Mass Constraints of the WASP-47 Planetary System from Radial Velocities
Evan Sinukoff, Andrew W. Howard, Erik A. Petigura, Benjamin J. Fulton, Howard Isaacson, Lauren M. Weiss, John M. Brewer, Brad M. S. Hansen, Lea Hirsch, Jessie L. Christiansen, Justin R. Crepp, Ian J. M. Crossfield, Joshua E. Schlieder, David R. Ciardi, Charles A. Beichman, Heather A. Knutson, Bjoern Benneke, Courtney D. Dressing, John H. Livingston, Katherine M. Deck, Sebastien Lepine, Leslie A. Rogers
DDraft version August 4, 2018
Preprint typeset using L A TEX style emulateapj v. 01/23/15
MASS CONSTRAINTS OF THE WASP-47 PLANETARY SYSTEM FROM RADIAL VELOCITIES
Evan Sinukoff , Andrew W. Howard , Erik A. Petigura , Benjamin J. Fulton , Howard Isaacson ,Lauren M. Weiss , John M. Brewer , Brad M. S. Hansen , Lea Hirsch , Jessie L. Christiansen , Justin R.Crepp , Ian J. M. Crossfield , Joshua E. Schlieder , David R. Ciardi , Charles A. Beichman , HeatherA. Knutson , Bjoern Benneke , Courtney D. Dressing , John H. Livingston , Katherine M. Deck ,Sébastien Lépine Leslie A. Rogers Draft version August 4, 2018
ABSTRACTWe report precise radial velocity (RV) measurements of WASP-47, a G star that hosts three tran-siting planets in close proximity (a hot Jupiter, a super-Earth and a Neptune-sized planet) and anon-transiting planet at 1.4 AU. Through a joint analysis of previously published RVs and our ownKeck-HIRES RVs, we significantly improve the planet mass and bulk density measurements. For thesuper-Earth WASP-47e ( P = 0.79 days), we measure a mass of . ± . M ⊕ , and a bulk densityof . ± . g cm − , consistent with a rocky composition. For the hot Jupiter WASP-47b ( P = 4.2days), we measure a mass of ± M ⊕ (1.12 ± M Jup ) and constrain its eccentricity to < . at 3- σ confidence. For the Neptune-size planet WASP-47d ( P = 9.0 days), we measure a mass of . ± . M ⊕ , and a bulk density of . ± . g cm − , suggesting it has a thick H/He envelope.For the outer non-transiting planet, we measure a minimum mass of ± M ⊕ (1.29 ± M Jup ),an orbital period of . ± . days, and an orbital eccentricity of . ± . . Our new measurementsare consistent with but 2–4 × more precise than previous mass measurements. INTRODUCTION
Approximately 1% of Sun-like stars host giant plan-ets on short-period orbits (P <
10 days), known as hotJupiters (HJs, Howard et al. 2012; Wright et al. 2012).These planets are thought to have migrated to their ob-served locations from beyond the ice-line at several AU.One proposed migration mechanism involves dynamicalinteraction between the planet and protoplanetary disk Institute for Astronomy, University of Hawai‘i at M¯anoa,Honolulu, HI 96822, USA Cahill Center for Astrophysics, California Institute of Tech-nology, 1216 East California Boulevard, Pasadena, CA 91125,USA Division of Geological and Planetary Sciences, California In-stitute of Technology, 1255 East California Blvd, Pasadena, CA91125, USA Astronomy Department, University of California, Berkeley,CA, USA Institut de Recherche sur les Exoplanètes, Dèpartement dePhysique, Universitè de Montrèal, C.P. 6128, Succ. Centre-ville,Montréal, QC H3C 3J7, Canada Department of Astronomy, Yale University and 260 WhitneyAvenue, New Haven, CT 06511, USA Department of Physics & Astronomy and Institute of Geo-physics & Planetary Physics, University of California Los Ange-les, Los Angeles, CA 90095, USA NASA Exoplanet Science Institute, California Institute ofTechnology, 770 S. Wilson Ave., Pasadena, CA, USA Department of Physics, University of Notre Dame, 225Nieuwland Science Hall, Notre Dame, IN, USA Department of Astronomy & Astrophysics, University ofCalifornia Santa Cruz, 1156 High St., Santa Cruz, CA, USA Department of Astronomy, The University of Tokyo, 7-3-1Bunkyo-ku, Tokyo 113-0033, Japan Department of Physics and Astronomy, Georgia State Uni-versity, GA, USA Department of Astronomy & Astrophysics, University ofChicago, 5640 South Ellis Avenue, Chicago, IL 60637, USA NSERC Postgraduate Research Fellow Hubble Fellow NSF Graduate Research Fellow NASA Sagan Fellow NASA Postdoctoral Program Fellow (e.g. Lin et al. 1996). In this case, the planet main-tains a low eccentricity. Other “high-eccentricity mi-gration” (HEM) modes have been proposed includingplanet-planet scattering (e.g. Rasio & Ford 1996), Kozaioscillations induced by either a nearby star (e.g. Wu &Murray 2003) or planet (e.g. Naoz et al. 2011), and secu-lar interactions (e.g. Wu & Lithwick 2011). In the HEMscenario, gravitational perturbations excite planets ontoeccentric orbits, which subsequently shrink and circular-ize due to stellar tides. Other proposed dynamical effectsinclude misalignment between the orbital axis of the HJand the stellar spin axis, as well as the destabilization ofclose-in planets encountered upon migration.Observations of systems with HJs are difficult to recon-cile with HEM theory. For example, Schlaufman & Winn(2016) found that HJ host stars are no more likely to hostadditional giant planets than stars with giant planets at
P > days. Knutson et al. (2014) found no differencebetween the occurrence of additional giant planets at 1–20 AU in systems with HJs whose orbits are eccentricor misaligned versus circular and aligned with the stellarspin. Moreover, Dawson et al. (2015) concluded that thenumber of migrating Jupiters on highly eccentric orbitsis lower than predicted by HEM theory (Socrates et al.2012).In support of HEM theory, Steffen et al. (2012) foundan absence of HJs in close proximity to smaller planets(0.7–5 R ⊕ ) discovered by Kepler . However, it remainsunclear whether HJs are intrinsically lonely or if theirclose neighbors have merely evaded detection. For ex-ample, Batygin et al. (2016) proposed a mechanism forin-situ formation of HJs, which predicts a population ofsmall planets mutually inclined to the HJ, and thereforeunlikely to transit. While HJs are observed to be lonely,Huang et al. (2016) found that roughly half of transiting“warm-Jupiters" ( P = 10–200 days) are accompanied bytransiting planets ∼ R ⊕ on interior orbits P < a r X i v : . [ a s t r o - ph . E P ] D ec Sinukoffdays. They proposed that the warm-Jupiters in thesemulti-planet systems formed in-situ and that occasion-ally this same mechanism might produce a very smallfraction of HJs. These latest theories add to the diver-sity of theories to explain HJ formation.WASP-47 is the first star known to host a Jovian-sizeplanet with
P <
10 days and additional close-in planets—proof that not all HJs are isolated and strengtheningthe argument that HEM alone cannot produce the en-tire population of HJs. The Jovian-size planet WASP-47b orbits the star every 4.2 days. It was first reportedand confirmed by Hellier et al. (2012) who detected bothits transit and radial velocity (RV) signatures. Beckeret al. (2015) detected two additional transiting planetsusing K2 photometry. One of these planets, WASP-47e, is an ultra-short-period (USP) super-Earth ( P =0.79 days). WASP-47d is Neptune-size ( P = 9.0 days).Becker et al. (2015) detected transit timing variations(TTVs) of both planets. Their TTV signals are anticor-related and have a super-period consistent with 52.67-days — the expected super-period for two such planetsnear 