Secluded Dark Matter Scenario

• Secluded dark matter is a framework where DM does not directly interact with Standard Model fields but instead annihilates into lighter mediators.
• The mediator decays produce distinctive indirect detection signals, such as broad gamma-ray and cosmic-ray spectra that differ from typical WIMP models.
• The scenario decouples the thermal relic freeze-out process from detection rates, reconciling null direct search results with cosmological abundance constraints.

Secluded dark matter scenarios constitute a paradigm wherein the dark matter (DM) particle does not interact directly with Standard Model (SM) fields at an appreciable rate, but rather annihilates or depletes its abundance predominantly through the production of lighter dark mediators that subsequently decay to SM particles. This framework generalizes the WIMP mechanism by decoupling the processes that control the cosmological relic abundance from those that set detection rates, and provides a natural explanation for the null results in direct detection experiments. The resulting phenomenology is characterized by unique indirect detection signals, a modified connection between thermal relic abundance and couplings to the SM, and novel cosmological features relevant to both early- and late-universe constraints.

1. Theoretical Foundations and Model Structure

At its core, the secluded dark matter scenario introduces (at minimum) three essential ingredients:

• A stable DM particle χ (often a Dirac or Majorana fermion, or a scalar/vector),
• A lighter mediator φ (scalar or vector, potentially CP-even/odd or with vector/axial assignments),
• A symmetry (typically ℤ₂ or U(1)_D) guaranteeing χ's stability.

The minimal secluded DM Lagrangian (scalar mediator case) is: $\begin{aligned} \mathcal{L} =&\ \mathcal{L}_{\mathrm{SM}} + \bar{\chi}(i \slashed{\partial} - m_\chi) \chi\ &+ \frac{1}{2} (\partial_\mu \phi) (\partial^\mu \phi) - \frac{1}{2} m_\phi^2 \phi^2 \ &- g_\chi \phi\ \bar{\chi}\chi - \sum_f g_f \phi\ \bar{f}f + \cdots \end{aligned}$ Vector mediators employ analogous terms, replacing scalar couplings. When $m_\chi > m_\phi$, the dominant process for setting the relic density is DM annihilation into a pair of mediators, $\chi\chi \to \phi\phi$, followed by $\phi \to \mathrm{SM}$ SM decays (Siqueira et al., 2021).

The nonrelativistic annihilation cross section typically has the form: $\sigma v \simeq a + b v^2 + \cdots$ For scalar mediators and Dirac $\chi$,

$$a = \frac{g_\chi^4}{8\pi m_\chi^2}\sqrt{1 - \frac{m_\phi^2}{m_\chi^2}}$$

If $\chi$ is Majorana or the portal is CP-odd, the $s$-wave contribution may vanish, yielding $p$-wave dominated scenarios.

2. Relic Abundance: Boltzmann Framework and Non-Thermal Corrections

The secluded scenario solution to the relic abundance is governed by a set of coupled Boltzmann equations for number densities $n_\chi$, $n_\phi$, and, in more elaborate models, for the full phase-space distributions $f_i(p,t)$. In the standard thermal approximation,

$$\frac{dn_\chi}{dt} + 3H n_\chi = -\langle \sigma v \rangle [n_\chi^2 - (n_\chi^{\mathrm{eq}})^2]$$

With the comoving yield $Y = n_\chi/s$ and $x = m_\chi/T$,

$$\frac{dY}{dx} = -\frac{s(m_\chi)}{x^2 H(m_\chi)} \langle \sigma v \rangle [Y^2 - (Y^{\mathrm{eq}})^2]$$

The analytic freeze-out abundance is

$$\Omega_\chi h^2 \approx 0.12\, \frac{3 \times 10^{-26} \mathrm{cm}^3 \mathrm{s}^{-1}}{\langle \sigma v \rangle}$$

in the $s$-wave regime (Siqueira et al., 2021).

