Galaxy with no dark matter? (NGC1052-DF2)

[mfb]

Direct detection experiments would be ... difficult.

Not that they would be easy here...

To whoever may know - is there any evidence that dark matter has the same G (gravitational constant) as visible matter? I think the reasoning in this thread all assumes that it does, I'm wondering if that is a default assumption or if there some way to draw that conclusion from cosmological observations.

I am admittedly over my head in asking this question - I intend to be asking if the ratio of inertia to gravitational attraction for dark matter is the same as for visible matter, or if we have any evidence to say one way or the other. I think we just assume that it does, and that drives our calculation of halo's etc of dark matter.

Which G would you use for the attraction between dark matter and regular matter?
No, you can't make a reasonable theory out of that. Especially as normally regular matter and dark matter stay together.

 

[Jonathan Scott]

I don't buy for an instant that MOND can explain the variation in observed dark matter between different galaxies. At least not in anything approaching a reasonable manner (that is, no parameters that are tuned per-galaxy). MOND is basically dead now anyway. Has been for a long time.

This seems to be the wrong way round. MOND in its original form has only a single universal acceleration parameter which applies to all galaxies and is amazingly successful in explaining or predicting individual galaxy rotation curves with no additional parameters, whereas in contrast different galaxies seem to require distinctly different distributions of dark matter to explain their rotation curves, so this is the area where MOND excels. MOND however is extremely unsatisfactory as a "theory" as it violates basic principles such as conservation of momentum, and has difficulty explaining motion above the scale of individual galaxies. There are more sophisticated modified gravity theories which approximate MOND but they have far more parameters and seem quite arbitrary especially compared with the neatness of General Relativity.

 

[PAllen]

This seems to be the wrong way round. MOND in its original form has only a single universal acceleration parameter which applies to all galaxies and is amazingly successful in explaining or predicting individual galaxy rotation curves with no additional parameters, whereas in contrast different galaxies seem to require distinctly different distributions of dark matter to explain their rotation curves, so this is the area where MOND excels. MOND however is extremely unsatisfactory as a "theory" as it violates basic principles such as conservation of momentum, and has difficulty explaining motion above the scale of individual galaxies. There are more sophisticated modified gravity theories which approximate MOND but they have far more parameters and seem quite arbitrary especially compared with the neatness of General Relativity.

I think the issue for this galaxy with MOND is that it simply fails to explain the rotation curve for this galaxy. This galaxy is different from most so you can choose:

1) To rescue MOND, assume there is an unknown counter effect for this galaxy. Since MOND is a gravity law, changing the law for one galaxy doesn’t make sense, so you are left with ... repulsive dark matter ?? that so far exists for only one known galaxy??

2) To rescue dark matter models, just assume little or no dark matter for this galaxy, leaving the problem of how the separation might have occurred. Such separation would be expected to be rare, consistent with observation.

To me, this galaxy finding clearly works against MOND due to implausibility of what is needed to explain this galaxy.

 

[Jonathan Scott]

I think the issue for this galaxy with MOND is that it simply fails to explain the rotation curve for this galaxy. This galaxy is different from most so you can choose:

1) To rescue MOND, assume there is an unknown counter effect for this galaxy. Since MOND is a gravity law, changing the law for one galaxy doesn’t make sense, so you are left with ... repulsive dark matter ?? that so far exists for only one known galaxy??

2) To rescue dark matter models, just assume little or no dark matter for this galaxy, leaving the problem of how the separation might have occurred. Such separation would be expected to be rare, consistent with observation.

To me, this galaxy finding clearly works against MOND due to implausibility of what is needed to explain this galaxy.

I agree that if the interpretation of the observations is correct in this case, this particular galaxy leads to something like the above options. There are probably other possible explanations too, perhaps about a very unusual line of sight giving misleading results.
But the curious success of MOND in the vast majority of cases suggests that something systematic that we don't understand is going on to make the results fit the MOND pattern, even if it somehow involves dark matter.
And my main point was simply that MOND doesn't need extra parameters to match different dark matter distributions for different galaxies.

