Removing biases on the density of sub-Neptunes characterised via transit timing variations

Abstract

Transit timing variations (TTVs) can provide useful information on compact multi-planetary systems observed by transits by setting constraints on the masses and eccentricities of the observed planets. This is especially helpful when the host star is not bright enough for a radial velocity (RV) follow-up. However, in the past decade, a number of works have shown that TTV-characterised planets tend to have lower densities than planets characterised on the basis of RVs. Re-analysing 34 Kepler planets in the super-Earth to sub-Neptunes range using the RIVERS approach, we show that at least some of these discrepancies were due to the way transit timings were extracted from the light curve, as a result of their tendency to underestimate the TTV amplitudes. We recovered robust mass estimates (i.e. with low prior dependency) for 23 of the planets. We compared these planets the RV-characterised population and found that a large fraction of those that previously had unusually low density estimates were adjusted, allowing them to occupy a place on the mass-radius diagram much closer to the bulk of known planets. However, a slight shift toward lower densities remains, which could indicate that the compact multi-planetary systems characterised by TTVs are indeed composed of planets that are different from the bulk of the RV-characterised population. These results are especially important in the context of obtaining an unbiased view of the compact multi-planetary systems detected by Kepler, TESS, and the upcoming PLATO mission.