Astron.Nachr./AN334,No.5,789–801(2013)/DOIpleasesetDOI! Long-term photometry of three active red giants in close binary systems: V2253 Oph, IT Com and IS Vir K.Ola´h1⋆,A.Moo´r1,K.G.Strassmeier2,T.Borkovits1,3,4,andT.Granzer2 1 KonkolyObservatoryMTACsFK,KonkolyThegeM.u.15/17,H-1121Budapest,Hungary 2 LeibnizInstituteforAstrophysicsPotsdam(AIP),AnderSternwarte16,D-14482Potsdam,Germany 3 BajaAstronomicalObservatory,Szegediu´tKt.766,H-6500Baja,Hungary 3 4 ELTEGothard-Lendu¨letResearchGroup,H-9700Szombathely,Hungary 1 0 2 Received2012,accepted2012 Publishedonlinelater n a J Keywords Binaries:spectroscopic,stars:fundamentalparameters, stars:late-type,stars:rotation,techniques:photom- 5 etry,starspots 1 Wepresent and analyze long-term optical photometric measurements of thethree activestarsV2253Oph, ITComand ISVir.Allthreesystemsaresingle-linedspectroscopicbinarieswithanearlyKgiantasprimarycomponentbutindifferent ] R stagesoforbital-rotationalsynchronization.Ourphotometryissupplementedby2MASSandWISEnear-IRandmid-IR magnitudesandthenusedtoobtainmoreaccurateeffectivetemperaturesandextinctions. ForV2253OphandITCom, S wefoundtheirspectralenergydistributionsconsistentwithpurephotosphericemission.ForISVir,wedetectamarginal . h mid-IRexcesswhichhintstowardsadustdisk.TheorbitalandrotationalplanesofITComappeartobecoplanar,contrary p topreviousfindingsintheliterature.Weapplyamultiplefrequencyanalysistechniquetodeterminephotometricperiods, - andpossiblychangesofperiods,rangingfromdaystodecades.Newrotationalperiodsof21.55±0.03d,65.1±0.3d,and o 23.50±0.04dweredeterminedforV2253Oph,ITCom,andISVir,respectively.Splittingoftheseperiodsledtotentative r t detectionsofdifferentialsurfacerotationsofδP/P ≈0.02forV2253Ophand0.07forITCom.Usingatime-frequency s technique based on short-term Fourier transforms we present evidence of cyclic light variations of length ≈10yrs for a V2253Ophand5-6yrsforISVir.Asingleflip-flopeventhasbeenobservedforITComofduration2–3yrs.Itsexchange [ of the dominant active longitude had happened close toatime of periastron passage, suggesting some response of the 1 magneticactivityfromtheorbitaldynamics.The21.55-drotationalmodulationofV2253Ophshowedphasecoherence v alsowiththeorbitalperiod,whichis15timeslongerthantherotationalperiod,thusalsoindicatingatidalfeedbackwith 5 thestellarmagneticactivity. 3 4 (cid:13)c 2013WILEY-VCHVerlagGmbH&Co.KGaA,Weinheim 3 . 1 1 Introduction theuseofplatearchivesspanningdecadesorevenacentury. 0 3 As exampleswe mentionthatPhillips& Hartmann(1978) 1 Duringthelasttwodecadesnumeroussmallandspecialized determinedfirsttimelong-termstarspotactivityofBYDra : surveyswerecarriedoutinordertoidentifynewcoolstars andCCErifromHarvardplatesfromhalfacenturybetween v i acrosstheH-Rdiagramthatshowphotosphericand/orchro- 1900-1950,and Fro¨hlich et al. (2002,2006) used the Son- X mosphericsignsofmagneticactivity(e.g.Lockwoodetal. nebergplatecollectionforstudyingthelong-termbehaviour r 1997,Baliunasetal. 1995,Henryetal.1996,Strassmeieret ofthesolar-typesingle activestarEK Dra andthe K-giant a al. 2000,Henryetal.2000,DeMedeirosetal.2002,Wright primaryoftheclosebinaryHKLac,respectively.Vogtetal. etal.2004,Whiteetal.2007,Strassmeieretal.2012).The (2004)foundseveralnewvariablesontheSonnebergplate nextandstillongoingstepwastoobtainmoreaccurateba- collection of 34 years, most of them are early type stars sic astrophysicalparametersforthese stars and to conduct from late-B to A, showing slow long-term changes on the time-resolvedphotometricandspectroscopicmonitoringin timescale of a few thousanddays.Recently Balona (2012) orderto determine rotationalperiodsand orbital elements. found rotational modulation on A stars from Kepler data The seemingly most interesting objects were followed up whichmightoriginatefromstarspotsasonepossibility,and withdedicatedcampaignswithautomatictelescopesateven ifthiswouldbeverifiedthanlonger-termmagneticvariabil- higher cadence in order to sample their starspot evolution ity could also be behind the slow changes shown by Vogt (e.g. Ola´h et al. 1997, Ja¨rvinen et al. 2008, Berdyugina & etal.’s(2004)earlytypenewvariables.TheDASCH(Dig- Henry2007,Strassmeieretal.2008,Savanov&Strassmeier ital Access to a Sky Century at Harvard) project recently 2008,Korhonenetal.2010,Lehtinenetal.2012). resultedinthediscoveryofthreeK-giants’long-termvari- A valuable toolto extend the time-base of the datasets abilitybyTangetal(2010),whichareveryprobablyactive forstudyingthelong-termbehaviorofmagneticactivityis stars,similartothosestudiedinthispaper. ⋆ Correspondingauthor:e-mail:[email protected] (cid:13)c 2013WILEY-VCHVerlagGmbH&Co.KGaA,Weinheim 790 Ola´hetal.:Long-termphotometryofthreeactiveredgiantsinclosebinarysystems Ourmain aimis to findevidencefora relationofpho- shouldbean earlyK giant.Strassmeier(1994) andStrass- tosphericstarspotactivitywiththeevolutionarystateofthe meier et al. (1994) had this object on their target list and starortheorbitalpropertiesofthebinarysystem.Doesthe detectedstrongCaIIH&KemissionlineswhileROSATde- rearrangement of the internal structure at the base of the tectedthesysteminX-rays(Dempseyetal.1993).Thepho- red-giantbranchreflectontochangesof the dynamoprop- tometricvariabilitywithaperiodof64±1dayswasdiscov- erties?Or,cantidalinteractiondissipateorbitalenergyinto eredbyHenryetal.