ebook img

WASP-36b: A new transiting planet around a metal-poor G-dwarf, and an investigation into analyses based on a single transit light curve PDF

1.4 MB·English
Save to my drive
Quick download
Download
Most books are stored in the elastic cloud where traffic is expensive. For this reason, we have a limit on daily download.

Preview WASP-36b: A new transiting planet around a metal-poor G-dwarf, and an investigation into analyses based on a single transit light curve

Draftversion January27,2012 PreprinttypesetusingLATEXstyleemulateapjv.5/2/11 WASP-36b: A NEW TRANSITING PLANET AROUND A METAL-POOR G-DWARF, AND AN INVESTIGATION INTO ANALYSES BASED ON A SINGLE TRANSIT LIGHT CURVE A. M. S. Smith1, D. R. Anderson1, A. Collier Cameron2, M. Gillon3, C. Hellier1, M. Lendl4, P. F. L. Maxted1, D. Queloz4, B. Smalley1, A. H. M. J. Triaud4, R. G. West5, S. C. C. Barros6, E. Jehin3, F. Pepe4, D. Pollacco6, D. Segransan4, J. Southworth1, R. A. Street7, and S. Udry4 Draft version January 27, 2012 ABSTRACT 2 1 We report the discovery, from WASP and CORALIE, of a transiting exoplanet in a 1.54-d orbit. 0 The host star, WASP-36, is a magnitude V =12.7, metal-poor G2 dwarf (Teff =5959 134 K), with ± 2 [Fe/H]= 0.26 0.10. We determine the planettohavemassandradiusrespectively2.30 0.07and − ± ± 1.28 0.03 times that of Jupiter. n ± We have eight partial or complete transit light curves, from four different observatories, which a J allows us to investigate the potential effects on the fitted system parameters of using only a single lightcurve. We find thatthe solutionsobtainedby analysingeachofthese lightcurvesindependently 6 are consistent with our global fit to all the data, despite the apparent presence of correlated noise in 2 at least two of the light curves. ] Subject headings: planetary systems – planets and satellites: detection – planets and satellites: fun- P damental parameters – stars: individual (WASP-36) – techniques: photometric E . h 1. INTRODUCTION WASP-36 was observed in 2009 and 2010 by WASP- p - Of the 171 confirmed transiting planetary systems8, South, which is located at the South African Astro- o themajorityhavebeendiscoveredfromtheground,from nomicalObservatory(SAAO), near Sutherland in South r surveys such as WASP (Pollacco et al. 2006) and HAT- Africa, and by SuperWASP at the Observatorio del t Roque de los Muchachos on La Palma, Spain. Each in- s net (Bakos et al. 2004). Although the Kepler space mis- a strument consists of eight Canon 200mm f/1.8 lenses, sion is discovering an increasing number of planets and [ each equipped with an Andor 2048 2048 e2v CCD even more candidate planets (e.g. Borucki et al. 2010; × camera, on a single robotic mount. Further details 2 Borucki et al. 2011), the ground-based discoveries have of the instrument, survey and data reduction proce- v the advantage that the host stars are generally brighter. dures are described in Pollacco et al. (2006) and details 3 Thisallowsradial-velocitymeasurementstomeasurethe of the candidate selection procedure can be found in 1 planetarymass,andisconducivetofurthercharacterisa- Collier Cameron et al.(2007) andPollacco et al.(2008). 3 tion observations, such as measuring occultations in the 5 infrared to probe atmospheric temperature and struc- A total of 13781 measurements of WASP-36 were made . ture. between 2009 January 14 and 2010 April 21. 0 WASP-South 2009 data revealed the presence of a 1 Manyofthe currentquestionsin exoplanetscienceare transit-likesignalwithaperiodof 1.5daysandadepth 1 being addressed by analysing the statistical properties ∼ of 15 mmag. The WASP light curve is shown folded 1 of the growing ensemble of well characterised transiting ∼ on the best-fitting orbital period in Figure 1. : planetary systems. Here we report the discovery of a v transitingplanetorbitingtheV 12.7starWASP-36(= Xi 2MASS J08461929-0801370)in t∼he constellation Hydra. 