Numerical Simulations of Planetesimal Formation

Abstract

A long-standing question in planet formation is the origin of planetesimals, the kilometre-sized precursors to protoplanets. Asteroids and distant Kuiper Belt objects are believed to be remnant planetesimals from the beginnings of our Solar system. A leading mechanism for explaining the formation of these bodies directly from centimetre-sized dust pebbles is the streaming instability (SI). Using high resolution numerical simulations of protoplanetary discs, we probe the behavior of the non-linear SI and planetesimal formation in previously unexplored configurations. Small variations in initial state of the disc can lead to different macroscopic outcomes such as the total mass converted to planetesimals, or the distribution of planetesimal masses. These properties can vary considerably within large simulations, or across smaller simulations re-run with different initial perturbations. However, there is a similar spread in outcomes between multiple smaller simulations and between smaller sub-regions in larger simulations. In small simulations, filaments preferentially form rings while in larger simulations they are truncated. Larger domains permit dynamics on length scales inaccessible to the smaller domains. However, the overall mass concentrated in filaments across various length scales is consistent in all simulations. Small simulations in our suite struggle to resolve dynamics at the natural filament separation length scale, which restricts the possible filament configurations in these simulations. We also model discs with multiple grain species, sampling a size distribution predicted from theories of grain coagulation and fragmentation. The smallest grains do not participate in the formation of planetesimals or filaments, even while they co-exist with dust that readily forms such dense features. For both single-grain and multiple-grain models, we show that the clumping of dust into dense features results in saturated thermal emission, requiring an observational mass correction factor that can be as large as 20-80\%. Finally, we present preliminary work showing that the critical dust-to-gas mass ratio required to trigger the SI can vary between 3D and 2D simulations.