2:1 orbital mean-motion-resonance (Lithwick et al.2012). Becker et al. (2015) reported planet mass con-straints M b = +73 − M ⊕ , M d = . ± M ⊕ , and M e < M ⊕ based on dynamical fits to the observed transittimes. Measurements of the Rossiter-McLaughlin effectby (Sanchis-Ojeda et al. 2015) ruled out orbits that arestrongly misaligned with the stellar spin axis. Crossfieldet al. (2016) independently validated the planetary sys-tem by demonstrating that the star is unlikely to be ablend of multiple stars, via Keck-NIRC2 adaptive opticsimages and a search for secondary lines in the stellarspectrum.A fourth planet, WASP-47c, was detected with anorbital period of ± days by Neveu-VanMalleet al. (2016) from 32 RV observations with the Eu-ler/CORALIE instrument spanning almost 3 years .Neveu-VanMalle et al. (2016) measure a minimum mass M c sin i = 394 ± M ⊕ . WASP-47c joins the populationof giant planets beyond 1 AU that have been found insystems with HJs (Knutson et al. 2014).WASP-47d and WASP-47e are examples of super-Earth- and Neptune-size planets, which are commonaround Sun-like stars (Howard et al. 2012; Fressin et al.2013; Petigura et al. 2013; Burke et al. 2015). Only ahandful of these planets have precisely measured massesand bulk densities. Compositional trends have emergedfrom this limited sample. Planets smaller than ≈ R ⊕ typically have high densities consistent with Earth-likebulk compositions, while most larger planets have lowdensities that require thick envelopes of H/He (Weiss &Marcy 2014; Marcy et al. 2014; Lopez & Fortney 2014;Rogers 2015; Dressing et al. 2015). However, there issignificant scatter about the mean mass-radius relation-ship, indicating compositional diversity, even for a fixedplanet radius. Due to the limited number of known sub-Neptunes with bright host stars, mass measurements arescarce, and this compositional diversity has yet to befully explored.Dai et al. (2015) obtained 26 RVs of WASP-47 with the WASP-47d and WASP-47e were published before WASP-47c,which was named while the work of Neveu-VanMalle et al. (2016)was still under revision.
Carnegie Planet Finder Spectrograph (PFS), measuring M b = ± M ⊕ , M e = . ± . M ⊕ , and M d = . ± . M ⊕ , consistent with TTV measurements byBecker et al. (2015). Dai et al. (2015) measure a bulkdensity of WASP-47e of 11.2 ± − , consistentwith a rocky and potentially iron-rich composition. Their ∼
80% measurement uncertainty on the mass of planet dis insufficient to constrain the planet’s bulk composition.Here we present improved mass constraints of all fourplanets in the WASP-47 system by combining Keck-HIRES RVs with the previously published RVs of Hellieret al. (2012), Dai et al. (2015), and Neveu-VanMalle et al.(2016). This work is part of a NASA “Key Project” tomeasure K2 planet masses using Keck-HIRES. Section2 of this manuscript summarizes our Doppler observa-tions and spectroscopic constraints of stellar parameters.Our analysis of the RV time-series and resulting planetmass measurements are detailed in Section 3. In Section4, we discuss possible compositions of WASP-47e andWASP-47d, eccentricity constraints of the Hot-Jupiter,and interpret these in the context of planet formationand evolution. OBSERVATIONS
K2 Photometry
WASP-47 was observed by the Kepler Telescope for 69consecutive days in Campaign 3 (C3) of NASA’s K2 mis-sion (Howell et al. 2014). It was one of only 55 targetsin K2 Campaign 3 that was observed in short-cadencemode (60 sec), enabling precise measurement of tran-sit parameters. We adopt the orbital ephemerides, andtransit depths reported by Becker et al. (2015).