 

[kimbyd]

This seems to be the wrong way round. MOND in its original form has only a single universal acceleration parameter which applies to all galaxies and is amazingly successful in explaining or predicting individual galaxy rotation curves with no additional parameters, whereas in contrast different galaxies seem to require distinctly different distributions of dark matter to explain their rotation curves, so this is the area where MOND excels. MOND however is extremely unsatisfactory as a "theory" as it violates basic principles such as conservation of momentum, and has difficulty explaining motion above the scale of individual galaxies. There are more sophisticated modified gravity theories which approximate MOND but they have far more parameters and seem quite arbitrary especially compared with the neatness of General Relativity.

I have a hard time believing that MOND can accurately describe the rotation curves of this galaxy. My understanding is that it has had problems with the diversity of rotation curves in visible galaxies ever since we started measuring a large number of them in detail. And it's never satisfactorily explained the behavior of galaxy clusters.

 

[nikkkom]

A medium-velocity galaxy collision with particularly suitable geometry might do it. Say, two spiral galaxies colliding edge-on would leave most of their gas and dust piled up at the site of the collision, while DM and stars would pass through and fly away.

 

[phinds]

A medium-velocity galaxy collision with particularly suitable geometry might do it. Say, two spiral galaxies colliding edge-on would leave most of their gas and dust piled up at the site of the collision, while DM and stars would pass through and fly away.

But would they not remain in the neighborhood and swirl back and collide again? You seem to imply that they would not remain gravitationally bound. Seems unlikely

 

[nikkkom]

But would they not remain in the neighborhood and swirl back and collide again? You seem to imply that they would not remain gravitationally bound. Seems unlikely

Obviously, depends on the velocity of the collision.

 

[JMz]

To one point there:

gravitational properties of antimatter, specifically if antimatter is repelled gravitationally by matter.

This will contradict GR, won't it? That is, positrons (for an example of antimatter) are as much concentrations of energy as electrons, and GR would therefore treat them identically. And, of course, both have the same momentum per unit velocity, so even the gravitational-mass/inertial-mass ratio would be different. So we would even need to give up the Equivalence Principle. Right?

No problem, if CERN shows it's truly necessary. But that's a very high hurdle.

 

[PeterDonis]

So called External Field Effects (EFE) can come to the rescue of MOND in the case of galaxy clusters.

MOND fits all galactic systems perfectly with its one universal parameter a0.

Please give references (textbooks or peer-reviewed papers) for these statements.

 

[Vanadium 50]

Please give references (textbooks or peer-reviewed papers) for these statements.

The 2nd one is: McGaugh et al. Phys. Rev. Lett. 117, 201101 (2016)

The first one will be harder to find that exact thing, but it certainly stands to reason: MOND's mechanics assumes that what matters is the total force on the object, not just the force from the galaxy of interest. In that regard it is identical to Newton and Einstein.

 

[Olorin]

Please give references (textbooks or peer-reviewed papers) for these statements.

Regarding the EFE here are two links for explanations: http://astroweb.case.edu/ssm/mond/EFE.html and http://astroweb.case.edu/ssm/mond/milgromonefe.html.
For peer rewiewed papers there are a few actually, just putting one here, but please, feel free to do your own research on the subject: https://arxiv.org/pdf/1404.2202.pdf.

 

[JMz]

Yes, that's correct. I guess it is fair to assume that GR won't survive a direct experimental violation of the weak equivalence principle. If antimatter falls up, that's the end of "space-time geometry" as a valid theory of gravity. My gut guess is that the quantum vacuum as a gravitational and electric dipolar medium is a much more profound and sound starting point to rethink the way gravity works. It actually naturally allows sweet coupling effects between electromagnetic and gravitational phenomena!
Not necessarily if that explains a lot of other things we fail to grasp while willing to keep GR as a viable theory of gravity at all costs, i.e. dark matter, dark energy, inflation, black hole and big bang singularities, information paradox etc...if breaking the weak equivalence principle has the power to explain all of it, which it seems to do when you delve into the consequences of anti-gravitational antimatter, so be it. But we must not wait till CERN results are published to develop the full consequences of the theory, which can have rather large implications for our understanding of the universe. Mark my words: I bet that GR won't survive the next decade of observational and experimental evidence, and depending on the cunning and openness of our best minds, we might have a new and better theory of gravity by then.

