My Dark Galaxy Is Darker Than Your Dark Galaxy

From xkcd : So far Voyager 1 has 'left the Solar System' by passing through
the termination shock three times, the heliopause twice, and once each through
the heliosheath, heliosphere, heliodrome, auroral discontinuity, Heaviside layer,
 trans-Neptunian panic zone, magnetogap, US Census Bureau Solar System
statistical boundary, Kuiper gauntlet, Oort void, and crystal sphere holding
the fixed stars

Neil Armstrong was the first man on the Moon, but that hardly makes Buzz Aldrin a loser. The media love "firsts", naturally enough. Everyone does. But sometimes being first isn't the most important thing. And often, just like with Voyager 1, it's very hard to even say which one was first.

A galaxy called Dragonfly 44 is currently having its 15 minutes of fame in the media, sometimes described as a "new class of galaxy". It seems to have about the same mass as the Milky Way but a hundred times fewer stars. And don't get me wrong, it's a very interesting object indeed, and it's good thing it's getting so much press coverage. But while it may be helpful to sell this as the "first" such detection, in reality things are a bit more complicated than that. So without further ado, let's look at a short selection* of other really dark galaxies, some of them much more extreme than Dragonfly 44.

* This is by no means a complete list. For a reasonable attempt at a complete catalogue, see my recent paper, especially appendix B. I also already covered the Smith Cloud in some detail here and Keenan's Ring here .


1) HI1225+01

Sometimes known as the Giovanelli & Haynes cloud, after its discovers. This was detected way back in 1989 by sheer luck in an Arecibo survey, and was billed as the possible first detection of a "protogalaxy". The cloud is a bloody enormous band hydrogen around 650,000 light years across with the mass of around 4 billion suns. And it seemed to be rotating, too quickly to be held together by the mass of the gas alone. Like nearly all galaxies, it would need some unknown "dark matter" to keep it from flying apart.

Then things got complicated.

See that fuzzy blue patch near the top of the hydrogen ? That's a dwarf galaxy that was spotted soon after the gas was detected. So part of the cloud contains a more-or-less normal galaxy. Could the rest of the gas just be some sort of "tidal" feature of the gas, pulled out of the dwarf galaxy by the gravity of some other passing galaxy ? Unlikely. First, there's nothing else nearby that could have done this, and second, a galaxy as small as the dwarf shouldn't have anything like this much gas.

So while HI1225+01 is one of the first discoveries in this list, no-one really knows what it's a discovery of. We don't know for sure if the cloud is rotating or not. Some people think the southern part of the cloud could still be a dark galaxy in its own right. But no-one really knows.


2) Malin 1

From Bad Astronomy, with another artist's impression of the Milky Way shown
for scale.

Malin 1 is a staggeringly large galaxy, probably the largest spiral galaxy known. It's about the same size as HI1225+01 but a lot more massive, with a mass of hydrogen getting on for 100 billion times the mass of the Sun. As far as I know it was never mistaken for a dark galaxy, but it is very very faint considering its enormous size and mass - but it's hard to say how faint, because that varies. The central parts are as bright as normal spiral galaxies, whereas the outer parts are much fainter.

Overall, it's probably about as bright as normal spirals, but its area is around 100 times larger. In contrast, the gas density is about normal - it's only got a lot of gas because it's so bloody big. Which is weird, because gas density is normally a good measure of star formation. What's going on ? No-one knows. You couldn't call Malin 1 a dark galaxy, but it's obviously a very dim one and it's certainly an extremely strange object.


3) HIJASS J1021+ 6842 (great name, guys)

Those early discoveries of the 1980's were somewhat lucky. Few other dark or weirdly dim galaxy candidates turn up until the early 2000's, when new instruments made it possible to map the hydrogen sky as never before. These early surveys had very poor resolution and sensitivity compared to today's standards, but they did reveal a few weird objects. One such object turned up in the M81 galaxy group.

Image from sky-map.org.

You've probably seen M82 lots of times, because it's a spectacular object. But in this case the interesting stuff is happening around the more obscure IC 2574, which is none too shabby either.

Image from helixgate.net.

But right next to IC 2754 there's a cloud of dark hydrogen about the same size as the galaxy, but of very low mass, probably just a few tens of millions of solar masses (compare that to the other objects we've looked at so far with tens of billions of solar masses). Like the other clouds, it might be rotating. If so it would need around five billion times the mass of the Sun to hold itself together.

Left : low resolution image from the discovery paper, showing the cloud connected to the (much brighter) IC 2754. Right : higher resolution image of the cloud itself, from a follow-up paper from 2005.

But is it really rotating ? Ah, well, that's the thing. Since IC 2574 is pretty disturbed, it makes sense that a massive object might be interacting with it - and the cloud seems to fit that description reasonably well. On the other hand, whatever it interacted with might have pulled out the gas stream and created the illusion of rotation (all we can measure is velocity along the line of sight, not true rotation). More on that problem later. For now, the jury's out on this object.


4) GEMS_N3783_2

From a 2006 paper.