2.2. Spectroscopy Spectroscopic observations of WASP-36 were made ar 2. OBSERVATIONS with the CORALIE spectrograph of the 1.2-m Euler- 2.1. WASP photometry Swiss telescope. Simultaneous spectra of a thorium- argon emission line lamp were obtained in order to cal- ibrate the stellar spectra. A total of nineteen spectra [email protected] were taken between 2010 March 11 and 2011 January 1Astrophysics Group, Keele University, Staffordshire, ST5 5BG,UK 11, and processed using the standard CORALIE data 2SUPA,SchoolofPhysics&Astronomy,UniversityofStAn- reduction pipeline (Baranne et al. 1996). The resulting drews,NorthHaugh,Fife,KY169SS,UK radial-velocity data are given in Table 1, and plotted in 3Institut d’Astrophysique et de G´eophysique, Universit´e de Figure 2. In order to rule out non-planetary causes for Li`ege,All´eedu6Aouˆt,17,Bˆat. B5C,Li`ege1,Belgium 4Observatoire de Gen`eve, Universit´e de Gen`eve, 51 Chemin the radial-velocity variation, such as a blended eclips- desMaillettes,1290Sauverny, Switzerland ing binary system, we examined the bisector spans (e.g. 5Department of Physics & Astronomy, University of Leices- Queloz et al.2001),whichexhibitnocorrelationwithra- ter,Leicester,LE17RH,UK 6Astrophysics Research Centre, School of Mathematics & dial velocity (Figure. 2), as expected. Physics,Queen’sUniversity,UniversityRoad,Belfast,BT71NN, UK 2.3. Follow-up photometry 7Las Cumbres Observatory, 6740 Cortona Drive Suite 102, Goleta, CA93117, USA We have a total of eight high-precision follow-up light 8 http://www.exoplanet.eu, 2011October13 curvesofthe transitofWASP-36b,summarisedinTable 2 A. M. S. Smith et al. is approximately the quadratic sum of vsini and v TABLE 1 mac Radial-velocity (RV) andline bisector span(BS) ( 4.9 0.8 kms−1 ). ≈ ± measurementsof WASP-36 3.2. Neighbouring objects BJD(UTC) RV σRV BS –2450000 (kms−1) (kms−1) (kms−1) The Two Micron All Sky Survey catalogue 5266.6926 −12.843 0.027 0.079 (Skrutskie et al. 2006) reveals the presence of four 5293.5807 −13.591 0.021 −0.017 fainter stars close on the sky to WASP-36. There is 5304.6409 −13.212 0.057 0.025 no evidence from analysis of catalogue proper motions 5305.5493 −13.423 0.025 −0.046 that any of these stars are physically associated with 5306.6464 −12.847 0.025 0.014 5315.5600 −12.981 0.023 0.013 WASP-36. The stars are separated from WASP-36 by 5316.5339 −13.624 0.028 −0.005 4′′, 9′′, 13′′and 17′′, meaning that they fall well within 5317.5642 −12.977 0.033 −0.046 the WASP photometric aperture, which has a radius of 5320.4833 −12.821 0.030 −0.034 48′′(3.5 pixels), but outside of the 1′′ CORALIE fibre. 5359.4568 −13.527 0.040 0.002 Intheabsenceofreliableopticalcataloguemagnitudes 5547.8347 −12.842 0.024 0.007 forallofthese objects,itwasnecessaryto measuretheir 5561.8393 −12.866 0.026 0.003 fluxes to quantify the effects of blending in the photom- 5562.8651 −13.306 0.025 −0.018 etry. The fluxes were measured from images taken dur- 5563.8184 −13.402 0.033 −0.070 ingthe twotransitsobservedwiththe 1.2-mEuler-Swiss 5564.7341 −12.920 0.024 −0.011 Telescope (see Table 2). The fluxes relative to that of 5565.8211 −13.467 0.029 0.056 WASP-36 are as follows: 0.012 (object at 4′′ separation 55557607..78908848 −−1122..983710 00..002233 −00..002081 fromWASP-36),0.00771(9′′),0.00558(13′′)and0.00827 5572.7427 −12.944 0.023 −0.068 (17′′). Using these flux ratios, we corrected the WASP photometry to account for all four objects, and the high precision photometry to account for the object at 4′′, 2. Differential aperture photometry was performed us- which is within the photometric apertures used. The ing the IRAF/DAOPHOT package for TRAPPIST and magnitude of this correctionis minimal, and had no sig- FTNdata,andthe ULTRACAMpipeline(Dhillon et al. nificant ( 1-σ) effect on the values of our best-fitting 2007; Barros et al. 2011) for the LT data, with aperture ≪ system parameters. radii and choice of comparison stars optimised to give the lowest RMS of the out-of-transit photometry. 3.3. Planetary system parameters 3. DETERMINATIONOFSYSTEMPARAMETERS CORALIEradial-velocitydatawerecombinedwithall 3.1. Stellar parameters our photometry and analysed simultaneously using the MarkovChainMonteCarlo(MCMC)method. Thebest- The individual CORALIE spectra of WASP-36 were fitting system parameters are taken to be the median co-added to produce a single spectrum with a typ- values of the posterior probability distribution. Linear ical S/N of around 60:1. The standard CORALIE functions of time were fitted to each light curve at each pipeline reduction products were used in the analy- step of the MCMC, to remove systematic trends. We sis. The spectral analysis was performed using uclsyn use the current version of the MCMC code described (Smith & Dworetsky1988;Smith1992)andthemethods in Collier Cameron et al. (2007); Pollacco et al. (2008) given in Gillon et al. (2009). The parameters obtained and Enoch et al. (2010). The MCMC proposal parame- from the analysis are listed in Table 3. The elemen- ters we use are: the epoch of mid-transit, T ; the orbital tal