New gravitational-lensing study hints at problems for dark matter models

This site is reader-supported. When you click through links on our site, we may be compensated.

Image of a large group of galaxies, with inset magnified views of some of them.
Enlarge / The massive galaxy cluster MACSJ 1206. Embedded within the cluster are the distorted images of distant background galaxies, seen as arcs and smeared features.


While the idea of dark matter was originally proposed to explain the structure of galaxies, one of its great successes was explaining the nature of the Universe itself. Features of the Cosmic Microwave Background can be explained by the presence of dark matter. And models of the early Universe produce galaxies and galaxy clusters by building on structures formed by dark matter. The fact that these models get the big picture so right has been a strong argument in their favor.

But a new study suggests that the same models get the details wrong—by an entire order of magnitude. The people behind the study suggest that either there's something wrong with the models, or our understanding of dark matter may need an adjustment.

Under a lens

The new study, performed by an international team of researchers, took advantage of a phenomenon called gravitational lensing. Gravity warps space itself, and it can do so in a way that bends light, analogous to a lens. If a massive object—say, a galaxy—sits between us and a distant object, it can create a gravitational lens that magnifies or distorts the distant object. Depending on the precise details of how the objects are arranged, the results can be anything from a simple magnification to circular rings or having the object appear multiple times.

Because dark matter's effects are detectable via gravity, we can “see” the presence of dark matter via its gravitational-lensing effects. In a few cases, we've even detected lensing where little matter is present. That's one of the many pieces of evidence in favor of dark matter.

The researchers used gravitational lensing to set up a test that, at least conceptually, was very simple. We've built models of the early Universe that indicate how dark matter helped structure the first galaxies and drew them into clusters of galaxies. These models, when run forward, provide a description of what that dark matter distribution should look like at different points in the Universe's history up to the present. So, the researchers decided to use gravitational lensing to determine whether the dark matter distribution seen in the models matched where we see it via gravitational lensing.

According to these models, the Universe was built hierarchically. Via gravitational interactions with itself, dark matter formed filaments that intersected in a complex, three-dimensional meshwork. The additional gravitational pull at the points where filaments intersected would draw in regular matter, leading to the first galaxies. Over time, the continued draw of gravity pulled galaxies together, forming large clusters. By examining the output of these models, we can get a look at the expected distribution of dark matter around clusters. And, by zooming in, we can see how dark matter should be distributed in the area of individual galaxies.

That distribution of dark matter can be viewed as a prediction of the models.

Meanwhile, in the actual Universe…

To test those predictions, the researchers used images from the Hubble space telescope to map out all the objects in and around a large collection of galaxy clusters. Follow-up imaging using the Very Large Telescope helped identify the distance of those objects based on how much their light was shifted to the red end of the spectrum by the expansion of the Universe—the larger the redshift, the more distant the object. This allowed the researchers to determine which objects must be behind the galaxy cluster and, thus, potential candidates for gravitational lensing.

A software package then used the data to create a mass distribution for each galaxy cluster. This included the overall lensing effects of the entire cluster, as well as the sub-lensing driven by individual galaxies within the cluster. The researchers found a strong agreement between the appearance of lensed objects and the location of individual galaxies, which allowed them to validate their mass-distribution calculations.

The researchers then used the Universe simulator to build 25 simulated clusters and performed a similar analysis with the clusters. They did so in order to identify the sites of possible lensing and the locations that could create the greatest distortions.

The two didn't match. There were significantly more areas that generated high distortion in the real-Universe galaxy than there were in the model. This would be the case if the distribution of dark matter were a bit more lumpy than the models would predict—the dark matter halos around galaxies were more compact than the models would predict.

This isn't the first discrepancy of the sort we've seen. Dark matter models also predict that there should be more dwarf satellite galaxies around the Milky Way and that they should be more diffuse than they are. But if we were to adjust our models to make these galaxies more diffuse, then we'd be less likely to see more compact structures in galaxy clusters. So, rather than finding two problems that could both be solved by making one adjustment, the two issues appear to need adjusting in opposite directions.

Two possibilities

The researchers suggest there are two likely explanations for this discrepancy: either we don't appreciate all the properties of dark matter, or there's something missing from our simulations of the Universe's evolution. Since both of those get the big picture of the Universe largely right, however, the issue is going to be a subtle one and consequently difficult to identify, should these results get an independent confirmation. One possibility is that the problems seem to be in the area of galaxies, where there would be a lot of matter-dark matter interactions. If there's something more complicated going on there, it could easily throw off the models.

For now, however, there are already likely to be teams with additional data in hand that could perform similar analysis, so we'll have to wait for those to be done. Theoretical cosmologists, being the impatient sort, will undoubtedly be testing out dark matter variants long before any additional reanalyses are out.

Science, 2020. DOI: 10.1126/science.aax5164  (About DOIs).

              <div id="action_button_container"></div>

Leave a Reply

This site uses Akismet to reduce spam. Learn how your comment data is processed.