some blog level discussion of massive gravity theories and their history is found here.  The criticism is a mass-energyful gravity expressed in the comments is that: 
There is no such thing as the energy density of the curvature. This is an old result from GR --- one cannot define the concept of energy for a gravitational field, locally. It can be done only in a suitable global sense (spacetime with asymptotic global time-translation symmetry).  
It is indeed an old result, but I continue to think that it is almost certainly wrong and probably the main cause of the deviation between GR and what is observed in Nature. Even if this is true of GR, that may just mean that GR is wrong.

As an aside, an excellent pdf defining the key terms and concepts in GR and the Einstein-Gordon-Klein equation (which describes the gravitational force generated by a scalar field in GR) is found here.  The Einstein-Gordon-Klein equation is intimately related to wave dark matter models and scalar field dark matter models.


Maju said...

I'd wish I could understand or get a good "explanation for dummies". I was reading the El Gordo press release yesterday (Basque research incidentally) but I really do not grasp well the implications, other than WIMPs seem to be discarded as Dark Matter because they are too massive, right?

When you deal with the graviton issue, you got me even more perplex, because, yeah, by definition gravitons have zero mass. Yet you go on to discuss the "mass-energy" of the hypothetical particle. But wait: mass is a form of energy but energy is not the same as mass. Are you implying that every particle for the simple fact of having energy has some sort of implicit mass m=E/c²? Photons too? Why then nobody thinks of photons as dark matter candidates?

andrew said...

Easy questions first:

Yes, photons too.

Photons are affected by gravity just like anything else (this is called gravitational lensing). And, photons give rise to gravitational fields proportional to their mass-energy at E=mc^2.

The right hand side of the general relativity equation (Einstein's field equations) is 8pi*G*T(uv). 8, pi and Newton's constant (G) are numbers and in the case of Newton's constant physical units (General Relativity does not have "dimensionless constants" the way that the Standard Model does). T is a four by four number matrix called the stress-energy tensor.

The top-left element is energy density which is equal to matter-energy of both matter and photons at E=mc^2 at the relevant point, and the remaining elements describe four kinds of motion that different sources of gravity can have in particular directions (like pressure and energy flux and momentum).

The left hand side of Einstein's field equations describe the curvature of space that is created by the stress-energy components on the right hand side (a.k.a. gravity).

2. Photons are not dark matter candidates first of all because they aren't dark (of course), and more importantly because physicists, being clever, have measured the total weight of all the photons in the universe and in places where we observe dark matter effects (which they can do because it isn't dark), and there aren't nearly enough of them to fit the bill. Photons make up something on the order of 1% or less of the mass-energy in the universe.

andrew said...

One more addition to #2 above:

We also know that photons are not dark matter because it moves too fast. One of the few things we know about dark matter (if it exists) in addition to how heavy it is, and roughly where it is located, is that it moves quite slowly relative to the speed of light. If it moves really, really slowly, we call it cold dark matter, and if it moves a bit faster, we call it "warm dark matter".

If dark matter moved at or near the speed of light, we would call it "hot dark matter" and one of the things we know for sure is that dark matter is not "hot" because we have observed its effects which are inconsistent with something moving close to the speed of light. If dark matter were "hot" then there would be no galaxies and there would be no galactic clusters, and the universe would be a lot less "lumpy" and instead would be far more of a homogeneous soup (imagine one huge nebula full of obscuring matter evenly spread out everywhere and you have an idea of what a hot dark matter universe would look like). An example of hot dark matter would be neutrinos, but we've measured how many of them there are too, and neutrinos contribute even less to the mass of the universe than photons do because they are each so light weight (and most neutrinos in space move at close to the speed of light).

Now for the harder question:

3. Mass v. mass-energy.

Mass, as opposed to mass-energy, also known as rest mass, quantifies the amount of inertia that a particle (either a fundamental particle or a composite one) has. This is to say, rest mass tells you how much energy it takes to accelerate the particle to a greater velocity. At speeds much smaller than the speed of light, the formula is F=ma.

At speeds approaching the speed of light, special relativity is required because it takes a bit more energy to shift a particle from v+1 to v+2 than it does to shift a particle from v to v+1, and the extra energy required to add the same amount of velocity increased infinitely as you approach the speed of light.

Particles with zero rest mass, in contrast, always move at exactly the speed of light, without having any energy applied to them.

But, both particle with rest mass and particles with zero mass are both equally affected by gravity according to E=mc^2.

andrew said...

To understand the El Gordo results, allow me to use some fake numbers:

El Gordo is about 10 billion year old light (which we known through something called "red shift" an issue too technical for now) (warning this is a fake, but order of magnitude right number).

If you do simulations on computers of what the universe should look like when it is 3.5 billion years old, give or take, there is only a one in million chance (warning, this is another fake, but order of magnitude right number), that we would be able to see even one object as big or nearly as big as as El Gordo at that general age of the universe.

But, in real life there are half a dozen or a dozen (warning, yet another fake but approximately right number) El Gordo scale objects that we can see in the sky at approximately the right age.

Ergo: (1) Cold dark matter cosmology is wrong, or (2) the simulations of cold dark matter cosmology are screwed up in some very material way.

andrew said...

Final bottom line re: El Gordo.

The El Gordo conclusion is that you need some kind of dark matter that gives rise to many extremely large galactic clusters and allows them time to collide into each other very rapidly after the Big Bang.

There is more than one possible dark matter theory that could produce this outcome (at least in theory) but it is hard to know which one is the best fit because good simulations exist only for a few of the many possibilities for dark matter. The researchers did not exhaustively consider the alternatives and only mentioned in passing that their result was at odds with the most popular dark matter cosmology in existence because that model is called the "Standard Model of Cosmology" and is the one to beat.

Maju said...

Thanks for your patience, Andrew. Very nice read.

"Ergo: (1) Cold dark matter cosmology is wrong, or (2) the simulations of cold dark matter cosmology are screwed up in some very material way".

So no WIMPs, no MACHOs and no RAMBOs. The universe is much less machista than many physicists though?

You also said that hot is impossible, so then only warm DM is possible with these results. Reading Wikipedia's entry on this issue (which is quite short as cold DM was the favored solution until now), I read that they mention sterile neutrinos and gravitinos as candidate particles, not the graviton as such. They also mentioned WIMPs but only if produced non-thermally (not sure of the implications).

Gravitinos seem to pose some problems, because without some very specific solutions of SUSY they would decay and not be available for DM in sufficient numbers.

I don't see any reference to the graviton as related to DM but your suggestion is interesting no doubt. However the existence of gravitons itself seems to be a confirmation of String Theory, what I know you hate.

So I wonder why aren't you rather favoring sterile neutrinos instead.