(1995).ThisperiodimpliedthatITCom magneticenergy?Inthispaper,wereportonfollow-upob- rotatesslightlyasynchronouswithrespecttotheorbitalpe- servationsofthree chromosphericallyactivebinaries,each riodofP =59.054d.Henryetal.(1995)alsonoticedthe orb containingaK-giantstar. strongly variable photometricamplitude between zero and The spectral types, effective temperatures, and radii 0m. 2. Glebocki & Stawikowski (1997) stated that the orbit (when known) of the primary stars in these binaries are is not coplanar, iorb = 13◦ while irot = 34◦; the authors similar to each other. However, their degree of rotational foundageneralmisalignmentoftheprimariesoftheasyn- synchronismtotheorbitalrevolutionismarkedlydifferent. chronsystemstotheirorbitsfromobservationsandstatisti- V2253Ophisashort-periodasynchronously-rotatingstarin calanalysis.However,theerrorofthe inclinationdetermi- abinarywithanorbitalperiod15timeslongerthanthestel- nations are between 5-9 degrees for systems like ITCom, lar rotation period. Its orbit is only slightly eccentric, if at whichmayquestionsomeoftheindividualresults.Finally, all,becausezeroeccentricityiswithin3σofthevaluegiven ITCom was observed with the ELODIE spectrograph at byFekeletal.(2002).Thus,theperioddifferencecouldnot OHP in the survey of Soubiran et al. (2003), who derived be due to pseudosynchronization.ITCom is also an asyn- atmospheric parameters, distance, and space velocities to- chronouslyrotatingstarbutitsrotationperiodislongerthan getherwithalargesampleofmostlyclumpgiantstars. theorbitalperiodbyabout10%.Itsinclinationoftherota- Thethirdtarget,ISVir(HD113816)isthebeststudied tional axis is coplanar to the orbit, which itself is highly ofthethreetargetsinthispaper.Fekeletal.(2002)summa- eccentric(Griffin1988).Thethirdsystem,ISVir,isasyn- rized the variousmeasurementsof this objectfromX-rays chronizedgiantinamarginallyeccentricorbit(Fekeletal. toopticalphotometryandspectroscopythatwaspublished 2002). until 2002. They obtained a new, well-determined orbital Inthefollowing,wefirstsummarizetherelevantlitera- solutionwithsynchronousrotationandfoundaverylowor- ture for the three binary systems in Sect. 2. In Sect. 3, we bitalandrotationalinclinationofthisSB1binary.From12 describeournewobservationsandpresentthetoolsthatwill yearsof photometry,cycliclightvariabilityofthe orderof later be appliedin Sect. 4 andSect. 5. Sect. 6 summarizes 7–8yrswas discovered.Detailed photosphericabundances anddiscussesourresults. ofISVirwerepublishedbyKatzetal.(2003)andbyMorel etal.(2003).Bothanalysisfoundnear-solarabundancewith perhapsaslightoverabundanceofNa,Mg,AlandCawith 2 The program stars:V2253Oph, ITCom, respectto Fe, by 0.1-0.2dex;butthe derivedFe/H ratio is and ISVir quitedifferent(0.04–0.09and−0.11±0.09,respectively. Table 1 summarizessome of the relevantbasic parameters forthethreesystemsforquickreference. 3 Observationsand data-analysismethods The light variability of V2253Oph (HD152178) was discovered by Hooten & Hall (1990) who determined a 3.1 NewAPTobservations rotational period of 22.35±0.05d from photometry. Their observationswere initiatedbythespectralclassification of We present altogether 3,191 new photometric VI data C Houk(1982)whonotedCaIIH&Kemissionandfoundthe points. All observations were taken with the 0.75m star to be a single-lined (SB1) spectroscopic binary. She Vienna-PotsdamAutomaticPhotoelectricTelescope(APT) alsoderivedaspectraltypeofG8/K0(buterroneouslyoflu- Amadeus at Fairborn Observatory in southern Arizona minosityclassV).Thecorrectspectraltypeoftheprimary (Strassmeier et al. 1997). For V2253Oph, 1,883measures and the firstorbitalelementswere publishedin the second wereavailablefrombetweenJune1997andJune2011.For CABS catalog (Strassmeier et al. 1993) showing a signifi- ITCom,698measuresfromDecember1999throughApril cantlylongerorbitalperiodof314.5d.Themostrecentor- 2010 and for ISVir 610 measures from December 1999 bitalelementsaregiveninFekeletal.(1999)andbasically throughJanuary2011couldbeused.Thesedataareplotted confirmedtheoriginalorbit.Strassmeier(1994)andStrass- in Fig. 1 for reference. Observations of V2253Oph from meieretal.