Kepler-432: a red giant interacting with one of its two long-period giant planets

TLDR: In this paper, the authors reported the discovery of Kepler-432b, a giant planet (M_b = 5.41^(+0.12)_(-0.036)_−0.039)R_Jup) transiting an evolved star with an eccentricity of e=0.

Abstract: We report the discovery of Kepler-432b, a giant planet (M_b = 5.41^(+0.32)_(-0.18)M_Jup, R_b = 1.145^(+0.036)_(-0.039)R_Jup) transiting an evolved star (M_* = 1.32^(+0.10)_(-0.07)M_⊙, R_* = 4.06^(+0.12)_(-0.08)R_⊙) with an orbital period of P_b = 52.501129^(+0.000067)_(-0.000053) days. Radial velocities (RVs) reveal that Kepler-432b orbits its parent star with an eccentricity of e=0.5134^(+0.0098)_(-0.0089), which we also measure independently with asterodensity profiling (AP; e=0.507^(+0.039)_(-0.114)), thereby confirming the validity of AP on this particular evolved star. The well-determined planetary properties and unusually large mass also make this planet an important benchmark for theoretical models of super-Jupiter formation. Long-term RV monitoring detected the presence of a non-transiting outer planet (Kepler-432c; M_c sin i_c = 2.43^(+0.22)_(-0.24) M_Jup, P_c = 406.2^(+3.9)_(-2.5) days), and adaptive optics imaging revealed a nearby (0".87), faint companion (Kepler-432B) that is a physically bound M dwarf. The host star exhibits high signal-to-noise ratio asteroseismic oscillations, which enable precise measurements of the stellar mass, radius, and age. Analysis of the rotational splitting of the oscillation modes additionally reveals the stellar spin axis to be nearly edge-on, which suggests that the stellar spin is likely well aligned with the orbit of the transiting planet. Despite its long period, the obliquity of the 52.5 day orbit may have been shaped by star–planet interaction in a manner similar to hot Jupiter systems, and we present observational and theoretical evidence to support this scenario. Finally, as a short-period outlier among giant planets orbiting giant stars, study of Kepler-432b may help explain the distribution of massive planets orbiting giant stars interior to 1 AU.

Figures

Table 2 Kepler-432 Relative Radial Velocities

Table 5 Kepler-432 Planetary Properties

Figure 8. Échelle diagram of Kepler-432 showing observed frequencies in white. Modes are identified as ℓ = 0 (circles), ℓ = 1 (triangles), ℓ = 2 (squares), and ℓ = 3 (diamonds). Frequencies of the best-fitting MESA and ASTEC models are indicated by the red and blue open symbols, respectively. For reference, a gray-scale map of the power spectrum is shown in the background. Numbers to the right of the plot indicate the radial order of the ℓ = 0 modes.

Figure 9. MCMC marginalized posterior distributions of the radial-velocity orbital parameters for the Kepler-432 system. While the inner planet (top row) has a wellconstrained solution, it is clear from the marginalized distributions that there is a tail of high-eccentricity outer orbits (bottom row) that cannot be ruled out by the current RV data. This tail can also be seen in the joint mass-eccentricity posterior shown in the stability analysis of Figure 11.

Figure 12. Stability of systems in the mass–orbital inclination plane of the outer planet (m i,c c), with black points going unstable within 104 yr and yellow points within 105 yr, and blue points representing stable systems after integrating for 105 yr. The results mirror those of the coplanar case—at all inclinations, the low-mass (low-eccentricity) systems remain stable, while the high-mass (high-eccentricity) systems do not survive long integrations. As some systems remain stable even for highly inclined orbits ( ~ i 0c ), we are unable to put strong constraints on the mutual inclination of the planets.

Table 3 Stellar Properties of Kepler-432

Frequently asked questions

Q1. What are the contributions in "Kepler-432: a red giant interacting with one of its two long-period giant planets" ?

The authors report the discovery of Kepler-432b, a giant planet ( = + M M 5. 41 b 0. 18 0. 32 Jup, = + R R 1. 145 b 0. 039 0. 036 Jup ) transiting an evolved star (   = = + +   M M R R 1. 32, 4. 06 0. 07 0. 10 0. 08 0. 12 ) with an orbital period of = + P 52. 501129 b 0. 000053 0. 000067 days. 5 day orbit may have been shaped by star–planet interaction in a manner similar to hot Jupiter systems, and the authors present observational and theoretical evidence to support this scenario. Finally, as a short-period outlier among giant planets orbiting giant stars, study of Kepler-432b may help explain the distribution of massive planets orbiting giant stars interior to 1 AU. Analysis of the rotational splitting of the oscillation modes additionally reveals the stellar spin axis to be nearly edge-on, which suggests that the stellar spin is likely well aligned with the orbit of the transiting planet. 

Q2. Why is Kepler able to find companions that are intrinsically rare or otherwise difficult?

Because of its unprecedented photometric sensitivity,duty cycle, and time coverage, companions that are intrinsically rare or otherwise difficult to detect are expected to be found by Kepler, and detailed study of such discoveries can lead to characterization of poorly understood classes of objects and physical processes. 

Q3. What are the parameters that were included in the fit?

Twelve parameters were included in the fit: for each planet, the times of inferior conjunction T0, orbital periods P, radial-velocity semi-amplitudes K, and the orthogonal quantities the authors sin and the authors cos , where e is orbital eccentricity and ω is the longitude of periastron; the systemic velocity, grel, in the arbitrary zero point of the TRES relative RV data set; and the FIES RV offset, DRVFIES. 

Q4. What was used to correct for systematic velocity shifts between runs?

The nightly observations of RV standards were used to correct for systematic velocity shifts between runs and to estimate the instrumental precision. 

Q5. How many time steps per innermost orbit is sufficient to ensure numerical stability?

Rauch & Holman (1999) demonstrated that ∼20 time steps per innermost orbit is sufficient to ensure numerical stability in symplectic integrations. 

Q6. What is the fit to the derived stellar properties?

The stellar model best fit to the derived stellar properties provides color indices that may be compared against measured values as a consistency check and as a means to determine a photometric distance to the system. 

Q7. Why do the authors think the v isin measurement is biased?

The authors do caution that their v isin measurement for this slowly rotating giant could be biased, for example, due to the unknown macroturbulent velocity of Kepler-432. 

Q8. How can the authors mitigate the effect of the global MCMC fit?

This means that the heights of the m = ±1 components relative to the m = 0 component will change in opposite directions, so the effect can be mitigated by forcing the m = ±1 components to have the same height in the fit, as well as by performing a global fit to all modes, as the authors have done.