Fair enough. My own bet, though, would be that, if GR doesn't "survive" the next decade, it will only be because something came along that fully agrees except where quantum effects become important: more properly an extension of GR than a contradiction of it. That's not entirely a foundational statement (i.e., that it embodies all the correct non-quantum insights, such as special relativity), but partly an expectation that, if anything else about it is amiss, we won't have the right equipment or do the right experiments to recognize it until long after that. Of course, I recognize that people are willing to live with some current problems with GR partly because there isn't an alternative that both agrees better with experiments and has foundations that are at least as simple and appealing.

 

[kimbyd]

The 2nd one is: McGaugh et al. Phys. Rev. Lett. 117, 201101 (2016)

The first one will be harder to find that exact thing, but it certainly stands to reason: MOND's mechanics assumes that what matters is the total force on the object, not just the force from the galaxy of interest. In that regard it is identical to Newton and Einstein.

Arxiv link to that article:
https://arxiv.org/abs/1609.05917

Note that in a response, these authors argue that the relation described above is not something new, but rather a function of the well-known baryonic Tully-Fisher relation:
https://arxiv.org/abs/1803.01849

As for MOND explaining these galaxies "perfectly", that's a matter open to interpretation. There's substantial scatter.

Regardless, MOND still fails to explain galaxy cluster behavior.

 

[Vanadium 50]

As for MOND explaining these galaxies "perfectly", that's a matter open to interpretation. There's substantial scatter.

I would say that scatter is no better and no worse than many other astronomical measurements, e.g. SNe as standard candles.

Regardless, MOND still fails to explain galaxy cluster behavior.

Agreed. MOND works on galactic scales and nowhere else. I believe that when the dust settles, the outcome will be MOND tells us little about gravity and more about galaxy formation.

 

[kimbyd]

Well it falls to Occam's razor. The best theory is the one that predicts the most with minimum amount of assumptions. Modified Gravity don't give a toss about this effects in order to produce the Tully Fischer relation, as gravitational effects are always tied to visible mass. DM is akin to epycyclic models, always in need of more ad-hoc and unecessary tweaking in order to even be able to start to make a rigorous sense of it all.

Except feedback effects are not assumptions. They are a part of reality whether we include them in our models or not. Failing to include them doesn't mean that you've got a simpler theory: it means you're ignoring pieces of the puzzle.

The right way to deal with this is not to say, "Well, MOND doesn't need this!" but rather, "Time to work out the consequences of this feedback effect in the MOND model, to make sure it makes sense."

 

[mfb]

We know antimatter falls down.

Atoms are about 0.03% electrons, 1% quark masses and 99% QCD binding energy, the exact fractions depend on the atom. We know from comparisons of countless atoms that all atoms fall down at the same rate, this is only possible if all three components satisfy the weak equivalent principle.

Anti-atoms are about 0.03% positrons, 1% antiquark masses and 99% QCD binding energy. We already know the last part satisfies the weak equivalence principle. For antimatter to fall up, the other components would have to do something completely crazy, and no matter what they do different antiatoms would fall up at different rates. While this is not yet ruled out by experiment, it doesn't sound plausible at all.

To make antimatter fall up you would have to assign a "matterness" to binding energies. Not just QCD, but also QED which I neglected above. And that doesn't sound plausible either. Which fraction of the QCD binding energy in a pentaquark is matter? ;)

 

[kimbyd]

We know antimatter falls down.

This has yet to be experimentally demonstrated, though it is very true that there would be some pretty extreme theoretical challenges with explaining how anti-matter could possibly interact with gravity differently from normal matter while normal matter still obeys the equivalence principle.

