The GAPS Programme with HARPS-N@TNG V. A comprehensive analysis of the XO-2 stellar and planetary systems

2.4 MB·English

by  M. Damasso

Astronomy&Astrophysicsmanuscriptno.xo2_system_arx cESO2015
(cid:13)
January8,2015
The GAPS Programme with HARPS-N@TNG
⋆
V. A comprehensive analysis of the XO-2 stellar and planetary systems
M.Damasso1,2,K.Biazzo3,A.S.Bonomo1,S.Desidera4,A.F.Lanza3,V.Nascimbeni4,5,M.Esposito6,7,G.
Scandariato3,A.Sozzetti1,R.Cosentino3,8,R.Gratton4,L.Malavolta5,9,M.Rainer10,D.Gandolfi3,11,E.Poretti10,R.
ZanmarSanchez3,I.Ribas12,N.Santos13,14,15,L.Affer16,G.Andreuzzi8,M.Barbieri4,L.R.Bedin4,S.Benatti4,A.
Bernagozzi2,E.Bertolini2,M.Bonavita4,F.Borsa10,L.Borsato5,W.Boschin8,P.Calcidese2,A.Carbognani2,D.
Cenadelli2,J.M.Christille2,17,R.U.Claudi4,E.Covino18,A.Cunial5,P.Giacobbe1,V.Granata5,A.Harutyunyan8,M.
5 G.Lattanzi1,G.Leto3,M.Libralato4,5,G.Lodato19,V.Lorenzi8,L.Mancini20,A.F.MartinezFiorenzano8,F.
1 Marzari4,5,S.Masiero4,5,G.Micela16,E.Molinari8,21,M.Molinaro22,U.Munari4,S.Murabito6,7,I.Pagano3,M.
0 Pedani8,G.Piotto4,5,A.Rosenberg6,7,R.Silvotti1,J.Southworth23
2

n
1 INAF–OsservatorioAstrofisicodiTorino,ViaOsservatorio20,I-10025,PinoTorinese,Italy.
a
J E-mail:[email protected]
2 OsservatorioAstronomicodellaRegioneAutonomaValled’Aosta,Fraz.Lignan39,I-11020,Nus(Aosta),Italy
7
3 INAF–OsservatorioAstrofisicodiCatania,ViaS.Sofia78,I-95123,Catania,Italy

4 INAF–OsservatorioAstronomicodiPadova,Vicolodell’Osservatorio5,I-35122,Padova,Italy
]
R 5 Dip.diFisicaeAstronomiaGalileoGalilei–UniversitàdiPadova,Vicolodell’Osservatorio2,I-35122,Padova,Italy
6 InstitutodeAstrofísicadeCanarias,C/ViaLácteaS/N,E-38200LaLaguna,Tenerife,Spain
S
7 DepartamentodeAstrofísica,UniversidaddeLaLaguna,E-38205LaLaguna,Tenerife,Spain
h. 8 FundaciónGalileoGalilei-INAF,RamblaJoséAnaFernandezPérez7,E-38712BreñaBaja,TF-Spain
p 9 Obs.Astronomiquedel’Univ.deGeneve,51ch.desMaillettes-Sauverny,CH-1290,Versoix,Switzerland
- 10 INAF–OsservatorioAstronomicodiBrera,ViaE.Bianchi46,I-23807Merate(LC),Italy
o 11 LandessternwarteKönigstuhl,ZentrumfürAstronomiederUniversitatHeidelberg,Königstuhl12,D-69117Heidelberg,Germany
r 12 Inst.deCienciesdel’Espai(CSIC-IEEC),CampusUAB,FacultatdeCiencies,08193Bellaterra,Spain
t
s 13 InstitutodeAstrofísicaeCiênciasdoEspaço,UniversidadedoPorto,CAUP,RuadasEstrelas,4150-762Porto,Portugal
a 14 CentrodeAstrofísica,UniversidadedoPorto,RuadasEstrelas,4150-762Porto,Portugal
[ 15 DepartamentodeFísicaeAstronomia,FaculdadedeCiências,Univ.doPorto,RuadoCampoAlegre,s/n,4169-007Porto,Portugal

16 INAF–OsservatorioAstronomicodiPalermo,PiazzadelParlamento,1,I-90134,Palermo,Italy
1
17 Dept.ofPhysics,UniversityofPerugia,viaA.Pascoli,06123,Perugia,Italy
v
18 INAF–OsservatorioAstronomicodiCapodimonte,SalitaMoiariello16,I-80131,Napoli,Italy
4
19 DipartimentodiFisica,UniversitàdiMilano,ViaCeloria16,I-20133Milano,Italy
2
20 Max-Planck-InstitutfürAstronomie,Königstuhl17,D-69117,Heidelberg,Germany
4
21 INAF-IASFMilano,viaBassini15,I-20133Milano,Italy
1
22 INAF–OsservatorioAstronomicodiTrieste,viaTiepolo11,I-34143Trieste,Italy
0
23 AstrophysicsGroup,KeeleUniversity,Staffordshire,ST55BG,UK
.
1
0
5
ABSTRACT
1
:
v Aims.XO-2isthefirstconfirmedwidestellarbinarysystemwherethealmost twincomponents XO-2NandXO-2Shaveplanets,
i anditisapeculiarlaboratorytoinvestigatethediversityofplanetarysystems.Thisstimulatedadetailedcharacterizationstudyofthe
X stellarandplanetarycomponentsbasedonnewobservations.
r Methods.Wecollectedhigh-resolutionspectrawiththeHARPS-Nspectrographandmulti-bandlightcurves.Spectralanalysisledto
a anaccuratedeterminationofthestellaratmosphericparametersandcharacterizationofthestellaractivity,andhigh-precisionradial
velocitiesofXO-2Nweremeasured.WecollectedfourteentransitlightcurvesofXO-2Nbusedtoimprovethetransitparameters.
Photometryprovidedaccuratemagnitudedifferencesbetweenthestarsandameasureoftheirrotationperiods.
Results.TheironabundanceofXO-2Nwasfound+0.054dexgreater,withinmorethan3σ,thanthatofXO-2S.Theexistenceofa
long-termvariationintheradialvelocitiesofXO-2Nisconfirmed,andwedetectedaturn-overwithrespecttopreviousmeasurements.
Wesuggestthepresenceofasecondmassivecompanioninanouterorbitorthestellaractivitycycleaspossiblecausesoftheobserved
acceleration. Thelatterexplanationseemsmoreplausiblewiththepresent dataset.Weobtainedanaccuratevalueoftheprojected
spin-orbitanglefortheXO-2Nsystem(λ=7 11 ),andestimatedthereal3-Dspin-orbitangle(ψ=27+12degrees).Wemeasuredthe
XO-2rotationperiods,andfoundavalueofP◦±=41◦.6 1.1daysinthecaseofXO-2N,inexcellentagreem−2e7ntwiththepredictions.The
±
periodofXO-2Sappearsshorter,withanambiguitybetween26and34.5daysthatwecannotsolvewiththepresentdatasetalone.
Theanalysisof thestellaractivityshows thatXO-2N appears tobe moreactivethanthecompanion, and thiscould bedue tothe
factthatwesampleddifferentphasesoftheiractivitycycle,ortoaninteractionbetweenXO-2NanditshotJupiterthatwecouldnot
confirm.