Recent studies have assessed the impact of non-thermal phase-space distortions on the relic calculation. Solving the full Boltzmann equation for $f(p,t)$ during freeze-out in a secluded system (e.g., $$\phi_A\,\phi_A^\dagger \leftrightarrow \phi_B\,\phi_B$$) demonstrates temporary $O(20\%)$ excesses over the Maxwell–Boltzmann estimate, but the prolongation of annihilation at lower temperatures results in a modest $\sim 6\%$ residual difference in the final DM density, typically subdominant compared to current theoretical uncertainties (Beauchesne et al., 2024).

3. Indirect Detection: Gamma-Ray and Cosmic-Ray Signatures

3.1 Differential Gamma-Ray Flux

The distinctive property of secluded models is their characteristic indirect detection signals, predominantly from the decay products of the mediators. The gamma-ray flux from an astrophysical source is

$$\frac{d\Phi_\gamma}{dE} = \frac{1}{8\pi m_\chi^2} \langle \sigma v \rangle \frac{dN_\gamma}{dE} \times J$$

where $dN_\gamma/dE$ is the photon yield per annihilation (convolved through the $\chi\chi \to \phi\phi \to$ SM chain), and $J$ is the $J$-factor: the integral over the line-of-sight of the DM density squared (Siqueira et al., 2021).

3.2 Cascade Spectra from Mediator Decay

Secluded annihilations produce four SM particles per DM annihilation. The resulting $\gamma$-ray spectrum is broader and softer than that of direct $\chi\chi \to$ SM SM annihilations, due to the intermediate step and the boost imparted to the $4$-body SM final state (Siqueira et al., 2021, Profumo et al., 2017). For leptonic mediator decays, the spectrum is especially soft; hadronic mediators generate harder spectra via $\pi^0 \to \gamma\gamma$.

Dark-sector showering effects—in scenarios with strong interactions or additional structure—can produce multiplicities of $\mathcal{O}(10-100)$ soft photons or leptons, significantly enhancing low-energy γ-ray and $e^+$ yields, with phenomenological signatures such as broad spectral bumps and smoother cutoffs relative to standard WIMP scenarios (Li et al., 2023).

3.3 Astrophysical and Solar Signatures

Long-lived mediators can escape astrophysical bodies before decaying, leading to unique solar γ-ray and positron signals with box-like energy spectra and pronounced directionality, a hallmark signature for the presence of secluded dark sectors (0910.1567).

4. Experimental Constraints and Future Sensitivities

4.1 Current Indirect Detection Bounds

• Fermi-LAT (dwarf spheroidals): For secluded annihilation to hadronic final states ($\phi \to b\bar{b}$), $\langle \sigma v \rangle$ is constrained to $\sim 2 \times 10^{-26}$ cm³/s for $m_\chi \lesssim 100$ GeV. Limits for leptonic channels are weaker, but present constraints in the $\tau\tau$ mode reach $4 \times 10^{-27}$ cm³/s at $m_\chi \sim 10$ GeV (Profumo et al., 2017, Siqueira et al., 2021).
• H.E.S.S. (Galactic Center): Exclusion at $\langle \sigma v \rangle \gtrsim 10^{-25}$–$10^{-24}$ cm³/s for $1$–$30$ TeV. Bounds weaken by $O(1$–$10)$ in secluded scenarios compared to direct annihilations, due to the softer photon energy distribution (Siqueira et al., 2021, Profumo et al., 2017).
• Planck (CMB): For annihilations at recombination, $$f_{\mathrm{eff}} \langle \sigma v \rangle / m_\chi < 4.1 \times 10^{-28}$$ cm³ s⁻¹ GeV⁻¹, providing significant constraints at low $m_\chi$ for leptonic final states (Profumo et al., 2017).

