Self-Interacting Dark Matter from a Non-Abelian Hidden Sector

TL;DR

This work presents a concrete realization of self-interacting dark matter in non-Abelian hidden sectors, showing that glueball DM from a pure gauge theory can achieve the required cross sections with a confinement scale around 100 MeV. It extends to a supersymmetric framework where dark gluinos yield glueballino DM with strong self-interactions mediated by light glueballs, naturally accommodated by anomaly-mediated supersymmetry breaking via the WIMPless miracle. The paper explores both decoupled and connector-coupled hidden sectors, including multicomponent DM scenarios with varying glueball/glueballino fractions and even mechanisms to deplete glueball density when necessary. The results emphasize rich astrophysical implications, potential impacts on early black hole growth, and the need for dedicated simulations to capture the phenomenology across dwarfs, LSBs, and clusters. Overall, the work offers simple, testable models of SIDM rooted in well-motivated high-energy theories with clear cosmological and astrophysical consequences.

Abstract

There is strong evidence in favor of the idea that dark matter is self interacting, with the cross section-to-mass ratio $σ/ m \sim 1\,\mathrm{cm^2/g} \sim 1\,\mathrm{barn/GeV}$. We show that viable models of dark matter with this large cross section are straightforwardly realized with non-Abelian hidden sectors. In the simplest of such models, the hidden sector is a pure gauge theory, and the dark matter is composed of hidden glueballs with a mass around $100\,\mathrm{MeV}$. Alternatively, the hidden sector may be a supersymmetric pure gauge theory with a $\sim 10\,\mathrm{TeV}$ gluino thermal relic. In this case, the dark matter is largely composed of glueballinos that strongly self interact through the exchange of light glueballs. We present a unified framework that realizes both of these possibilities in anomaly-mediated supersymmetry breaking, where, depending on a few model parameters, the dark matter may be composed of hidden glueballinos, hidden glueballs, or a mixture of the two. These models provide simple examples of multicomponent dark matter, have interesting implications for particle physics and cosmology, and include cases where a subdominant component of dark matter may be extremely strongly self interacting, with interesting astrophysical consequences.

Figures (8)

• Figure 1: Example timeline of events in the supersymmetric pure SU($N$) theory without connectors, in terms of the hidden- and visible-sector temperatures $T^h$ and $T$. The hidden-sector coupling $\alpha_h$ is shown as a function of these temperatures. It is weak at gluino freeze-out but grows as the temperature drops, leading to confinement and the formation of glueballino and glueball dark matter at a temperature $\sim \Lambda$. The scenario is described in detail in Sec. \ref{['sec:noconnectors']}, and the chosen parameters are represented by the yellow dot in Fig. \ref{['fig:AMSB-pure-relic']}.
• Figure 2: As in Fig. \ref{['fig:timeline1']}, but for supersymmetric pure SU($N$) theory with connectors. Since the hidden and visible sectors communicate efficiently in the early Universe, they share a temperature, $T$. The gluon population is depleted through annihilations to and the subsequent decays of the $\nu_R$ in the visible sector, and the resulting scenario has pure glueballino dark matter. The scenario is described in detail in Sec. \ref{['sec:connectors']}, and the chosen parameters are represented by the yellow dot in Fig. \ref{['fig:AMSB-pureConn-relic']}.
• Figure 3: Glueball dark matter in the case of a nonsupersymmetric pure gauge SU($N$) hidden sector. The self-interaction cross section and relic density are given in the $(\xi_{\Lambda}, \Lambda)$ plane, where $\Lambda$ is the confinement scale in the hidden sector, and $\xi_{\Lambda} \equiv T^h/T$ is the ratio of hidden to visible sector temperatures at the time that $T^h = \Lambda$. The self-interaction cross section is in the range $\langle \sigma_T \rangle / m_X=0.1 - 1~\mathrm{cm^2/g}$ in the shaded region. The glueball relic density is $\Omega_{\text{gb}} = \Omega_{\text{DM}} \simeq 0.23$ on the diagonal contours for the number of colors $N$ indicated.
• Figure 4: The ratio of the thermally averaged transfer cross section to the dark matter mass $\langle \sigma_T \rangle /m_X$ in the $(m_X, \Lambda)$ plane for $\alpha = 1$ and three different astrophysical systems: dwarf galaxies ($V_\textrm{max} = 40~\mathrm{km/s}$, solid), LSBs ($V_\textrm{max} = 100~\mathrm{km/s}$, dashed), and clusters ($V_\textrm{max} = 1000~\mathrm{km/s}$, dotted). For each system, three values of the cross section are shown: $0.1~\mathrm{cm^2/g}$ (top), $1~\mathrm{cm^2/g}$ (middle), and $10~\mathrm{cm^2/g}$ (bottom). The region below the straight magenta lines shows where inelastic processes may modify the picture based on elastic scattering for each type of system.
• Figure 5: Mostly glueballino dark matter in AMSB models with pure SU($N$) hidden sectors without connectors. Glueballinos make up 90% (top) or 99.99% (bottom) of the dark matter, and glueballs make up the remaining portion. For a point in the $(m_X, \Lambda)$ plane, these constraints on the relic densities determine $N$ and $\xi_f$; contours of constant $N$ and $\xi_f$ are shown. The gray shaded bands are from Fig. \ref{['fig:AMSB-pure-scattering-params']} and give the regions where the glueballino self-interaction cross section is in the preferred range. The red shaded region is excluded by null searches for visible-sector winos at LEP2. The yellow dot in the top panel defines a representative model with $m_X \simeq 14~\mathrm{TeV}$, $\Lambda \simeq 0.35~\mathrm{MeV}$, $N=2$, and $\xi_f \simeq 0.02$.