abundances were determined from equivalent width c period, P; the transit duration, T ; the fractional flux measurements of severalclean and unblended lines. The 14 deficit that would be observed during transit in the ab- lines used are those listed in Gonzalez & Laws (2000), sence of stellar limb-darkening, ∆F; the transit impact Gonzalez et al. (2001), and Santos et al. (2004a). A parameter, b; the stellar reflex velocity semi-amplitude, value for microturbulence, ξ , was determined from Fe i t K ; the stellar effective temperature, T ; the stellar using the method of Magain (1984). The quoted error 1 eff metallicity,[Fe/H];and√ecosω and√esinω,whereeis estimates account for the uncertainties in T , logg and eff the orbital eccentricity, and ω is the argument of perias- ξ , as well as for the scatter due to measurement and t tron(Anderson et al.2011). The stellar masswasdeter- atomic data uncertainties. minedaspartoftheMCMCanalysis,usinganempirical The projected stellar rotation velocity, vsini, was de- termined by fitting the profiles of severalunblended Fe i fitto[Fe/H],Teff,andthestellardensity,ρ∗(Enoch et al. 2010; Torres et al. 2010). lines in the wavelength range 6000–6200˚A,using the ro- Thetransitlightcurvesweremodelledusingtheformu- tation broadening function of (Gray 2008, Ch. 18). We lation of Mandel & Agol (2002) and limb-darkening was used an instrumental FWHM of 0.11 0.01 ˚A, deter- accounted for using a four-coefficient, non-linear model, ± mined from the telluric lines around 630 nm. The mea- employing coefficients appropriate to the passband from suredvsiniissensitivetotheadoptedvalueofv . The mac the tabulations of Claret (2000, 2004). The coefficients appropriatevalueofvmacis4.0kms−1accordingto(Gray were determined using an initial interpolation in logg∗ 2008, p. 507), but 2.9 kms−1 according to Bruntt et al. and [Fe/H] (values from Table 3), and an interpolation a(2n0d10v)s.inTih=es3e.7valu1e.s1ikmmplsy−1vsrienspie=cti2v.e9ly.±W1e.3takkmest−he1 irnesTp∗o,neffdiantgetaochthMeCbMesCt-fisttetpin.gTvhaeluceoeoffifcTie∗n,etffvaalrueesgicvoern- ± weighted average of these two values as the best-fitting in Table 3.3. Because some of our photometry was ob- one, vsini = 3.3 1.2 kms−1 . The quantity measured servedinpassbandsnottabulatedbyClaret(2000,2004), ± WASP-36 b 3 TABLE 2 Observing log forfollow-up transitphotometry Light Date/UT Telescope/instrument Band Nobs texp fullor Seeingor Aperture Airmass curve /s partial defocus(′′) radius(′′) range (i) 2010Dec13 Eulera /EulerCam Gunnr 94 120 partial 1.1–2.2 4.3 1.54–1.07–1.09 (ii) 2010Dec13 TRAPPISTb /TRAPPISTCAM clear 756 10 partial 3 8.3 1.84–1.11 (iii) 2010Dec25 FTNc /Spectral camera PSz 176 60 full 4.3 2.4 1.52–1.14–1.22 (iv) 2011Jan02 TRAPPIST/TRAPPISTCAM I+z 296 25 full 2 6.4 1.75–1.09 (v) 2011Jan05 TRAPPIST/TRAPPISTCAM clear 179 18 partial 3 7.7 1.18–1.07–1.14 (vi) 2011Jan08 TRAPPIST/TRAPPISTCAM clear 269 18 partial 3.5 9.0 1.10–1.44 (vii) 2011Jan15/16 LTd /RISEe V +R 1290 9 full 6 9.0 1.95–1.25 (viii) 2011Jan21 Euler/EulerCam Gunnr 167 60 full 0.45–1.0 4.1 1.46–1.20 a1.2-mEuler-SwissTelescope,LaSilla,Chile bTransitingPlanetsandPlanetesimalsSmallTelescope, LaSilla,Chile(Jehinetal.2011;http://www.astro.ulg.ac.be/Sci/Trappist) cFaulkes TelescopeNorth,HaleakalaObservatory, Maui,Hawaii,USA dLiverpoolTelescope, ObservatoriodelRoquedelosMuchachos, LaPalma,Spain eRapidImagingSearchforExoplanets camera(Steeleetal.2008;Gibsonetal.2008) TABLE 3 TABLE 4 Stellar parametersandabundancesfromanalysisof Limb-darkeningcoefficients CORALIE spectra. Claretband Lightcurves a1 a2 a3 a4 Parameter Value Parameter Value CousinsR WASP,i,ii,v,vi,viii 0.466 0.294 0.070 -0.128 RA(J2000.0) 08h46m19.30s [Fe/H] −0.26±0.10 Sloanz′ iii,iv 0.555 -0.099 0.348 -0.213 Dec(J2000.0) −08◦ 01′ 36.7′′ [Na/H] −0.33±0.08 JohnsonV vii 0.389 0.523 -0.066 -0.082 Teff 5900±150K [Mg/H] −0.08±0.08 logg(cgs) 4.5±0.15 [Si/H] −0.17±0.06 body in the system. The best-fitting radial acceleration ξta 1.0±0.2kms−1 [Ca/H] −0.15±0.11 is consistent with zero, indicating there is no evidence vsini 3.3±1.2kms−1 [Sc/H] −0.11±0.12 for an additional body in the system based on our RVs, logA(Li)b 1.69±0.13 [Ti/H] −0.16±0.11 whichspan10months. Theorbitalparameterswereport Sp. Type G2 [V/H] −0.20±0.15 Distance 450±120pc [Cr/H] −0.28±0.09 are the result of a fit which does not allow for a linear Age 1–5Gy [Mn/H] −0.44±0.10 trend in radial velocity. Mass 1.01±0.08 M⊙ [Co/H] −0.19±0.12 The system parameters derived from our best-fitting Radius 0.94±0.17 R⊙ [Ni/H] −0.30±0.08 circularmodelarepresentedinTable5. Thecorrespond- AdditionalidentifiersforWASP-36: ingtransitandRVmodelsaresuperimposedonourdata USNO-B1.00819-0221838 in Figures 1 and 2. 