Constraining the initial entropy of directly-detected exoplanets

Abstract: The post-formation, initial entropy S_i of a gas giant planet is a key witness to its mass-assembly history and a crucial quantity for its early evolution. However, formation models are not yet able to predict reliably S_i, making unjustified the use solely of traditional, 'hot-start' cooling tracks to interpret direct imaging results and calling for an observational determination of initial entropies to guide formation scenarios. Using a grid of models in mass and entropy, we show how to place joint constraints on the mass and initial entropy of an object from its observed luminosity and age. This generalises the usual estimate of only a lower bound on the real mass through hot-start tracks. Moreover, we demonstrate that with mass information, e.g. from dynamical stability analyses or radial velocity, tighter bounds can be set on the initial entropy. We apply this procedure to 2M1207 b and find that its initial entropy is at least 9.2 kB/baryon, assuming that it does not burn deuterium. For the planets of the HR 8799 system, we infer that they must have formed with S_i > 9.2 kB/baryon, independent of uncertainties about the age of the star. Finally, a similar analysis for beta Pic b reveals that it must have formed with S_i > 10.5 kB/baryon, using the radial-velocity mass upper limit. These initial entropy values are respectively ca. 0.7, 0.5, and 1.5 kB/baryon higher than the ones obtained from core accretion models by Marley et al., thereby quantitatively ruling out the coldest starts for these objects and constraining warm starts, especially for beta Pic b.