Nearly four-dozen gas giant planets have been detected in orbit around nearby stars, most of them discovered by DTM's Paul Butler and his colleagues. These discoveries have highlighted the unsatisfactory state of our knowledge about how gas giant planets like Jupiter form. Two mechanisms have been suggested for forming these bodies: core accretion and disk instability. Core accretion is the generally accepted mechanism, but if a disk instability is possible, it would occur well before core accretion could even get started.

In order to learn whether the disk-instability mechanism is plausible, I have calculated the evolution of a number of three-dimensional (3D), gravitational hydrodynamical models of protoplanetary disks starting from realistic initial temperature and density profiles, with an unprecedented degree of spatial resolution. This is accomplished with the Carnegie Alpha Cluster of state-of-the-art workstations. The models show (Fig. 1) that a clump-forming disk instability could occur in a marginally unstable disk with a mass as low as 10% that of a solar mass inside a radius of 20 AU (1 AU = Earth-Sun distance = 93 million miles). Disk instability thus seems to be possible in disks with total masses comparable to that inferred for the solar nebula, and indeed comparable to the disk masses that seem to be required in order for core accretion to produce giant planets in a few million years. Because a disk instability occurs within a few thousand years, it will handily outrace core accretion. The terrestrial and outer planets could then form much later by the usual process of collisional accumulation of solids, which has been studied in great detail by DTM's George Wetherill.

I have also been calculating 3D radiative hydrodynamical models of the fragmentation mechanism. Fragmentation, the break-up of molecular cloud cores during their self-gravitational collapse to form stars, is the leading explanation for the formation of binary and multiple protostars. Molecular cloud cores appear to be supported against collapse in large part by magnetic fields. However, most protostellar fragmentation calculations have either ignored the effects of magnetic fields, or found that in the presence of frozen-in magnetic fields, fragmentation is prohibited. I have shown previously that allowing for magnetic field loss by ambipolar diffusion prior to collapse leads again to fragmentation, but these earlier calculations did not take into account magnetic-field tension, which effectively dilutes the self-gravitational forces once a thin disk forms. Because self-gravity drives fragmentation, other theorists have hypothesized that magnetic tension might then prevent fragmentation. However, my new 3D calculations show that because magnetic tension also helps in avoiding a central density singularity during protostellar collapse, the net effect is to enhance fragmentation of collapsing magnetic cloud cores. Magnetic clouds can thereby fragment into multiple protostar systems (Fig. 2), which are likely to be unstable to subsequent orbital decay.

Figure 1. This image shows an equatorial density after 373 years for a protoplanetary disk with a mass 10% that of the Sun; note the formation of a multiple-Jupiter mass clump (tiny white dot at 12 o'clock). This region has a radius of 20 AU. The surface density in this relatively low-mass disk is comparable to that required for gas giant planet formation by the core-accretion mechanism.

Figure 2. This image depicts the equatorial density after 2.5 free-fall times for a dense molecular cloud core that begins to collapse after losing its magnetic field support through ambipolar diffusion. The cloud core is initially oblate, and collapses to form a ring, which then fragments into a quadruple protostar system. The region shown is two AU in radius. (Red corresponds to high density.)