Keywords. (Stars:)individual:XO-2S,XO-2N—Stars:fundamentalparameters,abundances—planetarysystems—techniques:
radialvelocities,photometric
Articlenumber,page1of26
A&Aproofs:manuscriptno.xo2_system_arx
Articlenumber,page2of26
M.Damassoetal.:GAPSV:GlobalanalysisoftheXO-2system
1. Introduction for the case of a circumbinary disk). If the warp is located in
theplanetformationregion,itmightaffecttheprocess,although
The diversity of the more than 1,700 extrasolar planets dis- sucheffectshavenotyetbeenstudied.
covered so far1 ( 850 of which are in multi-planet systems;
∼ Several examples of planets orbiting only one of
Roweetal. 2014), from their orbital architectures to the astro-
the binary components, defined as S-type planets, are
physicalenvironmentswheretheyreside,representsaverycom-
known (e.g. Roelletal. 2012; Mugraueretal. 2014, and
plex issue to be investigated to understand the mechanisms of
Thebault&Haghighipour 2014 for a recent review). The
theirformationanddynamicalevolution.
existence of S-type planets rises several questions: how the
Intheeraofcomparativeexo-planetology,afundamentalap- presence of a stellar companion can affect the formation,
proachfortheunderstandingofthedifferentexoplanetaryprop-
survivalanddynamicalevolutionofsuchplanets?Isitpossible
ertiesreliesonthecharacterizationofplanet-hoststarsandstatis-
to identify which properties of a stellar system most influence
ticalanalysisoftheplanetarysystemfrequency.Severalstudies
the physical characteristics of the hosted planets? Does any
shown that the propertiesof the planetary system architectures
constraint exist with respect to the case of isolated stars, or
and physical characteristics of the planets depend upon stellar
any dependence from the physical separation of the binary
properties, such as the mass (Johnsonetal. 2010; Bonfilsetal. components? Observational difficulties exist in searching for
2013) and the metallicity (Sozzettietal. 2009; Santosetal.
S-typeplanetsinmultiplestellar systems(Eggenberger&Udry
2011; Mortieretal. 2012). Moreover, during the last years the
2010), in particular those with a very small angular separation
search for exoplanets was extended in environments like giant
(less then 2 ), since, for instance, the components cannot be
′′
stars (e.g.Kepler-91,Lillo-Boxetal. 2014b,a),openandglobu- ∼
observed as two isolated targets in the spectrograph’s fiber or
larclusters,andonlyrecentlyplanetarycompanionswerefound
slit.Inmanycasesofconfirmedextrasolarplanetsthepresence
(see,e.g.,Quinnetal.2012;Meibometal.2013,andreferences
of a stellar companionto the hoststar was discoveredafter the
therein).
detectionofthe planet(e.g.Roelletal. 2012). Binarity posesa
Binarysystemsareinterestingastrophysicalenvironmentsto major challengein understandinghow such planetscould exist
searchforplanets.Consideringthatnearlyhalfofthesolar-type andtheirevolutionaryhistoryinthestellarsystem.
starsaregravitationallyboundwithatleastanotherstellarcom- Verywidebinaries,withsemi-majoraxesoftheorderof103
panion(e.g.Raghavanetal.2010;Duchêne&Kraus2013),bi- AUandskyprojectedangularseparationsofseveralarcsec,of-
narysystemsnaturallyrepresentatypicalenvironmenttobeex- ferthebestopportunityandlessobservationalcomplicationsto
ploredforstudyingtheprocessesofplanetformationandevolu- search for planets orbiting each component. At the same time
tionleadingtoverydifferentplanetaryarchitectures. theyallowanaccuratedeterminationofthepropertiesofthein-
Planet formation in a binaries has been considered by var- dividual stars. Despite the large separation between the stellar
ious authors. For coplanar orbits between the disk and the components,thesesystemsareaninterestingsubjectfordynam-
binary the general expectation (e.g., Marzari&Scholl 2000; icalstudiesalsorelatedtothepresenceofplanets.
Kley&Nelson 2008; Marzarietal. 2012) is that the tidal ef- Kaibetal. (2013) and Kaib&Raymond (2014) studied the
fect of the companion would induce an eccentricity growth in perturbationsproducedby passing stars and the tidal effects of
the planetesimalpopulationthat mightinhibitplanet formation theMilkyWayontheorbitsofverywidebinarystars.Interest-
if the binary separation is . 50 AU. In this configuration even ingly they foundthat occasionally the orbits could become ex-
the formation of icy planetesimals is jeopardized by the high tremely eccentric, resulting in the collision of the binary com-
temperatures in the discs and the formation of breaking waves ponents or, less dramatically, they could evolve toward close
(Nelson 2000; Picogna&Marzari2013). The situation is more orcontactsystems.Moreover,simulationsbyKaibetal.(2013)
complicatedformisalignedsystemswheretheplanetesimalsand indicate that very wide binary companions may often strongly
gas discs precess around the binary orbital plane. Planetesi- reshapeplanetary systems after their formation,by causing the
malsnaturallydevelopstrongdifferentialprecessionthatwould ejectionofplanetsfromthe system orheavilychangingtheor-
inhibit planet formation by increasing the velocity dispersion bitaleccentricitiesof those thatsurvive.Anotherintriguingex-
(Marzarietal. 2009). On the other hand, if the gas disk is ra- ample of a plausible fate for planets in wide binaries is the
dially narrow (which is expected for separation of the order of ’bouncing’ scenario investigated by Moeckel&Veras (2012).
50 AU because of tidal truncation), it will precess rigidly and They found that in 45 to 75 per cent of the cases a planet, ini-
would drag the planetesimals into rigid precession, mitigating tially orbiting one of the binary components and then ejected
thepreviouseffect(Fragneretal.2011),unlesstheinclinationis from its native system by planet-planet scattering, could jump
so large for Lidov-Kozai effects to become important. Also in upanddownpassingfromthegravitationalinfluenceofonestar
thiscase,theeffectsshouldberelevantmostlyforbinarysepara- totheother,formorethanaMyr.Thissituationmaytriggeror-
tionssmallerthan50-100AU.Forwider,misalignedbinariesthe bital instability among existing planets around the companion,
gas disk on the other hand is expected to reach a quasi-steady, andinsomecasesresultinanexchangeofplanetsbetweenthe
non precessing warped configuration (see Facchinietal. 2013 twostars,whenbothhostmultipleplanetarysystems.
Atpresent,fewverywidebinariesareknownwhereaplan-
⋆ Based on observations made (i) with the Italian Telescopio etary system was discovered around one of the stellar compo-
Nazionale Galileo (TNG), operated on the island of La Palma by the
nents.ThestarHD20782isthebinarycompanionofHD20781,
INAF - Fundacion Galileo Galilei (Spanish Observatory of Roque de
withaprojectedseparationof 9000AU,andithostsaJupiter-
losMuchachos oftheIAC);(ii)withtheCopernicoandSchmidttele- ∼
mass planet on a very eccentric orbit at 1.4 AU (Jonesetal.
scopes(INAF-OsservatorioAstrofisicodiPadova,Asiago,Italy);(iii) 2006)2. The very high eccentricity report∼ed in the literature (e
with the IAC-80 telescope at the Teide Observatory (Instituto de As-
trofísicadeCanarias,IAC);(iv)attheSerralaNave"M.G.Fracastoro" = 0.97) could be the result of significant perturbations experi-
Astronomical Observatory (INAF - Osservatorio Astrofisico di Cata-
nia);(v)attheAstronomicalObservatoryoftheAutonomousRegionof 2 HD20781 should host two Neptune-mass planets within 0.3 AU,
∼
theAostaValley(OAVdA). asreportedbyMayoretal.(2011),butthediscoveryhasnotyetbeen
1 NASAexoplanetarchive,http://exoplanetarchive.ipac.caltech.edu/ confirmedandpublished.
Articlenumber,page3of26
A&Aproofs:manuscriptno.xo2_system_arx
encedbytheplanetduringtheevolutionofthesystemandpos- cantellapartofthatstory.WhiletheanalysisoftheXO-2sys-
siblyduetothestellarcompanion.Inthisbinarysystem,where tem was still in progress,the discoveryof a hot Jupiter around
boththecomponentscanbewellanalysedseparately,acompar- eachofthetwins,metal-rich([Fe/H] 0.25dex),F-typecompo-
∼
ativestudyofthephysicalpropertiesofthepaircouldbeofhelp nentsoftheWASP-94widebinary(projectedseparation 2700
∼
forexplainingthenatureoftheplanetarysystem.Toinvestigate AU)wasannounced(Neveu-VanMalleetal.2014).Thissystem,
anyexistinglinkbetweentheplanetformationmechanismsand withhoststarpropertiesandplanetarysystemarchitecturesdif-
the chemical composition of the host stars, Macketal. (2014) ferent from those of XO-2, indeed makes more intriguing the
performedadetailedelementalabundanceanalysisoftheatmo- understandingoftheroleofstellarmultiplicityontheplanetfor-
spheresofHD20781/82,byconsideringbothstarstohostplan- mationandevolution.
