Formation of satellites in circumplanetary discs generated by disc instability

TL;DR

This study models satellite formation in circumplanetary discs formed by gravitational instability, revealing that most satellites are similar to Galilean moons, with rapid migration leading to significant planetary accretion of moon material.

Contribution

It introduces a population synthesis approach to simulate satellite formation in gravitational instability-generated discs, exploring effects of initial conditions and planet-star distance.

Findings

Most satellites are similar in mass to Galilean moons.

Rapid migration causes most moons to be engulfed by the planet.

Detection probability of such satellites is very low (<3%).

Abstract

We investigated the formation and evolution of satellite systems in a cold, extended circumplanetary disc around a 10 $M_{\rm{Jupiter}}$ gas giant which was formed by gravitational instability at 50\,AU from its star. The disc parameters were from a 3D global SPH simulation. We used a population synthesis approach, where we placed satellite embryos in this disc, and let them accrete mass, migrate, collide until the gaseous disc is dissipated. In each run we changed the initial dust-to-gas ratio, dispersion- and refilling time-scales within reasonable limits, as well as the number of embryos and their starting locations. We found that most satellites have mass similar to the Galilean ones, but very few can reach a maximum of 3 $M_{\rm{Earth}}$ due to the massive circumplanetary disc. Large moons are often form as far as 0.5 $R_{\rm{disc}}$. The migration rate of satellites are fast, hence during the disc lifetime, an average of 10 $M_{\rm{Earth}}$ worth of moons will be engulfed by the planet, increasing greatly its metallicity. We also investigated the effect of the planet's semi-major axis on the resulting satellite systems by re-scaling our model. This test revealed that for the discs closer to the star, the formed moons are lighter, and a larger amount of satellites are lost into the planet due to the even faster migration. Finally, we checked the probability of detecting satellites like our population, which resulted in a low number of $\leq$ 3\% even with upcoming powerful telescopes like E-ELT.