We report the detection of Kepler-47, a system consisting of two planets orbiting around an eclipsing pair of stars. The inner and outer planets have radii 3.0 and 4.6 times that of Earth, respectively. The binary star consists of a Sun-like star and a companion roughly one-third its size, orbiting each other every 7.45-days. With an orbital period of 49.5-days, 18 transits of the inner planet have been observed, allowing a detailed characterization of its orbit and those of the stars. The outer planet's orbital period is 303.2-days, and although the planet is not Earth-like, it resides within the classical "habitable zone", where liquid water could exist on an Earth-like planet. With its two known planets, Kepler-47 establishes that close binary stars can host complete planetary systems.
Kepler has discovered hundreds of systems with multiple transiting exoplanets which hold tremendous potential both individually and collectively for understanding the formation and evolution of planetary systems. Many of these systems consist of multiple small planets with periods less than ~50 days known as Systems with Tightly spaced Inner Planets, or STIPs. One especially intriguing STIP, Kepler-80 (KOI-500), contains five transiting planets: f, d, e, b, and c with periods of 1.0, 3.1, 4.6, 7.1, and 9.5 days, respectively. We provide measurements of transit times and a transit timing variation (TTV) dynamical analysis. We find that TTVs cannot reliably detect eccentricities for this system, though mass estimates are not affected. Restricting the eccentricity to a reasonable range, we infer masses for the outer four planets (d, e, b, and c) to be 6.75_-0.51_^+0.69^, 4.13_-0.95_^+0.81^, 6.93_-0.70_^+1.05^, and 6.74_-0.86_^+1.23^ Earth masses, respectively. The similar masses but different radii are consistent with terrestrial compositions for d and e and ~2% H/He envelopes for b and c. We confirm that the outer four planets are in a rare dynamical configuration with four interconnected three-body resonances that are librating with few degree amplitudes. We present a formation model that can reproduce the observed configuration by starting with a multi-resonant chain and introducing dissipation. Overall, the information-rich Kepler-80 planets provide an important perspective into exoplanetary systems.
Following the works of Ford et al. (2011, Cat. J/ApJS/197/2; 2012ApJ...756..185F) and Steffen et al. (2012ApJ...756..186S) we derived the transit timing of 1960 Kepler objects of interest (KOIs) using the pre-search data conditioning light curves of the first twelve quarters of the Kepler data. For 721 KOIs with large enough signal-to-noise ratios, we obtained also the duration and depth of each transit. The results are presented as a catalog for the community to use. We derived a few statistics of our results that could be used to indicate significant variations. Including systems found by previous works, we have found 130 KOIs that showed highly significant times of transit variations (TTVs) and 13 that had short-period TTV modulations with small amplitudes. We consider two effects that could cause apparent periodic TTV -- the finite sampling of the observations and the interference with the stellar activity, stellar spots in particular. We briefly discuss some statistical aspects of our detected TTVs. We show that the TTV period is correlated with the orbital period of the planet and with the TTV amplitude.
We produce and analyse eclipse time variation (ETV) curves for some 2600 Kepler binaries. We find good to excellent evidence for a third body in 222 systems via either the light-travel-time (LTTE) or dynamical effect delays. Approximately half of these systems have been discussed in previous work, while the rest are newly reported here. Via detailed analysis of the ETV curves using high-level analytic approximations, we are able to extract system masses and information about the three-dimensional characteristics of the triple for 62 systems which exhibit both LTTE and dynamical delays; for the remaining 160 systems, we give improved LTTE solutions. New techniques of pre-processing the flux time series are applied to eliminate false positive triples and to enhance the ETV curves. The set of triples with outer orbital periods shorter than ~2000d is now sufficiently numerous for meaningful statistical analysis. We find that (i) there is a peak near i_m_=~40{deg} in the distribution of the triple versus inner binary mutual inclination angles that provides strong confirmation of the operation of Kozai-Lidov cycles with tidal friction; (ii) the median eccentricity of the third-body orbits is e_2_=0.35; (iii) there is a deficit of triple systems with binary periods <~1d and outer periods between ~50 and 200d which might help guide the refinement of theories of the formation and evolution of close binaries; and (iv) the substantial fraction of Kepler binaries which have third-body companions is consistent with a very large fraction of all binaries being part of triples.