ets. This kind of study, when at least one component hosts a Thispaperisorganizedasfollows.WefirstdescribeinSect.2
close-in giant planet, could result in the evidence of chemical thespectroscopicandphotometricdatasetsusedinthisstudy.We
imprintsleftintheparentstarsuggestingtheingestionofmate- presentinSect.3anewdeterminationofthebasicstellarparam-
rial from the circumstellar disk (and possibly also of planetary etersbasedonHARPS-Nspectra,aswellasresultsofadetailed
origin)drivenbythedynamicalevolutionoftheplanetorbit.In and homogeneous analysis aimed at searching for differences
fact, whena star with a close-ingiantplanetis foundto be en- intheironabundancebetweentheXO-2componentswithhigh
richedwithelementsofhighcondensationtemperature,assug- levelofconfidence.InSect.4wepresentanddiscussresultsfrom
gestedbySchuleretal.(2011b),thiscanberelatedtotheinward aphotometricfollow-upofthetwostars,includingthemeasure-
migrationoftheplanetfromtheouterregionsofthecircumstel- mentof the stellar rotationperiodsand the modellingof a new
lar disk, where it formed, to the present position closer to the datasetof14transitsoftheplanetXO-2Nb.Sect.5isdedicated
star (Ida&Lin2008;Raymondetal. 2011). Macketal. (2014) totheXO-2Nplanetarysystem.Therewepresentanewanalysis
found for both stars a quite significant positive trends with the oftheradialvelocitytimeseries,thankstowhichweconfirmthe
condensationtemperatureamongtheelementalabundances,and existenceofanaccelerationwithanotyetestablishedcause.In
suggestthat the hoststars accreted rockybodieswith mass be- thesame Sectionwe also presentnew observationandanalysis
tween 10 and20 M initially formedinteriorto the locationof oftheRossiter-McLaughlineffect,betterconstrainingthevalue
thedetectedplanets.⊕Searchesforabundanceanomaliespossibly oftheprojectedspin-orbitangleλ.Adetailedanalysisofthestel-
causedbytheingestionofplanetarymaterialbythecentralstar laractivityfortheXO-2componentsispresentedanddiscussed
werealsoconductedbyDesideraetal.(2004,2007)(andrefer- in Sect.6. We finally concludewith Sect.7, where we speculate
encestherein)forasampleofwidebinaries.Theseauthorsfound abouttheevolutionandlong-termstabilityoftheXO-2planetary
that the amountof iron accreted by the nominally metal richer systems.
companion (in binaries with components having T >5500 K)
eff
iscomparabletotheestimatesoftherockymaterialaccretedby
the Sun during its main-sequence lifetime, and therefore con- 2. Observationsanddatareductionmethods
cludedthatthemetalenrichmentduetotheingestionofmaterial
2.1.Spectroscopy
ofplanetaryoriginshouldnotbeacommonevent.
A representativecase of a very wide binary with one com- The spectra of the XO-2 components analysed in this work
ponenthostingaclose-ingiantplanetisthatoftheXO-2system werecollectedwiththehigh-resolutionHARPS-Nspectrograph
(projectedseparation 31 ),wherethestarXO-2Nisorbitedby (Cosentinoetal. 2012) installed at the Telescopio Nazionale
′′
the transiting hot Jup∼iter XO-2Nb (Burkeetal. 2007). An ele- Galileo (TNG) on La Palma (Canary islands). The observa-
mentalabundanceanalysisofXO-2NanditscompanionXO-2S tionswerecarriedoutintheframeworkofthelargeprogramme
wasperformedbyTeskeetal.(2013),whodeterminedthestel- GlobalArchitectureofPlanetarySystems(GAPS;Covinoetal.
larabundancesofcarbonandoxygen,aswellasironandnickel. 2013;Desideraetal.2013).
Their goalwas to probethe potentialeffects thatplanetforma- ForthestarXO-2Nwecollected43spectrabetweenNovem-
tionandevolutionmighthavehadonthechemicalcomposition ber20,2012andOctober4,2014,whilethecompanionXO-2S
of XO-2N, where the companionXO-2S was treated as a non- was observed at 63 individual epochs between April 21, 2013
hostingplanetstar.Later,Desideraetal.(2014)discoveredthat andMay10,2014.TheTh-Arsimultaneouscalibrationwasnot
the star XO-2S actually hosts a planetary system formed by a usedtoavoidcontaminationbythelamplines.Thishasnosig-
planet slightly more massive than Jupiter orbiting at 0.48 AU nificant impact on the measurementsof the radial velocity, be-
andaSaturn-massplanetat0.13AU(withcautiousevidenceof cause the drift correction with respect to the reference calibra-
along-termtrendintheradialvelocitiestimeseriespossiblydue tionshowsa dispersionof 0.8m s−1, whichis smaller thanthe
toanoutercompanion,yetofunknownnature).Thediscoveryis medianradialvelocity(RV)uncertaintyduetophoton-noise,i.e.
of particularrelevance because it representsthe first confirmed 2.0 m s−1 for XO-2N and 2.2 m s−1 for XO-2S. The reduction
caseofaverywidebinaryinwhichbothcomponentshostplan- ofthespectraandtheRVmeasurementswereobtainedforboth
ets(seeFig.1forasketchoftheXO-2planetarysystems).The componentsusingthelatestversion(Nov.2013)oftheHARPS-
XO-2 binary offers a unique opportunity to explore the diver- NinstrumentData ReductionSoftwarepipelineandapplyinga
sityofplanetformationmechanisms,byconsideringthat,while K5mask.ThemeasurementoftheRVsisbasedontheweighted
the parentstars are almostequalin their main physicalproper- cross-correlation function (CCF) method (Baranneetal. 1996;
ties,theplanetarysystemsaredifferent.Thisfindinghasconse- Pepeetal.2002).
quentlymotivatedadetailedcomparativestudyofthewholesys-
tem, which is the subject of this paper,throughthe use of new
2.2.Photometry
high-resolutionand high S/N HARPS-N spectra and dedicated
differentialphotometry.Itiscrucialtohighlightanyexistingdif- Wecollectedandanalysednewphotometriclightcurvesforboth
ferencebetweenthepropertiesofthetwostarswhichcouldhave componentsoftheXO-2systemwiththreedifferentfacilities.In
playedadistinctiveroleintheformationprocessesoftheplan- particular,XO-2Swas intensivelymonitoredwith the HARPS-
etary systems:anyobserveddifference,if correctlyinterpreted, N spectrograph. When it became clear that the star had plane-
Articlenumber,page4of26
M.Damassoetal.:GAPSV:GlobalanalysisoftheXO-2system
2.2.1. TheAPACHEdataset
A follow-upofthe XO-2fieldwasconductedfor42 nightsbe-
tweenDecember2,2013andApril8,2014withoneofthe40cm
telescopescomposingtheAPACHEarray(Sozzettietal.2013),
basedattheAstronomicalObservatoryoftheAutonomousRe-
gionoftheAostaValley(OAVdA,+45.7895N,+7.478E,1650
ma.s.l.).EachtelescopeisaCarbonTrussf/8.4Ritchey-Chrtien
equipped with a GM2000 10-MICRON mount and a FLI Pro-
linePL1001E-2CCD Camera,with apixelscale of1.5 /pixel
′′
and a field of view of 26x26. The observations were carried
′ ′
out using a Johnson-CousinsI filter. The images were reduced
with the standard pipeline TEEPEE written in IDL3 and regu-
larlyusedfortheAPACHEpurposes(seeGiacobbeetal.2012).
The APACHE data were used to tentatively derive the rotation
periodofthetwostarsandtomonitortheirmagnitudedifference
inI band,asdescribedinSect.4.