2MASSJ08461929-0801370 1SWASP J084619.30-080136.7 3.4. System age The measured vsini of WASP-36 gives an upper limit Note: Thespectral type was estimated from Teff using the table totherotationalperiod,P 14.4 5.9days. Thiscor- othfaeM(GTicroraroyrteu2sr0eb0tu8l,aelpn..t(25v00e17lo0).c)itTcyahliebmraatsiosna.ndradiuswereestimatedusing ruessinpgontdhsetogyarnocuhprponerolloimgicirtaotlo≃nretlahteioan±geofofB∼arn1e.s8+−(212..0730G7)y, blogA(Li)=log(cid:16)NNLHi(cid:17)+12,whereNLi andNH arethenumber and a B-magnitude of 13.3 derived from V=12.7 and densitiesofLiandHrespectively. B V = 0.60 0.04 (estimated using Gray 2008, p. − ± 507). we also tried using coefficients corresponding to nearby InFigure3weplotWASP-36alongsidethestellarevo- passbands. None of our best-fitting system parameters lution tracks of Marigo et al. (2008). From this we infer was significantly affected by our choice of Claret pass- an age of 2.5+3.5 Gy. The age determined from the −2.2 band; values changedby muchless than their 1-σ uncer- lithium abundance of WASP-36 is poorly constrained, tainties. but the work of Sestito & Randich (2005) suggests that An initial MCMC fit for an eccentric orbit found e = the most likely age is 2 to 5 Gy. 0.012+0.014 (ω =55+55 degrees),with a 3-σ upper-limit We searched the WA∼SP photometry for periodic vari- −0.008 −118 to the eccentricity of 0.064, but we found this eccentric- ations indicative of star spots and stellar rotation, but ity is not significant. Following the F-test approach of no significant variation was detected. We place an up- Lucy & Sweeney (1971), we find that there is a 58 per per limit of1.5mmag atthe 95per centconfidence level centprobabilitythattheapparenteccentricitycouldhave on the amplitude of any sinusoidal variation. This null arisen if the underlying orbit were actually circular. We result is consistent with the low levels of stellar activity therefore present here the model with a circular orbit, expected from a main-sequence G2 star. A lack of stel- noting that the values of the other model parameters, lar activity is also indicated by the absence of calcium and their associated uncertainties, are almost identical II H+K emission in the spectra. The uncertainties on to those of the eccentric solution. theCaIIemissionindex,logR′ ( 4.5)aretoolargeto HK ≈ We tried fitting for a linear trend in the RVs with the allow meaningful constraints to be placed on the system inclusion of an additional parameter in our MCMC fit. age by using an activity - rotation relation such as that Such a trend (such as that found in the RVs of WASP- of Mamajek & Hillenbrand (2008). 34, Smalley et al. 2011) would be indicative of a third There is no evidence of any discrepancy between the 4 A. M. S. Smith et al. TABLE 5 Systemparameters Parameter Symbol Unit Value Orbitalperiod P d 1.5373653±0.0000026 Epochofmid-transit Tc HJD,UTC 2455569.83731±0.000095 Transitduration T14 d 0.07566±0.00042 Ingress/egressduration T12=T34 d 0.01540±0.00054 Planet-to-stararearatio ∆F =R2P/R2∗ - 0.01916±0.00020 Transitimpactparameter b - 0.665±0.013 Orbitalinclinationangle i ◦ 83.61±0.21 Stellarorbitalvelocitysemi-amplitude K∗ kms−1 0.3915±0.0083 Systemvelocity γ kms−1 −13.2169±0.0024 Orbitaleccentricity(adopted) e - 0 Orbitaleccentricity(3-σ upper-limit) - 0.0663 Stellarmass M∗ M⊙ 1.040±0.031 Stellarradius R∗ R⊙ 0.951±0.018 log(stellarsurfacegravity) logg∗ (cgs) 4.499±0.012 Stellardensity ρ∗ ρ⊙ 1.211±0.050 Stellareffective temperature T∗,eff K 5959±134 Metallicity [Fe/H] dex −0.26±0.10 Planetmass MP MJup 2.303±0.068 Planetradius RP RJup 1.281±0.029 Planetdensity ρP ρJ 1.096±0.067 log(planetsurfacegravity) loggP (cgs) 3.507±0.018 Scaledorbitalmajorsemi-axis a/R∗ - 5.977±0.082 Orbitalmajorsemi-axis a AU 0.02643±0.00026 Planetequilibriumtemperature(uniformheatredistribution) TP,A=0 K 1724±43 Systemage(fromfigure3) Gy 2.5+3.5 −2.2 . Thefollowingconstantvaluesareused: AU=1.49598×1011 m;R⊙=6.9599×108 m;M⊙=1.9892×1030 kg;RJup=7.1492×107 m; MJup=1.89896×1027 kg;ρJ=1240.67kgm−3 examine in detail the potential effects on the system pa- TABLE 6 Transit times rameters of using only a single light curve. ForsurveyphotometrywithlowSNR,thedurationsof Lightcurve E TC σTC O−C ingressandegressareill-defined, leading to