Radial Velocity Measurements
We collected RV measurements of WASP-47 usingHIRES (Vogt et al. 1994) at the W. M. Keck Observa-tory from 2015 August 10 UT to 2016 October 7 UT (424days). We followed standard procedures of the Califor-nia Planet Search (CPS; Howard et al. 2010). For eachRV observation, we used the “C2” decker ( . (cid:48)(cid:48) × (cid:48)(cid:48) slit), which yields a spectral resolution R = 55,000 andis long enough for sky subtraction. Before the starlightentered the spectrometer slit, it first passed through acell of iodine gas, which imprints a dense set of molecu-lar absorption lines on the stellar spectrum. These iodinelines were used for wavelength calibration and PSF refer-ence. We used an exposure meter to terminate exposuresafter reaching a SNR per pixel of ∼
100 at 550 nm (typ-ically ∼
15 min). A single iodine-free spectrum was ob-tained as a stellar template using the “B3” decker ( . (cid:48)(cid:48) × (cid:48)(cid:48) slit). RVs were measured by forward modelingeach observed spectrum as the product of an RV-shiftediodine-free spectrum and a high-resolution/high-SNR io-dine transmission spectrum. The latter was first con-volved with an instrumental PSF, modeled as the sum of13 Gaussians with fixed centers and widths but variableamplitudes (Marcy & Butler 1992; Valenti et al. 1995;Butler et al. 1996; Howard et al. 2009). Our measuredRVs are listed in Table 3. Stellar Parameters
We measured the effective temperature ( T eff ), surfacegravity ( log g ), and metallicity ( [ Fe/H ] ) of WASP-47ASP-47 3 TABLE 1RV datasets
Reference a Instrument N RV Median Unc. ∆ t [m s − ] [days]This study HIRES 47 b a V16: Neveu-VanMalle et al. (2016), D15: Dai et al.(2015), H12: Hellier et al. (2012) b We made 74 RV measurements with Keck-HIRES, butomit 17 RVs measured during the WASP-47b transit eventon 2015 August 10 UT. We binned the remaining 12 RVsfrom that night into two measurements for a total of 47 RVs from our iodine-free HIRES spectrum using the updatedSME analysis of Brewer et al. (2016). This new method-ology yields log g values that are accurate to 0.05 dex, asdetermined from careful comparisons against stars with log g determined from asteroseismology (Brewer et al.2015). We find T eff = 5475 ± K, log g = 4 . ± . dex,and [ Fe/H ] = 0 . ± . dex. To estimate the stellarmass and radius, we fit our spectroscopic measurementsof T eff , log g , & [ Fe/H ] to a grid of models from the Dart-mouth Stellar Evolution Database (Dotter et al. 2008)using the isochrones Python package (Morton 2015)with uncertainties determined by the emcee
MarkovChain Monte Carlo (MCMC) package (Foreman-Mackeyet al. 2013). The derived stellar mass and radius are0.99 ± M (cid:12) and . ± . R (cid:12) . These are consis-tent with the measurements of 1.04 ± M (cid:12) and 1.15 ± R ⊕ by Mortier et al. (2013). Following Sinukoffet al. (2016), we conservatively adopt uncertainties of5% on stellar mass to account for the intrinsic uncer-tainties of the Dartmouth models estimated by Feiden &Chaboyer (2012).Following the prescription of Isaacson & Fischer(2010), we measure S HK indices from the HIRES spec-tra, which serve as a proxy for stellar activity. Our S HK measurements are listed in Table 3. The median S HK index of 0.132 is consistent with other inactive stars inthe California Planet Search (Isaacson & Fischer 2010).Consistent with this picture, we measure the stellar jitterto be . ± . m s − (Table 2). ANALYSIS
Radial Velocity Data Analysis
We analyzed the RV time-series using the RV fittingpackage
RadVel (Fulton & Petigura, in prep.), which ispublicly-available on GitHub . We fit our Keck-HIRESRVs along with previously published RV datasets (Hel-lier et al. 2012; Dai et al. 2015; Neveu-VanMalle et al.2016), summarized in Table 1. We omit the six RV mea-surements reported by Neveu-VanMalle et al. (2016) thatwere taken after a CORALIE instrument upgrade. Thesewould have added two free parameters to our RV model,which was not worth the negligible gain in RV measure-ments. After omitting the 17 HIRES observations JD =2457244.9366–2457245.07451, taken during a WASP-47btransit, we still have 12 out-of-transit observations fromthat night. RVs have astrophysical and instrumental er- https://github.com/California-Planet-Search/radvelhttp://radvel.readthedocs.io/en/master/index.html rors that manifest on a variety of timescales from min-utes to year. Therefore, the consecutive measurementsduring the same night don’t constitute independent mea-surements. To guard against these data from having adisproportionate influence influence on the fit, we bin the8 pre-transit RV measurements and bin the 4 post-transitmeasurements. We note that an analysis of our HIRESRVs alone gives the same planet masses to within 