2021MNRAS.505..869W

The exotic range of known planetary systems has provoked an equally exotic range of physical explanations for their diverse architectures. However, constraining formation processes requires mapping the observed exoplanet population to that which initially formed in the protoplanetary disc. Numerous results suggest that (internal or external) dynamical perturbation alters the architectures of some exoplanetary systems. Isolating planets that have evolved without any perturbation can help constrain formation processes. We consider the Kepler multiples, which have low mutual inclinations and are unlikely to have been dynamically perturbed. We apply an adaption of previous modelling efforts, accounting for the two-dimensionality of the radius (R_pl =0.3-20 R_⊕) and period (P_orb = 0.5-730 d) distribution. We find that an upper limit in planet mass of the form M_lim ∝a_pl^β^ exp(-a_in/a_pl), for semimajor axis a_pl and a broad range of a_in and β, can reproduce a distribution of P_orb, R_pl that is indistinguishable from the observed distribution by our comparison metric. The index is consistent with β = 1.5, expected if growth is limited by accretion within the Hill radius. This model is favoured over models assuming a separable PDF in P_orb, R_pl. The limit, extrapolated to longer periods, is coincident with the orbits of RV-discovered planets (a_pl > 0.2 au, M_pl> 1 M_J) around recently identified low density host stars, hinting at isolation mass limited growth. We discuss the necessary circumstances for a coincidental age-related bias as the origin of this result; such a bias is possible but unlikely. We conclude that, in light of the evidence suggesting that some planetary systems have been dynamically perturbed, simple models for planet growth during the formation stage are worth revisiting.