considerable (HJD,UTC) (min) (min) uncertainty in the transit impact parameter and hence (i) −17 2455543.70602 5.72 5.65 (ii) −17 2455543.70378 1.18 2.42 tolargeuncertaintiesinthestellardensityandplanetary (iii) −9 2455556.00221 0.62 1.71 radius. So-called ‘follow-up’transit light curves are gen- (iv) −4 2455563.68807 0.33 0.32 erallyincludedintheanalysisofnewground-basedtran- (v) −2 2455566.76718 7.73 6.63 siting planet discoveries, and are of significantly higher (vi) 0 2455569.83686 0.86 −0.62 photometric precision than the light curves produced by (vii) 5 2455577.52412 0.21 −0.02 survey instruments such as WASP. Such follow-up light (viii) 9 2455583.67344 0.27 −0.23 curvesaretypicallytheresultofobservationswitha1–2- m classtelescope, and are ofcritical importance to mea- ages derived from lithium abundance, gyrochronology suring precisely basic system parameters. and isochrone fitting. This suggests that the star has Any light curve may suffer from correlatednoise, such undergonelittle ornotidalspin-up,despite the presence as from observational systematics or from astrophysical of a massive planet in a close orbit. sources such as stellar activity. To assess the levels of correlated noise in our follow-up light curves, we plot 3.5. Transit timing (Figure 4) the RMS of the binned residuals to the fit of We measured the times of mid-transit for each of the eachlightcurveasafunctionofbinwidth,alongwiththe eightfollow-uplightcurves,byanalysingeachlightcurve white-noise expectation. For six of our light curves, the separately, without any other photometry (see Section RMS of the binned residuals follows closely the white- 4.1). The times are displayed in Table 3.3, along with noise expectation, indicating that little or no correlated thedifferences,O C betweenthesetimesandthosepre- noise is present in the data. Light curves ii and vii show − dicted assuming a fixed epoch and period (Table 5). No deviation from the white-noise model, however,suggest- significantdeparture froma fixed ephemeris is observed. ing the presence of noise correlated on timescales of 1 ∼ and 10minutes,respectively. Wesuggestthatthisred 4. DETAILEDANALAYSISOFFOLLOW-UPLIGHT noise∼maybeduetothehighairmassofthetargetatthe CURVES start of these observations. Because we have several follow-up light curves of WASP-36 from different telescopes / instruments, whereas many planet discovery papers rely on only a 4.1. Method single such light curve, we take the opportunity here to WASP-36 b 5 -0.2 400 -0.1 mag 0.0 -1m s 300 ∆ 0.1 y / 200 cit 100 o 0.2 el 0.5 0.6 0.7 0.8 0.9 1.0 1.1 1.2 1.3 1.4 1.5 al v 0 (i) adi -100 1.00 e r ativ -200 Rel -300 0.98 (ii) -400 0.5 0.75 1 1.25 1.5 0.96 Orbital phase (iii) 60 0.94 -1s 30 m 0 C / -30 0.92 (iv) O- -60 5300 5400 5500 0.90 BJD(UTC) - 2450000 (v) 200 0.88 100 0.86 -1m s (vi) p. / 0 S 0.84 s. x Bi -100 u e fl (vii) Relativ 0.82 -200 -400 -300 -200 -100 0 100 200 300 400 Relative radial velocity / ms-1 0.80 (viii) Fig.2.— Radial velocities. Upper panel: Phase-folded radial- velocitymeasurements (Table1. Thecentre-of-mass velocity, γ = 0.78 -13.2169 km s−1, has been subtracted. The best-fitting MCMC solution is over-plotted as a solid line. Middle panel: Residuals from the radial-velocity fit as a function of time. Lower panel: 0.76 (i) Bisector span measurements as a function of radial velocity. The uncertaintiesinthebisectorsaretakentobetwicetheuncertainty intheradialvelocities. 0.74 (ii) (iii) 0.72 1.05 (iv) 12 Gy 0.70 1 (v) 3 0.68 (vi) -1/ρ) O• 0.95 (vii) ρ/* 0.66 ( (viii) 0.9 0.3 Gy 0.64 0.95 1.00 1.05 1.10 0.85 Orbital phase 6500 6000 5500 5000 T / K Fig. 1.—Photometry. Upper panel: WASP-36bdiscoverylight *,eff curvefoldedontheorbitalperiodofP =1.5373653d. Fordisplay purposes,pointswithanerrorgreaterthanthreetimesthemedian Fig.3.— Modified Hertzsprung-Russell diagram. WASP-36 uncertainty are not shown. Lower panel: High-precision transit is plotted alongside isochrones from the evolutionary models of photometry, over-plotted withour best-fitting model (solidlines). Marigoetal. (2008). The isochrones are for Z =0.01044 ([Fe/H] Each individualdataset is offset influxforclarity, andis labelled =−0.26). Theleftmostcurvecorrespondstoanageof0.3Gy,and withanumeralcorrespondingtothat inthefirstcolumnofTable the adjacent curve to 1.0 Gy. Thereafter the curves are regularly 2. The residuals to the best-fitting models arelabelled and offset spacedatintervalsof1Gy. inthesameway,andareshowninthelowerpartofthepanel. 