2.2.2. TheTASTEdataset
SinceFebruary2011severaltransitsofthehotJupiterXO-2Nb
have been observed at high temporal cadence with the 182cm
and the Schmidt 92/67 telescopes at the Asiago Astrophysical
Observatory (+45.8433 N, +11.5708 E, 1366m a.s.l.), in the
frameworkoftheTASTEProject(TheAsiagoSearchforTransit
timevariationsofExoplanets;Nascimbenietal.2011).Asubset
ofthesemeasurementswasusedtoestimatethemagnitudedif-
ferences of the two XO-2 binary components in the R band.
c
TwoadditionaltransitswerealsocollectedwiththeIAC-80tele-
scopeattheTeideObservatory,usingtheCAMELOTCCDim-
ager(seeNascimbenietal.2013foradescriptionoftheset-up),
andonesupplementarylightcurvewas obtainedwith the same
telescopebutusingtheTCPcamera(TromsöCCDPhotometer;
Østensen&Solheim2000).Alltheseobservations,foratotalof
14 light curves, were carried out in the Cousin/Bessell R band
andwereanalysedtogethertoimprovethedeterminationofthe
parametersofXO-2Nb.
2.2.3. TheSerralaNavedataset
The XO-2 system was also observed for five nights, between
May6,2013andMay7,2014withtherobotic80cmf/8Ritchey-
Chrétien telescope APT2, operated by the INAF-Catania As-
trophysical Observatory and located at Serra la Nave (SLN,
+37.692N,+14.973E,1725ma.s.l.).Thetelescopeisequipped
with a CCD detector Apogee Alta U9000 (3Kx3K pixels) op-
erated in binning 2x2, correspondingto a pixel scale 0.76 arc-
Fig. 1. Comparison between the planetary systems orbiting the XO-2 sec/pixel. The data were reduced with the IRAF package4 us-
binarystarsandthearchitectureoftheinnerSolarSystem.Thecaseof ingthestandardprocedureforoverscan,biasanddarksubtrac-
XO-2Nisshownintheupperpanel,whilethatofXO-2Sisrepresented tion and flat fielding. The aperture photometry was done with
inthelowerone.Inbothplots,theredellipses(solidlines)representthe thesoftwareSExtractor(Bertin&Arnouts1996).Thedataof
orbitsoftheXO-2planets.TheorbitsofMercury,VenusandtheEarth
SLNwerecollectedintheBVRI bandsandusedtomeasurethe
areindicatedwithblack,blueandgreendottedlinesrespectively,also
magnitudedifferencesbetweenthetwostars.
labelledwiththeplanetnames.Thecrosssymbolinthemiddleofeach
plotindicatesthestarlocation.
3. Stellarparameters
3.1.Spectroscopicanalysisoftheindividualstars
The components of the XO-2 binary system have a separation
of 31 arcsec (correspondingto a projected distance of 4600
tarycompanions,we starteda dedicatedphotometricfollow-up AU∼assumingadistanceof 150pc,asestimatedbyBurk∼eetal.
ofthetargettoestimatethestellarrotationalperiod,furtherchar- ∼
acterizingitslevelofactivityandlookforpossibletransitsofits 3 RegisteredtrademarkofExelisVisualInformationSolutions.
exoplanets. 4 http://iraf.noao.edu/
Articlenumber,page5of26
A&Aproofs:manuscriptno.xo2_system_arx
2007), and share, within the uncertainties, the proper-motion version 2013 of the MOOG code (Sneden 1973) and consid-
vectoras listed in the UCAC4 catalogue.Essentialinformation ered the abfind driver, assuming local thermodynamic equilib-
abouttheXO-2systemispresentedinTable1.Astandardanaly- rium(LTE).Initialstellarparametersweresettothesolarvalues
sisoftheHARPS-Nspectrawasperformedwithdifferentmeth- (T =5770 K, logg = 4.44, and ξ = 1.10 km/s) for both
eff
odstoderivethestellaratmosphericparametersofthetwocom- com⊙ponents.Then,th⊙efinaleffectivet⊙emperature(T )wasde-
eff
ponents,with those forXO-2Sfirstpresentedin Desideraetal. terminedbyimposingtheconditionthattheabundancefromthe
(2014). Table 2 summarizes our results, which represent the Feilines(logn(Fei))wasindependentonthelineexcitationpo-
weightedaveragesoftheindividualmeasurements.Weusedim- tentials;thefinalmicroturbulence(ξ)byminimizingtheslopeof
plementationsofboththeequivalentwidthandthespectralsyn- logn(Fei)versusthereducedequivalentwidth(EW/λ);andthe
thesismethods,asdescribedinBiazzoetal.(2012),Santosetal. finalsurfacegravity(logg) byimposingionizationequilibrium
(2013), and Espositoetal. (2014). It is interesting to note that (i.e.logn(Fei)=logn(Feii)).Plotsofironabundance(logn(Fe))
the effective temperature difference between the two stars de- versus excitation potential (χ) and reduced EW for both com-
rivedspectroscopicallyisingoodagreementwiththeresultob- ponentsare shown in Fig. 2, which shows that the correlations
tained from photometric measurements. In fact, following the areclosetozero,asrequired.Usingthesameprocedure,weob-
sameanalysisperformedbyMunarietal.(2014)forhalf-million tainedlogn(Fei) = logn(Feii) = 7.53 0.05fortheSun.We
ofstarsoftheRAVEspectroscopicsurvey,ifweassumeacolor point out that the⊙ exact values⊙of the so±lar parameters are not
excess E(B V)=0.02 mag (as explained in Sect.3.3) the XO- crucialas we are performinga differentialstudy for both com-
−
2S star results 60 33K hotter than the companion.The stellar ponentswithrespecttotheSun.Table3liststhefinalresultsof
±
mass, radius and age were determined by comparing our mea- thisprocedure,wheretheerrorsinironabundanceincludeuncer-
suredeffectivetemperature,ironabundance,andsurfacegravity taintiesin stellar parametersandin EWmeasurements.Results
withtheYonsei-Yale(Y-Y)evolutionarytracks(Demarqueetal. fromthisfirsthomogeneousanalysisseemstobeinaccordance
2004)throughtheχ-squarestatistics(Santerneetal.2011).The with those obtainedwith differentmethodsof spectral analysis
adopted errors include an extra 5% in mass and 3% in radius (Table2).
addedinquadraturetotheformalerrorstotakesystematicuncer- ii.Line-by-lineanalysis.Sinceweareinterestedinahomo-
taintiesinstellarmodelsintoaccount(Southworth2011).Taking geneous analysis, we applied the same iterative method as in
advantageofthe factthatthe planetXO-2Nbtransitsits parent the first step, but here we considered a strict line-by-line anal-
star,wederivedasecond,slightlymoreaccurateestimateforthe ysis, i.e. the same lines were used for both components, with-
mass, radiusandageofXO-2N,usingthe stellar densityinfor- out any implementation of line rejection criteria. This allowed
mationasdeterminedfromtheanalysisofthetransitlightcurves usto avoid uncertaintiesrelated to bothatomic parametersand
(e.g.,Sozzettietal.2007). measurementsofequivalentwidths,dueto,e.g.,blendsandcon-
tinuumlevel. In the end, the analysiswas based on 75 Fei and
10 Feii lines. Thanks to such analysis, the errors in T were
eff
3.2.Differentialspectralanalysis reducedsubstantially,asreportedinTable3.
iii. Differentialanalysis.Here,asstellar parametersofXO-
Thanks to the high-quality of the HARPS-N spectra, we per-
2N, we considered the values derived in the line-by-line anal-
formedadetailedandhomogeneousspectralanalysisofXO-2N
ysis, while XO-2S was analysed differentially with respect to
and XO-2S to search for possible differencesin the iron abun-
the companion, strictly using the same line set and EWs as in
dance between the components. The average values for [Fe/H]
thepreviousstep.Weappliedthedifferentialabundancemethod
obtained from the analysis of the individual stars (Table 2) in-
widely described in Grattonetal. (2001) and Desideraetal.