1 σ .We adopt a four-planet model that is the sum of fourKeplerian components. For each of the four datasets, ourmodel includes an RV offset, γ , as well as an RV “jitter”parameter, σ jit , to account for additional Doppler noiseof astrophysical or instrumental origins.Our likelihood function for this analysis follows that ofHoward et al. (2014): ln L = − (cid:88) i ( v i − v m ( t i )) (cid:16) σ i + σ (cid:17) + ln (cid:114) π (cid:16) σ i + σ (cid:17) (cid:35) , (1)where v i and σ i are the i th RV measurement andcorresponding uncertainty, and v m ( t i ) is the Keplerianmodel velocity at time t i . To increase the rate ofconvergence and to to counter the bias toward non-zero eccentricity (Lucy & Sweeney 1971), we adoptthe following parametrization of our model RV curve:{ P, T c , √ e cos ω, √ e sin ω, K }, where P is orbital period, T c is the time of conjunction, e is the orbital eccentric-ity, ω is the longitude of periastron and K is the RVsemi-amplitude.We first find the maximum-likelihood model using theminimization technique of Powell (1964), then perturbthe best-fitting free parameters by up to 3% to start100 parallel MCMC chains. The free parameters ofthe RV model are adopted as the MCMC step param-eters. RadVel incorporates the affine-invariant samplerof the emcee package (Foreman-Mackey et al. 2013). TheGelman-Rubin (Gelman & Rubin 1992) and T z statistics(Ford 2006) are checked in real-time during the MCMCexploration and the chains are deemed well-mixed andthe MCMC run is halted when the Gelman-Rubin iswithin 3% of unity and T z > for all free param-eters.We assume circular orbits for WASP-47d and WASP-47e while allowing the eccentricities of WASP-47b andWASP-47c to vary freely. An N-body dynamical stabil-ity analysis by Becker et al. (2015) showed that the orbitsof the inner three planets are unstable when eccentrici-ties of the three inner planets exceed ∼ . . For the ∼ − RV signals of WASP-47d and WASP-47e, oursignal-to-noise is too low to distinguish between eccen-tricities of 0.00 and 0.05. The orbital periods and orbitalphases of WASP-47b, d, and e were locked at the val-ues reported in Becker et al. (2015) from transits. Weadopt uninformed priors (i.e. no priors) on all free stepparameters and step in linear parameter space. The me-dian values and the 68% credible intervals are reportedin Table 2. The best-fitting RV model is shown in Figure1 We searched for additional companions at large or- Sinukoffbital distances by testing RV models with and without aconstant radial acceleration term, d v/ d t . We comparedthese two models using the Bayesian Information Cri-terion (BIC), with the RV jitter fixed at the values inTable 2. We compute ∆ BIC = BIC d v/ d t − BIC d v/ d t =0 =3.8, indicating that the simpler model is preferred, so weadopt d v/ d t = 0.We investigated whether the Keplerian orbit approxi-mation is valid for our RV model, given the dynamicalinfluences of the three inner planets on each other. First,we considered the TTV amplitudes, which indicate theorder of magnitude of non-Keplerian effects. The TTVamplitudes of planets b, d, and e measured by Beckeret al. (2015) of 0.63 min, 7.3 min, and < 1.2 min are0.01%, 0.06%, and < K and assum-ing a phase shift equal to the TTV amplitude ∆ T , thedeviation of RV( t ) is: ∆ RV( t ) = ∂ RV ∂t ∆ T = 2 πKP cos (cid:18) πtP (cid:19) ∆ T. (2)The maximum ∆ RV is πKP − ∆ T , which evaluates to0.09 m s − , 0.01 m s − , and < − for planets b,d, and e respectively. These represent upper bounds tothe orbit-averaged deviations from Keplerian over the K2time baseline. These deviations are much smaller thanour RV measurement uncertainties (1.5–2.0 m/s).Since the RV time-series is much longer than the K2baseline, one may wonder if there are large amplitude de-viations from Keplerian orbits that build up over longertimescales. To verify that the TTVs remain small overthe timescale of RV observations, we used the symplecticN-body integrator TTVFast (Deck et al. 2014) to numer-ically integrate the planet orbits over 2000 days. The or-bital elements were initialized at the maximum-likelihoodsolution obtained from RVs. The TTV amplitudes ofplanets b, d, and e remained at 0.6 min, 7 min and < P by fittinga linear ephemeris to the K2 transits. Since the K2 pho-tometry only spans one TTV super-period, the Beckeret al. (2015) orbital periods could be different from theaverage orbital periods over the time baseline of our RVmeasurements, which spans many TTV super-periods.To quantify the additional uncertainties of average or-bital periods, we used the 2000-day baseline of transittimes obtained with TTVFast. For each planet, we per-formed a linear fit to every unique set of N consecu-tive transit times, where N is the number of