Using Gravitational Lensing to Hunt for Hidden Knots of Dark Matter

Oct 18, 2016

Warren Morningstar

by Warren Morningstar

By modeling the warped images of a gravitational lens observed with one of the most powerful telescopes in the world, KIPAC scientists have made the dramatic discovery that there is a clump of dark matter with no currently visible normal-matter counterpart in a far-away galaxy. Such unaccompanied clumps are incredibly difficult to detect and only a small handful of them have ever been discovered, but a concerted effort to find them and determine how and why they form could pay off significantly in the long term by giving us new insights as to the nature of dark matter.

We look out at the universe in a myriad of different ways, and all of them are telling us the same story: the universe is thirteen and change billion years old, it is made up of roughly 70 percent dark energy, 26 percent dark matter, and 4 percent normal matter, and the dark matter is probably a particle that is cold (that is to say, it moved fairly slowly in the early universe—and still does, at least with respect to the speed of light).

This picture has been confirmed exceptionally well on large scales using radio observations of the cosmic microwave background, optical observations of Type Ia supernovae, and observations of the baryon acoustic oscillations. Comparisons between galaxy surveys and numerical simulations of dark matter also support the picture. These simulations find that dark matter naturally forms into large clumps which we call “halos,” because they surround galaxies in much the same way of that ethereal glow that always surrounds angels in artists' depictions (which is my best guess as to why they’re called halos). Halos are much more massive than the galaxies inhabiting them (about 6 times more, on average), and because dark matter doesn’t condense like normal matter does, they are also significantly larger in size.

When we simulate individual halos, we find that because the dark matter is cold, it should further collapse into tens of thousands of even smaller clumps, which we call subhalos. These subhalos are massive enough that, using what we know about conventional galaxies, we might expect that they host small galaxies of their own. These dwarf galaxies should thus be plentiful around galaxies like the Milky Way. When we look for these dwarf galaxies, we find a few dozen of them, but not nearly as many as the theory predicts should be found.

The apparent lack of dwarf galaxies could be explained in a number of ways, but two explanations in particular have gained a lot of ground. The first is that dark matter is not as cold as we once thought, meaning that it had a little bit of random motion in the early universe. This motion would have made the dark matter particles spread out more and clump together less overall, and thus could have provided enough dispersion of the particles to prevent most subhalos from being able to form, leading to an expected number of dwarfs that agrees with observations. Alternatively, dark matter could be cold as we think it is, but low-mass galaxies simply cannot form in the subhalos (and there are multiple unknowns relating to how and why stars form etc. that might explain why this might be the case). Because the second option predicts that most subhalos do not have galaxies, it also says that most subhalos also cannot be seen with telescopes. This implies that what we really need is a means of looking for subhalos that doesn’t require them to host galaxies.

One way to do this is by exploiting a phenomenon called gravitational lensing. Gravitational lensing is caused by light being deflected by a massive object (the lens) located between an observer (us) and a source of light (a distant galaxy). Through the deflection, an image of the source can be stretched, warped, magnified, and multiplied before we observe it, producing stunning images like those seen in the image below.

Gravitational lens SPD81

Above: By analyzing distortions in this ALMA/Hubble composite image of the gravitational lens SDP.81 (red arcs are the background galaxy; the central blue blob is the nearby lensing galaxy that provided the magnification), astronomers have determined that a dark dwarf galaxy (the white dot at lower left) is lurking nearly 4 billion light-years away. (Credit: Y. Hezaveh, Stanford Univ.; ALMA (NRAO/ESO/NAOJ); NASA/ESA Hubble Space Telescope.)

By modeling these images of the background source, the structure of the lens can be determined without actually looking at the lens itself. It all sounds very straightforward, but the techniques to do this are very challenging and require exquisitely accurate observations of the lensed sources as well as the use of immense supercomputers to handle the analysis. This becomes even more difficult when looking for subhalos, since their light-bending effects are thousands of times (or more) smaller than the deflections from full-sized dark matter halos.

Fortunately, the tools necessary for the exquisite observations that can be used to detect subhalos exist. In particular, an array of 50 twelve-meter telescopes called the Atacama Large Millimeter/Submillimeter Array (ALMA), located 16,500 feet above sea level in one of the driest places on Earth, the Atacama Desert of Chile, has recently begun taking some of the highest-resolution observations of systems such as these ever attempted. ALMA achieves this remarkable resolution by tightly integrating all of the telescopes together, mimicking the capabilities of a telescope that is sixteen kilometers across. In effect, this telescope could be used to resolve a quarter located in New York if the observer were located in San Francisco, or to resolve a person standing on the Moon.

The ALMA radio telescope

Above: The Atacama Large Millimeter/submillimeter Array (ALMA) on the Chajnantor Plateau in Chile. (Courtesy: Clem & Adri Bacri-Normier/ESO)

In early 2015, ALMA was used to observe the lensed galaxy SDP 81. This was the first time that ALMA had observed a gravitational lens using the fully extended array described above. With these observations, a team led by KIPAC researcher and Hubble Fellow Yashar Hezaveh and including myself used a software package we had collectively been developing for a few years to attempt to detect subhalos in the galaxy. To do this, we first reconstructed the un-lensed image of the background source and determined the smooth distribution of matter in the lens. Then we tested our model to determine whether or not it could be improved by the addition of clumpy substructure.

By doing this, we found that the data prefer a subhalo with a mass approximately equivalent to a billion solar masses that is unseen in any visible light band images of the lens. We were also able to say conclusively that the data favor a lack of subhalos in a number of locations and across a range of subhalo masses. This is also important because both detections and exclusions are necessary for determining lower and upper limits on the numbers of subhalos contained in the lens. These limits are then used to tell us whether or not there are an appropriate amount of subhalos that just cannot be seen, or if there is something wrong with our understanding of dark matter. The image below isn’t as pretty, or as easy to understand as the first one, but it illustrates how the goodness of the model was affected by the placement of a subhalo at various locations in the lens, with red being a strong improvement, and black being a strong decrement (Upshot: At a particular location this improvement was strong enough to claim a detection!).

Maps of SPD.81 with added clump of dark matter

Above: Maps of the gravitational lens SDP.81 with added subhalo (indicated by contour lines in the insets), based on two different ALMA datasets (left, middle) and the datasets combined (right). The colors indicate the strength of detection or exclusion of a subhalo, with black being strongly excluded, and red being strongly detected. A subhalo is strongly favored in the region shown in the box, and looks weakly favored in a number of other locations. When the detected subhalo is added to the model, the other potential detections go away.(Credit: Hezaveh, et. al.)

Perhaps more importantly, these results have demonstrated the potential for using ALMA to detect these subhalos and to make these constraints. This is a highly significant result, because there are a hundred or so of these lensed galaxies that ALMA could observe in the future. With ALMA observations of even a fairly small subset of these lenses, we could measure the abundance of subhalos precisely enough to be able to determine where all the missing dwarf galaxies have gone.


Detection of lensing substructure using ALMA observations of the dusty galaxy SDP.81, Hezaveh, et. al

The hunt for invisible dwarf galaxies   Deep dive; 10-min read.

Astronomers Spot Dwarf Dark-Matter Dominated Galaxy Four Billion Miles Away   Just the facts; 3-min read.

Dwarf dark galaxy hidden in ALMA gravitational lens image   Good overview; 5-min read.

Elusive Dark Matter Galaxy Revealed by Cosmic Lens   Very accessible; 5-min read.