We present a new method for confirming transiting planets based on the combination of transit timing variations (TTVs) and dynamical stability. Correlated TTVs provide evidence that the pair of bodies is in the same physical system. Orbital stability provides upper limits for the masses of the transiting companions that are in the planetary regime. This paper describes a non-parametric technique for quantifying the statistical significance of TTVs based on the correlation of two TTV data sets. We apply this method to an analysis of the TTVs of two stars with multiple transiting planet candidates identified by Kepler. We confirm four transiting planets in two multiple-planet systems based on their TTVs and the constraints imposed by dynamical stability. An additional three candidates in these same systems are not confirmed as planets, but are likely to be validated as real planets once further observations and analyses are possible. If all were confirmed, these systems would be near 4:6:9 and 2:4:6:9 period commensurabilities. Our results demonstrate that TTVs provide a powerful tool for confirming transiting planets, including low-mass planets and planets around faint stars for which Doppler follow-up is not practical with existing facilities. Continued Kepler observations will dramatically improve the constraints on the planet masses and orbits and provide sensitivity for detecting additional non-transiting planets. If Kepler observations were extended to eight years, then a similar analysis could likely confirm systems with multiple closely spaced, small transiting planets in or near the habitable zone of solar-type stars.
Eighty planetary systems of two or more planets are known to orbit stars other than the Sun. For most, the data can be sufficiently explained by non-interacting Keplerian orbits, so the dynamical interactions of these systems have not been observed. Here we present four sets of light curves from the Kepler spacecraft, each which of shows multiple planets transiting the same star. Departure of the timing of these transits from strict periodicity indicates that the planets are perturbing each other: the observed timing variations match the forcing frequency of the other planet. This confirms that these objects are in the same system. Next we limit their masses to the planetary regime by requiring the system remain stable for astronomical timescales. Finally, we report dynamical fits to the transit times, yielding possible values for the planets' masses and eccentricities. As the timespan of timing data increases, dynamical fits may allow detailed constraints on the systems' architectures, even in cases for which high-precision Doppler follow-up is impractical.
We present a new transit timing catalog of 2599 Kepler Objects of Interest (KOIs), using the PDC-MAP long-cadence light curves that include the full 17 quarters of the mission (ftp://wise-ftp.tau.ac.il/pub/tauttv/TTV/ver_112). The goal is to produce an easy-to-use catalog that can stimulate further analyses of interesting systems. For 779 KOIs with high enough S/N, we derived the timing, duration, and depth of 69914 transits. For 1820 KOIs with lower SNR, we derived only the timing of 225273 transits. After removal of outlier timings, we derived various statistics for each KOI that were used to indicate significant variations. Including systems found by previous works, we have detected 260 KOIs that showed significant TTVs with long-term variations (>100 days), and another 14 KOIs with periodic modulations shorter than 100 days and small amplitudes. For five of those, the periodicity is probably due to the crossing of rotating stellar spots by the transiting planets.
Transit timing variations provide a powerful tool for confirming and characterizing transiting planets, as well as detecting non-transiting planets. We report the results of an updated transit timing variation (TTV) analysis for 1481 planet candidates based on transit times measured during the first sixteen months of Kepler observations. We present 39 strong TTV candidates based on long-term trends (2.8% of suitable data sets). We present another 136 weaker TTV candidates (9.8% of suitable data sets) based on the excess scatter of TTV measurements about a linear ephemeris. We anticipate that several of these planet candidates could be confirmed and perhaps characterized with more detailed TTV analyses using publicly available Kepler observations. For many others, Kepler has observed a long-term TTV trend, but an extended Kepler mission will be required to characterize the system via TTVs. We find that the occurrence rate of planet candidates that show TTVs is significantly increased (~68%) for planet candidates transiting stars with multiple transiting planet candidates when compared to planet candidates transiting stars with a single transiting planet candidate.