transitsobserved in the K2 photometry. The resulting distribu-tion of slopes (orbital periods) provides an estimate ofthe uncertainty of the average orbital period attributedto the limited timescale of K2 observations. The 1-sigmauncertainties obtained from these orbital period distribu-tions are ± ± ∼ × larger than the uncertainties reported by Becker et al. (2015). Werefit our RV time-series using these larger orbital perioduncertainties, but there was no change in the RV solu-tion or corresponding uncertainties. The scale of theseuncertainties is still a tiny fraction of the RV phase. Nev-ertheless we recommend that future studies adopt theselarger uncertainties on average orbital period, which arelisted in Table 2. DISCUSSION
Figure 2 shows the mass-radius distribution of all con-firmed planets with R p < R ⊕ whose mass and radiusare measured to better than 50% precision ( σ ) eitherby RVs or TTVs . Previous studies of small planetsfrom the prime Kepler mission revealed a transition inthe typical composition of planets from mostly rocky toplanets having thick envelopes of low density H/He at ≈ R ⊕ (Weiss & Marcy 2014; Marcy et al. 2014; Lopez& Fortney 2014; Rogers 2015; Dressing et al. 2015). Animportant open question is if and how this transition de-pends on incident stellar flux. Jontof-Hutter et al. (2016)illustrate that the population of planets < M ⊕ tend tohave fewer volatiles as incident flux increases, consistentwith atmospheric loss via photoevaporation. WASP-47eis among the most highly irradiated small planets with awell-measured mass, and thus helps us to probe the mass-radius relationship at extreme values of incident stellarflux, in a regime similar to Kepler-10b, Kepler-78b, and55 Cnc e.The measured mass of WASP-47e ( . ± . M ⊕ ) isconsistent with the measurement of Dai et al. (2015)( . ± . M ⊕ ) at the 1 σ level. We improve the fractionaluncertainty from 30% to 13%, allowing for a more de-tailed interpretation of composition. The measurementsof Dai et al. (2015) favored an admixture of 50% iron and50% rock. Assuming an iron-rock admixture, we sampleour planet mass and radius posterior distributions andcompute an iron mass fraction (IMF) using Equation 8of Fortney et al. (2007). From 100,000 independent sam-ples, we obtain a median IMF of 13% and a 1 σ upperlimit of 24%, suggesting that WASP-47e is mostly rock.Its IMF is lower than Earth’s IMF (33%) at 80% con-fidence. Alternatively, WASP-47e could have an IMFsimilar to Earth but possess a significant atmosphere ofa high mean molecular weight species, such as water orsulfur.The measured mass and radius of WASP-47d ( . ± . M ⊕ and . ± . R ⊕ ) are consistent with severalother planets, including Kepler-94b, Kepler-95b, Kepler-30b, KOI-142b, and GJ 3470b. With an incident flux S inc = ± S ⊕ , the atmosphere of WASP-47d might haveundergone significant photoevaporation. Nevertheless, itmust still have an atmosphere containing some amount ofH/He. There are a number of degenerate planet compo-sitions in this region of the mass-radius diagram with dif-ferent fractions of rock, iron, water, and H/He (Rogers &Seager 2010; Valencia et al. 2013). Possible compositionsinclude a small iron-rich or rocky core with an extendedH/He or steam envelope, or a water-world with a modestH/He envelope. Future transmission spectroscopy obser-vations would help to break these degeneracies. NASA Exoplanet Archive, UT 24 September 2016,http://exoplanetarchive.ipac.caltech.edu
ASP-47 5 R V [ m s - ] a) CORALIE H12CORALIE V16HIRESPFS2011 2012 2013 2014 2015 2016500 1000 1500 2000 2500BJD
TDB - 245483325025 R e s i d u a l s b) R V [ m s - ] c) P b = 4.16 days K b = 142.4 ± -1 e b = 0.0059 ± R V [ m s - ] d) P c = 596.1 ± K c = 32.6 ± -1 e c = 0.271 ± R V [ m s - ] e) P d = 0.79 days K d = 6.31 ± -1 e d = 0.00 0.4 0.2 0.0 0.2 0.4Phase201001020 R V [ m s - ] f) P e = 9.03 days K e = 4.0 ± -1 e e = 0.00 Fig. 1.—
Four-planet RV model of WASP-47, assuming circular orbits for WASP-47d and WASP-47e a) The RV time-series. Filled redcircles indicate Keck-HIRES data. Orange squares represent CORALIE data published by Hellier et al. (2012). Purple pentagons representCORALIE data published by Neveu-VanMalle et al. (2016). Green diamonds indicate PFS data published by Dai et al. (2015). The solidblue line corresponds to the most likely model. Note that the orbital parameters listed in Table 2 are the median values of the posteriordistributions. Error bars for each independent dataset include an RV jitter term listed in Table 2, which are added in quadrature to themeasurement uncertainties. b) Residuals to the maximum-likelihood fit. c-f)
The RV time-series phase folded at the orbital periods ofeach of the four planets after subtracting the other three planet signals.
