Could the Emptiest Places in the Universe Reveal Dark Matter’s Identity?

Imagine a region of space so underdense that a galaxy survey finds very few galaxies there. Cosmic voids are not just blank spots in the cosmic web. They are the most underdense regions of the cosmic web, where matter is sparse and nonlinear structure formation is less advanced, and they may be unusually clean witnesses to the nature of dark matter.

That is the promise of a new preprint by Alexander Spencer London and colleagues, posted on arXiv on 26 February 2026. The team asks a simple but deep question: if dark matter includes an ultra-light axion component rather than being purely cold and sluggish, would the universe’s emptiest places look different? Their simulations say yes, and their forecasts suggest upcoming galaxy surveys could test the idea.

Why a wave-like dark matter particle would change the furniture

Dark matter shapes the universe’s large-scale structure, but no one knows what it is. The standard cosmological model, cold dark matter, treats it as a slow-moving particle that clumps readily. An ultra-light axion is a hypothetical particle so light that its quantum wavelength stretches across astrophysical scales. Instead of behaving like tiny grains, it behaves more like a wave. On large scales it can mimic cold dark matter, but on smaller scales its wave nature suppresses the growth of structure. The lighter the axion, the larger the affected scales.

Voids are attractive targets because they are less touched by the messy, nonlinear physics that shapes galaxies and clusters. Baryonic feedback—the astrophysical effects associated with galaxy formation—also has less purchase in these rarefied regions. The paper uses voids as a low-noise channel for a dark matter signal.

Smaller voids disappear into larger ones

To see what axions would do, the researchers ran dark-matter-only simulations using a fast method called mass-peak patch. They generated boxes 2,048 megaparsecs on a side, compared a standard cold dark matter universe with universes where axions made up 10% of dark matter, and varied axion mass from 10^-26 to 10^-23 electronvolts.

In the simulations, lighter axions suppress the formation of low-mass halos. Those halos are tracers used to define voids; when they are suppressed, smaller voids merge into fewer, larger ones. Small voids with radii below about 25 Mpc/h, a cosmological distance unit, merge into larger voids above that scale. For a 10^-26 eV axion, the average void radius rose from 31.6 to 39.2 Mpc/h compared with cold dark matter.

Larger was not fuller. The voids also became emptier of halos. The team measured a stronger anticorrelation between voids and halos—fewer halos than average near void centers—and deeper void density profiles. The void mass function, which counts voids by mass, dropped in amplitude, while the void size function shifted toward larger radii. These effects strengthened as the axion mass decreased.

A forecast, not a detection

The second half of the paper turns these simulation signatures into a forecast. The team modeled the void size function—the number density of voids as a function of their radius—and asked how well three galaxy surveys could constrain axions. They considered eBOSS-like, DESI-like, and Euclid-like surveys. For the sharpest forecast, they combined a Euclid-like survey, with an effective volume of 73 cubic gigaparsecs, with an expected prior on other standard cosmological parameters from the Simons Observatory’s cosmic microwave background measurements.

For an axion mass of 10^-25 eV, that combination could limit the axion energy density to less than 4.6% of dark matter at 95% credibility. That is about twice as strong as current best limits for this mass, the authors report. A DESI-like survey reaches comparable sensitivity. If axions made up 10% of dark matter, as some string theory scenarios motivate, then for that same 10^-25 eV axion mass a Euclid-like survey could recover the signature with roughly a 2 sigma preference; a DESI-like survey would do so with about 1 sigma.

The authors also find that voids and galaxies constrain dark matter in different ways. The parameter degeneracies—the trade-offs that let one cosmological parameter mimic another—are not the same for void size data and galaxy clustering. That makes future combined analyses more powerful than either probe alone.

What the emptiness could still hide

These are simulations and forecasts, not an observation of axions. The paper is a preprint, so its results have not completed peer review. The survey forecast makes several simplifying assumptions: it treats mock voids as if they were all in the present-day universe, ignores the way galaxy motions distort distances, and does not model the galaxy-halo connection needed to compare with real data. It also models axions mainly through their effect on the linear matter power spectrum, leaving out additional wave effects on small scales. The authors note that lower-mass halos and smaller voids, which upcoming surveys may resolve better, could make the real constraints stronger.

The value of nothing

Cosmic voids are often treated as absence—empty rooms in the architecture of the universe. This work suggests they may be more like quiet instruments. If dark matter includes an ultra-light wave component, it would leave a signature not by adding light or mass, but by changing how emptiness organizes itself. Ongoing and future galaxy surveys will scan those quiet regions, turning the universe’s blank spaces into a test of what dark matter really is.