The Milky Way's Gamma-Ray Excess Looks Stellar. What Does That Leave for Dark Matter?

At the heart of the Milky Way, the Fermi Large Area Telescope sees more gamma rays than expected. These photons, at energies of a few billion electron volts (GeV), form the Galactic Center Excess. For years, one leading explanation was annihilating dark matter. But the excess does not look like a simple dark-matter signal; its shape is better described by the boxy Galactic bulge, a flattened stellar structure. A new preprint asks how much room is left for particle dark matter if the excess is mostly stellar. The answer, from Fermi-LAT data, is not much for weakly interacting massive particles below a few hundred GeV—if the Milky Way's dark halo is as concentrated as some models predict.

A Glow That Refused to Fit

Fermi-LAT maps the sky in gamma rays. To find a faint signal near the Galactic center, astronomers subtract bright foregrounds: cosmic rays interacting with gas and starlight, plus known sources such as pulsars. A surplus remains at a few GeV.

For a while, that surplus looked like a possible sign of thermal relic dark matter. In many models, weakly interacting massive particles (WIMPs) annihilate in the dense Galactic center and produce gamma rays. The rate depends on the annihilation cross section, a measure of how often the particles collide and destroy each other.

The problem is shape. The excess follows the boxy Galactic bulge better than a spherical dark-matter cusp. That has pushed many researchers toward millisecond pulsars, rapidly spinning neutron stars, and other dim stellar sources.

Two Suspects, One Mixed Model

The new study by Silvia Manconi, Christopher Eckner, Francesca Calore, and Fiorenza Donato tests a mixed model: some GeV emission comes from a stellar bulge, and some could come from annihilating dark matter.

That is a harder accounting problem. Both components would be brightest near the Galactic center and could contribute at similar energies. The authors fit them together, varying dark-matter masses from 10 GeV to 1 TeV. They test two annihilation channels, hadronic bottom quarks and leptonic tau particles, and four plausible dark-matter distributions, from a steep central cusp to a flatter core.

Counting Photons Below the Detection Threshold

The analysis combines two statistical tools. The first, skyFACT, adaptively fits Fermi-LAT maps, allowing the diffuse Galactic emission and other foregrounds to adjust to reduce the chance that mismodeling masquerades as dark matter.

The second, a one-point probability distribution function method, counts photons in small patches of sky. Many faint sources, the suspected pulsar-like population, are too dim to appear individually in catalogs but still contribute photons. The method is sensitive to sources about ten times fainter than catalog detection limits.

Together, the two methods fit bright sources, unresolved sources, diffuse emission, a stellar bulge, and a possible dark-matter component. The authors then derive upper limits on how strongly dark matter could be annihilating without showing up.

What the Fermi Data Leave for Dark Matter

The result is no detection. In the skyFACT fits, very few dark-matter models show evidence above 3 sigma for being needed on top of the stellar bulge. In a representative fit, the stellar component outshines the dark-matter template by more than eight orders of magnitude.

The one-point statistics agree. Across all tested masses and channels, the fit never finds a significant dark-matter component. Instead, the authors set 95% confidence upper limits on the annihilation cross section.

The limits depend heavily on the assumed dark-matter distribution. For a contracted Navarro-Frenk-White profile, with a steep central cusp, their upper limits drop below the thermal relic cross section, the benchmark rate for a simple thermal WIMP, up to about 300 GeV for the hadronic channel and about 80 GeV for the leptonic channel. For a standard NFW profile, those masses fall to about 100 GeV and 30 GeV. A cored Burkert profile, flatter in the center, leaves much more room: its limits stay about an order of magnitude above the thermal benchmark across the tested mass range.

These limits are competitive with other indirect searches. For the hadronic channel and contracted NFW profile, they are about three times more sensitive than combined dwarf-galaxy constraints up to 500 GeV. They also beat antiproton limits below 200 GeV for that profile, though antiproton searches are stronger above 300 GeV.

Why Simulations Matter

The Galactic center is a minefield for systematic error. If diffuse emission is mis-modeled, a fake dark-matter signal can appear or a real one can hide. The authors ran injection-and-recovery tests on simulated Fermi-LAT data.

They created mock skies in which the excess was 100% stellar bulge, 90% bulge plus 10% dark matter, a 50-50 mix, or 100% dark matter. The pipeline recovered the background and known sources reliably and did not miss dark-matter contributions at the level of 10% or more of the excess. Residual mismodeling was about 2 sigma, a level that did not break the limits.

A null-hypothesis test compared the real-sky result with simulations containing no dark matter. The real data were compatible with the no-dark-matter expectation at 68% confidence and within the 95% interval. If the background is mismodeled, the limits generally become weaker, not stronger, so the result is conservative in that respect.

The Halo Problem and the Road Ahead

The biggest caveat is the dark-matter halo. The inner Milky Way's density is uncertain, and the four profiles considered can shift the limits by a factor of about 100. The analysis assumes spherical profiles, while simulations suggest the real halo may be non-spherical. The photon-count fit also uses a narrow energy band, 1.6 to 5.9 GeV, which sacrifices some constraining power but keeps the result conservative.

This is a preprint, not yet peer-reviewed. It does not rule out dark matter. It tightens the space for a particular class of WIMPs, particles that annihilate into gamma rays at GeV energies. If those particles exist, they are quieter in the Galactic center than earlier hopes suggested.

The Galactic center remains a crowded stage. The excess that once looked like a dark-matter beacon now appears increasingly stellar, and that stellar explanation leaves less room for a simple particle signal. The next word will come from better halo models, higher-energy observations, and the Cherenkov Telescope Array Observatory, expected to probe dark matter above 100 GeV. For now, the Fermi data say: whatever dark matter is doing at the heart of the Milky Way, it is not making itself obvious.