6 A. M. S. Smith et al. 0.01 qualityfollow-upphotometry(suchasasinglelightcurve (i) which covers only part of the transit), and there is no evidence that the star is evolvedor otherwise non-main- 0.001 sequence in nature. We also performed analyses where theonlyphotometryincludedwasasinglefollow-uplight 0.0001 curve,i.e. theWASPphotometrywasexcludedfromthe 0.01 (ii) analysis. The purpose of this is to determine whether the measured depth of transit is biased by inclusion of 0.001 the WASP photometry. For these runs only, the orbital period was fixed to the value determined as part of our 0.0001 global solution, since this parameter is very poorly con- 0.01 strained by a single transit light curve and a few RVs. (iii) The epoch of mid-transit was treated as normal, and al- lowed to float freely. Finally, we performed an analysis 0.001 exlcudingallfollow-upphotometry;theonlyphotometry s al analysed was the WASP data. u 0.0001 sid 0.01 4.2. Results e (iv) d r We producedcorrelationplots betweenseveralparam- ne 0.001 eters, but choose to present here only plots showing im- n bi pact parameter against planet radius and stellar radius of 0.0001 versus stellar mass (Figures 5 and 6, respectively). Such MS 0.01 (v) plots, whilst representative of the ensemble correlation R plots,areparticularlyinstructivesinceb andRP aretwo of the major quantities we wish to measure, are largely 0.001 constrainedbyfollow-uplightcurvesratherthanbysur- veyphotometryorbyradialvelocities,andcanbesignif- 0.0001 icantly correlated with each other, indicating a strongly 0.01 (vi) degenerate solution. The stellar density is measured di- rectly from the transit light curve, and the stellar mass 0.001 and radius, whilst interesting in themselves, are key in determining the values of several other system parame- 0.0001 ters of interest. 0.01 SeveralconclusionscanbedrawnfromstudyofFigures (vii) 5 and 6, and similar plots, namely: (1)eachanalysisincludingonlyasinglefollow-uplight 0.001 curvegivesresultsthatareconsistentwithourglobalso- lution, albeit with larger uncertainties. To measure the 0.0001 dispersion in the best-fitting parameter values obtained 00..0011 ((vviiiiii)) from each single follow-up light curve analysis, we cal- culated the weighted standard deviation. The standard 00..000011 deviationsof b, RP, R∗ andM∗ are 0.05,0.08RJup, 0.05 R⊙ and 0.006 M⊙, respectively. 00..00000011 (2)Thelargestuncertaintiesareobtainedforfollow-up 1 10 100 lightcurvesthatcoverthesmallestfractionofthetransit (light curves i and v), as expected. Bin width / minutes (3)TheanalyseswhichexcludetheWASPphotometry give larger uncertainties, but these are only significantly so when the follow-up photometry is poor. This indi- cates that the WASP photometry only makes a signifi- Fig.4.— Correlated noise in follow-up light curves – RMS of cantcontributiontoconstrainingtheshapeanddepthof binned residuals versus bin width, for light curves i - viii (solid the transit when the follow-up light curve is incomplete. lines). The white-noise expectation, where the RMS decreases in proportion to the square root of the bin size, is indicated by the (4)Evenapartialtransitlightcurveimprovesthe pre- straight,dottedlines. cision of the measured system parameters enormously compared to those derived solely from the WASP pho- After modelling all available data in a combined tometry and the RVs. MCMC analysis (see Section 3.3), our ‘global solution’, (5)The impositionofa main-sequenceconstraintdoes wealsoranseveralMCMCseachwithjustasinglefollow- notsignificantlyaltertheparametersoruncertaintiesfor up light curve in addition to the radial velocities and high-precision light curves that are complete, thus indi- WASPphotometry. Additionally,were-raneachofthese catingthatWASP-36isamain-sequencestar. Whenthe MCMCs applying a Gaussian prior to the stellar radius follow-uplightcurvedoesnotwellconstraintherangeof to impose a density typical of a main-sequence star (the possible models, however, limiting the star to the main- ‘main-sequence constraint’). Such a constraint is usu- sequencecansignificantlyreducethelargedegeneracyin allyappliedwhenanalysinganewplanetwhichhaspoor the possible solutions. This is best illustrated by light WASP-36 b 7 Fig.5.