 

[JMz]

The idea that antimatter should fall up seems like an idea based on nothing more than etymology. We call certain particles "anti" because people noted that they have opposite charge, charm, or whatever compared to particles we commonly encounter.

That doesn't prove that they don't have anti-mass as well, but the idea that they do (a) starts with much less plausibility, because they don't even behave that way for momentum, and (b) as @Jonathan Scott notes, photons, which are their own "anti"-particles, are observed to fall down in gravitational lenses.

 

[jerromyjon]

If for instance antimatter falls up consistently in the 3 CERN experiments, then GR is DEAD

What if antimatter and matter are more strongly attracted than either are to themselves? I only ask this naive question because my basic understanding is that matter and antimatter always seem to attract and annihilate so the premise in my mind is that there is something MORE powerful than typical physics involved.

 

[kimbyd]

The idea that antimatter should fall up seems like an idea based on nothing more than etymology. We call certain particles "anti" because people noted that they have opposite charge, charm, or whatever compared to particles we commonly encounter.

I honestly agree. It's highly speculative, and wouldn't match with existing theory in a wide variety of ways. However, it still has some potential value for two reasons:
1) It is a testable prediction. As I understand it, the current best method involves creating neutral anti-Hydrogen and using a photon trap to cool it and then observe its motion. This was attempted a few years ago, but the error bars were still too large (https://www.nature.com/articles/ncomms2787). A refinement of that experiment may produce a definitive answer (I believe they're currently trying for a 1% measurement error on the intertial properties of anti-Hydrogen).
2) However theoretically absurd it may be, there is always some possibility that our theories are incorrect, and so this kind of experiment is worth doing. Any discrepancy in the behavior of anti-matter compared to normal matter would provide tremendously valuable insights into how to improve our current theories.

 

[JMz]

If they are more strongly attracted to each other through electric charge, sure. But that's exactly how they are "anti's" (with respect to each other), and there's no new physics there: Opposite electric charges attract, always.

Being differently gravitationally attracted is the relevant concern here: If it were observed, then GR would be qualitatively wrong. Contrapositively, if GR, or anything like it, is right, then that doesn't happen. (Most of us would bet on the latter.)

 

[JMz]

I honestly agree. It's highly speculative, and wouldn't match with existing theory in a wide variety of ways. However, it still has some potential value for two reasons:
1) It is a testable prediction. As I understand it, the current best method involves creating neutral anti-Hydrogen and using a photon trap to cool it and then observe its motion. This was attempted a few years ago, but the error bars were still too large (https://www.nature.com/articles/ncomms2787). A refinement of that experiment may produce a definitive answer (I believe they're currently trying for a 1% measurement error on the intertial properties of anti-Hydrogen).
2) However theoretically absurd it may be, there is always some possibility that our theories are incorrect, and so this kind of experiment is worth doing. Any discrepancy in the behavior of anti-matter compared to normal matter would provide tremendously valuable insights into how to improve our current theories.

I would not question the value of running the experiments. However, I might make the same remark about measuring the gravitational attraction of a mountain range: a reassuring demonstration, and one that might teach us much about how to conduct experiments of that nature, teaching that might be valuable in other experiments. But this is as nearly settled science as we can expect, not a major uncertainty that keeps people from knowing how to make further progress in gravitation.

The odds are overwhelmingly against this observation -- it seems to me -- but that's not a reason not to look.

 

[kimbyd]

I would not question the value of running the experiments. However, I might make the same remark about measuring the gravitational attraction of a mountain range: a reassuring demonstration, and one that might teach us much about how to conduct experiments of that nature, teaching that might be valuable in other experiments. But this is as nearly settled science as we can expect, not a major uncertainty that keeps people from knowing how to make further progress in gravitation.

The odds are overwhelmingly against this observation -- it seems to me -- but that's not a reason not to look.

Right, if there's a discrepancy at all it will most likely not be of the simple form "anti-matter falls up". That's why it's so valuable that they're trying to get 1% measurement error on the equivalence principle applied to anti-Hydrogen: even a small deviation from equivalence would be a truly dramatic finding.