(1994)determinedabsoluteemission-linefluxes the first half of 1997 were published earlier (Strassmeier ofthestrongCaIIH&KemissionsandROSATdetectedthe et al. 1999) but are added to the present dataset. An inde- X-rayfluxofthesystem(Dempseyetal.1993). pendentcontinuous12-yrphotometricdatasetforISVirwas ITCom (HD118234)wasfirstmentionedasa spectro- published and analyzedby Fekel et al. (2002) who started scopic binaryby Griffin (1988) who publishedan SB1 or- observing 10 years prior to our APT campaign with two bital solution with a high eccentricity (e=0.589).Together seasons overlapping. We combine these two data sets and withunpublishedphotometry,heconcludedthattheprimary use it forparts of the presentanalysis. Our AmadeusAPT (cid:13)c 2013WILEY-VCHVerlagGmbH&Co.KGaA,Weinheim www.an-journal.org Astron.Nachr./AN(2013) 791 a b c V2253 Oph IT Com IS Vir 1998 2000 2002 2004 2006 2008 2010 2000 2002 2004 2006 2008 2010 2000 2002 2004 2006 2008 2010 7.45 8.30 8.65 7.50 8.35 7.55 8.40 8.70 7.60 8.45 V 8.75 V 7.65 V 8.50 7.70 8.55 8.80 7.75 8.60 8.65 7.80 8.85 8.70 7.85 8.75 7.35 6.30 7.10 6.35 7.15 7.40 6.40 7.20 6.45 7.25 IC 7.45 IC 6.50 IC 7.30 7.35 7.50 6.55 7.40 6.60 7.55 7.45 6.65 7.50 1.25 1.10 1.15 1.15 V-IC 1.30 V-IC V-IC 1.20 1.20 1.25 1.35 1.25 1.30 51000 52000 53000 54000 55000 52000 53000 54000 55000 52000 53000 54000 55000 J.D. J.D. J.D. Fig.1 OverviewoftheVienna-PotsdamAPTlightcurvesofthethreebinariesinthispaper;a.V2253Oph,b.ITCom andc.IS Vir.TheAPTtookatleastoneVI datapointeveryclearnightforthepast15years.Inthefigure,targetstars C arearrangedincolumns.V magnitudeisplottedinthetoprow,I magnitudesinthemiddlerow,andV −I colorinthe C C bottomrow. photometry had larger than usual scatter in the years be- 3.2 Determiningphotometricperiods tween 2007 and 2009 due to target acquisition problems (see Strassmeier et al. 2012 for the details). However, the Throughout this paper, we use the program package typical externalprecisionof Amadeushas been around4– MuFrAn (Multiple Frequency Analysis; Kolla´th 1990) to 6mmagdependingonthebrightnessofthetarget.Photom- quantify rotational periods and its errors, which are deter- etryfromliteraturesourcesaresparsefortheothertwosys- minedincreasingtheresidualscatteroftheleast-squaresso- tems, V2253Oph and ITCom, and are not taken into ac- lutionsbyabout10%oftheprecisionofthedata.Another countbecausewewouldscrutinizethehomogeneityofthe packageTiFrAn(TimeFrequencyAnalysis;Kolla´th&Ola´h APT data. Note that the ISVir data of Fekel et al. (2002) (2009)allowsatime-frequencyanalysiswithseveralmeth- weretakenwiththeneighboringTSU0.4mT3APTatFair- ods if the datasets are long enough in order to check for bornObservatory(butwithdifferentcomparisonandcheck possible cycles of length of a few years. In this paper we stars)anditsbasicdatareductioncoefficientsaretherefore apply a Short-Term Fourier Transform (STFT), numerical practicallyidenticalandthetwodatasetsstraightforwardto detailsoftheprocedurearegiveninKolla´th&Ola´h(2009), combine. includingtheinputofdatawithvariable(rotational)modu- lationandinthepresenceofseasonalgaps.Thisalgorithm All data were taken differentially with respect wasalreadyusedforthedetectionofcyclicbehaviorofsev- to a nearby comparison star and a check star. A eralotheractivestars(e.g.Ola´hetal.2009). nearby bright star is used for (automatic) navigation. The comparison/check stars were HD152501/HD151179, HD118670/HD118244, and HD113448/HD112805 for 3.3 Fittingspectralenergydistributions(SEDs) V2253Oph,ITComandISVir,respectively.Asinglemea- surementisthemeanofthreereadingsbetweenthevariable We determine effective temperatures by fitting ATLAS-9 andthecomparisonstar.Skyreadingsthroughthesame30′′ models (Castelli & Kurucz 2003) to optical and infrared apertureasforallotherreadingsaretakenusuallybetween fluxes.OpticalfluxesatJohnsonBV andCousinsI wave- C thecomparisonandthevariablestarmeasuretoensuretime- lengthswerederivedfromourlightcurvesforthetimesof liness. The detailed data reduction and Johnson-Cousins maximum stellar brightness (presumably the least spotted transformationsforAmadeusweredescribedinGranzeret state). In orderto take into accountthe large variabilityof al.(2001).Wereferthereadertothisandsubsequentpapers stars,weassignedanuncertaintyof0.1magtoeachphoto- aswellasreferencestherein. metricvalue. www.an-journal.org (cid:13)c 2013WILEY-VCHVerlagGmbH&Co.KGaA,Weinheim 792 Ola´hetal.:Long-termphotometryofthreeactiveredgiantsinclosebinarysystems Table1 Theprogramstars. Parameter V2253Oph ITCom ISVir HD 152178 118234 113816 hVi(mag) 8.5 7.4 8.3 M-Kclass K0III K0III K2III Teff (K) 4900±100 4700 4700±77 logg (2.5) 3.1 2.45±0.21 vsini(kms−1) 28.8±1 5.3 5.9±1 Rsini(R⊙) 12.7 7 2.7 i(◦) 63 13 13±4 [Fe/H] ... −0.19 −0.11±0.09 Orbit SB1 SB1 SB1 Porb(days) 314.47±0.18 59.054±0.004 23.655±0.001 e 0.024±0.009 0.589±0.010 0.022±0.008 K(kms−1) 14.32±0.15 9.66±0.15 6.598±0.054 f(m) 0.096±0.003 0.0029±0.0002 0.00070±0.00002 a1sini(106km) 61.92±0.64 6.34±0.12 2.146±0.017 d(pc) 360±105 185±25 250±50 Sourceoforbitalelements:Fekeletal.(1999)forV2253Oph,Griffin(1988)forITComandFekeletal.