Dissipative SIDM simulations invert conduction direction and shorten collapse timescales for compact halo objects
Preprint simulations demonstrate that dissipation in SIDM qualitatively alters gravothermal evolution by inverting conduction and enabling compact objects at lower cross sections. The work connects halo microphysics to observed perturbers and highlights limitations of fluid approximations at high densities. Stronger cosmological tests are required to confirm the signatures.
The arXiv preprint (Schmidt et al., 2026) extends the fSIDM N-body formalism to include tunable radiative cooling alongside heat conduction. Isolated halo simulations vary dissipation strength independently while holding elastic scattering fixed. Fluid-model comparisons validate the approach but highlight its breakdown once central densities exceed 10^10 solar masses per cubic parsec. Dissipation does not merely accelerate gravothermal collapse; it inverts the usual temperature gradient so that outer regions beyond the scale radius cool and infall coherently rather than being heated.
This produces smoother density profiles lacking the classic core-halo indentation and allows weakly dissipative models to reproduce the observed mass and size of the JVAS B1938+666 strong-lens perturber with evolution times reduced by factors of three to five. The result links microphysical dark-sector cooling rates directly to galactic-scale observables, bridging SIDM elastic-collapse studies with compact-object formation channels previously attributed only to baryonic or primordial processes.
Future work must embed these isolated-halo findings in full cosmological volumes to test whether the inverted conduction signature survives tidal stripping and mergers. Multi-messenger constraints from stellar-stream gaps and strong-lensing flux ratios will be decisive within the next three years.
HELIX: Within 24 months, at least two additional strong-lens systems will exhibit central densities exceeding 10^9 M⊙ kpc^-3 at radii <100 pc, matching dissipative SIDM predictions but exceeding elastic-only models.
Sources (2)
- [1]Primary Source(https://arxiv.org/abs/2606.19428)
- [2]Supporting Source(https://arxiv.org/abs/2110.00017)