 

[jerromyjon]

But that's exactly how they are "anti's" (with respect to each other), and there's no new physics there: Opposite electric charges attract, always.

Yes, I understand that. Electrons are repelled from protons and positrons are repelled from antiprotons. But when we are referring to say hydrogen and antihydrogen, there would be no net charge imbalance and therefore no attraction or repulsion due to charge. So what would make all the pieces annihilate unless electrons only attract positrons and not repel antiprotons? Or is there a sequence where fermions and bosons interact in succession?

 

[kimbyd]

Yes, I understand that. Electrons are repelled from protons and positrons are repelled from antiprotons. But when we are referring to say hydrogen and antihydrogen, there would be no net charge imbalance and therefore no attraction or repulsion due to charge. So what would make all the pieces annihilate unless electrons only attract positrons and not repel antiprotons? Or is there a sequence where fermions and bosons interact in succession?

If you mix a neutral Hydrogen and anti-Hydrogen gas, they'll still annihilate. They'll just do so more slowly. As they don't have a long-range attraction, you'll have to wait until two atoms randomly get close enough that their electron/positron shells start attracting one another. Once the electron/positron annihilate, there will be a bare proton/anti-proton which will attract one another strongly, and they'll annihilate pretty rapidly.

What happens next depends upon whether the energy from that annihilation escapes the gas or not. If the energy escapes the gas, then the remaining interactions will remain pretty slow. However, if it ionizes the gas, then that may result in an increase in reaction rates (but this may also cause the gas to disperse, making it less dense and slowing the reaction back down).

 

[jerromyjon]

If you mix a neutral Hydrogen and anti-Hydrogen gas, they'll still annihilate. They'll just do so more slowly.

What if they are larger atoms? Then there is no "outward dissipation" of the particles and they all annihilate quickly and completely? What if this galaxy is a rare mix of nearly equal proportions and that is why there is no "dark matter component" to be observed? I mean in that case GR is perfectly fine if annihilation and symmetry don't conflict with it, right? We'd be looking at sides of a coin in most galaxies and the rare spread of proportions... like this, where no dark matter or cosmological expansion is neccesary to model it...

 

[mfb]

What if they are larger atoms? Then there is no "outward dissipation" of the particles and they all annihilate quickly and completely? What if this galaxy is a rare mix of nearly equal proportions and that is why there is no "dark matter component" to be observed? I mean in that case GR is perfectly fine if annihilation and symmetry don't conflict with it, right? We'd be looking at sides of a coin in most galaxies and the rare spread of proportions... like this, where no dark matter or cosmological expansion is neccesary to model it...

Whatever you put together, as long as there is both matter and antimatter in it you will get annihilation. If it is dense enough to form a galaxy, this galaxy (which has no plausible mechanism of forming in the first place) will be immediately obvious to us due to its gamma ray emission.

> What if this galaxy is a rare mix of nearly equal proportions and that is why there is no "dark matter component" to be observed?

Why exactly would you expect such a galaxy, which cannot even form, to have no dark matter?

 

[Mike Johnson]

<Moderator's note: Post approved as alternative to another thread. @Mike Johnson: Please read this discussion first to find out, whether your question has already been answered.>

Could dark matter fill 'empty' space, strongly interact with visible matter and be displaced by visible matter?

Could the reason for the mistaken notion the galaxy is missing dark matter is that the galaxy is so diffuse that it doesn't displace the dark matter outward and away from the galaxy to the degree that the dark matter is able to push back and cause the stars far away from the galactic center to speed up?

What if it's not that there is no dark matter connected to and neighboring the visible matter; it's that the galaxy is not well defined enough to displace the dark matter to such an extent that it forms a 'halo' around the galaxy?

Could a galaxy's halo be displaced dark matter and these types of galaxies are not coalesced enough to displace the dark matter into forming a halo?

 

[PeterDonis]

[Moderator's note: response provided since this poster's question was moved here from a separate thread.]