(2002)forISVir.Thedistance is from the revised Hipparcos parallax (van Leeuven 2007). The other parameters are for orientation only and were taken from the CABS-IIIcatalog(Ekeretal.2008)andreferencestherein. Near-infraredJHK dataweretakenfromtheTwoMi- thegeneralapproachforclearexcessdetection,werequire s cronAll-SkySurvey(2MASS;Skrutskieetal.2006).These χ ≥3.0(e.g.Kennedy&Wyatt2012). λ dataweresupplementedbyphotometryfromtheWide-field InfraredSurveyExplorer(WISE)satellite.WeusedtheW1 band (centered on 3.4µm) photometry from the All-Sky Database (Wright et al. 2010) but rejected the W2-band 3.4 Estimatingstellarages (4.6µm)measurementsbecausetheseweresaturatedforall ofourtargets. Wefixedboththesurfacegravity(logg)andthemetal- Agesareestimatedfromacomparisonofthelocationinthe licity([Fe/H])duringthefittingprocedure.ForITComand H-R diagram with evolutionary tracks from Pietrinferni et for ISVir these values were set to the literature data cited al.(2006).StellarluminosityisbasedontheHipparcospar- in Table 1 (Morel et al. 2003). We then selected model allax (vanLeeuwen2007) andthe brightestmagnitudesof atmospheres from the ATLAS-9 grid that have the clos- ourV-bandlightcurves.Bolometriccorrectionswerecom- est metallicity and logg value to the quoted ones. In the putedbasedonT valuesusingtheformulaoutlinedinTor- eff caseofV2253Ophnoliteraturedataforgravityandmetal- res(2010), andwe adoptedan absolutebolometricmagni- licity were found. Therefore, we adopted solar metallic- tudeof4.73fortheSun.Luminosityerrorsaresimplyprop- ity andestimated the surfacegravityfrom its spectraltype agatedparallaxandmagnitudeerrors.Thetemperaturesand and luminosity class (see Table 2). All of our targets have itserrorsarethosefromtheSEDfits. good quality photometry in the mid-infrared W3 (12µm) and W4 (22µm) bands with signal-to-noise ratio >20:1. There is an indirect dynamical method to estimate the By utilizing the W3 and W4 bands, we are also able to ageofabinarysystem,oratleasttogivesomeconstraints search for dust emission around our targets. None of our forit.Themethodisbasedonacomparisonofthepresent- sourcesareaffectedbyanyknownartifacts.Asafirststep, dayorbitalconfigurationwiththetheoreticaltidalcircular- the predicted photospheric flux densities were derived for izationtimescales.AtthispointwerefertoVerbunt&Phin- the WISE wavelengths using the best-fit ATLAS-9 mod- ney (1995) who investigatedthe circularizationin binaries els. Then the significance levelof any infraredexcess was wheretheprimarycomponentisanevolved,redgiantstar. calculated for both mid-infrared WISE bands by χ = Fromtheequilibrium-tidetheoryofZahn(1977,1989)these λ (F −F )/σ ,whereF isthefluxinaspecific authors derived a practical formula (their Eq. [7]) which WISE pred λ WISE WISE band and F is the correspondingpredicted flux givestherateofcircularizationforeveryevolutionarystage pred valuefromATLAS-9.σ isthequadraticsumoftheuncer- of an evolved primary as a function of the mass-ratio and λ taintyofthemeasuredfluxdensity,theabsolutecalibration theorbitalperiod.Wenotethatthisformalismdependsonly uncertainty(takenfromtheExplanatorySupplementofthe weaklyonthe(uncertain)mass-ratiobutstronglyontheac- WISEAll-SkyDataReleaseProducts),andtheuncertainty ceptedevolutionarymodel,andthusonthemassofthepri- ofthepredictedfluxdensityinthespecificband.Following mary. (cid:13)c 2013WILEY-VCHVerlagGmbH&Co.KGaA,Weinheim www.an-journal.org Astron.Nachr./AN(2013) 793 Table2 Revisedstellarparameters. 10.0 V2253 Oph Parameter V2253Oph ITCom ISVir Vmax 8.61 7.45 8.30 1.0 y] Vmin 8.87 7.86 8.70 F [Jν 10-1 PhVrot−(dIaCyis) 12.12.55-51±.305.03 615.1.12±-10.2.34 213.1.570-1±.208.04 Cycle? 10yr flip-flop 5-6yr,2yr Teff = 5050 K Teff (K) 5050±200 4750±130 4600±90 10-2 E(B-V) = 0.24 E(B−V) 0.24±0.08 0.05±0.05 0.07±0.04 10.0 IT Com logL1(L⊙) 1.97±0.37 1.68±0.15 1.66±0.21 R1(R⊙) 12.7±5.4 10.3±1.8 10.7±2.6 M1(M⊙)a ≈3 ≈1.4 ≈1.3 1.0 M2(M⊙) 1.35: 0.29: 0.55: y] i(◦) ≈65 40±12 15±5 J F [ν a1(106km) 68.3 9.38 8.87 10-1 Constantvaluesusedwhencalculatingtherevisedparameters loggb 2.5 2.5 2.5 T = 4750 K 10-2 Ee(fBf-V) = 0.05 Fe/Hc 0.0 −0.2 0.0 distance(pc)b 360±105 185±25 250±50 10.0 IS Vir a Possible mass ranges are given in the text. b For IT Com and IS Vir Fe/H is from the literature, for IS Vir we suppose solar 1.0 30 abundance because of the contradicting (but close to the solar) y] 25 literature values. The logg of IT Com of Soubiran et al (2003) J F [ν 10-1 Excess [mJy]1120505 choarsrencotedpuHbilpisphaerdcoesrproarraallnadx.is considered uncertain. cFrom the 0 T = 4600 K -5 eff 10-2 10 15 20 25 E(B-V) = 0.07 Wavelength [µm] The position in the H-R diagram in Fig. 3a shows that 1 10 V2253Ophisquitemassive,about3M withamassrange Wavelength [µm] ⊙ between2.4–3.8M .Hence,thesystemcannotbeveryold, ⊙ maybe1GyrorevenlessaccordingtothemodelsofPietrin- Fig.2 Spectralenergydistributionsof V2253Oph(top), fernietal.