Astronomy&Astrophysicsmanuscriptno.tx-psc (cid:13)c ESO2013 December11,2013 Catching the fish – Constraining stellar parameters for TXPsc ⋆ using spectro-interferometric observations D.Klotz1,C.Paladini1,J.Hron1,B.Aringer1,S.Sacuto1,2,P.Marigo3,andT.Verhoelst4,5 1 DepartmentofAstrophysics,UniversityofVienna,Tu¨rkenschanzstrasse17,A-1180Vienna e-mail:[email protected] 2 DepartmentofPhysicsandAstronomy,DivisionofAstronomyandSpacePhysics,UppsalaUniversity,Box516,75120,Sweden 3 DepartmentofPhysicsandAstronomyG.Galilei,UniversityofPadova,Vicolodell’Osservatorio3,I-35122Padova,Italy 4 BelgianInstituteforSpaceAeronomy(BIRA-IASB),Ringlaan-3-AvenueCirculaire,B-1180Brussels,Belgium 3 5 InstituutvoorSterrenkunde,KULeuven,Celestijnenlaan200D,3001Heverlee,Belgium 1 Received;accepted 0 2 ABSTRACT n a Context.Stellar parameter determination is a challenging task when dealing with galactic giant stars. The combination of dif- J ferent investigation techniques has proven to be a promising approach. We analyse archive spectra obtained with the Short- 3 Wavelength-Spectrometer(SWS)onboardofISO,andnewinterferometricobservationsfromtheVeryLargeTelescopeMID-infrared Interferometricinstrument(VLTI/MIDI)ofaverywellstudiedcarbon-richgiant:TXPsc. ] Aims.Theaimofthisworkistodeterminestellarparametersusingspectroscopy andinterferometry.Theobservationsareusedto R constrainthemodelatmosphere,andeventuallythestellarevolutionarymodelintheregionwherethetracksmapthebeginningof S thecarbonstarsequence. . Methods.Twodifferentapproachesareusedtodeterminestellarparameters:(i)the‘classic’interferometricapproachwheretheeffec- h tivetemperatureisfixedbyusingtheangulardiameterintheN-band(frominterferometry)andtheapparentbolometricmagnitude; p (ii)parametersareobtainedbyfittingagridofstate-of-the-arthydrostaticmodelstospectroscopicandinterferometricobservations. - Results.Wefindagood agreement betweentheparameters of thetwomethods. Theeffectivetemperatureand luminosityclearly o place TXPscinthe carbon-rich AGB star domain intheH-R-diagram. Current evolutionary trackssuggest that TXPscbecame a r t C-starjustrecently,whichmeansthatthestarisstillina‘quiet’phasecomparedtothesubsequent strong-windregime.Thisisin s agreementwiththeC/Oratiobeingonlyslightlylargerthan1. a [ Keywords.Stars:AGBandpost-AGB-Stars:atmospheres-Stars:carbon-Stars:fundamentalparameters-Techniques:interfero- metric-Techniques:spectroscopic 1 v 4 0 1. Introduction Catalogue of Variable stars (Samusetal. 2009) with a mean 4 brightnessof ∼5mag and a peak-to-peakamplitudeof 0.4mag 0 Theasymptoticgiantbranch(AGB)isthelateevolutionarystage in the V-band (Jorissenetal. 2011). Distance estimates range 1. of low- to intermediate- mass stars (1 − 8M⊙). On the early- from 275 to 315 pc (vanLeeuwen 2007; Claussenetal. 1987; AGBthecarbon-to-oxygen-ratioissmallerthanone.Aftersev- 0 Bergeat&Chevallier 2005). Different ISO/SWS spectra of 3 eralthermalpulses,theatmospheresofobjectswithmassesbe- TXPsc(Jørgensenetal.2000;Gautschy-Loidletal.2004)show 1 tween 1 − 4M⊙ will very likely turn from oxygen-rich into that there is a difference in the 3µm feature between 1996 : carbon-rich because of the third dredge-up (Iben&Renzini v and 1997. With the help of plane-parallel, hydrostatic models 1983). Good estimates of stellar parameters are needed for a i Jørgensenetal. (2000) interpret this difference as a change in X profound understanding of the evolution of this stage. Their temperature of ∼100K. The photometry and spectra were suc- determination is a challenging task because of the complex- r cessfully modeled by Gautschy-Loidletal. (2004) with dust- a ity of the atmospheres of these objects. It is demonstrated that free dynamical models that reproduce the region between 1 − the combined use of spectroscopic and interferometric observ- 5µm. They claim that the region between 8−9µm is affected ing techniques can efficiently help to ascertain stellar parame- almostsolelybyCS.Theauthorssuggestthatobservationsfrom ters (e.g. Wittkowskietal. 