A rocky composition for an Earth-sized exoplanet

Abstract

Planets with sizes between that of Earth (with radius R ⊕ ) and Neptune (about 4 R ⊕ ) are now known to be common around Sun-like stars 1,2,3 . Most such planets have been discovered through the transit technique, by which the planet's size can be determined from the fraction of starlight blocked by the planet as it passes in front of its star. Measuring the planet's mass-and hence its density, which is a clue to its composition-is more difficult. Planets of size 2-4 R ⊕ have proven to have a wide range of densities, implying a diversity of compositions 4,5 , but these measurements did not extend down to planets as small as Earth. Here we report Doppler spectroscopic measurements of the mass of the Earth-sized planet Kepler-78b, which orbits its host star every 8.5 hours (ref. 6). Given a radius of 1.20 ± 0.09 R ⊕ and mass of 1.69 ± 0.41 M ⊕ , the planet's mean density of 5.3 ± 1.8 g cm -3 is similar to the Earth's, suggesting a composition of rock and iron.Kepler-78 is one of approximately 150,000 stars whose brightness was precisely measured at 30-minute intervals for four years by the NASA Kepler spacecraft 7 . This star is somewhat smaller, less massive, and younger than the Sun (Table 1). Every 8.5 hours the star's brightness declines by 0.02% as the planet Kepler-78b transits (passes in front of) the stellar disk. The planet's radius was originally measured 6 to be 1.16 !!.!" !!.!" R ⊕ . The mass could not be measured, although masses >8 M ⊕ could be ruled out because the planet's gravity would have deformed the star and produced brightness variations that were not detected.We measured the mass of Kepler-78b by tracking the line-of-sight component of the host star's motion (the radial velocity, RV) due to the gravitational force of the planet. The RV analysis is challenging not only because the signal is expected to be small (~1-3 m s -1 ) but also because the apparent Doppler shifts due to rotating starspots are much larger (~50 m s -1 peak-to-peak). Nevertheless the detection proved to be possible, thanks to the precisely known orbital period and phase of Kepler-78b that cleanly separated the timescale of spot variations (P rot ≈ 12.5 days) from the much shorter timescale of the planetary orbit (P ≈ 8.5 hours). We adopted a strategy of intensive Doppler measurements spanning 6-8 hours per night, long enough to cover nearly the entire orbit and short enough for the spot variations to be nearly frozen out.We measured RVs using optical spectra of Kepler-78 that we obtained from the High Resolution Echelle Spectrometer (HIRES) 8 on the 10-m Keck I Telescope. These Doppler shifts were computed relative to a template spectrum with a standard algorithm 9 that uses a spectrum of molecular iodine superposed on the stellar spectrum as a reference for the wavelength scale and instrumental profile of HIRES (Supplementary Table 1). Exposures lasted 15-30 minutes depending on conditions and produced RVs with 1.5-2.0 m s -1 uncertainties. The time series of RVs spans 38 days, with large velocity offsets between nights ...

SECURE MASS MEASUREMENTS FROM TRANSIT TIMING: 10 KEPLER EXOPLANETS BETWEEN 3 AND 8 M ⊕ WITH DIVERSE DENSITIES AND INCIDENT FLUXES

“…Radial velocity (RV)spectroscopy has provided the bulk of mass measurements of transiting exoplanets (Batalha et al 2011;Winn et al 2011;Fressin et al 2012;Gillon et al 2012;Gautier et al 2012;Howard et al 2013;Pepe et al 2013;Weiss et al 2013;Haywood et al 2014;Marcy et al 2014;Dressing et al 2015). However, among subNeptune-mass planets, RV mass detections are limited to planets on short orbital periods, because the RV signal depends on the strength of planet-star interactions and declines with orbital distance.…”

Deformation and tidal evolution of close-in planets and satellites using a Maxwell viscoelastic rheology

“…Here we apply our model to the planets 55 Cnc e (Gillon et al 2012) and Kepler-78 b (Howard et al 2013). The main difference between these two planets is their masses, only 1.7 M ⊕ for Kepler-78 b, and about 8.6 M ⊕ for 55 Cnc e (Table 1).…”

Ground-based near-UV observations of 15 transiting exoplanets: constraints on their atmospheres and no evidence for asymmetrical transits

“…The BIC has been used extensively in many other exoplanet transit studies (e.g. Gibson et al 2010Gibson et al , 2013bKipping et al 2010;Croll et al 2011;Sing et al 2011;Crossfield et al 2013;Demory et al 2013;Howard et al 2013;Rogers et al 2013;Murgas et al 2014;Stevenson et al 2014a;Zellem et al 2014 to find a best fit to the data and a bootstrap Monte Carlo technique (Press et al 1992) where σ p are the photometric (observational) error bars for each data point in the light curve, N(μ, σ 2 ) is a random Gaussian distributed variable (N) with a mean μ = 0 and a standard deviation σ = 1. (3) We add σ n to the flux measurements in the light curve.…”