Resonant Pseudo-Dirac Dark Matter as a Sub-GeV Thermal Target

TL;DR

This paper explores a resonant pseudo-Dirac dark matter model with sub-GeV mass, showing that the excited state persists without violating CMB bounds and could lead to new detection signatures.

Contribution

It demonstrates that in the resonant regime, the excited state remains thermally populated without conflicting with CMB constraints, expanding the potential detection avenues for pseudo-Dirac dark matter.

Findings

Resonance reduces couplings needed for relic abundance

Excited state not thermally depopulated in this regime

Excited state can produce detectable signals today

Abstract

Dark matter (DM) could be a pseudo-Dirac thermal relic with a small mass splitting that is coupled off-diagonally to a kinetically mixed dark photon. This model, particularly in the sub-GeV mass range, is a key benchmark for accelerator searches and direct detection experiments. Typically, the presence of even a tiny fraction of pseudo-Dirac DM in the excited state around the time of recombination would be excluded by DM annihilation bounds from the cosmic microwave background (CMB); thus, viable thermal histories must typically feature an exponential suppression of the excited state. We revisit assumptions about the thermal history in the resonant regime, where the dark photon mass is slightly more than twice the DM mass (to within $\sim10\%$), leading to an $s$-channel resonance in the annihilation cross section. This resonance substantially reduces the couplings required for achieving the observed relic abundance, implying that in much of the parameter space, the DM kinetically decouples from the Standard Model well before the final DM relic abundance is achieved. We find that the excited state is not thermally depopulated in this regime. In spite of this, we find that the presence of the excited state does $\textit{not}$ violate CMB bounds, even for arbitrarily small mass splittings. The present-day abundance of the excited state opens up the possibility of signatures that are usually not relevant for pseudo-Dirac DM, including indirect detection, direct detection, and self-interacting DM signatures.