(2006). ITCom(middle)andIS Vir(bottom).Thedotsaretheob- served fluxesfrom ourAPT data, from 2MASS, and from FromtheformulaofVerbunt&Phinney(1995)wefind WISE.ThelinesarethebestfitATLAS-9models.Effective that in case of the ≈314-dorbitalperiod, we could expect temperaturesand interstellarextinctionvaluesare given in a circularized orbit only as late as during the AGB phase eachpanel. of stellar evolution. Apparently, the star is not in a such late phase of its evolution. A possible solution is offered byVerbunt&Phinney(1995).Theysuggest,thatforsome 4 Masses andages circularizedverywidebinaries,apossibilitytoexplainthe circular orbit with such a massive primary is, that the un- seensecondarycouldbeawhitedwarf.Itisthoughtthatthe The spectral energy distributions (SEDs) of V2253 Oph, white dwarfforcesthecircularizationduringtheprimary’s IT Com and IS Vir are plotted in Fig. 2 and the result- redgiantphaseeitherduetothestrongertidalfrictiondur- ing effective temperatures and extinction values are listed ingtheepochwhentheprimaryradiusbecomesverylarge inTable2.DetailedH-RdiagramsaregiveninFig.3. or/and a more rapid mass-transferin case of a Roche-lobe fillingconfiguration. 4.1 V2253Oph The orbit of the V2253 Oph system is fairly wide, so thesuggestionofVerbunt&Phinney(1995)couldbecon- For V2253Oph, we found the near-IR and mid-IR fluxes sidered.Theminimumradiusoftheprimaryof12.7R to- ⊙ to be consistentwith pure photosphericemission of T = getherwiththerotationalperiodof21.55dandameasured eff 5050±200K.Thisvalueisclosetotherecentdetermination vsini of 28.8 kms−1 (Fekel et al. 1999) and 24 kms−1 of Bailer-Jones(2011) of4944+293 and4814+174K, using (Strassmeieretal.1994)yieldmoderatelyhighinclinations −336 −251 BVJHK photometryand parallax with two differentmeth- of the rotationalaxis of i ∼ 70◦ and 54◦, respectivelyfor ods. Our quite large E(B-V)=0.24±0.08 is also supported thesevsinivalues.Assumingacoplanarorbit,thissuggests by Bailer-Jones’ (2011) values of 0.18+0.11 and 0.14+0.06 amassratiobetweenq∼0.43–0.51resultinginasecondary −0.12 −0.10 from the two methods, respectively, Errors mark the 90% with a mass of about 1.3 M orbiting the 3 M primary. ⊙ ⊙ upper-andlowerboundsforbothparameters. Ontheotherhand,iftheprimarymassisjust2.5M ,then ⊙ www.an-journal.org (cid:13)c 2013WILEY-VCHVerlagGmbH&Co.KGaA,Weinheim 794 Ola´hetal.:Long-termphotometryofthreeactiveredgiantsinclosebinarysystems a) Table 3 Newly determined galactic coordinatesand ve- M=2.3 locities M=2.5 M=3.0 par.(unit) V2253Oph ITCom ISVir M=4.0 2.5 vrad(km/s)a −36.75±0.13 −19.13±0.08 21.25±0.04 U(km/s) −34.9±0.7 −63.8±9.0 12.4±1.5 V(km/s) −13.8±7.1 −88.6±13.2 −26.8±4.2 W(km/s) −19.9±5.7 −5.3±2.1 7.2±2.5 X(pc) 352.1 37.8 85.1 2 Y(pc) −29.9 1.9 −101.5 Z(pc) 67.5 180.6 210.5 b(deg.)b 10.81 78.18 57.82 1.5 aGamma velocitiesformthe spectroscopic orbits. bGalacticlati- tude. In thecalculation of theGalactic space velocity weused a right-handedcoordinatesystem(UispositivetowardtheGalactic 3.76 3.74 3.72 3.7 3.68 3.66 3.64 center, V ispositive in the direction of Galacticrotation, and W b) ispositivetowardthenorthGalacticpole)andfollowedthe gen- 2 eral recipe described in the Hipparcos and Tycho catalogs (ESA M=1.7 M=1.5 1997). X, Y, Z are the physical space coordinates, centered on M=1.3 1.9 M=1.2 the Sun, in the same directions as U, V, W.For the calculation M=1.1 of galacticspace motion components the coordinate, proper mo- 1.8 tion and trigonometric distance informationwere taken fromthe Hipparcoscatalogue(vanLeeuwen2007). 1.7 1.6 the secondary mass would be slightly less than 1.2 M . ⊙ More aboutthe mass of the secondary is in Sect. B in the 1.5 appendix. Theexistenceofsuchahighmasswhitedwarfhasbeen 1.4 proven by Kalirai et al. (2005), who studied white dwarfs 1.3 in the young (650 Myr) cluster NGC 2099. For WD 24 3.7 3.69 3.68 3.67 3.66 3.65 3.64 a present-day mass of 1.11 M was claimed and an ini- ⊙ c) tial mass of 4.43 M computed, although with consider- ⊙ 2 ableerrors.Thus,ascenarioofV2253Ophwithared-giant M=1.7 M=1.5 primary of 2.4 M⊙ and a white-dwarf secondary of about 1.9 MM==11..32 1.1 M , which should have been originally over4 M , is M=1.1 ⊙ ⊙ notimpossible,takingintoaccounttheyoungageofthesys- 1.8 tem,whichcouldhavebeenstilllongenoughforthepresent 1.7 secondarytoevolvetoahigh-masswhitedwarf. 1.6 4.2 ITCom 1.5 For IT Com, we found the measured mid-infrared fluxes to be consistentwith pure photosphericemission of T = 1.4 eff 4750±130K.Bailer-Jones’(2011)resultsof4821+337 and −333 1.3 4735+237K with two different methods are practically the 3.7 3.69 3.68 3.67 3.66 3.65 3.64 −213 sameasourresult,althoughwithquitehighuncertainty.The Fig.3 Positions of the three primary stars in the H-R nearzeroE(B-V)=0.05±0.05wefoundislowerthanthatof diagram (dots with error bars). From top to bottom, a. Bailer-Jones’s(2011)of0.14+0.12and0.11±0.09fromthe V2253Oph,b.ITCom,andc.ISVir.Theverticalaxisplots −0.13 twomethods,butstillarewellwithintherespectiveerrors. logarithmicluminosityinsolarunitsandthehorizontalaxis The position in the H-R diagram suggests a compara- islogarithmiceffectivetemperatureinKelvin.Thelinesare blelowmassof1.4M (Fig.3b)withafullrangeof1.2– tracks for different masses as indicated in the inserts and ⊙ 1.6M .ThePietrinfernietal.