Stellar-like Galactic center excess challenges particle dark matter

TL;DR

This work reassesses the Galactic Center as a DM search target by jointly fitting a DM annihilation template and a stellar bulge template to Fermi-LAT gamma-ray data, using an adaptive skyFACT template fitting stage to minimize diffuse mismodeling and a subsequent 1pPDF photon-count analysis to constrain DM while accounting for unresolved sources. The authors test a range of DM masses (10 GeV–1 TeV), annihilation channels ($b\bar{b}$, $\tau^+\tau^-$), and density profiles (NFW100, NFW126, Einasto, Burkert), and they propagate the DM signal through both the skyFACT and 1pPDF analyses to obtain 95% CL upper limits on $\langle \sigma v \rangle$. Across most DM configurations, no significant DM detection is found; the strongest constraints arise for a contracted NFW profile (NFW126), excluding thermal relic cross sections up to $\sim 3\times10^{-27}$ cm$^3$/s for $m_\mathrm{DM}\lesssim 300$ GeV in the $b\bar{b}$ channel, while Burkert profiles yield weaker limits. The results are robust to simulated data tests and competitive with limits from dwarf galaxies and other messengers, highlighting the continued power of GC gamma-ray observations in probing light DM, albeit with remaining halo-uncertainty systematics and a narrow energy window. The methodology is extensible to broader DM models and higher energies, with future synergy anticipated from CTA for heavier masses.

Abstract

The Galactic Center (GC) is potentially hosting the largest indirect signal from particle dark matter (DM), which in many well-motivated models would produce gamma rays as their final states. However, this region has often been dismissed for DM studies because of the evidence for an unexpected gamma-ray component over astrophysical backgrounds at GeV energies, firstly discovered in the data of the \textit{Fermi} Large Area Telescope (LAT), the so-called Galactic Center Excess (GCE). While this was initially considered to hint at GeV thermal relics, recent work supports a GCE interpretation in terms of a stellar population of millisecond pulsar-like sources in the Galactic bulge. Building on this preference, we re-evaluate the GC as a powerful target for indirect DM searches via gamma rays. This is achieved by combining adaptive template fitting and pixel-count statistical methods to assess the role of sub-threshold point sources in the observed \textit{Fermi}-LAT gamma-ray counts, while minimizing the mismodeling of Galactic diffuse emission backgrounds. In a fully self-consistent way, the gamma-ray data are fitted with a mixed model comprising a DM signal and a stellar bulge, both potentially contributing to the GCE. The space left for signals from weak-scale DM particle annihilations is quantified by extracting 95\% C.L. upper limits on the annihilation cross section, which, depending on the DM density profile, result in stringent limits for masses $\lesssim 300$ GeV. The robustness of our results is supported by tests on simulated data.

Figures (7)

• Figure 1: Stellar bulge (left) and DM (right) gamma-ray flux maps integrated in the energy bin 1.6 -- 5.9 GeV as obtained with skyFACT for the benchmark case of NFW126 density profile and DM of 40 GeV annihilating in the $b\bar{b}$ channel. The flux maps correspond to the best-fit model, and are reported in cartesian coordinates. The full region of interest (40$\times$40 square degrees) is the one considered for the skyFACT analysis, while the dashed lines delimitate the region used to analyze the data with the 1pPDF method and to derive the DM constraints. The colorbars for the two panels are different, and they both span about five orders of magnitude.
• Figure 2: Residuals of the null hypothesis skyFACT fit in the energy range from 1.6 -- 5.9 GeV. The residuals are defined in terms of (data - model)/$\sqrt{\mathrm{data}}$ and smoothed with a Gaussian kernel of size $0.75^{\circ}$.
• Figure 3: Constraints to thermal relic DM obtained in this work using Fermi-LAT gamma rays from the GC. The upper limits (95% C.L) on the annihilation cross section $\langle \sigma v \rangle$ as a function of the DM mass are shown for the $b \bar{b}$ ($\tau^+ \tau^-$) channel in the left (right) panel. We show the results obtained when varying the Milky Way DM halo profile among the four benchmarks considered: a contracted NFW profile (NFW126, solid blue), a classical NFW (NFW100, dashed magenta), Einasto (dotted green) and the cored Burkert profile (yellow dot dashed). The thermal relic cross section as estimated in Steigman:2012nb is also indicated with a thin black line.
• Figure 4: Comparison of the results obtained in this work with complementary multimessenger DM constraints. The left (right) panel illustrates the upper limits obtained using the GC gamma rays and assuming contracted NFW profile ($\gamma=1.26$, solid blue) and the classical NFW profile ($\gamma=1$, dashed magenta) for the $b \bar{b}$ ($\tau^+ \tau^-$) channel. These are compared with the upper limits obtained with a similar contracted profile, but using the local cosmic ray $\bar{p}$ fluxes in Calore+22 Calore:2022stf, as well as the constraints obtained combining multiple gamma ray observations of dwarf galaxies in Abdollahi+25 Fermi-LAT:2025fst.
• Figure 5: Performance of skyFACT on simulated data: residual background mismodeling. We illustrate the spectral residuals per energy bin shown as red data points. These residuals were derived for a single realization of the GCE scenario featuring 100% stellar bulge and 0% DM emission. The displayed residuals are defined as (data - model)/$\sqrt{\mathrm{data}}$. The thick red line marks the residual obtained for the integrated energy range we consider in the 1pPDF analysis. The gray band denotes the energy range considered in the 1pPDF analysis part.