Sub-MeV Electrophilic Dark Matter

• Sub-MeV electrophilic dark matter is defined as a class of light dark-sector particles below 1 MeV that preferentially interact with electrons, influencing both cosmological and experimental outcomes.
• Direct detection strategies employ low-threshold experiments, including semiconductor superlattices, Dirac materials, and doped semiconductors, to probe electron recoils with high sensitivity.
• A combination of cosmological, astrophysical, and laboratory constraints narrows the viable parameter space and motivates alternative models with light mediators and absorption mechanisms.

Sub-MeV electrophilic dark matter refers to hypothesized dark-sector particles with mass below 1 MeV that couple preferentially—or exclusively—to electrons rather than nuclei. This interaction channel is of particular theoretical and experimental interest given both the suppressed nuclear-recoil rates for such light dark matter (DM) and the unique cosmological implications of their possible electromagnetic couplings. The sub-MeV regime is characterized by strong cosmological, astrophysical, and laboratory constraints, and is the subject of active research in both direct detection methodologies and model building.

1. Effective Field Theory and Basic Model Structures

Sub-MeV electrophilic DM is typically described within an effective field theory (EFT) framework. The canonical scenario involves a single DM particle χ, stabilized by a Z₂ symmetry and singlet under the Standard Model (SM) gauge group. DM–electron interactions are assumed to be mediated by a heavy new particle (mass $m_\phi \gg \mathrm{MeV}$), which, when integrated out, yields a leading four-fermion operator: $$\mathcal{L}_\mathrm{int} = \frac{g_\chi g_e}{M^2} \, \bar\chi\chi \, \bar e e + \ldots$$ where $g_\chi$ and $g_e$ are the DM– and electron–mediator couplings, $M \simeq m_\phi$, and the ellipsis denotes other possible operator structures (scalar, pseudoscalar, vector, etc.) (Lehmann et al., 2020). The EFT description is valid for momentum transfers and plasma temperatures $T$, $q \ll m_\phi$.

In alternative formulations, the mediator can be a light vector (e.g., a kinetically mixed dark photon $A'$), leading to a form-factor–dependent coupling. In this case, the Lagrangian reads: $$\mathcal{L}_\mathrm{portal} \supset \epsilon F_{\mu\nu}F'^{\mu\nu} + g_\chi A'_\mu \bar\chi\gamma^\mu\chi + e A_\mu \bar e\gamma^\mu e$$ where $\epsilon$ is the kinetic-mixing parameter (Hochberg et al., 2017, Aboubrahim et al., 2021).

Electrophilic refers specifically to the hierarchy $g_e \gg g_q$, such that electron scattering dominates over nucleon scattering at achievable thresholds (Zhang et al., 2024).

2. Cosmological and Astrophysical Constraints

Stringent constraints arise from the early universe, predominantly from the following epochs:

A. Big-Bang Nucleosynthesis (BBN). If χ (or its mediator) equilibrates with electrons at $T \lesssim 1$ MeV, its additional energy density alters the Hubble rate, affecting helium and deuterium yields. BBN constraints require the DM–electron coupling satisfy (Lehmann et al., 2020): $$\frac{g_\chi g_e}{M^2} \lesssim 10^{-10} \, \mathrm{MeV}^{-2}$$ For $M \sim 10$ MeV and $m_\chi \sim 100$ keV: $g_e \lesssim 10^{-5}$.

B. Relic Abundance (Freeze-In/Out-of-Equilibrium Production). Overproduction occurs if annihilation $\sigma v$ is too small, restricting the coupling from below. For correct relic density ($\Omega_\chi h^2 \lesssim 0.12$), one needs (Lehmann et al., 2020): $$\frac{g_\chi^2 g_e^2}{M^4} \gtrsim 10^{-19} – 10^{-17}~\mathrm{GeV}^{-4}$$ C. Extra Radiation ($\Delta N_\mathrm{eff}$). Late-time entropy transfer to/from the electron-photon bath shifts the photon-to-neutrino temperature ratio. Planck and future CMB experiments tightly constrain $\Delta N_\mathrm{eff} \lesssim 0.3$ (2σ), limiting additional light degrees of freedom at $T_\mathrm{BBN} \sim 1$ MeV (Lehmann et al., 2020, Green et al., 2017). For typical mediator and DM parameter choices, the $\Delta N_\mathrm{eff}$ constraint is as restrictive as BBN.