2001, 2008, 2011; Neilson&Lester 8−9µmshowthedeepphotosphere,whiletheexpectedabsorp- 2008;Paladinietal.2011;Sacutoetal.2011a;Mart´ı-Vidaletal. tionofHCNandC H originatingfromthehigherlayersisnot 2 2 2011). At the same time these observationsprovideconstraints observed.Asthisobjectisalmost(carbon)dust-freewemayin- forexistingmodelatmospheres:e.g.COMARCS(Aringeretal. feritbecameacarbonstarquiterecently.Therefore,thisstaris 2009),Ho¨fneretal.(2003),PHOENIX(Hauschildtetal.1999), a perfectcandidate to constrain the regionwhere the transition ATLAS(e.g.Kurucz1993),CODEX(Irelandetal.2008,2011). fromoxygen-tocarbon-richoccurs. TXPsc is one of the brightest and closest carbon-rich AGB In this work we present a study of the atmosphere and a stel- stars. It is listed as an irregular variable in the General lar parameter determination for TX Psc. We combine spectro- interferometric observations of VLTI/MIDI and spectroscopic ⋆ Based on observations made with ESO telescopes at Paranal observations from ISO/SWS and compare them to geometric, Observatory under program IDs 74.D-0601, 60.A-9224, 77.C-0440, hydrostatic,andevolutionarymodels. 60.A-9006,78.D-0112,84.D-0805 1 Klotzetal.:ConstrainingstellarparametersforTXPsc calibratorobservationsis<0.2,(ii)thespectraltypeofthecali- bratorisnotlaterthanM0.Consideringthesecriteria,fivespec- traarederived(flaggedwith‘a’inTable1). TXPsc is almost unresolved for baselines shorter than 32m. Therefore, we expect that the star is unresolved by the single- dishUTs.Thus,mostofthemid-infrared(mid-IR)fluxislocated withinthefield-of-view(FoV)oftheUTsandconsequentlyalso in the FoV of the ATs and ISO, which makes the spectra fully comparabletoeachother. 2.2.ISOspectra Three spectra of TXPsc were observed with the Short- Wavelength-Spectrometer (SWS, deGraauwetal. 1996) on- board of ISO (Sloanetal. 2003; Jørgensenetal. 2000). Two spectra have a resolution of R ∼ 200 and range from 2.36 − Fig.1.N-bandspectrallydisperseduv-coverageoftheMIDIob- 45.35µm.Theotherspectrumrangesfrom2.45−45.20µmand servationsofTXPsc.Colourlevelsrangefrom8-11.5µm(black hasaresolutionofR ∼ 2000whichhasbeenbinnedtotheres- toblue,respectively)withastepsizeof0.5µm.Northisupand olutionoftheotherspectra.FortheISOspectraamultiplicative Eastisleft. error of ±10% is assumed from 2.38-4.05µm and ±5% after- wards(Sloanetal.2003). A description of observations and data reduction is given in Sect.2. Modelsand approachesto derive syntheticobservables 2.3.Interferometricandspectroscopicvariability are presented in Sect.3. The stellar parameters are derived in IntheleftpanelofFig.2thecalibratedMIDIspectraareover- Sect.4 and compared to state-of-the-art evolutionary tracks in plottedtotheISO/SWSspectra.ThefluxleveloftheMIDIob- Sect.5.AsummaryoftheresultsisgiveninSect.6. servations is the same (within the error bars) as the ISO/SWS spectra taken at 1996 Nov 26 and 1997 Dec 11. This suggests that no significant cycle-to-cycle variation is expected in the 2. ObservationsandDataReduction mid-IR. The right plot of Fig. 