dicatethattheNorthcomponentischaracterisedbyhigheriron
(2004,2006),whichyieldsveryaccurateresultswhen,inpartic-
abundance,butthetwoestimatesarecompatiblewithintheun-
ular, the two componentsare very similar in stellar parameters
certainties.In orderto validatethis differencewith a highlevel
(asfortheXO-2components).Insummary,themostimportant
of confidence,we determinedthe difference in [Fe/H] very ac-
ingredient in this analysis is the difference in temperature be-
curatelywitha dedicatedanalysis.Togetherwiththespectraof
tween the components. In our case, we take advantage of the
both targets, we also acquired three solar spectra through ob-
knowledgeof the luminosity difference between the targets, as
servations of the asteroid Vesta. This allowed us to perform a
we can assume theyare at the same distance fromus. We con-
differentialanalysisforeachcomponentwithrespecttotheSun,
sideredasmagnitudedifferencesindifferentbandsthemeanval-
thusavoidingthecontributionduetotheuncertaintiesinatomic
uesreportedinSect.4.1andasbolometriccorrectionsthosede-
parameters, such as the transition probabilities. All HARPS-N
rivedbyusingthecodeprovidedbyCasagrande&VandenBerg
spectrawereshiftedinwavelengthandthenco-addedinorderto
(2014), which allows to estimate the bolometric corrections in
obtainveryhighS/N spectra( 180 250atλ 6700Å).We several photometric bands using T , [Fe/H], and logg as pri-
∼ − ∼ eff
performedthespectralanalysisfollowingthreemainsteps:
maryinputs.Then,sincethestarsareonthemainsequence,we
i. Analysis based on equivalent widths. We measured the couldobtainaccurateestimatesofthedifference∆loggbetween
equivalentwidths(EWs)ofironlinesonone-dimensionalspec- theirsurfacegravities(seeEq.6inDesideraetal.2006,whereas
tra using the splot task in IRAF, paying attention to trace as stellar mass differencewe assumed 0.01 M , as given in Table
much as possible the same position in the continuum level for 2). This allowed us to accurately derive th⊙eir temperature dif-
thespectraofbothcomponentsandtheasteroidVesta.Wecon- ference(∆T )usingthe equilibriumoftheironionization,be-
eff
sidered the line-list from Biazzoetal. (2012), and adopted the causethedifferencebetweentheabundancesprovidedbytheFei
sameprocedureastheseauthorsforthedifferentialanalysisrel- and Feii lines is strongly sensitive to T (about 0.001 dex/K;
eff
ativetotheSun,therejectioncriteriaofbadlines,thelinebroad- seeGrattonetal.2001).Thefinalatmosphericparametersofthe
eningmechanisms,themodelatmospheres,themeasurementof secondary were derived as differences with respect to those of
stellarparametersandironabundances,andtheircorresponding the primary by means of an iterative procedure (because logg,
uncertainties.We refertothatpaperandtoBiazzoetal. (2011) ξ, and [Fe/H] depend on the assumed temperatures). We inter-
for detailed descriptions of the method. In brief, we used the ruptedtheprocedureconsideringreachedtheconvergencewhen
Articlenumber,page6of26
M.Damassoetal.:GAPSV:GlobalanalysisoftheXO-2system
the grid step-size was of 1 K in T , 0.001 dex in logg, and a lower iron content. To explain this deficiency, Ramírezetal.
eff
0.01km/sinξ.Nootherσclippingwasimplementedhere.The (2011) suggest thatan early depletion of metals happeneddur-
final adopted differences in stellar parameters are listed in Ta- ing the formation phase of the 16 Cyg Bb planet. An alterna-
ble3,whileFig.3shows∆[Fe/H]asafunctionoftheexcitation tive and suggestive possibility is that 16 Cyg A may have in-
potential.Theunder-abundancein iron andthe highereffective gested a massive planet that enriched the star with iron, while
temperatureofXO-2SwhencomparedtoXO-2Nareconfirmed thelargeorbitof16CygBb(withsemi-majoraxisof1.68AU)
by this procedure. All differential parameters we obtained are likely prevented any mass loss from the planet. We note that,
affectedbybothinternalandsystematicerrors.Internalerroron despite ∆[Fe/H] is similar with that we found for XO-2, the
∆T was derived from the line-by-linescatter of Fei and Feii amount of iron involved should be different. In fact, the XO-2
eff
andtheerrorsintheotheratmosphericparameters.Internalerror starsarecoolerthan16CygniAandBandhavethereforemore
in ∆logg includes the uncertainties in the differences in mass, massive convective envelopes, implying that more iron should
magnitude,bolometriccorrection,andeffectivetemperature.Er- benecessarytopollutetheXO-2Nphotosphereandproduceal-
rorinthemicroturbulencedifferencewasobtainedbysumming most the same ∆[Fe/H]. Also Laws&Gonzalez (2001) found
inquadraturetheerrorcontributionsinξofeachcomponent.The the primary of the 16-Cyg system enhanced in Fe relative to
error in the iron abundance difference was derived taking into the secondary. Similar studies conducted in binary stars host-
accounttheuncertaintiesinallstellarparametersaddedquadrat- ing planets (see, e.g., Grattonetal. 2001; Desideraetal. 2004,
ically.Allerrorsinthedifferentialanalysisareverylow,assum- 2006; Schuleretal. 2011a; Teskeetal. 2013; Liuetal. 2014;
marized in Table 3, and this demonstrates how this procedure Macketal. 2014) did notfindrelevantdifferencesin elemental
is efficient to unveil small differences in atmospheric parame- abundanceamong the components.All these results imply that
ters,removingmanysourcesofsystematicerrors.Infact,asre- the presence of giantplanets does notnecessarily imply differ-
markedbyDesideraetal.(2006),externaluncertainties,dueto, encesinthechemicalcompositionofthehoststar.
e.g.,wrongestimationsofparallaxandmass,orinadequaciesof Our results will produce a more clear picture when several
modelatmospheresandLTE deviations,havenegligibleeffects iron-peak,α-, s-process,and otherodd/even-Zelementalabun-
onourresults,becauseofthesimilarcharacteristicsofthecom- danceswill be analysedforbothstars andtheir behaviourwith
ponents.Intheend,thedifferenceinironabundanceofXO-2N thecondensationtemperaturestudiedwiththeaimtoinvestigate
withrespecttoXO-2Sisof+0.054dexatamorethan3σlevel. possibleselectiveaccretionofplanetarymaterial.Oneofthead-
ThedifferenceinironabundancebetweenthetwoXO-2stel- vantagesof comparingcoevalstars in wide binariesis that any
larcomponentsposesaninterestingissue. Theybelongtoa vi- differenceintheirabundancetrendcouldbemorerelatedtoac-
sual binary and, as normally assumed for such systems, they cretion of rocky-planetary material rather than to the Galactic
shouldsharethe same originandinitialbulk metallicity.A rel- chemicalevolution.Thisanalysiswillbethesubjectofaforth-
evant characteristic of this system is that both of the stars host comingpaperoftheGAPSseries(Biazzoetal.,inprep.),where
planets. Thus, for components of wide binaries where at least wewillinvestigateifthepresenceofhotJupiterscouldleadthe
one star has a planet, a reasonable hypothesis to explain any hoststarstoingestmaterial,whichinturnmayleavemeasurable
measuredandsignificantdifferenceintheirpresent-dayelemen- chemicalimprintsintheiratmosphericabundances.
tal abundancesis that most likely the planet formation process
had played a relevant role. The higher iron abundance of XO-
3.3.XO-2distanceandgalacticspacevelocities
2N when compared to XO-2S might be due to past ingestion
of dust-rich or rocky material, coming from the inner part of Usingourstellarparameters,wederivedanestimateofthespec-
theproto-planetarydiskandpushedintothehoststarbythehot troscopicdistancesof XO-2N and XO-2Sbymeansof the fol-
JupiterXO-2Nasitmigratedinwardtoitscurrentorbit.Asec- lowingprocedure.WegeneratedMonte-Carlo(MC)normaldis-
ondmechanism,whichactsonadifferenttimescaleandafterthe tributions for each spectroscopic parameter T , [Fe/H], and
eff
pre-main-sequencestage,ispresentedbyFabrycky&Tremaine logg, composed of 10,000 random values and centred on the
(2007),whodiscussthepollutionofthestellarphotospherewith best estimates (Table 2). By keeping the stellar radii fixed to
metalsproducedbymasslossofinwardmigratinghotJupiters. the values listed in Table 2 (for XO-2N we used the most ac-
Also,withtheirsimulationsKaibetal.(2013)showedthatvery curateestimate),foreachMCsimulationwefirstdeterminedthe
distant binary companions may severely affect planetary evo- stellarbolometricluminosityL (insolarunits)fromtheStefan-
lution and influence the orbits of any planet around the other Boltzmann law, and then we ∗derived the absolute bolometric
component of the system, thus maybe favouring the ingestion magnitude M from the relation M = 4.75 -2.5 log(L ).