(2006)tracksindicateanage were computed for solar metallicity for V2253 Oph and ⊙ of≈1.6Gyrs.Largedispersioninthegalacticspaceveloc- ISVir,butinterpolatedto[Fe/H]=−0.2forITCom. ities of IT Com (Table 3) impliesthatit may notbe a thin disc star. Indeed, using the kinematic criteria proposed by Reddyetal.(2006)wefoundthatITComprobablybelongs The orbit has a high eccentricity which, together with tothethickdiscpopulation. thedynamicagebasedonVerbunt&Phinney(1995),sug- (cid:13)c 2013WILEY-VCHVerlagGmbH&Co.KGaA,Weinheim www.an-journal.org Astron.Nachr./AN(2013) 795 gests that IT Com is in some stage of H-shellburningand notyeton the AGB because thenthe orbitshouldhaveal- 0.03 V2253 Oph Prot=21.55 d Fourier spectrum readybeencircularized. 22.1 d 0.02 The calculated radius of 10.3 R⊙ and vsini of 0.01 5.5±1kms−1 (Henryetal.1995)constraintheinclination wofitthhethroetamtioosntalliakxeilsytocobpeltawneaerno3r4b–it5a4l◦in,icnlignoatoidonagorfeeamboeuntt mplitude 0.00 0.04 0.045 0.05 40◦,derivedinSect.Aintheappendix.Themassofthesec- A 1 Spectral window ondarystarisabout0.55M ,seeSect.Bintheappendix. ⊙ 0.5 0 -0.005 0 0.005 4.3 ISVir Frequency [c/d] Fig.4 MuFrAnperiodogramofV2253Ophfromthe V- For IS Vir, we derived Teff = 4600±90K, very similar to banddatauntiltheendof2005.Twosignificantperiodsare therecentdeterminationofBailer-Jones(2011)of4521+−145150 marked in the upper panel, the spectral window is plotted and 4574+106K. Our E(B-V)=0.07±0.04is practically the below. −161 same as that of Bailer-Jones’ (2011) of 0.03+0.07 and −0.03 0.06+−00..0046fromthetwomethods. 5 Starspot activity TheSEDfitsforthemid-IRbandsachievedχ =1.5 W3 andχ =2.7,implyingamarginalexcessdetectioninthe 5.1 V2253Oph W4 22-µmW4band.Recently,acomparableamountofexcess The V-band amplitude due to rotational modulation var- emission at W4 was announced for HD139357, a planet ied between 0m. 15 in 1997 at a time of faint average light hosting K4III giant star (Morales et al. 2012). Generally, and 0m. 10 in around2002 at a time of brightaverage light Jura(2004)explainedexcessIRAS25µmemissionbydust (Fig.1a).Thissuggestsarelationbetweentheoverallspot- releaseduringicesublimationofKuiperbelt-likeobjectsas tednessandthesizeofindividualspotsorspotgroupssim- the starevolvesfromthemainsequenceontothe redgiant ilartowhathasbeenseenonotherover-activestars.How- branch.Furtherinfraredobservationsareneededtoconfirm ever, the shape of the light curve of V2253Oph changes theexcessandtoclarifywhetheritisrelatedtocircumstel- continuously and smoothly between subsequent rotations. lar dust or can be explained differently,e.g. by the contri- This is indicative of rapid spot rearrangements typically butionofabackgroundgalaxyorbyanapparentpositional observed for the more active stars. Not surprisingly, the coincidencewithaninterstellardustcloud. MuFrAn periodogram reveals two closely-space periods, ThepositionofISVirintheH-Rdiagramiscompatible but only until the end of 2005, see Fig. 4. The dominant with a mass of 1.3 M (Fig. 3c), i.e. of similar mass than 21.55±0.03-d period with an average 0m. 06 full V am- ⊙ ITCom.Themassrangeis1.1-1.5andtheageofthesystem plitude is seconded by a 22.05±0.04-d period with a full iscertainlybeyond5GyrsfromtheHRD. 0m. 025 amplitude. We tentatively interpret this due to dif- ferentialsurfacerotationandestimatearelativeshearfactor Fekeletal.(2002)derivedasimilarmassof1.5M and ⊙ δP/P of≈0.02,i.e.a lap time oftheequatorwith respect claimsthesystemisnotyoungsincethestarisabout250pc tothepolesof≈314d. abovethegalacticplane.Thetidalcircularizationtimescale The rotation of the spotted primary is strongly asyn- doesnotprovideanyfurtherconstraintsforitsagebecause chronouswithrespecttoits314-dorbitalperiod.Notethat theorbital-periodlimitforcircularizationislongerorcom- theeccentricityisverylowanddoesnotexceedits3σ un- parable(28d)totoday’sorbitalperiod(23.6d). certainty,andcannotaccountfortheperioddifference.Yet ISVirhasaloworbitalinclinationofabout13◦accord- the lightvariationof two subsets of ourAPT data seem to ingto(Fekeletal.2002).Fromitsnewlycalculatedradius produceafairlyhomogeneouslightcurvewhenplottedver- of10.7R ,itsrotationalperiodof23.5d,andtheprojected sustheorbitalperiod(Fig.5).Itsamplitudesaremarginally ⊙ rotationalvelocitiesfoundinthe literature.Theinclination largerthan the “scatter” but show extrema (maximumand lies between 10–20◦; the range of inclination comes from