—Analysisoffollow-uplightcurvesI.TheMCMCposteriorprobabilitydistributionsforRP andbforeachofthefollow-uplight curves. The numbering of each panel corresponds to the light curve numbering in Table 2 and the 1-σ and 2-σ contours are shown. In eachcaseredcorresponds toanalysisofasinglefollow-uplightcurveplusthe WASP photometry, blacktothe singlelightcurveplusthe WASP photometry withthe main-sequence constraint imposed, and blueto that of a singlelightcurve with no WASP photometry. The greencontours indicateourglobalsolution,andthegreycontours thesolutionwithoutfollow-upphotometry ,andarethereforeidentical ineachpanel. 8 A. M. S. Smith et al. Fig. 6.— Analysis of follow-uplight curves II. The MCMC posterior probability distributions for M∗ and R∗ for each of the follow-up lightcurves. ThenumberingofeachpanelcorrespondstothelightcurvenumberinginTable2andthe1-σand2-σcontoursareshown. In eachcaseredcorresponds toanalysisofasinglefollow-uplightcurveplusthe WASP photometry, blacktothe singlelightcurveplusthe WASP photometry withthe main-sequence constraint imposed, and blueto that of a singlelightcurve with no WASP photometry. The green contours indicate our global solution, and the greycontours the solution without follow-upphotometry, and are therefore identical in each panel. Also in each panel are dashed lines which are contours of constant stellar density, corresponding, from top to bottom, to 0.7,1.0,1.5,and3.0timessolardensity. WASP-36 b 9 curve (iv), where the effects of the constraint are to de- ([Fe/H] = 0.40 0.12, Simpson et al. 2011) and HAT- − ± creasethestellardensitywefindandconfinethesolution P-12 ([Fe/H] = 0.29 0.05, Hartman et al. 2009). − ± to a smaller area of parameter-space,close to the global Such systems will be critical in probing our under- solution, while largely resolving the degeneracy between standing of the planet–metallicity correlation; proposed b and R . explanationsforthecorrelationincludeinsufficientmate- P In summary, if only one of the follow-up light curves rial for proto-planetary cores to attain the critical mass had been available, we would have reached a solution needed for runaway accretion, and the suggestion that compatiblewiththecurrentbest-fittingmodel,although thehighdensityofmolecularhydrogenintheinnergalac- the uncertainties on the model parameters may have tic disk is responsible for the effect (Haywood 2009). been much greater, if the light curve was not of the WASP-36b may also play a key role in determining highest precision. Obtaining additional light curves is whetherstellarmetallicityisthekeyparameterinfluenc- clearly of benefit if one only has a light curve that par- ing whether or not a hot Jupiter’s atmosphere exhibits tially covers transit. It is also useful to have multiple a thermal inversion. Insolation was initially propounded high-precision light curves for systems where stellar ac- asthisparameter(Fortney et al.2008),butseveralplan- tivity may bias the observed transit depth by varying ets now appear to contradict this theory. XO-1b, for amounts at different epochs, as may be the case for instance, has a relatively low insolation, and was there- WASP-10b (Christian et al. 2009; Johnson et al. 2009; fore predicted to lack an inversion, but Machalek et al. Dittmann et al. 2010; Maciejewski et al. 2011b,a). (2008) report the presence of an inversion; TrES-3b does not exhibit an inversion (Fressin et al. 2010) de- 5. DISCUSSIONANDCONCLUSION spite a prediction to the contrary. More recently stel- WASP-36 is a metal-poor, Solar-mass star which is lar activity (Knutson et al. 2010) and stellar metallicity host to a transiting planet in a 1.54 d orbit. We find (Wheatley et al. 2011) have been advanced as alterna- the planet to have a mass of 2.30 M , and a radius tives to insolation; work aiming to resolve this issue is Jup 1.28 R , meaning it is slightly denser than Jupiter. ongoing. Jup There is an observed correlation between planetary ra- dius and insolation (e.g. Enoch et al. 2011), with the 6. ACKNOWLEDGMENTS more bloated planets generally receiving a greater flux WASP-SouthishostedbytheSAAOandSuperWASP fromtheirstar. WASP-36bissomewhatlargerthanpre- by the Isaac Newton Group and the Instituto de As- dictedbythemodelsofBodenheimer et al.