Fermi unassociated-source search finds no dark-matter satellites

The editorial temptation is to treat unidentified Fermi sources as hidden dark-matter satellites. The preprint does the opposite of a discovery claim. Starting from 2,428 unassociated sources in the 14-year 4FGL-DR4 catalog, the authors apply property filters expected for Galactic subhalos that never formed stars, then test the survivors with a spectral Fermi-LAT analysis. This is a preprint, not a journal result, and it should be read as a constraint paper. The corresponding author listed is Fernando Valenciano; the abstract does not name a journal.

They target weakly interacting massive particles whose annihilation would yield gamma rays, particularly in the bottom-antibottom channel, inside starless subhalos predicted by hierarchical assembly in the standard cosmological model. Filters shrink the unassociated list before a spectral fit asks whether a dark-matter spectrum beats ordinary source models. Methods note: this is a catalog selection plus spectral analysis of Fermi-LAT data drawn from 2,428 unassociated sources, not a spatially resolved subhalo detection, and the abstract omits the size of the sample that survived the filters, which is the number a reader needs to judge the search.

None of the remaining sources shows a statistically significant preference for dark matter. Conservatively, cross-section limits sit a factor of two to three above the thermal-relic benchmark, near 3×10^-26 cm^3/s, at the lowest masses. Only under a sensitivity-reach assumption do the authors exclude thermal WIMPs below about 50 GeV in the bottom-antibottom channel. Stacked dwarf searches, starting with the Fermi-LAT six-year analysis in Physical Review Letters in 2015, already exclude thermal WIMPs to higher masses. Earlier unassociated-source papers by Bertoni, Hooper and Linden and by Coronado-Blázquez and colleagues treated non-detections as population limits. Complementarity, not a candidate list, is the result.

The open question is whether the null result constrains the particle or the predicted satellite population. Subhalo counts, distances and J-factors come from cosmological simulations, not from this catalog, so a sparser population can mimic a particle-physics non-detection. A spectrum-only test can miss extended emission, and pulsars or faint active galaxies still fill unassociated samples until other wavelengths identify them. Stronger evidence would require peer review, the number of sources passing each filter, and a joint likelihood with dwarf galaxies instead of a sensitivity-reach exclusion.