
the annihilation particles or their annihilation products, the non-perturbative effects like
the SE or bound state effects [9], need to be considered.
In recent times, the freeze-in dark matter mechanism has gained popularity as several
novel models have proposed it as a viable explanation for the existence of dark matter.
These models may include axion-like particles (ALPs), dark photons, dark Z’ bosons, Feebly
Interacting Massive Particles (FIMPs), or sterile neutrinos [10]... DM was generated with
an initial zero density during freeze-in. The dark sector (DS) will undergo a heating and
then a subsequent cooling process. After decoupling, the DM relic density is proportional to
⟨σv⟩. Hence, the mechanism requires a small coupling to match the observed relic density.
The cross-section correction effect will change the relic density. If there are various particles
in the DS, there will be different “phases” according to the various couplings between the
Standard Model (SM) particles and the DS, as well as the coupling within DS.
The relic density of WIMPs is commonly determined through the freeze-out mechanism
[2,3], in which DM particles are initially in thermal equilibrium with the SM and then
decouple as their interaction rate becomes smaller than the Hubble expansion rate. Since
the DM relic density is proportional to 1/⟨σv⟩, a relatively large cross-section is needed to
get a proper DM density. The cross-section enhancement effect will suppress the final relic
density of DM, so it requires a larger mass or smaller coupling with the SM.
Non-perturbative effects, such as the SE, bound-state formation, resonance effects,
and co-annihilation, have been well-studied in the context of freeze-out [9,11–16], but
they are rare in the context of freeze-in. One reason for this is that freeze-out typically
occurs at a low temperature compared to the mass of the DM particles (x=m/T ∼25),
resulting in a low velocity for the DM particles before or after annihilation, which can lead
to rich non-perturbative phenomena. In contrast, at the end of freeze-in, the temperature
is approximately equal to the DM mass, resulting in a relatively high temperature during
DM production. At high temperatures, the annihilation cross-section becomes simpler and
is roughly proportional to 1/s, with negligible non-perturbative effects. However, we found
that the SE still makes a significant contribution to the cross-section in different phases of
the DM freeze-in. In this paper, we focus on the possible SE during the freeze-in process
and we analyze its correction to the DM abundance. Furthermore, we provide a preliminary
investigation of the SE influence on the boundary and parameter space via specific models.
We concentrate on the SE in the infrared (IR) freeze-in, rather than ultraviolet (UV)
freeze-in. DM is assumed to have a negligible initial abundance, and its interaction with
the particles in the bath can be so feeble1that it was never in thermal equilibrium with
the SM plasma. The feeble interaction leads to the continuous production of DM until
the reaction rate becomes smaller than the Hubble expansion rate, then DM abundance is
gradually fixed. The typical freeze-in temperature is about x=mr/T ∼2−5[5], where
mrrepresents the relevant mass for the Boltzmann suppression. The relevant particles
are moving at nearly non-relativistic velocities. Furthermore, it is natural for the relevant
particles to exchange light force mediators. In the Standard Model, particle-antiparticle
pairs exchange gauge bosons, and the same may happen in the dark sectors. This is the
1The feeble coupling between DM and standard model particles is about 10−7or less [6,17,18].
– 2 –