Testing self-interacting dark matter with galaxy warps

TL;DR

This study uses galaxy warps to set constraints on long-range self-interacting dark matter, finding no evidence of such interactions and establishing upper limits on their cross section based on observed galaxy shapes.

Contribution

First to use galaxy warp measurements to constrain long-range dark matter self-interactions with detailed modeling and analysis.

Findings

No evidence of SIDM contribution to galaxy warps

Upper bound on normalized Rutherford-like cross section: ~3×10⁻¹³ cm²/g at 300 km/s

Constraints on momentum transfer cross section: < 0.1 cm²/g at 300 km/s

Abstract

Self-interacting dark matter (SIDM) is an able alternative to collisionless dark matter. If dark matter does have self-interactions, we would expect this to cause a separation between the collisionless stars and the dark matter halo of a galaxy as it falls through a dark matter medium. For stars arranged in a disk, this would generate a U-shaped warp. The magnitude of this warping depends on the SIDM cross section, type of self-interaction, relative velocity of galaxy and background, halo structure, and density of the dark matter medium. In this paper, we set constraints on long-range (light mediator) dark matter self-interaction by means of this signal. We begin by measuring U-shaped warps in $3,213$ edge-on disk galaxies within the Sloan Digital Sky Survey. We then forward-model the expected warp from SIDM on a galaxy-by-galaxy basis by combining models of halo structure, density and velocity field reconstructions, and models for the dark matter interactions. We find no evidence for a contribution to the warps from SIDM. Our constraints are highly dependent on the uncertain velocities of our galaxies: for a normalized Rutherford-like cross section we find $\tilde{\sigma}/m_{\rm{DM}} \lesssim 3\times 10^{-13}~\rm{cm}^2/\rm{g}$ at fixed velocity $v = 300~\rm{km/s}$ -- a bound that scales roughly linearly with increasing $v$. In the appendix we translate these bounds into limits on the momentum transfer cross section, finding $\sigma_T(300~\rm{km/s})/m_{\rm{DM}} \lesssim 0.1~\rm{cm}^2/\rm{g}$. [abridged]