WIMP Shadows: Phenomenology of Secluded Dark Matter in Three Minimal BSM Scenarios

TL;DR

This work analyzes secluded dark matter in three minimal beyond-Standard-Model realizations, where χ annihilates predominantly to lighter mediators φ that couple feebly to the SM via a small portal ε. By solving single and coupled Boltzmann equations, it shows that the relic density can be set by hidden-sector dynamics χχ→φφ with mediator decays ensuring BBN safety, while direct and indirect detection signals are highly suppressed for ε ≪ 10^{-3}. The study maps viable regions in each model, highlighting resonance funnels (e.g., m_χ ≈ m_{A'}/2 or m_Z/2) and truly secluded dynamics, and emphasizes collider and intensity-frontier probes as key tests. It concludes that secluded DM remains a compelling, testable thermal relic scenario that can evade current and near-future direct-detection limits while remaining accessible to collider-based and cosmological probes.

Abstract

We present a comprehensive study of secluded dark matter (DM) $χ$, where the relic abundance is set by annihilations into lighter dark mediators $φ$ that couple only feebly to the Standard Model (SM). In contrast to canonical WIMPs, which are now strongly constrained by direct and indirect searches, secluded models still achieve the observed relic abundance via thermal freeze-out into hidden-sector mediators, while predicting highly suppressed present-day signals. We analyze three minimal models: (i) a $U(1)_X$ gauge boson ($A'$) with kinetic mixing; (ii) a scalar DM candidate $S$ with a scalar mediator $K$ that has a trilinear vertex; and (iii) a Dirac fermion $χ$ whose mass arises from a Higgs-mixed singlet $H_p$. For each model we derive annihilation and scattering rates in both WIMP-like and secluded regimes, and solve the Boltzmann equations: a single-species equation for the WIMP case and a coupled $χ$-$φ$ system for the secluded case to account for possible early departure of the mediator from thermal equilibrium with the SM bath. In this regard, we provide explicit lower limits on the portal coupling $ε$ required to keep the mediator in thermal equilibrium with the SM bath and to ensure mediator decay before BBN. We show that for portal couplings $ε\ll 10^{-3}$ the relic density is dominantly controlled by DM annihilation into mediator pairs, while spin-independent scattering lies well below current limits and remains viable even for future experiments approaching the irreducible neutrino background floor. Indirect constraints are typically weak due to $p$-wave suppression, off-resonance $s$-channels, and cascade spectra controlled by $ε^2$. Finally, we highlight the most promising collider tests, which remain sensitive despite tiny portal couplings.

Figures (6)

• Figure 1: Rates for the processes that control chemical and kinetic equilibrium. We show the rate for $A'$ decay $\langle \Gamma_{A'} \rangle$ (red solid), annihilation into fermions $\Gamma_{A'A'\leftrightarrow f\bar{f}}$ (orange dot--dashed) and elastic scattering $\Gamma_{\rm el}$ (blue dashed). We also display the rate for $A'A' \to \chi \bar{\chi}$ (green dotted), the Hubble rate (cyan band) and the temperature equivalent to the dark photon mass (yellow band).
• Figure 2: This plot shows the comoving number density $Y$ for the DM particle $S$ and mediator $K$ for Model II when fixing $m_K=10$ GeV and $m_S=100$ GeV and $g_K=0.8$ and $\beta=0.05$. We are thus in the secluded regime and the dominant annihilation channel for these parameters is $SS \to KK$. We show also the equilibrium $Y$ for the two particles and the rescaled equilibrium number density for the particle $K$ ($Y_{\rm{eq},K}$) but rescaled in order to be directly comparable with the DM one $Y_{\rm{eq},S}$.
• Figure 3: This figure shows the $(m_\chi,\epsilon)$ values that reproduce the observed relic abundance (blue solid). Also shown are the upper limits on $\epsilon$ inferred from the LZ spin–independent WIMP–nucleon bound (green dashed) and from Fermi–LAT dwarf–spheroidal $\gamma$–ray constraints (red dot–dashed).
• Figure 4: We show in this figure the secluded case for Model I with $m_A'=10$ GeV. We report the parameters $g_X$ and $m_\chi$ for which DM has the right relic density (blue solid curve) and the upper limits obtained when fixing different values of $\epsilon$ from $10^{-5}$ and $10^{-3}$. Finally we report the upper limits for $g_X$ obtained when fixing $\epsilon=10^{-4}$ from indirect detection (red dot-dashed curve).
• Figure 5: This figure shows the results obtained for the Model II. The top panels include the WIMP scenario with the values $m_S$ and $\beta$ for which DM has the right relic density (blue solid) compared with the upper limits obtained from LZ (green dashed) and dSphs bounds (red dot-dashed). The left (right) panel display the case with $m_K=200$ (1000) GeV. The bottom figures show the secluded case with $m_K=10$ GeV (100 GeV) in the left (right) panel. We show with a cyan band the unitarity bound for the coupling parameter $g_K$, the curves with values of DM mass and $g_K$ that provides the correct relic density (dashed curves) obtained for $\beta=[10^{-3}-10^{-1}]$. Next to each relic density curve we report if these model parameters are excluded or allowed by LZ bounds.