You might think that if a dark hydrogen cloud near a galaxy could just be some form of "tidal debris", then perhaps a better approach would be to look for dark hydrogen clouds far away from any galaxies. And you'd be right, you clever goose. The problem is deciding how far away a cloud needs to be to be sure it's not some form of debris*. That's the problem for this particular cloud, on the far left in the above image. It's at least 1.6 million light years away from the nearest gas-rich spiral galaxy, and contains around 400 million sun's worth of hydrogen. But is this far away enough ?

* And also, looking for hydrogen in this way is still a difficult, slow process. It will be many years - if not decades - before we're mapped the whole sky in this way.

It's not clear. I would say, "probably", but the original authors say, "probably not." There are a few other galaxies a bit closer, but they don't seem to be particularly disturbed, and the mass of gas in the cloud is rather large. Personally I don't think a tidal origin is very likely for this object. The authors main reason to reject this as a dark galaxy is that very few dark galaxy candidates are known... much fewer than models predict. But this is a bit of a self-fulfilling prophecy !

In my opinion this is an under-rated dark galaxy candidate. The really intriguing thing is its velocity width (which can indicate rotation), which is a very respectable 100 km/s. As I've shown in great, great detail, it's really frickin' hard to form tidal debris with properties like this. A more reasonable objection is that the optical data of this region isn't very sensitive - so there could be a faint galaxy there after all, hiding in the noise of the image. Still, that would be not dissimilar to Dragonfly 44, 10 years before that discovery.


5) GRB060505comp

From the 2015 discovery paper. The red contours show the
gas. 19 kiloparsecs is about 62,000 light years. The grey
ellipse shows the size of the telescope beam on the sky.

While we're on the subject of obscure dark galaxy candidates that don't make the headlines, here's one not even the authors suggested. This is another cloud apparently connected to a galaxy. It's extremely massive (about 5 billion solar masses) but much more compact than some of the other examples, probably no larger than the Milky Way, with a fairly high velocity width of 70 km/s. And there's nothing nearby, so a tidal encounter isn't likely. Pretty decent dark galaxy candidate !

But the authors of the paper don't think so. The closest they get is to note that given the sensitivity of the optical data, if this object does have any visible light at all it would be incredibly faint - thousands of times fainter than the Milky Way (or more) despite having about the same amount of gas. They favour that it's an infalling gas cloud, which isn't very helpful. Where has this cloud come from, and why do so few galaxies have associated clouds this massive ?


6) VIRGOHI21

Although I've covered VIRGOHI21 before (see also Robert Minchin's blog post) it's worth re-iterating because this still dominates search results for "first dark galaxy". It's a blob of hydrogen a bit smaller than the Milky Way but with perhaps a hundred times less gas, and no detectable stars whatsoever - despite incredibly sensitive observations with Hubble. If it's rotating it would be made up of at least 99.8% dark matter, with an extremely high line width of 200 km/s.

Long story short, again we don't know if the structure is really rotating or not. So is it a massive dark object which pulled that stream of hydrogen out of the spiral galaxy, or is it just tidal debris, with some other object pulling out the stream and causing the illusion of rotation ? We really don't know. Being inside a whole cluster of galaxies - not just a small group - this is a particularly complicated simulation. Based on the latest simulations (which I hope to publish later this year) I'd have a hard time picking the most likely explanation.


7) Crater 2

No pretty picture for this one because it was discovered by counting stars - it's so faint you can't even see it in the image if you know it's there. Which isn't all that unusual. Crater 2 is remarkable for its size in the Local Group (our patch of turf in the Universe), but we don't know how dark matter dominated it is yet.


8) The AGES Clouds

I've described these before in detail because they're my own research. Short summary : they're eight clouds of hydrogen of a few tens of millions of solar masses, with line widths of up to 180 km/s, floating around in the Virgo cluster. They're at least 300,000 light years from the nearest galaxy and (at most) the same size of the Milky Way (though they could also be smaller). We don't have a good explanation for these objects, but we're pretty confident they can't be tidal debris. Like most of these other candidates, they'd need more than 99% of their mass to be dark matter if they're really rotating as the measurements suggest they are.


9) VCC 1287

Now just to sow some extra confusion into the idea of which dark/dim galaxy candidate was the first, consider this pathetic smudge of starlight. VCC 1287 has been known since at least 1985, earlier than Malin 1 or HI1225+01 - but it was only earlier this year that its internal motions were measured. That found that it had far more dark matter than a typical galaxy of this size or brightness, being at least 99% dark matter but plausibly more like 99.99% (a more typical value might be 90%). It's quite a bit smaller than the Milky Way, having a diameter only about 15% as large. Unlike some of the other candidates it doesn't seem to have any gas at all.

This galaxy and the AGES clouds are all in the Virgo cluster, and all appear to have far more dark matter than expected given their mass of ordinary matter. Could there be a connection ? Could the gas-rich clouds have all their gas transformed into stars and become things like VCC 1287 ? Perhaps. The gas mass in the AGES clouds matches the stellar mass of VCC 1287 pretty well, as does the estimated dark matter mass in each system.


10) Like, all these hundreds of other objects

From the press release about 800 ultra-diffuse galaxies in the Coma Cluster.

