Thermal Relic of Self-Interacting Dark Matter with Retarded Decay of Mediator

TL;DR

This paper investigates a light mediator in self-interacting dark matter models, focusing on its decoupled decay, temperature evolution, and implications for relic density, considering constraints from BBN and CMB.

Contribution

It provides a detailed calculation of dark matter freeze-out with a retarded mediator decay, including bound state formation and temperature effects, within a Higgs-portal framework.

Findings

BBN and CMB constraints favor a not-too-hot hidden sector.

Positive cubic interactions are compatible with constraints, negative are ruled out.

The model explains relic density considering mediator decay and thermal effects.

Abstract

The existence of a light mediator is beneficial to some phenomena in astroparticle physics, such as the core-cusp problem and diversity problem. It can decouple from Standard Model to avoid direct detection constraints, generally realized by retard decay of the mediator. Their out-of-equilibrium decay process changes the dark matter (DM) freeze-out via temperature discrepancy. This type of hidden sector (HS) typically requires a precision calculation of the freeze-out process considering HS temperature evolution and the thermal average of the cross-section. If the mediator is light sufficiently, we can not ignore the s-wave radiative bound state formation process from the perspective of CMB ionization and Sommerfeld enhancement. We put large mass splitting between DM and mediator, different temperature evolution on the same theoretical footing, discussing the implication for DM relic density in this HS. We study this model and illustrate its property by considering the general Higgs-portal dark matter scenario, which includes all the relevant constraints and signals. It shows that the combination of BBN and CMB constraint favors the not-too-hot HS, $r_{\mathrm{inf}}<10^2$, for the positive cubic interaction of mediator scenario. On the other hand, the negative cubic interaction is ruled out except for our proposed blind spot scenario.