The most commonly accepted mechanism for the formation of Jupiter-like planets is the core accretion model. In this model a rocky core forms through the coagulation of planetesimals until it is sufficiently massive to accrete a gaseous envelope. Initially this envelope is in hydrostatic equilibrium, with most of the luminosity provided by the accreting planetesimals. Once the core reaches a critical mass, however, hydrostatic equilibrium is no longer possible, and a phase of rapid gas accretion occurs.

There are, however, a number of issues with this model. The dust grains from which the planetesimals form may undergo a phase of rapid inward migration when they reach a certain size, the core itself, once it reaches about an Earth mass, should also migrate inwards rapidly, and upper limits for Jupiter's core are smaller than most theoretical estimates. Detailed models also predict growth timescales that may easily exceed the lifetime of the gaseous disc from which the planet formed.

Recent work on self-gravitating discs has already suggested that the grain growth may be accelerated by concentrating the grains in the center of the self-gravitating structures, possibly overcoming the issue of the rapid grain migration. Other simulations of Earth mass-like cores in turbulent discs, by Richard Nelson and John Papaloizou at Queen Mary University in London, have also suggested that the core may undergo a random walk, rather than migrating in a single direction. What we find, when including this random walk in the standard core accretion model, is that the core growth accelerates and reaches a critical mass far more quickly than if the core is assumed to remain stationary. What is more, when the planetesimal density is reduced slightly, we find that the critical core mass is consistent with Jupiter's measured core mass.

On the formation time scale and core masses of gas giant planets, W.K.M. Rice and P.J. Armitage, ApJL, 598, L55 (2003)