bol bol
ofmaterialbythehoststar.FollowingtheresultsshowninFig. By estimating the appropriate bolometric correction·(BC),∗a
2 of Pinsonneaultetal. (2001), we roughly estimated that XO- value for the absolute magnitude in V-band M was then ob-
V
2N could have ingested an amount of iron slightly higher than tained.TheBC termwas evaluatedusing thecodeprovidedby
5 M to increase its photospheric iron content by 0.05 dex, Casagrande&VandenBerg (2014). An additional input is the
given⊕itsTeff 5300K.Thedifferenceintheironab∼undanceof color excess E(B V) of the star, which we derived through
the two XO-2∼stellar components is similar to what found by therelation E(B −V) =A (s)/3.1,whereA (s) istheinterstel-
V V
Ramírezetal.(2011)forthesolartwins16CygAand16CygB lar dust extinctio−nin V-band integratedat the distance s of the
(∆[Fe/H]=0.042 0.016dex).Thesestarshavemassescloseto star (in pc) and measured along the line-of-sight. We derived
±
those of the XO-2 companionsand, together with a third com- A (s) by adopting a simplified model of the local distribution
V
panion 16 Cyg C, are members of a hierarchicaltriple system, of the interstellar dust density (Drimmel&Spergel 2001), ex-
wheretheAandCcomponentsformaclosebinarywithapro- pressedbytherelationρ=ρ sech2(z/h ),wherezistheheightof
0 s
jectedseparation 70AU.ThecomponentBisknowntohosta thestarabovetheGalacticpl·aneandh isthescale-heightofthe
∼ s
giantplanetin a long-periodandhighlyeccentricorbit(P 800 dust, for which we adopted the value of 190 pc. The term z is
daysande 0.69)5,butunlikethecaseofXO-2itisthestar∼with relatedtothedistancesandtheGalacticlatitudeofthestarbby
∼
theformulaz=s sinb.Fromthismodelweobtainedtherelation
·
5 http://exoplanet.eu AV(s)=AV(tot)sinh(s sinb/hs)/cosh(s sinb/hs),whereAV(tot)is
· · ·
Articlenumber,page7of26
A&Aproofs:manuscriptno.xo2_system_arx
theinterstellarextinctioninV-bandalongtheline-of-sightinte- Table1.BasicinformationabouttheXO-2stellarbinarysystem
gratedthroughtheGalaxy,andcanbeestimatedfrom2-DGalac-
ticmaps.ForthispurposeweusedthevalueA (tot)=0.16mag Parameter XO-2N XO-2S Ref.
V
derivedfromthemapsofSchlafly&Finkbeiner(2011)6.Byas-
UCAC4ID 702-047113 702-047114 (1)
suming s=150 pc as a prior distance of the stars (Burkeetal.
2007), we obtained A (150pc) 0.06 mag, corresponding to R.A.(J2000) 07:48:06.471 07:48:07.479 (1)
V
∼
E(B V)=0.019mag.Thisisthevalueusedasinputtothecode DEC(J2000) +50:13:32.91 +50:13:03.25 (1)
−
of Casagrande&VandenBerg (2014) to obtain a first guess of
NearUV(mag) 18.153 0.047 17.988 0.044 (2)
theBCinV-band.Thisinturnwasusedinthedistancemodulus ± ±
formulaV-(M -BC )=5 log(s)-5-A (s),toobtainanewvalue B(mag) 12.002 0.015 11.927 0.015 (3)
bol V · V ± ±
forthestellardistances.Thenewdistancewasusedtorepeatthe V(mag) 11.138 0.026 11.086 0.025 (3)
procedureiteratively,bydeterminingateachstepanewvalueof ± ±
R (mag) 10.669 0.020 10.631 0.019 (3)
AV(s) and BCV(s), and finally another estimate of s. When the c ± ±
absolutedifferencebetweenthelastandpreviouscalculatedval- I (mag) 10.243 0.012 10.216 0.013 (3)
c ± ±
uesofswasbelow0.1pc,theiterativeprocesswasinterrupted
J(mag) 9.744 0.022 9.742 0.022 (4)
and the last derivedvalue fors was assumed as the distance of ± ±
H(mag) 9.340 0.026 9.371 0.027 (4)
thestarfortheN-thMonte-Carlosimulation.Theadoptedesti-
± ±
matesforthe distance ofthe XO-2componentsare the median K (mag) 9.308 0.021 9.272 0.021 (4)
s ± ±
ofthedistributionsofthe10,000MCvalues,andtheasymmet-
Distance(pc) 145.5+2.9 148.0+2.8 (5)
ric error bars defined as the 15.85th and 68.3th percentile (see 3.0 2.9
− −
Table1).Beingmodel-dependent,wedonotargueherewhether µα(masyr−1) −31.9±3.0 −29.9±3.6 (1)
the difference of 2.5 pc between the XO-2S and XO-2S dis- µ (masyr 1) 156.2 3.3 156.2 3.6 (1)
tances is real. We∼only note that the two values are compatible δ − − ± − ±
U(kms 1) 70.39+0.79 71.05+0.75 (5)
within the uncertainties and that our best estimate for the dis- − 0.81 0.76
− −
tanceofXO-2Nlocatesthestarfewparseccloserthanreported V(kms 1) 76.0+2.0 77.65+1.92 (5)
− − 2.0 − 1.88
byBurkeetal.(2007). − −
W(kms 1) 2.91+0.67 3.46+0.64 (5)
Distances, together with the stellar equatorial coordinates, − − −0.66 − −0.62
the proper motions, and the average radial velocities (Table 1 Relativemagnitudes(XO-2N-XO-2S)
and7),wereusedtoprovidenewestimatesofthegalacticspace ∆B(mag) 0.075 0.002 (6)
±
velocities(U,V,W)ofthestarswithrespecttotheLocalStandard
∆V(mag) 0.057 0.002 (6)
ofRest(LSR).Wethusassumed(i)aleft-handedsystemofref- ±
erence(i.e.thevelocitycomponentU ispositivetowardthedi- ∆Rc(mag) 0.040±0.002 (6)
rectionoftheGalacticanti-center),(ii)thetransformationmatrix 0.041 0.004 (7)
from equatorial to galactic coordinates taken from the Hippar- ±
∆r (mag) 0.038 0.001 (8)
cos catalogue (Perryman&ESA 1997), and (iii) the correction ′ ±
forthesolarmotionintheLSRderivedfromCos¸kunogˇluetal. ∆Ic(mag) 0.031±0.003 (6)
(2011).TheresultsarelistedinTable1.Burkeetal.(2007)dis- 0.030 0.006 (9)
cussedthekinematicsoftheXO-2systembyusingdataforXO- ±
2Nonly,andconcludedthat,despitethehighgalacticspaceve- References. (1) UCAC4 catalogue (Zachariasetal. 2013); (2)
locity,whichindicatesathickdiskmembership,thesuper-solar Bianchietal. (2011); (3) APASS all-sky survey Henden&Munari
metallicity is indeed representative of a thin disk membership. (2014) ; (4) 2MASS catalogue (Skrutskieetal. 2006); (5) This work;
Thissuggeststhatthe high-propermotionofthesystem canbe (6) This work (Serra La Nave Observatory); (7) This work (TASTE
related to an eccentric orbit with a maximum height above the project); (8) (Kundurthyetal. 2013), λ0=626 nm, 4epochs ; (9) This
galacticplaneof 100pc,thenconfinedtothegalacticdisk. work(APACHEproject).