We present the results of a survey aimed at discovering and studying transiting planets with orbital periods shorter than one day (ultra-short-period, or USP, planets), using data from the Kepler spacecraft. We computed Fourier transforms of the photometric time series for all 200000 target stars, and detected transit signals based on the presence of regularly spaced sharp peaks in the Fourier spectrum. We present a list of 106 USP candidates, of which 18 have not previously been described in the literature. This list of candidates increases the number of planet candidates with orbital periods shorter than about six hours from two to seven. In addition, among the objects we studied, there are 26 USP candidates that had been previously reported in the literature which do not pass our various tests. All 106 of our candidates have passed several standard tests to rule out false positives due to eclipsing stellar systems. A low false positive rate is also implied by the relatively high fraction of candidates for which more than one transiting planet signal was detected. By assuming these multi-transit candidates represent coplanar multi-planet systems, we are able to infer that the USP planets are typically accompanied by other planets with periods in the range 1-50 days, in contrast with hot Jupiters which very rarely have companions in that same period range. Another clear pattern is that almost all USP planets are smaller than 2 R_{earth}_, possibly because gas giants in very tight orbits would lose their atmospheres by photoevaporation when subject to extremely strong stellar irradiation. Based on our survey statistics, USP planets exist around approximately (0.51+/-0.07)% of G-dwarf stars, and (0.83+/-0.18)% of K-dwarf stars.
An accurate characterization of the known exoplanet population is key to understand the origin and evolution of planetary systems. The determination of true planetary masses through the radial velocity (RV) method is expected to experience a great improvement thanks to the availability of ultra-stable echelle spectrographs. We took advantage of the extreme precision of the new- generation echelle spectrograph ESPRESSO to characterize the transiting planetary system orbiting the G2V star K2-38 located at 194pc from the Sun with V~11.4. This system is particularly interesting because it could contain the densest planet detected to date. We carried out a photometric analysis of the available K2 photometric light curve of this star to measure the radius of its two known planets K2-38b and K2-38c with Pb=4.01593+/-0.00050d and Pc=10.56103+/-0.00090d, respectively. Using 43 ESPRESSO high-precision radial velocity measurements taken over 8 months along with the 14 previously published HIRES RV measurements, we modeled the orbits of the two planets through a Markov Chain Monte Carlo (MCMC) analysis, significantly improving their mass measurements. Using ESPRESSO spectra we derived the stellar paremeters, Teff=5731+/-66K, logg=4.38+/-0.11dex, and [Fe/H]=0.26+/-0.05dex, and thus the mass and radius of K2-38, Ms=1.03^+0.04^_-0.02_M_{sun}_ and Rs=1.06^+0.09^_-0.06_R_[sun}_. We determined new values for the planetary properties of both planets. We characterized K2-38b as a super-Earth with Rp=1.54+/-0.14R_{earth}_ and Mp=7.3^+1.1^_-1.0_M_{earth}_, and K2-38c as a sub-Neptune with Rp=2.29+/-0.26R_{earth}_ and Mp=8.3+/-1.3M_{earth}_. Combining the radius and mass measurements, we derive a mean density of rho_p_=11.0^+4.1^_-2.8_g/cm^3^ for K2-38b and rho_p_=3.8^+1.8^_-1.1_g/cm^3^ for K2-38c, confirming K2-38b as one of the densest planets known to date. The best description for the composition of K2-38b comes from an iron-rich Mercury-like model, while K2-38c is better described by an ice-rich model. The maximum collision stripping boundary shows how giant impacts could be the cause for the high density of K2-38b. The irradiation received by each planet places them on opposite sides of the radius valley. We find evidence of a long-period signal in the radial velocity time-series whose origin could be linked to a 0.25-3M_Jup_ planet or stellar activity.
