Cosmological Simulations with Self-Interacting Dark Matter II: Halo Shapes vs. Observations

TL;DR

Cosmological simulations reveal that previous constraints on dark matter self-interaction cross sections based on halo shapes were underestimated, allowing for larger cross sections than previously thought, with implications for future observational constraints.

Contribution

This study revises previous limits on dark matter self-interaction cross sections by incorporating effects of triaxiality and scatter, showing larger permissible cross sections.

Findings

Previous constraints underestimated cross sections by over an order of magnitude.

Observed ellipticity contributions from outside the core affect shape-based constraints.

Constraints from strong lensing may improve via core densities and sizes, not just halo shapes.

Abstract

If dark matter has a large self-interaction scattering cross section, then interactions among dark-matter particles will drive galaxy and cluster halos to become spherical in their centers. Work in the past has used this effect to rule out velocity-independent, elastic cross sections larger than sigma/m ~ 0.02 cm^2/g based on comparisons to the shapes of galaxy cluster lensing potentials and X-ray isophotes. In this paper, we use cosmological simulations to show that these constraints were off by more than an order of magnitude because (a) they did not properly account for the fact that the observed ellipticity gets contributions from the triaxial mass distribution outside the core set by scatterings, (b) the scatter in axis ratios is large and (c) the core region retains more of its triaxial nature than estimated before. Including these effects properly shows that the same observations now allow dark matter self-interaction cross sections at least as large as sigma/m = 0.1 cm^2/g. We show that constraints on self-interacting dark matter from strong-lensing clusters are likely to improve significantly in the near future, but possibly more via central densities and core sizes than halo shapes.