Strongly interacting dark matter: Self-interactions and keV lines

We consider a simple supersymmetric hidden sector: pure $\mathrm{SU}(N)$ gauge theory. Dark matter is made up of hidden glueballinos with mass $m_X$ and hidden glueballs with mass near the confinement scale $\mathrm{\Lambda }$. For $m_X\sim 1\mathrm{TeV}$ and $\mathrm{\Lambda }\sim 100\mathrm{MeV}$, the glueballinos freeze out with the correct relic density and self-interact through glueball exchange to resolve small-scale structure puzzles. An immediate consequence is that the glueballino spectrum has a hyperfine splitting of order ${\mathrm{\Lambda }}^2/m_X\sim 10\mathrm{keV}$. We show that the radiative decays of the excited state can explain the observed 3.5 keV x-ray line signal from clusters of galaxies, Andromeda, and the Milky Way.

Figure 1

Thermal strongly interacting dark matter with ${\mathrm{\Omega }}_G=0.8{\mathrm{\Omega }}_{\mathrm{DM}}$ and ${\mathrm{\Omega }}_{\tilde{g}}=0.2{\mathrm{\Omega }}_{\mathrm{DM}}$. For fixed $(m_X,\mathrm{\Lambda })$, $N$ and ${\xi }_f$ are determined by the relic densities; contours of $N=3$, 5, 10, 100 and ${\xi }_f=10^{-3}$, $5\times 10^{-3}$, $10^{-2}$ are shown. In the indicated bands, $\sigma /m=0.1$–$10\text{barn}/\mathrm{GeV}$ and $\mathrm{\Delta }E=0.356–35.6\mathrm{keV}$. Where these overlap, the model may explain both self-interactions and the 3.5 keV line through long-lifetime ${\tilde{G}}^{*}$ decays (see text). In the lower-right shaded regions, $\tilde{G}$ reannihilation may be significant for the values of $N$ indicated.

Figure 2

Pure glueballino strongly interacting dark matter with ${\xi }_f=1$. For fixed $(m_X,\mathrm{\Lambda })$, $N$ is determined by ${\mathrm{\Omega }}_{\tilde{g}}={\mathrm{\Omega }}_{\mathrm{DM}}$; contours of $N=3$, 10, 30 are shown. In the indicated bands, $\sigma /m=0.1–10\text{barn}/\mathrm{GeV}$, $\mathrm{\Delta }E=0.356–35.6\mathrm{keV}$, and short-lifetime ${\tilde{G}}^{*}$ decays give a keV line flux within an order of magnitude to explain the Perseus observations, assuming ${\sigma }_{{\tilde{G}}^{*}}={\sigma }_{\tilde{G}}$. The flux may also be explained by long-lifetime ${\tilde{G}}^{*}$ decays (see text). In the lower-right shaded regions, $\tilde{G}$ reannihilation may be significant for the values of $N$ indicated.