On The Possibility of Enrichment and Differentiation in Gas Giants During Birth by Disk Instability

We investigate the coupling between rock-size solids and gas during the formation of gas giant planets by disk fragmentation in the outer regions of massive disks. In this study, we use three-dimensional radiative hydrodynamics simulations and model solids as a spatial distribution of particles. We assume that half of the total solid fraction is in small grains and half in large solids. The former are perfectly entrained with the gas and set the opacity in the disk, while the latter are allowed to respond to gas drag forces, with the back reaction on the gas taken into account. To explore the maximum effects of gas-solid interactions, we first consider 10cm-size particles. We then compare these results to a simulation with 1 km-size particles, which explores the low-drag regime. We show that (1) disk instability planets have the potential to form large cores due to aerodynamic capturing of rock-size solids in spiral arms before fragmentation; (2) that temporary clumps can concentrate tens of $M_{\oplus}$ of solids in very localized regions before clump disruption; (3) that the formation of permanent clumps, even in the outer disk, is dependent on the grain-size distribution, i.e., the opacity; (4) that nonaxisymmetric structure in the disk can create disk regions that have a solids-to-gas ratio greater than unity; (5) that the solid distribution may affect the fragmentation process; (6) that proto-gas giants and proto-brown dwarfs can start as differentiated objects prior to the H$_2$ collapse phase; (7) that spiral arms in a gravitationally unstable disk are able to stop the inward drift of rock-size solids, even redistributing them to larger radii; and, (8) that large solids can form spiral arms that are offset from the gaseous spiral arms. We conclude that planet embryo formation can be strongly affected by the growth of solids during the earliest stages of disk accretion.
Accepted by ApJ. 55 pages including 24 figures. In response to comments from the referee, we have included a new simulation with km-size objects and have revised some discussions and interpretations. Major conclusions remain unchanged, and new conclusions have been added in response to the new run