4.2 Future Facilities

• CTA (Cherenkov Telescope Array): With anticipated sensitivity down to $\langle \sigma v \rangle \sim 1 \times 10^{-26}$ cm³/s at $m_\chi \sim 1$–$10$ TeV for hadronic secluded channels, improving over H.E.S.S. by a factor of $5$–$10$ (Siqueira et al., 2021, Siqueira, 2019).
• SWGO: Potential for probing $m_\chi \sim 10$–$100$ TeV with reach near the thermal relic cross section in the multi-PeV regime (Siqueira et al., 2021).
• AMS-02 (positrons), next-gen solar and neutrino telescopes: Set to further constrain models with long-lived mediators via directional $e^\pm$ and $\gamma$ signatures from the Sun or the Galactic Center (0910.1567, Li et al., 2023).
• KM3NeT/ANTARES: Providing first neutrino limits on secluded DM up to $m_\chi = 6$ PeV, exploiting neutrino signals from cascade annihilations $\chi\chi \to \phi\phi \to 4X$ (Albert et al., 2022).

5. Phenomenological Implications and Model Variations

5.1 Direct Detection Evasion

One of the critical features of secluded models is the suppression of DM–nucleon scattering. The relevant cross section is mediated by off-shell $\phi$ exchange and scales as $g_\chi^2 g_f^2 / m_\phi^4$. Small $g_f$ or heavy mediators ($m_\phi \gg 100$ MeV) suppress $\sigma_{\mathrm{SI}}$ below $10^{-46}$ cm², readily evading XENON1T, LUX, and PandaX bounds while maintaining a thermal $\langle \sigma v \rangle$ (Siqueira et al., 2021, Mauro et al., 27 Oct 2025).

5.2 Complementary and Novel Probes

• γ-line searches and rare decays are relevant if the mediator mixes with the Higgs or photon, e.g., $\phi \to \gamma\gamma$ (Siqueira et al., 2021).
• Collider searches: Missing energy plus φ resonance reconstruction (in visible decays) probes mediator couplings (Siqueira et al., 2021).
• Neutrino telescopes: Sensitive to scenarios where $\phi$ decays promptly to neutrinos; can constrain otherwise elusive regions of parameter space (0910.1567, Albert et al., 2022).
• Cosmological and astrophysical self-interactions: Models with light mediators can naturally induce DM self-interactions $\sigma/m$ in the range favorable to resolving small-scale structure anomalies (e.g., with $m_\phi \sim 10$ MeV, $\alpha_\chi \sim 10^{-3}$–$10^{-2}$) (Yamamoto, 2017).

6. Structure of Viable Parameter Space and Guidance for Model Building

Viable secluded DM models typically occupy regions where $m_\chi \sim 1$–$10$ TeV, $m_\phi/m_\chi \lesssim 0.1$, and $\langle \sigma v \rangle \sim 3 \times 10^{-26}$ cm³/s (Siqueira et al., 2021, Mauro et al., 27 Oct 2025). These models can accommodate a wide range of mediator masses and couplings, as the relic freeze-out is decoupled from the portal strength to the SM. For portal couplings below $10^{-3}$, the relic density is set entirely by $\chi\chi \to \phi\phi$, while detection rates are suppressed by $O(\epsilon^2)$ or higher powers (Mauro et al., 27 Oct 2025).

Future indirect detection instruments, especially TeV-scale γ-ray observatories, are poised to probe the last open window for secluded thermal relics. These include both the canonical regime and more exotic scenarios with complex dark sectors, strong dark showering, or feebly interacting metastable mediators (Siqueira et al., 2021, Li et al., 2023).

---

In summary, the secluded dark matter scenario provides a structurally robust extension of the WIMP paradigm, characterized by the dynamical separation of freeze-out and detection processes. It remains viable across a wide swath of the parameter space in light of null results from conventional DM searches and is subject to increasingly powerful constraints from indirect astrophysical and cosmological observations (Siqueira et al., 2021, Mauro et al., 27 Oct 2025, Profumo et al., 2017, Li et al., 2023, Beauchesne et al., 2024).