Light(ly)-coupled Dark Matter in the keV Range: Freeze-In and Constraints

TL;DR

This work assesses keV–GeV dark matter produced by freeze-in through two photonic portals: a heavy dark photon with kinetic mixing and DM coupled to photons via an electric or magnetic dipole moment. It computes freeze-in production from SM fermion annihilation and plasmon decays, and derives stringent stellar cooling constraints from red giants and horizontal-branch stars, mapping excluded regions in the model parameter space. The results show that the relic-density–required couplings are ruled out below roughly tens of keV for the dark photon portal and below a few keV for dipole-moment DM, while laboratory probes are generally unlikely to test these scenarios except for certain dipole-moment cases with masses above the reheating temperature. Collectively, the study highlights the complementary role of astrophysical constraints in probing ultra-weak DM interactions and clarifies the UV vs IR nature of freeze-in for these models.

Abstract

Dark matter produced from thermal freeze-out is typically restricted to have masses above roughly 1 MeV. However, if the couplings are small, the freeze-in mechanism allows for production of dark matter down to keV masses. We consider dark matter coupled to a dark photon that mixes with the photon and dark matter coupled to photons through an electric or magnetic dipole moment. We discuss contributions to the freeze-in production of such dark matter particles from standard model fermion-antifermion annihilation and plasmon decay. We also derive constraints on such dark matter from the cooling of red giant stars and horizontal branch stars, carefully evaluating the thermal processes as well as the bremsstrahlung process that dominates for masses above the plasma frequency. We find that the parameters needed to obtain the observed relic abundance from freeze-in are excluded below a few tens of keV, depending on the value of the dark gauge coupling constant for the dark photon portal model, and below a few keV, depending on the reheating temperature for dark matter with an electric or magnetic dipole moment. While laboratory probes are unlikely to probe these freeze-in scenarios in general, we show that for dark matter with an electric or magnetic dipole moment and for dark matter masses above the reheating temperature, the couplings needed for freeze-in to produce the observed relic abundance can be probed partially by upcoming direct-detection experiments.

Figures (8)

• Figure 1: The dominant processes relevant for the production of dark matter interacting with a dark photon (upper diagrams) and dipole moments (lower diagrams) in the early Universe: pair annihilation (left) and plasmon/$Z$-boson decay (right).
• Figure 2: Solid lines show the values of $\epsilon$ for which the correct dark matter relic abundance is obtained in a model with fermionic (left) and scalar (right) dark matter from freeze-in through a dark photon. We show various choices for the dark-matter-to-dark-photon mass ratio and dark-photon couplings $\alpha_D$. For $m'<2 m_{\chi/\phi}$ and $m'<2m_e$, the dark photon can make up the relic abundance, rather than the fermion $\chi$ or scalar $\phi$ (see text for details). For parameters above these "freeze-in lines", too much dark matter is produced in the early Universe. Above the dashed lines, the dark matter and SM sector are in chemical equilibrium. Below the chemical equilibrium lines the model is safe from constraints on the number of relativistic degrees of freedom in the early Universe.
• Figure 3: Solid lines in the left (right) plot show the values of the electric (magnetic) dipole moment needed to obtain the correct relic abundance from freeze-in for electric (magnetic) dipole dark matter for different reheating temperatures. For parameters above these "freeze-in lines", too much dark matter is produced in the early Universe. Above the dashed lines, the dark matter and SM sector are in chemical equilibrium, so that this parameter region is not compatible with any form of freeze-in production. Below the chemical equilibrium lines the model is safe from constraints on the number of relativistic degrees of freedom in the early Universe.
• Figure 4: The diagrams contributing to the bremsstrahlung process of producing dark matter in stars. A radiated photon during the electron-proton collision can produce dark matter either through mixing with the dark photon (top two diagrams) or through the electric or magnetic dipole moment (bottom two diagrams).
• Figure 5: Stellar cooling constraints derived in this work on Dirac fermion dark matter interacting with a dark photon with dark photon masses $m^\prime=3m_\chi$ and $\alpha_D=0.5$ ($\alpha_D=10^{-6}$) for the solid (dashed) lines. The cooling constraints are derived for stars on the horizontal branch (brown) and red giants (red). In green, we show the parameters for which freeze-in production provides the entire dark matter relic abundance (see also Fig. \ref{['fig:FIfermion']}); above the line too much dark matter would have been produced. In blue, we show the parameters for which thermal freeze-out production provides the entire dark matter relic abundance. Above the cyan lines, the dark sector was in chemical equilibrium with the SM bath and is constrained below $m_\chi=9.4$ MeV by $N_\text{eff}$. Below $\sim$1 keV dark matter is constrained from structure formation. Other relevant constraints and some projections from terrestrial searches are shown in Fig. \ref{['fig:summaryHP']}. The bounds on scalar dark matter coupling to a dark photon (not shown) are similar.