Circumbinary planets: migration, trapping in mean-motion resonances, and ejection

TL;DR

This study uses extensive n-body simulations with Stokes-like forces to analyze how binary star properties influence the migration and resonance trapping of circumbinary planets, revealing key parameters affecting their ejection and capture.

Contribution

It introduces a large-scale simulation approach combined with machine learning to identify main factors influencing mean-motion resonance capture in circumbinary planets.

Findings

4/1 MMR has the highest capture probability at 37%.

Binary eccentricity is the most critical factor for resonance capture.

Initial planetary eccentricity and binary mass ratio also significantly influence outcomes.

Abstract

Most of the planetary systems discovered around binary stars are located at approximately three semi-major axes from the barycentre of their system, curiously close to low-order mean-motion resonances (MMRs). The formation mechanism of these circumbinary planets is not yet fully understood. In situ formation is extremely challenging because of the strong interaction with the binary. One possible explanation is that, after their formation, the interactions between these planets and the surrounding protoplanetary disc cause them to migrate at velocities dependent on the nature of the disc and the mass of the exoplanet. Although extensive data can be obtained with direct hydrodynamical simulations, their computational cost remains too high. On the other hand, the direct n-body simulations approach allows us to model a large variety of parameters at much lower cost. We analyse the planetary migration around a wide variety of binary stars using Stokes-like forces that mimic planetary migration at a constant rate. Our goal is to identify the main parameters responsible for the ejection of planets at different resonances with the inner binary. We performed 4200 n-body simulations with Stokes-like forces and analysed their evolution and outcome as a function of the properties of each system. For each simulated exoplanet, we applied an ensemble learning method for classification in order to clarify the relationship between the inspected parameters and the process of MMR capture. We identify the capture probability for different N/1 MMRs, 4/1 being the most prone to capture exoplanets, with 37% probability, followed by MMR 5/1 with $\sim$ 23% of probability. The eccentricity of the binary is found to be the most important parameter in determining the MMR capture of each circumbinary exoplanet, followed by the mass ratio of the binary and the initial eccentricity of the planet.