2 shows part of the lightcurve Sections2.1 and 2.2 discuss the interferometric and spectro- of TXPsc in V taken from AAVSO. The spectra from 1996 scopic observations obtained with VLTI/MIDI and ISO/SWS, Nov 26/1997 Dec 11 were observed at a ’local’ visual maxi- respectively. Section2.3 discusses possible cycle-to-cycle and mum/minimum,respectively.Inthefollowingwewillusethese intra-cyclevariabilityofthedata. two spectra to derive the parameters of the star at different phases.AswedonothaveanyinformationontheV magnitude of the spectrum from 1997 May 24, this spectrum is not used 2.1.MIDIvisibilitiesandspectra in the subsequent sections. Additionally, the phase for all the TX Psc was observed in 2004 with the 8.2m Unit Telescopes MIDI observations is unknown. However, AAVSO visual esti- andin2005,2006and2011withthe1.8mAuxiliaryTelescopes matesfromonespecificobserverrevealedthatnoneoftheMIDI oftheVeryLargeTelescopeInterferometerMIDI(Leinertetal. observationsweretakenatvisualminimaormaxima. 2003). MIDI covers the N-band and provides spectrally dis- Simulationswithdust-freemodelatmospheresshowthattheef- persedvisibilities,differentialphasesandfluxes(resolutionR= fect of pulsation on the N-band UD-radius is smaller than 0.1 230forobservationsin2004,R = 30forobservationsin2005, masat1kpc(Paladinietal. 2009).Fora star atthedistanceof 2006and2011). TX Psc thiswouldbe ∼0.3mas. Thisvalueisbeyondthelimit The journalof available MIDI observationsis givenin Table 1 ofresolutionofMIDI.Therefore,wedonotexpectanyobserved (electronicversiononly).Theuv-coverageisplottedinFig.1. intra-cyclevariabilityeffectthatislargerthantheerrors.Thisal- Data are reduced using MIA+EWS 1.7.11 (Jaffe 2004; Ratzka lowstocombineallcalibratedvisibilitiesforthefurtheranalysis. 2005; Leinertetal. 2004). Observations are not used if one or more of the selection criteria discussed in Klotzetal. (2012a) are violated (observations that are not boldfaced in Table1). 3. Modeldescription Uniform-diskangulardiametersandIRAS12µmfluxofthecal- In thefollowingwe presentthe differentclasses ofmodelsand ibratortargetsaregiveninTable2(electronicversiononly). thederivationofsyntheticobservablesthatwillbecomparedto As only one suitable calibrator is available per observation, a observationsinSect.4. standard multiplicative error of 10% is assumed for the cali- brated visibilities (Chesneau 2007) . Some calibrated visibili- tiesatbaselinesshorterthan∼30maresignificantlynoisierand 3.1.Hydrostaticmodels sometimeslargerthanunityafter11.5µm.Therefore,inthefol- Observed spectra and visibilities are compared to the grid lowing,spectro-interferometricobservationsareconsideredonly of spherical hydrostatic model atmospheres and spectra forwavelengthsshorterthan11.5µm. of Aringeretal. (2009). These models are computed with Forthecalibratedspectraadditionalselectioncriteriaareapplied COMARCS andaregeneratedassuminghydrostaticlocalther- (Chesneau2007):(i)theairmassdifferencebetweenscienceand malandchemicalequilibrium.Themolecularandatomicopaci- 1 http://www.strw.leidenuniv.nl/∼jaffe/ews/MIA+EWS- tiesaretreatedintheopacitysampling(OS)approximation.The Manual/index.html parametersthat characterize a modelare: effective temperature 2 Klotzetal.:ConstrainingstellarparametersforTXPsc Fig.2. Left: MIDI flux (error bars) for differentdates. Overplottedare the three ISO/SWS spectra (full lines) of TXPsc. Right: AAVSOV-bandlightcurveofTXPsc.VerticallinesmarktheobservationdateoftheISOspectra.Colorsarethesameasintheleft figure. T ,metallicityZ,surfacegravityg,massM,andcarbontooxy- surfacechemicalabundancesthatmaysignificantlyvarydueto eff genratioC/O. thethirddredge-upepisodesandhot-bottomburning.Thetransi- Forthisworkwelimitthesampletomodelshavingsolarmetal- tionfromC/O <1toC/O >1isfollowedaccurately,inparticu- licityastheeffectofmetallicityisexpectedtobesmallforlow- larinthenarrowrangefrom0.95≈C/O≈1.05,whereanabrupt resolutionspectroscopy.Additionally,thereisnoindicationfor changeinthemolecularchemistryandopacityisexpectedtooc- largelynon-solarmetallicityfromotherpropertiesofthisstar. cur(seefigures11and16inMarigo&Aringer2009).Thispoint The spectra cover the following parameters: 2400 ≤ T ≤ isparticularlyrelevantinthecontextofthepresentwork,asTX eff 4000Kwithstepsof100K;Z/Z⊙ =1;−1.0≤log(g[cms−2])≤ PscisfoundtohaveasurfaceC/Oslightlyaboveunity. +0.0; M/M = 1,2; C/O= 1.05,1.10,1.40,2.00. In order to ⊙ get precise estimates of T , additional model atmospheres eff were producedresulting in a grid spacingof ∆T=10K. All the 4. Stellarparameterdetermination mainmolecularopacitiestypicalforC-starswereincluded:CO Stellar parametersforTXPsc were determinedbya numberof (Goorvitch&Chackerian 1994), C (Quercietal. 1974), HCN 2 authors.AsummaryisgiveninTable3. (Harrisetal. 2006), CN (Jørgensen 1997) in the form of line In the following two different approaches will be used to de- lists,whileC H andC (Jorgensenetal.1989)asOSdata.CS 2 2 3 termine the stellar parameters of TXPsc: (i) in Sect.4.1 we isnotincludedduetothelackoflinelistsandOSdata.Synthetic use the ‘classical approach’of interferometristswhere spectro- spectra with a resolution of 18000 are computed in the wave- interferometric observations are used to determine the ef- lengthrange0.8−25µm.Thespectraareconvolvedinorderto fective temperature T ; (ii) in Sect.4.2 we follow the ap- getthesameresolutionastheobserveddata. eff proach of Paladinietal. (2011, P11 hereafter) where spectro- AmongtheoutputofthesphericalradiativetransfercodeCOMA scopic/interferometric measurements are compared to hydro- is the monochromatic spatial intensity profile. This profile is static models to determine T and C/O ratio and to constrain used to calculate a syntheticvisibility profile in the mid-IRfor eff massandlogg. asubsetofthemodelsinthegrid.Adetaileddescriptionofthe computation of the visibility profiles is given in Paladinietal. (2009). 4.1.Approach1:Geometricmodels Interferometryis a powerfultoolforconstrainingthemorphol- 3.2.Evolutionarytracks ogyandsizeofstarsandtheirenvironments. A deviation from spherical symmetry for the circumstellar en- The luminosity and effective temperature that are determined vironmentof TXPsc was detected by several authorsat differ- fromhydrostaticmodelsarecomparedtothermally-pulsing(TP) ent spatial scales and wavelengths(e.g. Cruzale`besetal. 1998; AGB evolutionary tracks from Marigo et al. (in prep). We se- Raglandetal. 2006; Sacutoetal. 2011b; Jorissenetal. 2011, lectedTP-AGBsequenceswithan initialscaled-solarchemical Hron et al in prep.). Clearly, the circumstellar environment of composition (Z = 0.014, Y = 0.273), where Z and Y denote TXPsc isverycomplexandmostlikelyunrelatedeffectsshape themassfractionsofmetalsandhelium,respectively.TP-AGB themorphologyindifferentregions,resultingina largevariety evolutionary calculations are carried out from the first thermal ofstructures.To studythegeometryoftheinnermostregionof pulse - extracted from the PARSEC database of stellar models thestarintheN-band,thegeometricalmodelfittingtoolGEM- (Bressanetal. 