Giant planet formation in the near-infrared: single large telescopes are not enough

Giant planets are thought to form at large orbital separations (>3 au), which is why direct imaging is crucial to detect young gas giants and study their formation process. PDS 70 b remains the only fascinating case of such an object for which accretion is clearly evident.

Using Keck/NIRC2, we search for companions around 33 members of the 1 Myr old Taurus star-forming region. Fourier plane imaging (kernel phase) and PSF subtraction allow us to probe a large range of orbital separations down to Solar System scales ( 6 au = 0.5 λ/D), which has not been possible before in the near-infrared. Together with our simulation of giant planet formation via core accretion and the latest planet distribution from radial velocity surveys, our observations put some constraints on the peak luminosity and the timescale of the runaway accretion.

Considering future observations, we show that even with optimistic assumptions, the number of "normal" core accretion giant planets detectable by 8-10 m telescopes is of order 1. This is due to evidence for a turnover in planet frequency at 3 au, the small number of solar-mass young stars harbouring gas disks within 200 pc, and the insufficient brightness of truly "Jovian" planets. We make the case that in order to probe the formation of planets like Jupiter, a high-contrast interferometric instrument (such as Hi-5/VIKiNG) is required.