This project aims to fill this gap between advanced hydrodynamical simulations in the standard cold dark matter model and the alternative dark matter field that still lags behind in accuracy and statistics. Only a limited number of available alternative dark matter (ADM) hydrodynamical simulations are of cosmological scale. The AIDA-TNG project advances the field of ADM simulations by providing a consistent set of high-resolution cosmological boxes simulated with the same galaxy formation model and multiple warm and self-interacting dark matter models, in addition to standard cold dark matter. Each of the three boxes is created in multiple variations of dark matter scenarios (see below), as well as with and without baryonic physics.

All viable dark matter models predict that the cosmic structures in the Universe originate from small initial density perturbation; they grow via gravitational instability and hierarchically through the merging and accretion of small systems into larger structures. Comparing CDM simulations with observations has highlighted discrepancies at small scales, including but not limited to the missing satellites, the core-cusp and the diversity problems: this regime consequently has been the driving force behind the alternative dark matter field.

If dark matter particles are lighter (WDM) and thus have large free-streaming velocities, density fluctuations are erased on scales of 1 Mpc and below, preventing the formation of small mass objects: confirming the existence of such small objects and modelling their abundance is one of the most critical empirical tests for warm dark matter. The current constraints on the WDM particle mass are calculated from the number of satellites of the Milky Way, strong lensing and the Lyman-⍺ forest (or a combination of these) and vary in the range mDM≥2-6 keV - leaning towards colder models, where the differences with CDM become relevant only at scales M≤107 Msun. However, constraints in this regime are tricky because baryons or self-interactions can also cause variations at M≤109-10 Msun. Moreover, recent lensing results based on VLBI ratio observations measured an extremely smooth mass distribution on scales well below 1 kpc, and could pose challenges to the abundances in cold models. Physics beyond the standard model could also include forces between dark matter particles: if the probability of interacting is non-negligible on the time scale of millions of years, particle scatterings will allow energy and momentum to flow from one part of dark matter haloes to another beyond what is enabled by gravity only. Over the course of cosmic time, these interactions lead primarily to the formation of density cores at the centre of galaxies and satellite. Self-interacting dark matter (SIDM) can alleviate some of the above-mentioned tensions and produce a larger diversity of rotation curves, making it one of the most compelling alternatives to CDM. In particular, observational results point in the direction of a velocity-dependent cross-section so that dark matter particles behave as a collisional fluid on small scales but are essentially collisionless over large scales. The latter is built to be compatible with observational constraints on scales from galaxy clusters to dwarf galaxies (σ~100). Finally, particle scatterings could be inelastic, and the transition between different stats could produce strong effects even with a small cross-section. In AIDA-TNG we consider

Three thermal relic WDM models with dark matter particle mass equal to 1, 3 and 5 keV. The first is already excluded by observational constraints

Two (for now) SIDM models: one with constant cross-section (σ=1) and one with a velocity-dependent cross-section from Correa et al. 2021 .

We run each box at two resolution levels, where ``A'' is the highest. The 100/A(B) and 50/A(B) boxes start from the same initial conditions of the TNG100-2(3) and TNG50-2(3) boxes of the original IllustrisTNG project. The table below lists the properties of all other runs, and which combinations of box size, resolution and dark matter model have been computed. The choice was made with the aim of an economic use of computational resources and is based on which models produce significant differences compared to CDM at a given resolution -- an issue particularly relevant for the WDM models.

Some visualisations to a first visual impression about the differences in galaxies between the considered dark matter models. In each case, the system is matched across all runs at z=0.

If you are interested in using the simulations for your science, please get in touch

• giulia.despali@unibo.it