WASP-47e is among the few known USP planets > R ⊕ . Lopez (2016) explains the dearth of larger USPplanets as a consequence of photoevaporation of H/Heenvelopes of larger planets that formed water-poor. Theone potential counter-example noted by Lopez (2016) isthe 1.9 R ⊕ USP planet 55 Cnc e. The most recent massand radius constraints suggest the presence of a water-rich envelope, 8 ±
3% of the planet’s mass.55 Cnc has remarkable similarities to WASP-47. Ithosts a USP super-Earth (55 Cnc e), a non-transitinggiant planet (55 Cnc b) at P =15 days, and three addi-tional non-transiting planets at P = 44, 262, and ∼ ± M ⊕ , 1.92 ± R ⊕ ) are consistent with WASP-47e ( . ± . M ⊕ , . ± . R ⊕ ). Therefore,both planets could have water-rich envelopes. More well-characterized USP planets ≈ R ⊕ are needed to deter-mine if they represent a distinct population of USP plan-ets spawning from unique formation and/or evolutionaryprocesses. In particular, as proposed by Huang et al.(2016), WASP-47b and 55 Cnc b might represent therare close-in extremes of in-situ formation hypothesizedto produce the ∼
50% of warm-Jupiters ( P = 10–200days) that have small companions at shorter orbital dis-tances. This highlights the limitations of classifying HJsand warm-Jupiters based on orbital period alone, with-out taking the more complete system architecture intoaccount.One clue about the formation history of 55 Cnc e isthe fact that it transits whereas the outer planet (P Sinukoff TABLE 2WASP-47 system parameters
Parameter Value Units Ref.
Stellar Parameters T eff ± K A log g . ± . dex A [ Fe/H ] 0 . ± . dex A v sin i . +0 . . km s − C M (cid:63) . ± . M (cid:12) A R (cid:63) . ± . R (cid:12) A Planet Parameters
WASP-47b P . ± . days A, B T conj . ± . BJD B R p /R (cid:63) . ± . — B a . ± . AU A S inc ± S ⊕ A R p . ± . R ⊕ A e . +0 . − . — A ω +183 − deg A K . ± . m s − A M p ± M ⊕ A ρ p . ± . g cm − AWASP-47c P . ± . days A T conj ± BJD A a . ± . AU A S inc . ± . S ⊕ A e . ± . — A ω ± deg A K . ± . m s − A M p ± M ⊕ sin i AWASP-47d (circular orbit assumed) P . ± . days A, B T conj . ± . BJD B R p /R (cid:63) . ± . — B a . ± . AU A S inc ± S ⊕ A R p . ± . R ⊕ A K . ± . m s − A M p . ± . M ⊕ A ρ p . ± . g cm − AWASP-47e (circular orbit assumed) P . ± . days B T conj . ± . BJD B R p /R (cid:63) . ± . — B a . ± . AU A S inc ± S ⊕ A R p . ± . R ⊕ A K . ± . m s − A M p . ± . M ⊕ A ρ p . ± . g cm − A Other γ HIRES . ± . m s − A γ PFS , D15 . ± . m s − A γ CORALIE , H12 − . ± . m s − A γ CORALIE , V16 − . ± . m s − A σ jit , HIRES . ± . m s − A σ jit , PFS , D15 . ± . m s − A σ jit , CORALIE , H12 . ± . m s − A σ jit , CORALIE , V16 . ± . m s − A Note . — S inc = Incident flux, T conj = Time of conjunction,A: This study, B: Becker et al. (2015), C: Sanchis-Ojeda et al.(2015). H12: Hellier et al. (2012), D15: Dai et al. (2015), V16:Neveu-VanMalle et al. (2016). Orbital periods of planets b and dare those from Becker et al. (2015), but with larger uncertainties(See §3). Fig. 2.—
Radii and masses of all confirmed planets whose massand radius are measured to better than 50% (2 σ ) precision (bluetriangles). Solar System planets are represented as black squares.Red circles indicate our measurements of WASP-47d and WASP-47e. Green curves show the expected planet mass-radius curves for100% iron, 100% rock (Mg SiO ), 100% water (ice), and Earth-like(67% rock, 33% iron) compositions according to models by Fortneyet al. (2007). ∼ < . at . (3 σ ) confidence. The very low eccentricity andthe alignment between the orbital axis and stellar