∼
Weperformedanewanalysisofthegalacticorbits,basedon
theworkofBarbieri&Gratton(2002),andweconfirmthepre-
4. Analysisofthelightcurves
viousfindings.Resultsforasampleof1000orbitsshowthatthe
XO-2 system belongs to the thin disk, with a likely maximum
4.1.Out-of-transitphotometry
height above the galactic plane of 120 pc. This result does not
changesignificantlyunlessthedistanceofthestarsisatleast20 Figure 4 shows the relative photometryof the two components
pcmoreorlessourestimates.Ouranalysisalsoindicatesapos- oftheXO-2systemobtainedfromdifferentfacilities.Measure-
sibleeccentricorbit,withamodaleccentricityequalto0.42+0.09 ments were collected in the BVRI standard photometric pass-
0.06
(the asymmetric errorbarsare defined as the 10th and 90th−per- bandsandthe pointscorrespondingto the transitsof the planet
centile), and a minimum distance from the galactic center of XO-2Nbwere notincludedin the calculationof the intra-night
R =3.5 1.5kpc.IfweassumeR asaroughestimateofthe average values. The uncertainties associated to the measure-
min min
galactic r±egion where the two stars formed, this could explain mentsare the RMS of the pointscollected in each single night
theirsuper-solarmetallicity,becauseregionsclosetothegalactic of observation,and when only one measurementper nightwas
centeraremoremetalrichthanthesolarneighbourhood.More- available from SLN, we adopted the dispersion of all the data
over,this couldalso have modified the orbitof the binarywith astheerror.Theaveragevaluesforthemagnitudedifferencesin
timeduetotheinfluenceofthemoredensestellarenvironment each band are indicated in Table 1. We note that the indepen-
closetothegalacticcenter,withimportantconsequencesonthe dentmeasurementsinR(TASTEandSLN)andI(APACHEand
evolutionoftheorbitsoftheplanets. SLN) bandsagreewell each other.Our derivedrelativemagni-
tudeshavebeenusedinthedifferentialabundanceanalysisdis-
6 availableathttp://irsa.ipac.caltech.edu/applications/DUST/ cussedinSect.3.2.
Articlenumber,page8of26
M.Damassoetal.:GAPSV:GlobalanalysisoftheXO-2system
Table2.StellarparametersfortheXO-2componentsderivedfromthe
analysisoftheHARPS-Nspectraandfromstellarevolutionarytracks.
Theadoptedvaluesrepresenttheweightedmeanofindividualmeasure-
mentsobtainedwithdifferentmethods(exceptfortheprojectedvelocity
VsinI ,forwhichweadoptthemeasurementderivedfromtheRossiter-
⋆
McLaughlin effect), and their associated uncertainties are, for a more
conservativeestimate,theaverageoftheuncertaintiesoftheindividual
values.
Parameter XO-2N XO-2S Note
T [K] 5332 57 5395 54
eff ± ±
logg[cgs] 4.44 0.08 4.43 0.08
± ±
[Fe/H][dex] 0.43 0.05 0.39 0.05
± ±
Microturb.ξ[kms 1] 0.88 0.11 0.90 0.10
−
± ±
VsinI [kms 1] 1.07 0.09 1.5 0.3
⋆ − ± ±
Mass[M ] 0.97 0.05 0.98 0.05 (1)
⊙ ± ±
0.96 0.05 - (2)
±
Radius[R ] 1.01+0.1 1.02+0.09 (1)
⊙ 0.998−+00..00733 -−0.06 (2)
0.032
Age[Gyr] 7.9+−2.3 7.1+2.5 (1)
3.0 2.9
7.8+−1.2 -− (2)
−1.3 Fig.3.IronabundancedifferencebetweenXO-2NandXO-2Sversusχ.
Luminosity[L ] 0.70 0.04 0.79 0.14 forXO-2N:(2)
⊙ ± ± Dashedlineshowsthemean∆[Fei/H],whiledottedlinesrepresentthe
forXO-2S:(1) standarddeviation 1σfromthemean,whereσ=0.022dex.
±
Notes.(1)MatchingtheT ,[Fe/H],andloggtotheYonsei-Yaleevo-
eff
lutionarytracks.(2)MatchingtheT ,[Fe/H],andstellardensitytothe
eff
Yonsei-Yaleevolutionarytracks. very good instrumental stability and turned out to be useful to
strengthen the results obtained with broad-band measurements
Table3.Stellaratmospheric parametersandtheirdifferencesforXO- of the differentialmagnitudes.In fact, these are in goodagree-
2NandXO-2S.eediscussioninSect.3.2 mentwiththeAFOSCmeasurements,asshowninFig.5.Dueto
thehighscatterofthedataintheI-bandcausedbyfringing,we
Parameter XO-2N XO-2S ∆ did not include in the plot the measurementswith wavelengths
N S
− greaterthan 770nm.
Analysisbasedonequivalentwidths ∼
T [K] 5320 50 5330 50 10 71
eff
± ± − ±
logg[cgs] 4.46 0.08 4.41 0.12 +0.05 0.14 4.1.1. Rotationalperiods
± ± ±
ξ[kms 1] 0.89 0.14 0.93 0.04 0.04 0.15
−
[Fe/H]a[dex] 0.37±0.07 0.32±0.08 +−0.05±0.11 ThankstothetimespanoftheAPACHEI-bandobservationsand
± ± ± 800usefuldata pointsfor each component,we could analyse
Line-by-lineanalysis ∼
the photometric time series to search for periodic modulations
T [K] 5290 18 5322 25 32 31
eff ± ± − ± ascribabletotherotationofthestarsandproducedbybrightness
logg[cgs] 4.43 0.10 4.41 0.10 +0.02 0.14
± ± ± inhomogeneitiesdistributed on the stellar photospheres(which
ξ[kms 1] 0.86 0.06 0.93 0.05 0.07 0.08
[Fe/H]a[−dex] 0.37±0.07 0.32±0.08 +−0.05±0.11 areassumed to changeontime scales longerthanthe timespan
± ± ± ofthe observations).Takingintoaccountthatthe XO-2system
Differentialanalysis isquiteold(seeTable2)andthetwostarsaremagneticallyquiet
Teff (K] 5290 18 5325 37 35 8 (seediscussioninSect.6),theyareexpectedtohaverotationpe-
± ± − ±
logg[cgs] 4.43 0.10 4.420 0.094 +0.010 0.020 riodsofsometensofdaysandlowamplitudesinthelightcurve
± ± ±
ξ[kms−1] 0.86 0.06 0.93 0.03 0.07 0.07 variations. We used the Generalized Lomb-Scargle algorithm
[Fe/H]a[dex] 0.37±0.07 0.32±0.08 +−0.054±0.013 (GLS; Zechmeister&Kürster 2009) to search for periodicities
± ± ±
inourphotometrictimeseries,afterbinninginasinglepointthe
Notes.(a)Theironabundance[Fe/H]referstotheabundanceofFei. datacollectedduringeach night.The dataare availableon-line
attheCDS,eachcontainingthetimestamps,differentialmagni-
tudesandcorrespondinguncertainties.ResultsareshowninFig.
