Dark matter and galaxies normally go hand in hand. Dark matter seems to be needed to draw in sufficient material to form the galaxy and its stars, and halos of dark matter keep galaxies from spinning apart as they rotate. So scientists were more than a bit surprised to find a galaxy that has little to no dark matter at all. Confusing things further, the galaxy appears extremely similar to others that are nearly entirely composed of dark matter.
Slow motion
This was one of those cases where discovery began with the phrase "huh, that looks weird." The weirdness came courtesy of the Dragonfly Array, a collection of small telescopes designed to pick up faint objects. When observing a collection of galaxies called the NGC 1052 group, the array spotted an object that had shown up in other surveys of that region of the sky.
"It stood out to us because of the remarkable contrast between its appearance in Dragonfly images and Sloan Digital Sky Survey data," the authors of the paper describing the object write. Dragonfly saw it as a diffuse object with some structures in it; Sloan imaged it as a collection of distinct objects.
Closer observations using the Hubble and 10-meter Keck telescopes showed that the object, now termed NGC1052–DF2, was a member of a recently discovered class of objects called ultra-diffuse galaxies. "This thing is astonishing," said team member Pieter van Dokkum, "a gigantic blob that you can look through. It's so sparse that you see all of the galaxies behind it. It is literally a see-through galaxy."
Ultra-Diffuse galaxies had been identified previously in other galaxy clusters. Some of them appear to be normal galaxies that were stripped of most of their stars and gas by interactions with other members of the cluster, leaving them incredibly dark-matter rich.
To see whether this is the case with NGC1052–DF2, the researchers used the Keck telescope to image some of the structures that appeared to be present within the galaxy (the objects appeared to be bright versions of the globular clusters seen in other galaxies, including our own). Imaging them allowed the research team to determine their motions orbiting the galaxy, which could, in turn, provide a measure of the amount of mass that is involved in keeping the objects in place.
The answers were pretty surprising. The data on the 10 globular clusters the team tracked showed them moving much more slowly than would be expected. That led to an estimated mass that was extremely low for a galaxy—on the order of 108 solar masses. Using the amount of light emitted by the galaxy produced an estimate of the total mass of stars in the galaxy that was also in the neighborhood of 108. Normally, we infer that there's dark matter around because the galaxy appears to have a lot more matter than the amount provided by the stars we can see. But in this case, there's a minimal difference between the two.
Not so dark
In fact, the best fit for the motion of the globular clusters is a galaxy with no dark matter at all, although the uncertainties are high enough that there could be equal amounts of matter and dark matter present. But there's normally much more dark matter than ordinary stuff in most galaxies, so the authors conclude that, "NGC1052–DF2 is extremely deficient in dark matter, and a good candidate for a 'baryonic galaxy' with no dark matter at all."
The researchers seemingly anticipate the sorts of questions that will be posed by people who see this and wonder why we need any dark matter around anyway (a population that includes a number of regular Ars readers). It turns out that a galaxy without dark matter is incompatible with models that replace it using modified gravity. Since there's some normal matter here, any version of modified gravity would have that matter produce dark-matter-like effects. We don't see any indication of these effects, so modified gravity ideas must be wrong.
"Paradoxically, the existence of NGC1052–DF2 may falsify alternatives to dark matter," the authors conclude, noting that those alternatives include both variations of MOND and emergent gravity.
The obvious question is how you end up with a galaxy like NGC1052–DF2. More generally, how do you form a galaxy without dark matter? One idea is that the gas got concentrated as it was being ejected from a galaxy merger; another is that it formed from matter spewed out by quasars. The final option the authors suggest is that it formed as a large, nearby galaxy drew in a tremendous amount of gas, creating star-forming eddies.
But each of these would have to be examined in detail to determine if it could form the distinct features of NGC1052–DF2. "It's like you take a galaxy and you only have the stellar halo and globular clusters, and it somehow forgot to make everything else," noted van Dokkum.
The researchers suggest that the somewhat-odd appearance of the globular clusters is probably related to the galaxy's unexpected properties, and they announced they're working on a paper that will describe those. They also say that they've identified a number of other galaxies that look fairly similar. Of course, they point out that most of the things we've found that look similar have turned out to have an excess of dark matter, so there's a lot we still need to figure out here.
Nature, 2017. DOI: 10.1038/nature25767 (About DOIs).
[contf] [contfnew] 
Ars Technica
[contfnewc] [contfnewc]