spin(Sanchis-Ojeda et al. 2015) are consistent with disk mi-gration, in-situ formation, and the aforementioned secu-lar interaction scenario. In future, this eccentricity con-straint can be used to inform TTV models.WASP-47 has a high metallicity ( . ± . dex) whichhas been shown to be associated with HJ occurrenceand giant planet occurrence (e.g. Fischer & Valenti 2005;Buchhave et al. 2014). The Kepler sample of Earth-size planets were found around stars of widely varyingmetallicity (Buchhave et al. 2014). However, if USPsare associated with metal-rich stars, it suggests differentformation pathway than the bulk of known Earth-sizeplanets—one that may be more closely associated withHJs. Although it is beyond the scope of this study, acomparison between the metallicities of stars hosting HJsASP-47 7with those hosting USPs will provide a useful test of therelationship between the formation of USPs and HJs.We note that while this manuscript was under review,Almenara et al. (2016) reported mass and radius con-straints of the WASP-47 system using a photodynamicalmodel. They simultaneously fit the K2 photometry andthe RV measurements of Hellier et al. (2012), Dai et al.(2015), and Neveu-VanMalle et al. (2016). Their planetmass measurements are consistent with this study at the1- σ level. Future incorporation of our Keck-HIRES RVsinto a photodynamical analysis would further improveconstraints of the WASP-47 system.We thank the many observers who contributed to themeasurements reported here. We thank Geoff Marcy andTrevor David for helpful discussions. We thank TomGreene, Michael Werner, Michael Endl, and WilliamCochrane for participation in our NASA Key Project.We gratefully acknowledge the efforts and dedication ofthe Keck Observatory staff. This paper includes datacollected by the K2 mission. Funding for the K2 mis-sion is provided by the NASA Science Mission direc-torate. E. S. is supported by a post-graduate scholar-ship from the Natural Sciences and Engineering ResearchCouncil of Canada. E. A. P. acknowledges support byNASA through a Hubble Fellowship grant awarded bythe Space Telescope Science Institute, which is operatedby the Association of Universities for Research in As-tronomy, Inc., for NASA, under contract NAS 5-26555.B. J. F. was supported by the National Science Foun-dation Graduate Research Fellowship under grant No.2014184874. A. W. H. acknowledges support for our K2 team through a NASA Astrophysics Data Analysis Pro-gram grant. A. W. H. and I. J. M. C. acknowledge sup-port from the K2 Guest Observer Program. L. M. W.acknowledges the Trottier Family Foundation for theirgenerous support. This work was performed [in part] un-der contract with the Jet Propulsion Laboratory (JPL)funded by NASA through the Sagan Fellowship Programexecuted by the NASA Exoplanet Science Institute. Thisresearch has made use of the NASA Exoplanet Archive,which is operated by the California Institute of Technol-ogy, under contract with the National Aeronautics andSpace Administration under the Exoplanet ExplorationProgram. Finally, the authors extend special thanks tothose of Hawai‘ian ancestry on whose sacred mountainof Maunakea we are privileged to be guests. Withouttheir generous hospitality, the Keck observations pre-sented herein would not have been possible.
Facilities:
Kepler, Keck-HIRES.
TABLE 3Relative radial velocities, Keck-HIRES
BJD RV [m s − ] a Unc. [m s − ] b S HK a RVs do not include zero point offset ( γ HIRES , Table 2) b Uncertainties do not include jitter ( σ jit , HIRES , Table 2)