D. Stellar Cooling. Electron-coupled mediators can lead to excessive energy loss from stars (white dwarfs, red giants). White-dwarf cooling constrains $g_e$ below $8.4 \times 10^{-14}$ for $m_\phi \lesssim 400$ keV (Green et al., 2017).

E. Direct Astrophysical Observables. In dense environments (e.g., white dwarfs near high-density DM regions), sub-MeV DM can alter cooling via capture, scattering, and annihilation, providing complementary constraints in $m_\chi$–$\sigma_e$ space (Zhang et al., 2024).

Combined Parameter Space: Cosmology, BBN, and relic-abundance constraints carve out a narrow viable window: only $g_e \sim 10^{-7}$–$10^{-8}$ and $m_\chi \gtrsim$ few hundred keV evade all bounds in standard EFT. Lower $m_\chi$ values are practically excluded by order-of-magnitude (Lehmann et al., 2020).

3. Direct Detection Approaches and Projected Sensitivities

Sub-MeV electrophilic DM-induced electron recoils are sought in ultralow-threshold experiments, leveraging a variety of condensed-matter targets and detection strategies:

A. Semiconductor Superlattices and Quantum-Cascade Lasers (QCLs). Superlattice superstructures (SSS) engineered for gaps $E_g \sim 100$–300 meV enable detection of sub-MeV DM recoils via prompt mid-infrared photon emission, with QCL-based readout reaching practical thresholds $E_{\rm th} \sim 50$ meV and projected $90\%$ C.L. sensitivities $\bar\sigma_e \sim 10^{-42}$–$10^{-41}$ cm$^2$ for $m_\chi=0.1$–$1$ MeV—orders of magnitude beyond current semiconductor limits (Bora, 2022).

B. Dirac and Graphene-Based Materials. Three-dimensional Dirac semimetals ($\Delta \sim$ meV) or voltage-tunable bilayer graphene enable detection thresholds as low as $\sim$10 meV (Hochberg et al., 2017, Das et al., 2023, Geilhufe et al., 2019). Their unscreened in-medium response allows probing cross sections $\sigma_e \lesssim 10^{-38}$–$10^{-41}$ cm$^2$ for $m_\chi$ in the 4 keV–1 MeV range, exploiting daily modulation due to anisotropic response and Earth's rotation as a potential discriminant against background.

C. Doped Semiconductors and Skipper-CCD Technology. P-type or n-type semiconductors with dopant-induced shallow energy levels ($E_I \sim$10–100 meV) enable access to lower DM masses than pure (eV-gap) materials. Projected sensitivities with exposures of $\sim$100 g·day and dark counts $\lesssim$1/(g·day) can reach the freeze-in target cross sections ($\sigma_e \sim 10^{-44}$–$10^{-41}$ cm$^2$), contingent on improved noise and backgrounds (Du et al., 2022).

D. Plasmon-Enhanced and Quantum Materials Approaches. Plasmon excitations in e.g., silicon CC(D)s, particularly with cosmic-ray (CR)–boosted DM, can yield strong limits on $\sigma_e$ down to $10^{-35}$ cm$^2$ at $m_\chi \sim 1$ keV in the light-mediator scenario (Liang et al., 2024).

E. Boosted, Reflected, or Absorption Signals. Additional handles come from CR-boosted DM (Cao et al., 2020, Shang et al., 2024) or solar-reflected DM, which can produce signals well above the kinematic thresholds of halo DM and yield direct-detection constraints for $m_\chi$ as low as several keV (An et al., 2017).

Target Type	Minimum $m_\chi$, ($E_\mathrm{th}$)	Best Current/Projected Sensitivity ($\sigma_e$)	Unique Features
Dirac materials	$4$ keV (few meV)	$10^{-41}$–$10^{-38}$ cm$^2$ (3–100 events/kg·yr)	No in-medium suppression, daily modulation possible
Superlattice superstructure	$0.05$ MeV (50 meV)	$10^{-42}$ cm$^2$ (1 kg·yr)	Tunable gap, photon-based readout
Doped semiconductors	30 keV (10–100 meV)	$10^{-41}$ cm$^2$ (100 g·day, DC=0)	Leverages shallow levels, mature technology

No platform currently achieves sensitivity to the couplings allowed by cosmology in the minimal heavy-mediator models; see Section 5.

