Yo Dawg, Herd You Like Missing Matter

So we took your missing matter away so you can miss your missing missing matter.

Ahem.

Galaxies lacking dark matter are a reoccurring topic these days and they're back in the headlines again. Time to take a look at the paper behind the headlines.

To recap, there have been a few different galaxies reported all showing the same basic trend : their stars and/or gas are moving much more slowly, at any given radius, than normal galaxies. Whereas most galaxies rotate so quickly that they require a huge amount of extra, invisible dark matter to hold them together (or equally, their rotation can be said to imply that there's something wrong with our theory of gravity), these ones don't. Apparently, in these handful or peculiar cases, just ordinary stars and gas and good ol' Newton are more than enough.

Thus far, Six galaxies present especially nice examples. Unlike some other, more prominent cases (we'll get back to those later), these are too far away to attribute this rotational oddity to distance measurement errors. To determine the amount of dark matter, we need to know both the speed at which any part of it is moving and the distance of that point from the centre of the galaxy itself. In essence, it's the speed as a function of radial distance, not speed alone, which is the important number. We don't ever measure total dark matter content directly - instead we get the amount of dark matter enclosed within a certain radius.

Measurements of the speed don't directly depend on knowing the distance of the galaxy, but the radial (galactocentric) distance does. And in these six cases, the uncertainty on the distance is far too low to be a possible explanation for their unusual properties. Which is very helpful.

But, there's an added complication. We don't really get the speed directly either. Well, we do, but not the true rotational speed. We get how fast things are moving along our line of sight. To convert this to rotation, we need to correct for how the galaxy is inclined towards us : for edge-on galaxies the measurement is directly equivalent to rotation; for face-on galaxies we can't measure rotation at all. For galaxies somewhere in between, knowing the inclination angle allows us to make a correction. And that, arguably, is a potentially big difficulty.

In the current paper prompting the latest press release, the authors revisit one of their old favourites, using new high-resolution data of the atomic hydrogen content to estimate the inclination angle. They find good agreement with their old data, essentially supporting their original inclination angle estimate and conclusion that it lacks dark matter. They also go much further in their efforts to establish the inclination, looking at a wide variety of possibilities for the true nature of the galaxy, e.g. considering different thicknesses, orientations on the sky, and find a model which minimises the residuals : that is, when subtracting the model galaxy from the original data, it leaves behind nothing but noise.

All this seems to indicate that the galaxy is well-fitted by a model disc of pure gas and stars at an inclination of about 32 degrees. They go further, and try at some considerable length to find if there's even any dark matter model at all which could account for the observations. In short, there isn't : the dark matter would either have a bafflingly low concentration, or just gets the velocity measurement all wrong. The disagreement is so low that it could even be called non-physical.

Interestingly, this galaxy also poses difficulties for alternative theories of gravity. Now Modified Newtonian Dynamics does allow for galaxies which rotation curves like this, but only if they're near to massive galaxies. This one apparently isn't. It seems to be isolated and in nice stable equilibrium. And so MOND, just like the standard model, predicts this galaxy should be rotating much faster than it actually is.

(It's worth remembering that this is ordinarily an advantage of MOND in that it does just as well as the standard model in predicting rotation speeds. But whereas the standard model doesn't absolutely require the dark and normal matter to always be associated with each other, MOND says that in most cases, the same mass distribution should give rise to the same rotation speed. So galaxies like this one ought to be if anything more problematic for MOND than the standard model of dark matter.)

What's the catch ? Is this just a really, really weird galaxy, or is something more pragmatic at work ?

I've favoured different possibilities on this one in the past, but this paper unwittingly swings me (slightly) in favour of the latter. Here's their money figure. It shows an optical image of the galaxy with the hydrogen gas as a white contour. Their best-fit inclination (32 degrees) is shown by the black line, with the blue dashed line at the much lower 11 degrees needed for this galaxy to have a more normal sort of rotation speed :

And for all their hard work and sophisticated modelling, I'm just not convinced that they can really rule out the lower inclination angle possibility. They use the gas disc to estimate the inclination, but look how ragged that edge is... is it really such a stretch to imagine that deeper observations would reveal a better fit for the lower angle model ? Observational errors have a nasty habit of being larger than we would like (estimates of errors should always be treated with caution), and it seems like the galaxy might need to have only a marginally unusual morphology for it to be well-fitted with this lower angle. After all, the inclination angle of the gas and stellar discs can be different, and the angle measured is already very low.

But, this is not to say the case is settled. Far from it. Recall the similar case of NGC1052-DF2, which was also thought to lack dark matter. A lower distance would have settled all the anomalies of that galaxy, so from that perspective a lower distance seemed far more likely... but the best, most recent available distance measurement does appear to be the larger, weirder value (at least, no-one has challenged the claim yet, as far as I'm aware). So I'm going to be very cautious indeed about reaching any kind of firm conclusion here.

To credit the authors, they admit that the inclination angle is the most crucial measurement. And I remain persuaded that their previous arguments about how the low velocity dispersion variation (the random motions of the gas distinct from the rotation) are much more consistent with a nice stable disc than any large-scale disordered motions. So my instinct to say that this oddball galaxy is probably due to a weird inclination measurement problem is not a very strong one; maybe 60:40 odds. As they say, if they could get rotation estimates at even larger distances, this would be really interesting : their interpretation clearly predicts that here we should see declining rotation velocities. How they might get these, though, is hard to say.

Obviously, then, this isn't mystery solved. Rather, we're still at the stage of determining if there is even a mystery at all*.

* Which is a lot better than the recent approach of the Metropolitan Police in investigating crimes.