minimum) always near periastron. The very first seasonal theuncertaintiesoftheradiusandthedifferentvelocitydata data set from 1997, which shows the highest rotational- collected in the CABS-III catalog (Eker et al. 2008). The modulationamplitudeof all of ourdata, appearsto have a star is seen nearly pole-on. As a consequence of the low minimumatorbitalphasenear0.5,halfbetweentheformal inclinationtheamplitudeoftherotationalmodulationisal- times of periastron and apastron (note that one observing wayssmallandlessthan0m. 1inV,ascanbeseeninFig.1c. seasonusuallycovers140nights,i.e.onehalfoftheorbital Themassofthesecondarycomponentmustbe≈0.55M , period).Possibly,the 1997season belongedto yetanother ⊙ seeSect.Bintheappendix,formore. orbit-inducedvariabilitypattern.Duetothesmalleccentric- www.an-journal.org (cid:13)c 2013WILEY-VCHVerlagGmbH&Co.KGaA,Weinheim 796 Ola´hetal.:Long-termphotometryofthreeactiveredgiantsinclosebinarysystems a b Fig.6 Time-frequencydiagramsforV2253Oph.a.fromtheV magnitudesand,b.,fromtheV −I colorindex.The C toppanelsshowtheobservationsasred(grey)dots,blue(black)dotsshowthedataprewhitenedwiththemainrotational periodanditsharmonics,redandblue(greyandblack)lines showsplineinterpolationstotherespectivedata.Noticethe strongestsignalsareataround10yrsfromV and≈2.5yrsfromV −I . C ity it is not very likely that this variation originates from Unfortunately, we are left with an inconclusive situa- magneticinteractionofthetwobinarycomponentsatperi- tion. It is tempting to interpretthe long-term V-brightness astron but nevertheless there remains some evidence for a variation due to a dynamo-induced change of the over- relationofthespotactivityontheprimarywiththeorbitof all spottedness, like for the solar 11-year cycle, and the thesystem. long-term V −I index variation due to an average spot- C temperature change. However,the two tracers should vary Thetime-frequencydiagramfrom15yearsofphotom- with the same “cycle” period or, alternatively, are not re- etryofV2253OphisplottedinFig.6.Themostprominent lated with each other. A bit frustratingly we conclude that long-term variation is even seen by eye and stretches over after 15 years of photometrywe need another 15 years of 10yrswithapeak-to-peakV amplitudeof0m. 1.Despitethat photometry. thetime-frequencydiagramgivesastrongsignalaround10 years,wecannotbesureofitscyclicnatureduetotherela- 5.2 ITCom tivelyshorttimecoverage.Halfofthisperiod,avariability of about4.5 years,also appearsin the time-frequencydia- The best-fit average photometric period for IT Com from gram.ThesimilarlyconstructedplotfromtheV −I color the11-yrV-banddatais65.1±0.3d(Fig.7).However,the C indexdoesnotshowanytracesofthedominant10-yrvari- photometryshowssignificantlight-curvechangesfromone ation but shows a weak signal around4.5 years. However, stellar rotation to the next (note that only two stellar ro- the largest amplitude in the V −I color index variation tations can be covered during one observing season) that C isreconstructedonthetimescaleof2.5years.Allthesepe- mayaffectitsperioddetermination.Anamplitudeincrease riodicities seem to be related to each otherbecause within oftherotationalmodulationbynearlyafactorofthreehap- their uncertaintiesof about0.5–1yrthey appearto be har- penedbetween2005and2006.In2002,asuddenphaseshift monics.NotethatthecolorscalesinthetwoplotsinFig.6 by≈0p.4waswitnessedduringtheobservingseasonwhich are the same, which allows to directly comparethe ampli- thenrevertedbacktotheoriginalphaseinearly2005.This tudeof thesignals.We also notethatdespitethe scatterof couldbetermedaclassicalflip-flopphenomenon(e.g.Elst- themeasurementswaslargerbetween2007–2009duetoan ner&Korhonen2005,Ola´hetal.2006).Ithasstartednear instrumental problem, the comparison-check data proofed theperiastronpassageofthetwobinarycomponentsatJD that the mean light levelwas not affected,so we conclude 2,452,406.734,markedinFig.8toppanelwheretheobser- thatthelargelong-termvariationisreal. vationsbefore,during,andafterthe phaseflipsare plotted (cid:13)c 2013WILEY-VCHVerlagGmbH&Co.KGaA,Weinheim www.an-journal.org Astron.Nachr./AN(2013) 797 Fig.8 Upper panel: ∆V magnitudes of IT Com versus Fig.5 Evidencefora relation of the spot activity on the JD for 2000–2010.The plot symbols and colors mark the primaryofV2253Ophandthebinaryorbit.Thetoppanel dataembracedbytheflip-flop.Thefirstphaseflipstartedin shows again the long-term ∆V magnitudes plotted versus early 2002 and flopped back by early 2005. Lower panel: Juliandatewhilethebottompanelshowsthesamedataplot- The same data as abovebutphased with the rotationalpe- tedversusorbitalphase(P =314.17d).Symbolsonboth riod of 65.1d. The plot symbols and colors are the same orb panelsrefertothesamedatapoints. as in the top paneland indicate the phase shift of the data withintheflipflop.Notethatthelowerpanelnormalizedthe differentialmagnitudesto the meanlightlevelofthe three subsets. IT Com Prot=65.1 d Fourier spectrum with different symbols (and color). The back-flip in 2005 0.06 66.6 63.9 was actually notdirectlyseen becauseit happenedoutside 0.04 theannualobservingwindowbutthephasecoherencethere- afterwas thesame asbeforethe flip andconstantuntilthe 0.02 end of the APT observations in 2010. The time duration mplitude 0.00 0.01 0.015 0.02 fbreotmwefleniptthoefllaosptwobassebrveatwtioeneno2f.02–020.47aynedartsh,ei.efi.r,sittoennedeind A 1 Spectral window 2005.Sofar,thephenomenondidnotrepeatandwecould notspeakofaflip-flop“cycle”butstatethatflipflopsmay 0.5 occuraperiodicallyanditslengthisnotrelatedto itsreoc- currence. 