Disc fragmentation. II. Ejection of low mass Free Floating Planets from growing binary systems

TL;DR

Disc fragmentation in very young, often binary-forming systems can generate a population of sub-Jupiter-mass fragments that migrate and interact with a growing secondary. As the secondary contracts inward and grows via accretion, close planetary encounters produce a gravity-assisted ejection, yielding free-floating planets with efficiencies up to about $F_{ m ej}\sim 0.5$ for mass ratios $q\gtrsim 0.05$. The authors validate this mechanism with N-body toy tests, 3D SPH, and 2D hydrodynamic simulations, finding that low-mass planets are readily ejected while Jovian-mass planets tend to migrate inward instead of being ejected. The framework provides a natural explanation for the microlensing-detected muFFP population and predicts observable signatures such as FFP mass functions and possible debris discs around FFPs, offering concrete avenues for observational tests of disc-fragmentation–driven FFPs.

Abstract

Observations indicate that disc fragmentation due to Gravitational Instability (GI) is the likely origin of massive companions to stars, such as giant planets orbiting M-dwarf stars, Brown Dwarf (BD) companions to FGK stars, and binary stars with separations smaller than 100 au. Additionally, we have recently showed that disc fragmentation in young rapidly evolving binary systems ejects an abundant population of massive Jupiter-mass Free-Floating Planets (FFPs). In this model, a massive disc around an initially single protostar undergoes GI and hatches a number of fragments; the most massive oligarch grows by runaway accretion into the secondary star. As the system rearranges itself from a single to a binary star configuration, a dramatic "pincer movement" by the binary ejects planets through dynamical interactions with the stars. Here we propose that the same scenario applies to an even more abundant population of smaller FFPs discovered by the microlensing surveys. Although disc fragmentation is usually believed to form only massive objects, several pathways for forming small core-dominated planets at separations of tens of au exist. We present results from three complementary simulation approaches, all of which confirm planet ejection efficiency as high as 0.5 for secondaries more massive than $\sim 10$\% of the primary star mass. On the other hand, Jovian mass planets migrate through the region of tens of au too rapidly to eject planets from that region. We discuss implications of this scenario, concluding that microlensing FFPs may be the most convincing evidence yet that disc fragmentation forms planets much less massive than Jupiter.

Figures (10)

• Figure 1: The curves show initial fragment mass (§ \ref{['sec:initial_fragments']}), $M_{\rm frag}$, gap opening mass (§ \ref{['sec:fragment_migration']}), $M_{\rm gap}$, and isolation mass (§ \ref{['sec:planets_or_binary']}), $M_{\rm iso}$ for a star with mass $M_*=1{\rm M_{\odot}}$. Note that fragment mass is a strongly increasing function of distance. The points show initial fragment masses from literature, with approximate error bars. The shaded band on the bottom gives the estimated masses of dust cores forming by collapse of the dust component only.
• Figure 2: Results of three groups of numerical simulations Nayakshin17a, for fragments of four different initial masses (see legend) evolving in massive self-gravitating discs. Note that fragments with mass $M_{\rm frag}\lesssim 4 {\,{\rm M}_{\rm J}}$ usually remain in the planetary mass regime and migrate inward rapidly. In contrast, higher mass fragments grow in mass quickly, and open a gap. This slows down their migration and allows them to accrete gas in a runaway fashion. Such fragments evolve towards the isolation mass. See § \ref{['sec:preliminaries']} for detail.
• Figure 3: The toy secondary object migration experiments described in § \ref{['sec:toy']}. Three low mass non-migrating planets are strongly affected by interactions with the secondary. Top panels: Separation of the secondary and the planets to the central star versus time. Bottom panels: Separation between the planets and the secondary object vs time (which is in years rather than kyrs). Left: A planetary mass ($M_{\rm s}\approx 4{\,{\rm M}_{\rm J}}$) secondary pumps eccentricity but does not eject the planets. Right: A BD-mass ($M_{\rm s}\approx 50{\,{\rm M}_{\rm J}}$) secondary imparts larger velocity kicks to the planets, all of which are ejected.
• Figure 4: The initial configuration of planets in the phantom simulation. The secondary has a mass of 25 ${\,{\rm M}_{\rm J}}$ while the 8 planets have masses of 3 $\times$ 10$^{-3}$${\,{\rm M}_{\rm J}}$.
• Figure 5: Top panel: the star-planet distances for all of the ejected planets in the Phantom simulation, with planet ID given in the legend. The secondary (ID equal to 9) starts with $M_{\rm s}=25 {\,{\rm M}_{\rm J}}$ and grows in mass to $\approx 50{\,{\rm M}_{\rm J}}$ by the end of the simulation. Note how rapidly it migrates from its initial $R=140$ au orbit to $R\approx 30$ au, where it stops by opening a gap, and consuming most of the local gas there. Bottom panel: Planet-secondary distances versus time. Note that for some of the planets, the distances of the closest approach are rather large.