Make me a planet

Planet formation. Not so long ago one might have said that we had some pretty good ideas about how it worked. None of them were perfect for sure, but the general feeling was that somewhere in amongst the various studies we were edging towards a reasonable physical model. On the one side was core accretion - the coalescence of solids from a proto-planetary disk with giant planets slurping up gas after passing a critical core size and rocky planets coming along a little later. On the other was gravitational instability - density patterns in proto-planetary disks reaching critical points of instability where gravity would collapse gas to giant spherical blobs that could then hoover up solids to build a core. Later evaporation might produce ice-giants and the rocky planets would still form by direct accretion. A combination of these mechanisms seemed increasingly likely.

Now in the past couple of weeks a slew of new results and reconsidered old results seem to be calling for some serious rethinking of what we know about planet formation. The recent Kepler data release indicates a propensity for highly packed planetary systems - if you can build it, it will be there. It also demonstrates that Neptune-class worlds are incredibly common (actually confirming earlier microlensing results). It also further confirms an apparent 'pile-up' or discontinuity in very short period planets. Around 3 day orbital periods then something is going on between planets and their parent stars to hold them back from being tidally hauled to death in stellar atmospheres. New ultra-high fidelity imaging of young planetary systems confirms earlier sightings that proto-planetary disks of gas and dust can be far from symmetric or simple. A simple interpretation of the lop-sided annuli of circumstellar material is that we're watching as giant proto-planets scoop up matter during their multi-decade orbits. The extent of the disturbances is intriguing. 

Imaging of giant planets on very long orbits of 100 astronomical units, as in the Formalhaut system, are an immense and surprising challenge to planet formation models - especially if the orbits are not highly elliptical, which could be explained by dynamical scattering from an inner origin. And finally, there is the revised talk of an outer giant in our own solar system, perhaps 1 to 4 times the mass of Jupiter. If such a world, currently nicknamed Tyche, were to exist lurking beyond about 2,000 astronomical units (orbital periods of more than 90,000 years) then our solar system would have likely formed as a highly lopsided star-planet binary.

Where do we go from here? It's quite a challenge. There was already a long list of yet-to-be-fully-solved issues, from orbital inclinations and retrograde planets to orbital migration and tidal evolution. An inherent difficulty is the vast parameter space involved. Even with a single coherent model for planet formation then every individual system will evolve at the whim of non-linear dynamics and stochastic or random processes. It seems quite likely that we will end up having to adopt some type of classification scheme to just sort the wheat from the chaff. The big question is what classification scheme has the most physical meaning?

The one aspect that I personally glean from the new wealth of data is that it is almost overwhelmingly convincing that if there is an opportunity for planets to form then they will, with a vengeance. If even pulsars can host objects likely re-coalesced from post-supernova debris then there are some pretty potent mechanisms at play. For me this suggests that a fertile ground for investigation might be to flip the question around to ask exactly what the tipping point is? Do all stars initially form planets? Or is there a critical level of element abundance, dynamical environment, or radiation environment below which there cannot be planet formation of any kind? How many of those twinkling objects in the night sky are genuinely barren and alone?