Testing dark matter interactions with CMB spectral distortions

TL;DR

This paper refines the understanding of how dark matter interactions with Standard Model particles can cause spectral distortions in the CMB, providing tighter constraints and exploring future detection prospects.

Contribution

It offers an improved treatment of heat exchange in DM-SM interactions, resulting in more accurate bounds on DM properties from CMB spectral distortions.

Findings

Tighter upper bounds on DM-SM elastic scattering cross sections for DM masses below 0.1 MeV.

Future instruments like PIXIE could probe significantly smaller cross sections, down to orders of magnitude below current limits.

Spectral distortions can constrain DM electric and magnetic dipole moments, but are unlikely to surpass current bounds.

Abstract

Possible interactions of dark matter (DM) with Standard Model (SM) particles can be tested with spectral distortions (SDs) of the cosmic microwave background (CMB). In particular, a non-relativistic DM particle that scatters elastically with photons, electrons or nuclei imprints a negative chemical potential $\mu$ to the CMB spectrum. This article revisits the first study of this effect, with an accurate treatment of heat exchange between DM and SM particles. We show that the instantaneous-decoupling approximation made in the original study systematically and significantly underestimates the amplitude of SDs. As a consequence, we derive tighter upper bounds to the DM-SM elastic-scattering cross section for DM masses $m_\chi \lesssim 0.1$ MeV, from the non-detection of $\mu$-distortions by FIRAS. We also show that a future instrument like PIXIE, sensitive to $|\mu| \sim 10^{-8}$, would be able to probe DM-SM cross sections much smaller than first forecasted, and orders of magnitude below current upper limits from CMB-anisotropy data, up to DM masses of $\sim 1$ GeV. Lastly, we study the sensitivity of SDs to the electric and magnetic dipole moments of the DM. Although SDs can place non-trivial constraints on these models, we find that even future SD experiments are unlikely to improve upon the best current bounds. This article is accompanied by the public code DMDIST, which allows one to compute CMB SDs for generic particle-DM models, specified by their cross sections for elastic scattering with and annihilation into SM particles.