Kepler-10 is a historic system from the treasury of the NASA Kepler Mission‘s discoveries.  It contains the first definitively rocky Kepler planet, Kepler-10 b.  It also contains what has been mistakenly called a “mega-Earth,” Kepler-10 c.  In a new paper, my co-authors and I investigate whether Kepler-10 c could have a rocky or icy outer layer. Our new results from the Keck telescope show that Kepler-10 c, a planet which was previously thought to be the mass of Neptune even though it is only 2.35 times the radius of Earth, is not so massive.  Now weighing in at only 2/3 of Neptune’s mass, Kepler-10 c requires at least 10% of its radius (30% of its volume) to come from a gaseous envelope.

What could the gaseous envelope be?  According to the detailed simulations done by second author Leslie Rogers, both a hydrogen-helium envelope similar in composition to the sun and a steamy envelope of water are possible.  If the envelope is mostly hydrogen-helium, it makes up 0.2% of the planet’s mass and 16% of the radius.  If the envelope is pure steam, it makes up 30% of the mass and 30% of the radius.  (Helium balloons are waaaay less dense than water, and uncompressed hydrogen is waaay less dense than the helium in your balloon.)

The updated measurements of the masses and radii of exoplanets Kepler-10 b and Kepler-10 c. Kepler-10 b (“b”) has a mass and radius that place it between the constant-composition line of a ball of pure silicate (yellow line) and an Earth-like composition of silicate and iron (brown line). Kepler-10 c (“c”) has a mass and radius that allow a pure silicate ball (a composition that has never actually been observed in astronomy and would require the highest-density interpretation of the planet), but substantially favor lower-density alternatives that might include water layers and hydrogen layers. (A ball of pure water would fall on the blue line.)

In addition to clearing up the possible compositions for Kepler-10 c, we found a new planet candidate!  We call it KOI-72.X, in keeping with the convention that only confirmed planets get “Kepler” names.  The candidate has at least a 97% chance of being a real planet (99% is the threshold for confirmation).  The clue that led to this planet candidate’s discovery is that its neighbor, Kepler-10 c, has slight changes in its orbit each time it passes in front of the star.  The changes in Kepler-10 c’s orbit are slight, but if they are real, they almost certainly come from a third planet the system, KOI-72.X.  Co-authors David Kipping, Jason Rowe, and Eric Agol each found evidence for KOI-72.X through an independent analysis of the motion of Kepler-10 c, and I put their work together to determine possible orbits and masses for the new planet candidate.  So far as we can tell, KOI-72.X does not pass in front of the star from our point of view, and so it eludes confirmation.  Also, because we have not seen KOI-72.X transit its star, we don’t know its physical radius, which means that we cannot pinpoint its density.  Therefore, the composition of KOI-72.X remains a mystery.

Kepler-10 data
These panels show two data products from the Kepler-10 system. The left three panels show the transit timing variations for each planet (labeled O-C, for “observed minus calculated,” a.k.a. how late this transit was compared to our expectations) as a function of time in days. The top two panels show the observed transit timing variations of Kepler-10 b and c as colored points. The black curves show our best-fit model. The transit timing variations of planet c in this model are produced by a third planet, KOI-72.X, which has an orbital period of 101 days. We do not see KOI-72.X transit, so we cannot compare its predicted times (bottom left panel) to observations. The right three panels show the component of the star’s motion, in meters per second, due to the gravitational influence of each planet. The blue points are our Keck observations, the green points are the previously published HARPS-N data, and the red diamonds show weighted averages. The black curves are from the same best-fit model as what is shown on the left.