How do Most Planets Form? -- Constraints on Disk Instability from Direct Imaging

Abstract: Core accretion and disk instability have traditionally been regarded as the two competing possible paths of planet formation. In recent years, evidence have accumulated in favor of core accretion as the dominant mode, at least for close-in planets. However, it might be hypothesized that a significant population of wide planets formed by disk instabilities could exist at large separations, forming an invisible majority. In previous work, we addressed this issue through a direct imaging survey of B2--A0-type stars, and concluded that <30% of such stars form and retain planets and brown dwarfs through disk instability, leaving core accretion as the likely dominant mechanism. In this paper, we extend this analysis to FGKM-type stars by applying a similar analysis to the Gemini Deep Planet Survey (GDPS) sample. The results strengthen the conclusion that substellar companions formed and retained around their parent stars by disk instabilities are rare. Specifically, we find that the frequency of such companions is <8% for FGKM-type stars under our most conservative assumptions, for an outer disk radius of 300 AU, at 99% confidence. Furthermore, we find that the frequency is always <10% at 99% confidence independently of outer disk radius, for any radius from 5 to 500 AU. We also simulate migration at a wide range of rates, and find that the conclusions hold even if the companions move substantially after formation. Hence, core accretion remains the likely dominant formation mechanism for the total planet population, for every type of star from M-type through B-type.