Galaxy Cluster
© IllustrisTNGGalaxy clusters generated by the Universe simulator IllustrisTNG.
From our current perspective, the Universe seems to be dominated by two things we find frustratingly difficult to understand. One of these is dark matter, which describes the fact that everything from galaxies on up behaves as if it has more mass than we can detect. While that has spawned extensive searches for particles that could account for the visual discrepancy, it has also triggered the development of alternative theories of gravity, ones that can replace relativity while accounting for the discrepancies in apparent mass.

So far, these proposals have fallen well short of replacing general relativity. And they say nothing about the other big mystery, dark energy, which appears to be accelerating the expansion of the Universe. Instead, researchers have developed an entirely separate class of theories that could modify gravity in a way that eliminates the need for dark energy. Now, researchers have run simulations of galaxy and star formation using this alternative version of physics, and they found we might be on the cusp of testing some of them.

Gravitational alternatives

General relativity explains a broad range of phenomena, and it works well to describe the Universe as a whole, provided dark matter and dark energy exist as separate entities. Any alternatives to gravity have to account for everything that's explained by general relativity while also accounting for the additional effects of at least one of these two dark forces. A class of theories, collectively termed MOND (for Modified Newtonian Dynamics), is intended to do away with dark matter, but it struggles to account for things relativity handles with ease.

And, when it comes to dark energy, MOND is silent, in part because it was originally developed before dark energy was known to be an issue. Instead, an entirely separate class of theories has been developed that handles gravity while eliminating the need for a separate dark energy. These theories are known as f(R) models and are commonly described as having a "chameleon" mechanism. That's because they posit an additional force that changes its behavior based on its surroundings.

Where there's a lot of matter, the chameleon force is minimized, allowing it to blend in with its surroundings. As matter becomes sparse on larger scales, it starts to make its presence felt. That's why we can't detect any major deviations from relativity on Earth or near objects like neutron stars, but we do detect them when we start looking at the large-scale structure of the Universe. The net result is an acceleration of the expansion of the Universe that's only apparent at large scales — just like dark energy.

For any additions to physics to be successful, they have to make sense with what we know of relativity, plus handle details it can't. That makes it difficult to test, because the new models are already crafted to match existing data (and would be pretty pointless if they weren't). So, the trick is finding data we don't yet have but could show a difference between relativity and these new models.

To search for these sorts of discrepancies, a group of cosmologists at Durham University decided to plug some chameleon proposals into massive computational models that simulate the formation of structures ranging from stars to galaxy clusters.

A model Universe

The researchers worked with the IllustrisTNG model, a mini-Universe that can simulate galaxy formation and evolution. Under standard conditions, the model controls this evolution in part by having everything obey general relativity. But for this test, the research team also ran a version where relativity was replaced with a chameleon f(R) version of gravity. (They also ran an exaggerated version of f(R) in order to accentuate the differences.)

All models assumed the presence of large amounts of dark matter; remember it's MOND that hopes to get rid of that. Simulations were run under two conditions: with feedback from regular matter, and without. Unlike dark matter, regular matter ignites into stars and forms black holes, and those provide feedbacks that alter the behavior of nearby matter.

The simulations indicated that the gas in the inner regions of galaxies doesn't feel the effect of modified gravity, behaving much as it would with general relativity. This includes gas flowing into the area near supermassive black holes that power active galaxies. By contrast, the outer regions of galaxies should show some differences because of the changes caused by the chameleon force. Here, additional stars are expected to form due to changes in the dynamics of gravity under the chameleon model.

Too small

Unfortunately, most of these effects are too small to create detectable differences between f(R) and general relativity. There is, however, one exception. The changes to the gas in the outer region of galaxies causes higher densities of gas to form there, which in turn increases the efficiency of cooling of that gas. It turns out that an instrument of the Square Kilometer Array radiotelescope will be sensitive to the altered properties of the gas. As a result, it may be able to pick up any deviations from general relativity.

The other result that was significant here is the finding that, for chameleon models that are similar to general relativity, the effect of including feedbacks from regular matter is simply an additive effect; there are no further interactions with modified gravity. This would allow future computations to be considerably simplified.

So, while we're not yet ready to start ruling out alternatives to general relativity, the new work highlights the sorts of things we need to do to be able to test potential replacements. Because relativity has been so successful and seemingly explains so much of what we see, there's not much space for alternatives to stake out a distinct identity. By putting in the effort to figure out where that rare space may reside, research like this sets up the possibility of ultimately putting some of our ideas to the test.

And, well, if they fail the test, there's still dark energy.

Nature Astronomy, 2019. DOI: 10.1038/s41550-019-0823-y (About DOIs).