Astrophysical Tests of Dark Matter Self-Interactions

TL;DR

This review addresses how self-interacting dark matter (SIDM) models modify structure formation from galaxies to clusters by enabling energy and momentum transfer within halos. It outlines the microphysical cross sections, their velocity and angular dependence, and how these map onto macroscopic halo properties through simulations and analytic models, including core formation and gravothermal collapse. The authors compile current astrophysical constraints—from strong lensing, stellar streams, X-ray/weak lensing, mergers, and dwarf galaxies—emphasizing the need for velocity-dependent cross sections to reconcile small-scale cores with cluster-scale limits. They also discuss degeneracies with baryonic physics and alternative dark-sector/gravity models, and describe extensions to SIDM such as dissipation, subcomponents, and light mediators. Looking ahead, the paper highlights upcoming surveys and simulation efforts as crucial to robustly testing simple SIDM scenarios and potentially uncovering dark-sector physics through a combination of lensing, kinematics, and large-scale structure observables.

Abstract

Self-interacting dark matter (SIDM) arises generically in scenarios for physics beyond the Standard Model that have dark sectors with light mediators or strong dynamics. The self-interactions allow energy and momentum transport through halos, altering their structure and dynamics relative to those produced by collisionless dark matter. SIDM models provide a promising way to explain the diversity of galactic rotation curves, and they form a predictive and versatile framework for interpreting astrophysical phenomena related to dark matter. This review provides a comprehensive explanation of the physical effects of dark matter self-interactions in objects ranging from galactic satellites (dark and luminous) to clusters of galaxies and the large-scale structure. The second major part describes the methods used to constrain SIDM models including current constraints, with the aim of advancing tests with upcoming galaxy surveys. This part also provides a detailed review of the unresolved small-scale structure formation issues and concrete ways to test simple SIDM models. The review is rounded off by a discussion of the theoretical motivation for self-interactions, degeneracies with baryonic and gravitational effects, extensions to the single-component elastic-interactions SIDM framework, and future observational and theoretical prospects.

Figures (20)

• Figure 1: This set of figures from Rocha13 shows the results from a test without gravity where a DM sphere is moving with uniform velocity with respect to a uniform background using the formalism described in Eq. \ref{['eq:rho_ij_rocha']}. The top row shows the distribution in $\theta$ after scattering compared to the theory prediction for single elastic scatterings. The right panel shows better agreement when multiple scattering events are removed. The bottom row shows the distribution after scattering in speed and azimuthal angle. Note the excess in the speed distribution due to multiple scattering events.
• Figure 2: Density profile of a MW-like DM halo (Aquarius Aq-A) at three resolution levels for both CDM (black) and isotropic SIDM with a cross section of 10 ${\rm cm}^2/{\rm g}$ (red). The resolution levels correspond to particle masses of $4.9 \times 10^4 \,\mathrm{M}_\odot$ (solid), $3.9 \times 10^5 \,\mathrm{M}_\odot$ (dashed), and $3.1 \times 10^6 \,\mathrm{M}_\odot$ (dotted). The vertical lines mark the gravitational softening length (2.8 times the Plummer-equivalent softening length). The figure is adapted from Ref. Vogelsberger:2012ku.
• Figure 3: Figure from Sameie:2018chj showing the DM density (top) and velocity dispersions (bottom) in the central regions of halos in $N$-body simulation with SIDM and a central disk. The left panel shows the density and dispersion in the presence of a compact baryonic disk and the right panel corresponds to an extended disk. In the absence of a baryonic potential the main effect of introducing self-interactions is thermalization and the formation of a central core. A baryonic disk in the centre can shorten the period of core expansion and trigger core contraction.
• Figure 4: Left panel: Gravothermal evolution for the central density of a halo with an initial NFW profile (solid) and with truncated NFW profiles, with truncation radii at $r_s$ (dotted) and $3r_s$ (dashed). The evolution of the central density is normalized to the NFW scale density $\rho_s$ and shown as a function of time, normalized to an interaction time scale $t_0 \propto 1/(\rho_s V_{\rm max} \sigma/m_\text{DM})$. The top axis shows the self-interaction cross section per mass needed for the associated dimensionless time on the bottom axis to correspond to 13 Gyr of evolution. From Ref. nishikawa2019. Right panel: Evolution of the central density of a $10^{10.5} M_\odot$ halo, relative to the density at the start of the simulation, as a function of concentration $c$ and isolation criterion, for a constant cross section of $\sigma = 6\,$cm$^2$/g. Magenta lines denote halos evolved in isolation. Black lines show halos on a radial orbit within a group-scale halo, but with evaporation by the host turned off in order to highlight the effects of gravitational tides on the halo evolution. Turning off evaporation is also the correct thing to do for SIDM models with significant velocity dependence where the cross section falls below a few $\rm cm^2/g$ at velocities of $200\ \rm km/s$. From right to left, the concentrations are 45, 60, 75 and 90. Figure by Z. C. Zeng, adapted from Zeng:2021ldo.
• Figure 5: Figure from Kim:2016ujt showing the evolution of DM and galaxies densities in a merger of halos in SIDM. The top panel shows the distribution in both the halos while the bottom panel shows the distribution in one of the merging halos. The BCGs remain offset and do not merge due to reduced dynamical friction in the core. In both panels the black contours correspond to dark matter evolution and the red contours correspond to galaxy evolution.