4. Model Extensions, Absorption Channels, and Non-Minimal Cosmology

While elastic scattering signatures are tightly constrained by cosmology and freeze-in/out-of-equilibrium history, alternative mechanisms can open viable parameter space:

A. Fermionic Absorption by Electrons. Models where light fermionic DM is absorbed (e.g., $\chi$ + A $\to$ $e^-$ + A$^+$ + $\nu$) rather than scattered can evade some cosmological bounds (Dror et al., 2020). For vector-mediated absorption, XENON1T already explores $20$ keV $\lesssim m_\chi \lesssim 1$ MeV with projected reach to lower masses in future liquid-xenon TPCs.

B. Multi-sector or Dark-Sink Cosmologies. Scenarios with multiple hidden sectors or late-time entropy injection can alleviate BBN and $\Delta N_\mathrm{eff}$ constraints, allowing heavier couplings or different thermal histories (Aboubrahim et al., 2021, Bhattiprolu et al., 2024). For instance, the introduction of a "Dark Sink"—a bath of very light fermions interacting with DM—modifies freeze-in and allows present-day cross-sections up to several orders of magnitude above the canonical freeze-in line: $$\sigma_e \in [10^{-39}~\mathrm{cm}^2,~10^{-34}~\mathrm{cm}^2]$$ for $m_\chi \sim 10-500$ keV (Bhattiprolu et al., 2024).

C. Bosonic Absorption. Absorption of bosonic DM (e.g., dark photons) in targets with low excitation thresholds can provide observable signals independent of velocity distributions and with weaker cosmological model dependence (Hochberg et al., 2017, Geilhufe et al., 2019, Das et al., 2023), achieving sensitivity to kinetic-mixing parameters as low as $\epsilon \sim 10^{-12}$ for sub-eV–MeV dark photon masses.

5. Combined Parameter Space, Experimental Outlook, and Limitations

The convergence of cosmological, astrophysical, and laboratory data imposes a dramatic narrowing of the allowed parameter space for minimal sub-MeV electrophilic DM with heavy mediators. In the ($m_\chi$, $g_e$) or ($m_\chi$, $\sigma_e$) plane, combined BBN, $\Delta N_\mathrm{eff}$, and relic-density constraints restrict:

• $g_e$ (effective coupling): $10^{-7}$–$10^{-8}$,
• $m_\chi \gtrsim 200$ keV,
• $\sigma_e \lesssim 10^{-38}$ cm$^2$ (Lehmann et al., 2020, Green et al., 2017).

Proposed next-generation electron-recoil experiments with thresholds of few meV and exposures $\sim$ 1 kg·yr target $g_e$ values at least 1–3 orders of magnitude above the cosmological upper bounds. Thus, in a minimal EFT, no unconstrained parameter space remains for observable elastic scattering unless the cosmological background is non-standard (e.g., late-time phase transition, entropy injection, or strong number-changing processes in the dark sector).

However, models with:

• Light mediators (kinetically mixed dark photons) and non-minimal cosmological histories,
• Absorption-based signals (including bosonic or fermionic DM),
• CR- or solar-boosted detection channels,

remain viable and in several cases testable by current or forthcoming low-threshold experiments (Dror et al., 2020, Bhattiprolu et al., 2024, Liang et al., 2024, An et al., 2017).

6. Key Future Directions and Open Issues

Efforts continue on several fronts:

1. Experimental Development. Lowering detection thresholds toward single–electron or single–phonon sensitivity, expanding target materials (e.g., Dirac materials, bilayer graphene, doped semiconductors) (Das et al., 2023, Du et al., 2022, Geilhufe et al., 2019).
2. Theoretical Refinement. Incorporating full atomic and condensed-matter structure factors at low energy transfers, in-medium corrections for light mediators, and precise calculations of backgrounds and daily modulation signatures (Hochberg et al., 2017).
3. Astrophysical Complementarity. Using observations of stellar cooling and white-dwarf pulsation rates in regions of high DM density to extend sensitivity to parameter regions inaccessible to terrestrial detectors (Zhang et al., 2024).
4. Non-Minimal Cosmology. Exploring more complex cosmological histories (multisector, "Dark Sink" depletion, late entropy injection) that enlarge the allowed direct-detection parameter space (Bhattiprolu et al., 2024, Aboubrahim et al., 2021).

In summary, while minimal heavy-mediator sub-MeV electrophilic dark matter is stringently constrained by cosmological considerations, several well-motivated model extensions—especially those involving light mediators, absorption channels, or alternative production histories—remain open and will continue to motivate both theoretical and experimental advances in the field (Lehmann et al., 2020, Green et al., 2017, Bhattiprolu et al., 2024).