Could dark matter fill 'empty' space, strongly interact with visible matter and be displaced by visible matter?

No, because of the "strongly interact with visible matter" part. The whole point of dark matter is that it does not strongly interact with anything except through its gravity: no EM, no weak interaction, no strong interaction. If it did interact via any of those three mechanisms, we would have other ways of seeing that it was there besides its gravitational effect.

See post #8 for an example of a way dark matter and normal matter could be separated; note that it involves the normal matter interacting with other normal matter; it does not involve any interaction between normal matter and dark matter.

 

[ohwilleke]

Stacy McGaugh makes an important observation about a paper on NGC1052-DF2 (the "dark matterless galaxy"), in which van Dokkum et al. measure the rms velocity dispersion for DF2 to be 8.4 km/s, with a 90% confidence upper limit of 10 km/s.

It turns out that this is greatly influenced by a key methodological point that is doubtful:

On closer reading, I notice in the details of their methods section that the rms velocity dispersion is 14.3 km/s. It is only after the exclusion of one outlier that the velocity dispersion becomes unusually low. As a statistical exercise rejecting outliers is often OK, but with only 10 objects to start it is worrisome to throw any away. And the outlier is then unbound, making one wonder why it is there at all.

McGaugh also notes that:

I’ve seen plenty of cases where the velocity dispersion changes in important ways when more data are obtained, even starting from more than 10 tracers. Andromeda II comes to mind as an example. Indeed, several people have pointed out that if we did the same exercise with Fornax, using its globular clusters as the velocity tracers, we’d get a similar answer to what we find in DF2. But we also have measurements of many hundreds of stars in Fornax, so we know that answer is wrong. Perhaps the same thing is happening with DF2? The fact that DF2 is an outlier from everything else we know empirically suggests caution.

McGaugh (the leading authority on MOND) notes that the correctly calculated MOND prediction, including the External Field Effect is:

σ = 14 ± 4 km/s.

van Dokkum, et al., incorrectly calculated a MOND prediction of 20 km/s.

So, the evidence that this is really a no dark matter phenomena galaxy is not strong, and the evidence that it contradicts MOND is likewise weak.

[Moderator's note: This was posted in a separate thread while this thread was closed. Since the thread is reopened, this post and its responses have been moved here.]

 

[mfb]

σ = 14 ± 4 km/s.

That is not much better than no prediction at all (the 2 sigma range covers everything from 22 km/s to 6 km/s, a factor of nearly 4), and it is unclear how many of the assumptions that went into that number are included in the uncertainty.

van Dokkum, et al., incorrectly calculated a MOND prediction of 20 km/s.

I don't think it is fair to call it incorrectly. They used a different assumption (e.g. a larger separation from the other galaxy is sufficient).

The outlier has a very large uncertainty and a very large separation from the rest. Using an unweighted rms doesn't make sense. As far as I understand the original paper they don't remove it, they just assign a smaller weight to it according to the larger uncertainty. Which is the most reasonable thing to do.

 

[Vanadium 50]

That is not much better than no prediction at all

That is true. It is, however, key to understanding the dynamics of the galaxy. Ignore MOND. It's still important to the dynamics what the relationship between DF2 and NGC1052 is. And that's far from clear.

One way to look at this galaxy is that there is a discrepancy between the distance we infer from the dynamics, about 8 MPc, and the distance we infer from standard distance measures. The authors' favor the larger distance, which means the dynamics is odd, and it's interpreted as zero gas and zero dark matter (or worse, some gas and negative dark matter ). While I tend to agree with them, a closer distance would change their results (the authors say this as well) and would explain the anomalous brightness of the globulars. You would have to understand why DF2's light profile is 40-45% too smooth (the SBF distance measurement), but that's the only discrepant result left in this view.

 

[PeterDonis]

Moderator's note: A number of off topic posts have been deleted and the thread has been reopened. Please keep discussion within the bounds of peer-reviewed literature relevant to this particular galaxy.

 

[mitchell porter]

"Three papers today on the galaxy apparently lacking DM"