Constraints on decaying Dark Matter from XMM-Newton observations of M31

TL;DR

This work uses XMM-Newton observations of the outer M31 region to set stringent limits on radiatively decaying dark matter, focusing on sterile neutrinos in the Dodelson-Widrow framework. By modeling multiple M31 dark matter distributions, carefully subtracting backgrounds with ESAS and SBS, and employing both a Gaussian-line (statistical) and a full-flux approach, the authors derive robust upper limits on the decay width and mixing angle as a function of mass. The results strengthen prior bounds, yielding m_s < 4 keV for the DW scenario when combined with contemporary Lyman-α bounds, effectively ruling out the case where all DM is produced via DW, while leaving room for production via resonant or alternative mechanisms. The analysis demonstrates that M31, with proper DM-mass modeling and background treatment, provides a powerful diagnostic for decaying DM scenarios with observable X-ray signatures.

Abstract

We derive constraints on parameters of the radiatively decaying Dark Matter (DM) particles, using XMM-Newton EPIC spectra of the Andromeda galaxy (M31). Using the observations of the outer (5'-13') parts of M31 we improve the existing constraints. For the case of sterile neutrino DM, combining our constraints with the latest computation of abundances of sterile neutrino in the Dodelson-Widrow (DW) scenario, we obtain the lower mass limit m_s < 4 keV, which is stronger than the previous one m_s < 6 kev, obtained recently by Asaka et al. (2007) [hep-ph/0612182]. Comparing this limit with the most recent results on Lyman-alpha forest analysis of Viel et al. (2007) [arXiv:0709.0131] (m_s > 5.6 kev), we argue that the scenario in which all the DM is produced via DW mechanism is ruled out. We discuss however other production mechanisms and note that the sterile neutrino remains a viable candidate of Dark Matter, either warm or cold.

Figures (12)

• Figure 1: Selected regions in the central part of M31 (shown in linear scale). Small circles correspond to excluded point source regions, large circles have radius of 5 and 13 arcmin.
• Figure 2: M31 DM column density versus off-centre angle as result of our Monte Carlo integration, based on DM profiles of Sec. \ref{['sec:DM_calc']}. (Point sources are not excluded).
• Figure 3: Observed spectrum (top) and modelled instrumental background (bottom) MOS1 from ObsID 0112570101, region ring5-13. It can be seen that the spectrum and modelled background almost coincide for $E > 7 \:\mathrm{keV}$.
• Figure 4: Folded spectra from ring5-13 region (by ESAS method), with excluded point sources. The presence of two unsubtracted instrumental lines at 1.49 $\:\mathrm{keV}$ and 1.75 $\:\mathrm{keV}$ is clearly shown.
• Figure 5: Unfolded spectra and best-fit model from ring5-13 region (by ESAS method), with excluded point sources. The "line forest" at energies lower 2.0 keV is clearly visible.