0 Systematicseasonalchangesarealsoobviousfortheav- -0.005 0 0.005 Frequency [c/d] erage V-light level in Fig. 1b (replotted on a larger scale Fig.7 MuFrAnperiodogramofITComfromtheV-band in Fig. 8, top panel)andamountto almost0m. 15. Unfortu- data. Three significant periods are marked in the upper nately, the 11 years of photometryis insufficientto search panel,thespectralwindowisplottedbelow. for cyclic changes; a simple look at the data suggests a timescale of a cycle (if any) longer than the length of the datasetitself. Thelong-termV −I lightcurvetracesthe C www.an-journal.org (cid:13)c 2013WILEY-VCHVerlagGmbH&Co.KGaA,Weinheim 798 Ola´hetal.:Long-termphotometryofthreeactiveredgiantsinclosebinarysystems withtheemergence,ordecay,ofanactivelongitudeaspart IT Com of the flip-flop phenomenon. If we nevertheless assume a solar-likedifferentialrotationlaw,adoptalatituderangefor 7.50 thespotsof50◦,andassumetheshortestperiodbelongsto theequatorandthelongesttoalatitudeof50◦,wemayes- 7.60 timate a lower limit for the relative differential rotation of δP/P ≈ 0.07.Therefore,the lap time ofthe equatorwith V 7.70 respecttothehighlatitudesis≈90d. 7.80 5.3 ISVir 6.30 0.010 IS Vir Prot=23.50 d Fourier spectrum 6.40 C 0.005 I 6.50 6.60 mplitude 0.000 0.035 0.04 0.045 0.05 A 1 Spectral window 1.10 0.5 1.15 0 C -0.01 -0.005 0 0.005 0.01 V-I Frequency [c/d] 1.20 Fig.10 MuFrAnperiodogramofISVirfromthe V-band data. One period is markedin the upperpanel, the numer- 1.25 ouspeaksarounditindicateother,slightlydifferentperiods. Note the low amplitude of the variations, see the text for 0.0 0.1 0.2 0.3 0.4 0.5 0.6 0.7 0.8 0.9 1.0 more.Thespectralwindowisplottedbelow. rotational phase Fig.9 PhasedlightandcolorindexcurvesofITComwith the 65.1 days period, phase zero is at periastron. The plot Fekel et al. (2002) published 12 seasons of BV pho- symbolsarethe same as in Fig. 8. V −IC isredderwhen tometry from the TSU T3 APT of which the last two sea- thestarisfainter. sonsoverlapwithourfirsttwoofthe11seasonsofVienna- PotsdamAPTdata.Basedonthecomparisonofthe76com- mon nights during the two overlapping seasons, we com- average V-light level in the sense that redder V −I oc- C bined the two datasets with a shift of 2.07mag. in V (due curs when the system is brighter.The V −I variation is C to two differentcomparisonstars) to a 21-yeartime series inphasewiththecolorvariability,atthesametimetheam- forfurtheranalysis.Afterpre-whiteningwiththelong-term plitudeoftherotationalcolormodulationdoesnotrelateto high-amplitude variations with a 6250 days period and its theamplitudeoftheV (orI )modulation,whichindicates C fourharmonics(whichisapurelymathematicaltooltoget aconstantspottemperature,seeFig.9. ridofthelong-termvariation)wefindameanphotometric Apart from the dominant rotational signal of 65.1d periodof23.50±0.04dfromJohnsonV thatweinterpretto (with 0m. 11 full V amplitude), two more significant peri- be the stellar rotationperiod.Itis very close butnotequal odsshowup,seeFig.7.Oneisslightlylonger,66.6dwith tothe23.655±0.001dorbitalperiod(Fig10). anamplitudeof0m. 072,andoneslightlyshorter,63.9dwith Duringthefirstthreeandthelastthreeseasonswhenthe an amplitudeof0m. 068.Bothperiodshavean approximate starwasthebrightest,theanalysisdoesnotgiveclearperi- errorof±0.3dandarehighlysignificant. ods. The data at medium brightness, i.e., without the first Between2000–2002the65.1-dperiodisdominant(red and last three seasons and the season of the minimum,re- dotsinFig.8),whereasduringthenexttwoyearsthe66.6-d sultsintwoclosely-spacedperiodsof23.64and23.50days periodgavethebestfit(bluecrossesinFig.8).From2005 with an error of 0.3 days and a full Fourier amplitude of onwardsagainthe65.1-dperiodwasdominantandresulted 0.024mag.Finally,theyearofminimumlightshowsapho- intheleastscatteredlightcurve.Thisindicatesthattheindi- tometricperiodof22.87±0.80dayswithafullFourieram- vidualperiodsstemnotfromlongitudinalmigrationdueto plitudeof0.028mag.Consideringthelowinclinationofthe somesortofadifferentiallyrotatingsurfacebutarerelated star (≈ 15◦) this result suggests, that at maximum light (cid:13)c 2013WILEY-VCHVerlagGmbH&Co.KGaA,Weinheim www.an-journal.org