On the night between October 31, and November 1 2014, 6and7forXO-2NandXO-2S.Themajorpeaksatlowfrequen-
we took for 30 minutes a series of 60 s exposures of the XO- ciesintheperiodogramsareduetothelimitedtimebaselineof
2 field with the Asiago Faint Object Spectrographand Camera ourobservations.FortheNorthcomponentwefindapeakinthe
(AFOSC)atthe182cmAsiagoAstrophysicalObservatorytele- periodogramcorrespondingtoaperiodP=41.6 1.1days,witha
scope, while the field was close to the zenith. By placing both semi-amplitudeofthefoldedlightcurveof0.00±27 0.0004mag.
componentsina8′′ slitandusingagrismcoveringthespectral Becausethereis noevidenceofharmonicsin the±periodogram,
range3900-8000Å,weobtainedafinalsimultaneousdifferential the measurementscan be arrangedalmostperfectlywith a sine
spectrumbydividingeachotherthecombinedindividualspectra function.Toestimatethesignificanceofthebestperiod,weper-
ineachwavelengthresolutionelement.Thedifferentialspectrum formedabootstrapanalysis(withreplacement)byusing10,000
is shown in Fig. 5. These observationswere characterisedby a fake datasets derived from the original photometry, and in six
Articlenumber,page9of26
A&Aproofs:manuscriptno.xo2_system_arx
cases we founda powergreaterthanthatassociated tothe best Through the combined analysis of 14 new transit light
period in the real dataset. This corresponds to a False Alarm curves,we presenthererefined parametersfor the XO-2N sys-
Probability (FAP) equal to 0.06%, which makes the result for tem,togetherwith those in Table 2. Theindividualtransitlight
XO-2Nratherrobust. curves are shown in Fig. 8, and the full dataset combined in
Interestingly, to our knowledge other two systems with an a single, high-quality transit light curve is shown in Fig. 9.
old star have measured rotation periods thanks to Kepler data. The information about each dataset is summarized in Table 4.
The A and B components of the 16 Cygni system, which is Therawscientificframeswerefirstcorrectedforbias,darkand
estimated to be 6.87 2 Gyr old, have rotation periods 23.8+1.5 flat-fieldusingstandardsoftwaretools.TheSTARSKYpipeline
and 23.2+11.5 days r±espectively (Daviesetal. 2015). Anot−h1e.8r (see Nascimbenietal. 2013, and references therein) was em-
3.2 ployedto performdifferentialphotometrybetweenXO-2N and
example −is that of Kepler-77, a star-planet system 7 2 Gyr
± a set of optimally-weighted reference stars. The whole set of
old. The detected stellar rotation period is 36 6 days, with
± fourteen light curves was then simultaneously analysed with
sizeable changes in the shape of the light curve modulation
the JKTEBOP7 code (version 34, Southworthetal. 2004). The
(Gandolfietal. 2013). This case and especially that of 16 Cyg
adopted transit model parametrizes the stellar limb darkening
B, with uncertainties of several days, demonstrate that for old
(LD) with a quadratic law, i.e. with the two parameters u and
late-type stars the determinationof the true rotation periodcan 1
u . Assuming a circular orbit, seven parameters were left free
be particularly difficult. The intrinsic evolution of their surface 2
tovaryinthefittingprocedure:theperiodPandreferencetime
activeregionstakesplaceonatimescaleshorterthantherotation
T of the underlying transit ephemeris, the ratio k=R /R and
period,andthereforestarspotsbecomebadtracersofstellarro- 0 p ⋆
sum (R +R )/a of the fractional radii (a being the semi-major
tation.IntheSun,dependingonthephaseofthesolarcycle,the p ⋆
axisofthesystem),theorbitalinclinationi,andtheLDparam-
rotation period measured by fitting a sinusoid to the total solar
etersu andu .Leavingbothu andu freetovaryallowsusto
irradiance- a goodproxyfor the Sun-as-a-staropticalmodula- 1 2 1 2
avoidtoputanyaprioriassumptiononthestellar atmospheric
tion-variesbetween22.0and31.9days(thetruesynodicperiod
parameters.Ourestimatesareingoodagreementwiththosede-
being27.27days).Thevariationsareespeciallylargeduringthe
rivedbyinterpolatingthetheoreticaltablesofClaret&Bloemen
maximumofthesolarcyclebecauseseverallargeactiveregions
(2011) using our derived stellar atmospheric parameters (0.48
are simultaneously present on the solar surface and evolve on
and0.21foru andu ).Thetransitmodelwasfittedtoourdata
time scales of 1-2weeks, i.e. significantlyshorterthan the true 1 2
withanon-linearleast-squarestechnique;theuncertaintiesover
rotationperiod(e.g.,Lanzaetal.2003).
the best-fit parameters are then recovered with a bootstrap al-
Even if the rotation period of XO-2N appears well con-
gorithm, described by Southworth (2008). Other quantities of
strained, however the complications expected for old late-type
interest, such as the impact parameter b or the transit duration
starscouldactuallyhavepreventedustodetectanaccuratevalue.
T , were derivedthroughtheir analyticalexpressionsfromthe
ThisisindeedthecaseforthecompanionstarXO-2S,forwhich 14
fittedparameters.Thefinalbest-fitvalues,alongwiththeiresti-
a less convincingdetectionresulted in ourdata.The GLSperi-
matederrors,aresummarizedinTable5.Wegivetheestimates
odogram returned the highest peak signal at f=0.0384 days 1,
− asthemedianofthebootstrappeddistributions,withthe asym-
corresponding to P=26.0 0.6 days, with a semi-amplitude of
± metric errorbars defined as the 15.87th and 84.13th percentile.
0.0024 0.0005mag of the folded light curve. Also for XO-2S
Ourresultsareingoodagreementwiththoseintheliterature,in
±
thereisnoevidenceofharmonics.Thebootstrapanalysisinthis
particularwiththeresultsofCrouzetetal. (2012)basedonob-
caseledtoaFAP=2.6%,whichcastssomedoubtsonthereality
servationsmadewiththeHubbleSpaceTelescope.Wewereable
ofthesignal,eveniftheperiodisstillcompatiblewiththemea-
to provide an updated ephemeris (with an error on P of about
suredprojectedvelocityVsinI .Thesimilaritybetweentheob-
⋆ 0.025 s) that we also used as a prior for the radial velocity fit
servedperiodogramandtherecenteredspectralwindow(bottom
(Sect.5).Ourdeterminationofthestellardensityρ/ρ ,thanksto
panelintheupperpartofFig.7)strengthensourconfidenceon u andu beingleftfreetovary,iscompletelyindepe⊙ndentfrom
the stellar originof the periodicitywe found.However,it must 1 2
stellar models, and was used as input to interpolate within the
benoticedasecondpeakat0.029days 1 (P=34.5days)witha
− Y-Yevolutionarytracksandderiveanaccurateestimateforthe
slightly less power. This has some relevance in the discussion,
stellarmass,radiusandage(Sect.3).
sincearotationalperiodof34.5daysmatchesthegyrochronol-
ogyrelations(e.g.seeFig.13inBarnes2007)betterthantheone
of 26 days,and it is also closer to the estimates basedon mea- 5. RadialvelocitiesofXO-2N
surementsoftheactivityindexR (seeTable8).Moreover,in
analogywiththecaseoftheSun-′HasK-a-star,ifweassumeP=34.5 We observed the star XO-2N as part of a program aimed at
daysasthebestdetection,atrueperiodwhichisabithigherthan searchingforadditionalouterplanetsinthesystem,andwein-
that is nonethelessreliable. Thereforemore data are necessary, vestigated the Rossiter-McLaughlin (RML) effect by acquiring
possiblycoveringseveralactivitycycles,toreachamorerobust high-resolutionspectrawithHARPS-Nduringonetransitevent
conclusionconcerningthetruerotationalperiodofXO-2S. oftheplanetXO-2Nb.
5.1.Analysisoftheradialvelocitytimeseries
4.2.TransitphotometryofXO-2N
To refine the orbitalparametersof XO-2Nb, ournew HARPS-
FollowingthediscoveryofthehotJupiterXO-2Nb(Burkeetal.
N data were analysed along with the RVs obtained with the
2007),newtransitlightcurvesofXO-2Nwerecollectedtoim-
spectrographsHDS(HighDispersionSpectrograph)attheSub-
provethestellarandplanetaryparametersandsearchfortransit
aru telescope (Naritaetal. 2011), and HIRES (High Resolu-
timingvariations(Fernandezetal.2009;Kundurthyetal.2013;
tionEchelleSpectrograph)attheKecktelescope(Knutsonetal.
Crouzetetal. 2012), and the planet atmosphere was analysed
2014).TheoriginalspectroscopicdatafromBurkeetal.(2007)
with space-based (Machaleketal. 2009; Crouzetetal. 2012)
andground-basedobservations(Singetal.2011,2012). 7 http://www.astro.keele.ac.uk/jkt/codes/jktebop.html
Articlenumber,page10of26