(2003),which trof´ısicadeCanarias;wegratefullyacknowledgetheiron- predict radii between 1.08 R (for a planet with a core going support and assistance. Funding for WASP comes Jup at 1500 K) and 1.20 R (for a core-less planet at 2000 from consortium universities and from the UK’s Science Jup K). and Technology Facilities Council (STFC). TRAPPIST The close orbit and large radius of the planet make isaprojectfundedbytheBelgianFundforScientificRe- it a good target for measuring the planetary thermal search (FNRS) with the participation of the Swiss Na- emission, via infra-red secondary eclipse (occultation) tional Science Fundation. MG and EJ are FNRS Re- measurements with, for example, Spitzer. The expected search Associates. The RISE instrument mounted in signal-to-noiseratios of the occultations in Spitzer chan- the Liverpool Telescope was designed and built with re- nels 1 (3.6 µm) and 2 (4.5 µm) are around 10 and 9 sources made available from Queen’s University Belfast, respectively,assumingaplanetwithzeroalbedoanduni- Liverpool John Moores University and the University of form heat redistribution. Manchester. The LiverpoolTelescope is operatedon the One of the striking properties of the WASP-36 is the islandofLaPalmabyLiverpoolJohnMooresUniversity low stellar metallicity ([Fe/H] = 0.26 0.10). Giant intheSpanishObservatoriodelRoquedelosMuchachos − ± planetsareknowntoberarearoundsuchlow-metallicity ofthe Instituto de Astrof´ısicade Canariaswithfinancial stars (e.g. Santos et al. 2004b; Fischer & Valenti support from the STFC. We thank Tom Marsh for use 2005), although several other low-metallicity systems of the ULTRACAM pipeline. We thank the anonymous are known, including the transiting systems WASP-21 refereefor their thoroughcomments whichimprovedthe ([Fe/H] = 0.46 0.11, Bouchy et al. 2010), WASP-37 final manuscript. − ± REFERENCES Anderson,D.R.,etal.2011, ApJ,726,L19+ CollierCameron,A.,etal.2007,MNRAS,380,1230 Bakos,G.,Noyes,R.W.,Kova´cs,G.,Stanek, K.Z.,Sasselov, Dhillon,V.S.,etal.2007, MNRAS,378,825 D.D.,&Domsa,I.2004,PASP,116,266 Dittmann,J.A.,Close,L.M.,Scuderi,L.J.,&Morris,M.D. Baranne,A.,etal.1996,A&AS,119,373 2010,ApJ,717,235 Barnes,S.A.2007, ApJ,669,1167 Enoch,B.,CollierCameron,A.,Parley,N.R.,&Hebb,L.2010, Barros,S.C.C.,Pollacco,D.L.,Gibson,N.P.,Howarth,I.D., A&A,516,A33+ Keenan, F.P.,Simpson,E.K.,Skillen,I.,&Steele, I.A.2011, Enoch,B.,etal.2011,MNRAS,410,1631 MNRAS,416,2593 Fischer,D.A.,&Valenti,J.2005,ApJ,622,1102 Bodenheimer,P.,Laughlin,G.,&Lin,D.N.C.2003,ApJ,592, Fortney,J.J.,Lodders,K.,Marley,M.S.,&Freedman,R.S. 555 2008,ApJ,678,1419 Borucki,W.J.,etal.2010,Science, 327,977 Fressin,F.,Knutson,H.A.,Charbonneau, D.,O’Donovan,F.T., —.2011,ArXive-prints Burrows,A.,Deming,D.,Mandushev,G.,&Spiegel,D.2010, Bouchy,F.,etal.2010, A&A,519,A98+ ApJ,711,374 Bruntt,H.,etal.2010,MNRAS,405,1907 Gibson,N.P.,etal.2008, A&A,492,603 Christian,D.J.,etal.2009,MNRAS,392,1585 Gillon,M.,etal.2009,A&A,496,259 Claret,A.2000,A&A,363,1081 Gonzalez, G.,&Laws,C.2000, AJ,119,390 —.2004,A&A,428,1001 10 A. M. S. Smith et al. Gonzalez, G.,Laws,C.,Tyagi,S.,&Reddy,B.E.2001,AJ,121, Pollacco,D.L.,etal.2006,PASP,118,1407 432 Queloz,D.,etal.2001, A&A,379,279 Gray,D.F.2008,TheObservationandAnalysisofStellar Santos,N.C.,Israelian,G.,&Mayor,M.2004a, A&A,415,1153 Photospheres, 3rdEdition(CambridgeUniversityPress) —.2004b, A&A,415,1153 Hartman,J.D.,etal.2009,ApJ,706,785 Sestito,P.,&Randich,S.2005, A&A,442,615 Haywood,M.2009,ApJ,698,L1 Simpson,E.K.,etal.2011,AJ,141,8 Jehin,E.,etal.2011,TheMessenger,145,2 Skrutskie,M.F.,etal.2006,AJ,131,1163 Johnson,J.A.,Winn,J.N.,Cabrera,N.E.,&Carter,J.A.2009, Smalley,B.,etal.2011,A&A,526,A130+ ApJ,692,L100 Smith,K.C.1992,PhDThesis,UniversityofLondon Knutson,H.A.,Howard,A.W.,&Isaacson, H.2010,ApJ,720, Smith,K.C.,&Dworetsky,M.M.1988,inElementalAbundance 1569 Analyses,ed.S.J.Adelman&T.Lanz,32 Lucy,L.B.,&Sweeney, M.A.1971,AJ,76,544 Steele,I.A.,Bates,S.D.,Gibson,N.,Keenan, F.,Meaburn,J., Machalek,P.,McCullough,P.R.,Burke,C.J.,Valenti,J.A., Mottram,C.J.,Pollacco,D.,&Todd,I.2008,inSocietyof Burrows,A.,&Hora,J.L.2008, ApJ,684,1427 Photo-Optical Instrumentation Engineers(SPIE) Conference Maciejewski,G.,Raetz,S.,Nettelmann, N.,Seeliger,M.,Adam, Series,Vol.7014,SocietyofPhoto-OpticalInstrumentation C.,Nowak,G.,&Neuha¨user,R.2011a, A&A,535,A7+ Engineers(SPIE)ConferenceSeries Maciejewski,G.,etal.2011b, MNRAS,411,1204 Torres,G.,Andersen,J.,&Gim´enez, A.2010,A&ARev.,18,67 Magain,P.1984, A&A,134,189 Wheatley, P.,Harrington,J.,Fortney,J.,CollierCameron,A., Mamajek,E.E.,&Hillenbrand,L.A.2008, ApJ,687,1264 Smith,A.,Anderson,D.,&Pollacco,D.2011,inSpitzer Mandel,K.,&Agol,E.2002, ApJ,580,L171 ProposalID#80164, 80164 Marigo,P.,Girardi,L.,Bressan,A.,Groenewegen, M.A.T., Silva,L.,&Granato,G.L.2008,A&A,482,883 Pollacco,D.,etal.2008,MNRAS,385,1576

See more

The list of books you might like

Most books are stored in the elastic cloud where traffic is expensive. For this reason, we have a limit on daily download.