Or more probably thousands. I described these results already but it's worth a short summary. Recent advances in instrumentation have made it much easier to detect very faint galaxies, but everyone appears very taken aback by the results. We knew of a few giant "low surface brightness" galaxies like Malin 1, and a lot of smaller ones like VCC 1287. A few people had been saying for many years that we should expect to be able to detect many more faint giant galaxies, but as of last year they're being detected en masse.

So far these new discoveries have been in clusters. They can be as large as the Milky Way but a thousand times fainter. It's difficult to know how dark matter dominated they are - we need to measure their internal motions to properly estimate how much dark matter they contain, and that's very difficult because they're so faint. But given that they'd apparently survived inside a cluster without being torn apart, estimates said they've be made of at least 99% dark matter. And that's what Dragonfly 44 appears to be confirming.


Conclusions

So is Dragonfly 44 really the first of a new class of galaxy ? Not really. Various sorts of incredibly dim galaxies have been known for many years. Some might be purely gas and dark matter with no stars at all, others appear to have no gas in whatsoever, while a very few are very gas rich but have just a few stars too. Many of the pure gas candidates remain controversial - people seem to have this odd bias against believing that gas clouds can endure without forming any stars. Chuck in even just a few stars and it's, "ooooh, look at that faint galaxy"; without them they're all like, "that's just tidal debris". I'm not really sure why.

It's very hard to justify some of the statements in the Dragonfly 44 press release, even (especially) given the need for simplifications. It certainly isn't the first time a galaxy has been detected that needs dark matter to hold it together, because that's true of all galaxies ! Nor is it the first time a galaxy has been shown to have an unusually high dark matter content from its stars rather than gas, as VCC 1287 demonstrates. It's also not the first time the prospect of a really massive dim or dark galaxy has been raised, as we've seen.

But you might be wondering that maybe this is the first time something like this has been directly measured, and that's what all the fuss is about. Well, maybe - but it's far from clear that's the case. Never mind the press releases, in the original paper the authors say :

 ...in our study the total halo mass is an extrapolation of the measured mass by a factor of ∼ 100. A more robust and less model-dependent conclusion is that the dark matter mass within r = 4.6 kpc is similar to the dark matter mass of the Milky Way within the same radius.

So is this galaxy massive or not ? We don't know. The problem is that we can directly measure (assuming our theory of gravity is correct) the mass only where we have something to measure the motion, i.e. stars. But we don't know how much further the dark matter extends - the only way to measure that would be through gravitational lensing, which is tricky to do. Currently the total mass estimate is based on the results of numerical simulations, so I would treat the result with some caution. And for those very few other "ultra diffuse" galaxies for which a mass estimate is possible, it seems that most of them are dwarves rather than giants*.

* Also, one oddity struck me while comparing the VCC 1287 and Dragonfly 44 papers. In the VCC 1287 paper, the authors say they can use the total number of observed globular clusters to estimate the total mass. In the the Dragonfly 44 paper they use the same trick, but seem to think it only tells them about the enclosed mass at that radius - which is very much lower than their highest mass estimate. It's not clear to me what's going on here.

Let's assume that the high mass estimates are correct though. How important would this be ? Well, from the Dragonfly 44 article in the Washington Post :

"We thought that that ratio of matter to dark matter was something we understood. We thought the formation of stars was kind of related to how much dark matter there is, and Dragonfly 44 kind of turns that idea on its head," he continued.

"It means we don’t understand, kind of fundamentally, how galaxy formation works."

On the one hand yes, on the other no. I'd have said that there were more than enough problems with galaxy formation theory to give anyone pause for thought, as the above examples should illustrate. Does Dragonfly 44 change this ? Sort of, maybe. It was already clear that we don't really understand how the gas gets turned into stars very well at all, as we've seen in several examples. Dragonfly 44 takes things further : now it seems we don't understand how normal matter gets into dark matter halos in the first place too.

But... we knew that already ! As I've written about a great many times previously, observations don't show anything like as many galaxy as models predict. What Dragonfly 44 does mean though, possibly, is that the nature of the problem is worse than we thought. Significantly so.

The thing is, if these new discoveries are all dwarves, that doesn't solve the missing galaxy problem because they're aren't nearly enough of them. But it opens the door to a solution : if so many dwarves form just a few stars, maybe lots more form none at all. Yay ! We also know that some of the missing galaxies are so big there's no way they should be able to avoid forming stars. The D44 mass, if correct, would mean that at least some of them do form a very few stars. But why ? Why should some objects form a thousand times more stars than others despite having the same total mass and being in the same environment ? And worse... are there giant dark matter halos out there without any stars at all ?

That's the point. Dark galaxies were thought to be a way to solve the missing galaxy problem, but then that was only thought to apply to the smallest objects - and later, a few which were a bit bigger. There wasn't ever a problem of missing galaxies as large as the Milky Way - although there were clearly problems with star formation, at least the number of galaxies observed was roughly in line with the simulations. But now D44 indicates there might be a problem where we never knew there was one. Maybe there are now large numbers of massive galaxies out there our models didn't predict at all. If that's really the case, it could be very scary and exciting all at once.