Making Earths: How to Treat Collisions in Simulations

While planetary collisions are well described by state-of-the-art SPH simulations, the problem of including collisional debris in N-body simulators is far from being solved. Each collision can generate billions of particles ranging from dust to asteroids, and including these into simulations is computationally impracticable.

Several solutions have been proposed, the 4 most popular being: the “perfect merging” method, where collisions are assumed to be inelastic; the “fragmentation without debris” method, where fragmentation during collisions is allowed but the fragmented mass is removed from the simulation; the “same mass debris” method, where the fragmented mass is clustered into few bodies with mass usually around one lunar mass; the “unresolved debris” method, where the fragmented mass is included in the simulation as unresolved material that can be accreted by the fully resolved bodies. Each of these methods employs different degrees of approximation that can affect the simulation outcome.

In this work we present the results of an homogeneous comparative study aimed at testing how simulations of rocky planets formation are sensitive to the method employed. We used a set of 20 starting setups (similar to the one presented in Burger et al. 2020) and we run simulations with the 4 methods described above, for a total of 80 simulations. For solving collisions, we interpolated within a catalogue of SPH simulations (see Crespi et al. 2021). We tracked two materials: rocky material (basalt) and water (ice).

The preliminary results of our simulations are hereby presented.