2012) - to the completeejectionof the envelope FIND(Klotzetal.2012b)wasusedtofittheMIDIinterferomet- duetostellarwinds.TheTP-AGBtracksarebasedonnumerical ricobservationsofTXPsc.AsphericalUD-modelisabletore- integrationsofcompleteenvelopemodelsinwhich,forthefirst producethecalibratedvisibilities,i.e.TXPsccanbeassumedto time, molecular chemistry and gas opacities are computed on- besphericallysymmetricinthemid-IRatspatialscalesprobed the-flywiththeÆSOPUScode(Marigo&Aringer2009).This by our MIDI observations. The reason why the asymmetrical guaranteesa fullconsistencyofthe envelopestructurewith the structuresdetectedbyotherworksarenotobservedisthatthey 3 Klotzetal.:ConstrainingstellarparametersforTXPsc Table3.PublishedstellarparametersofTXPsc. Reference T logg Mass C/O θ λ d eff θ [K] [M ] [mas] [µm] [pc] ⊙ Laskeretal.(1973) 9.00 0.66 deVegt(1974) 8.00 0.71 Dunhametal.(1975) 10.20 0.69 Lambertetal.(1986) 3030 0.0 1.03 Claussenetal.(1987) 280 Quirrenbachetal.(1994) 2805 11.20 0.7-0.8 Richichietal.(1995) 8.38 0.55-3.60 Dycketal.(1996) 2921 11.20 2.2 Jørgensenetal.(2000) 3000 -0.5 1.02 Ohnakaetal.(2000) 3000 1.07 3100 1.17 Bergeatetal.(2001) 3115 Harrisetal.(2003) 3050 0.0 1.02 Gautschy-Loidletal.(2004) 3200 -0.3 1 1.10 Bergeat&Chevallier(2005) 3125 315 Raglandetal.(2006) 9.89 1.65 vanLeeuwen(2007) 275+34 −26 Fig.3. Left:Calibrated visibilities(symbols)versusbaselinelengthforthree differentwavelengths.Thelines representthe best- fittingUD-modelatthegivenwavelength.Right:Spectrallydispersedangulardiameterpluserrorsfromthebest-fittingUD-model (darkgreyshadedarea).Thelightgreyshadedareamarkstheregionthatisomittedforthemeandiameterestimation. wereeitherdetectedatdistancesoutsidetheFoVofMIDIorat F.Kerschbaumbyfitting acombinationof blackbodiesto near- smaller spatial scales beyondthe detectionlimitof MIDI. This IRandIRASdata2. is supported by the differential phase measurements of MIDI, Variousdefinitionsfortheradiuscanbefoundinliterature(c.f. whichdonotshowanydeviationfromzero.Thecalibratedvisi- reviews by Bascheketal. 1991; Scholz 2003), where the most bilitiesareplottedtogetherwiththebest-fittingUD-modelinthe commonlyusedradiusinatmosphericmodelingistheRosseland leftpanelofFig.3forthreedifferentwavelenghts. radius. It is defined by the distance between the center of the The right panel of Fig.3 shows the wavelength dispersed di- star and the layer having Rosseland optical depth τ = 2. ross 3 ameter calculated with the UD-model(dark grey shaded area). Thisradius,however,isnotanobservablequantityandobserved The star appears larger between 8 and 9µm. According to radii have to be converted by using model considerations. In Gautschy-Loidletal. (2004) CS is affecting this wavelength the following we will derive this conversion factor for hydro- range. Because of this molecular contamination, this region is staticC-starsbyusingasubsetofthehydrostaticmodelsinthe omittedandameanangulardiameterofθ=10.51±0.70masis gridof Aringeretal. (2009) to derivea mean UD-radiusin the calculatedbyaveragingthediameterfrom9-11.5µm. mid-IR (9 − 11.5µm). This mean UD-radius is plotted versus theRosselandradiusofthecorrespondinghydrostaticmodelin Fig.4.Thereisaclearcorrelationbetweenthetworadii,yielding 4.1.1. Effectivetemperature R =0.95R . (1) Ross UD Thetemperaturecanbedeterminedusingtheapparentbolomet- ricmagnitudembolandtheangularRosselanddiameterθross.The 2 Method described inKerschbaum&Hron(1996a) andreferences apparentbolometricmagnitudembol =2.26magwasderivedby therein;near-IRdatafromtheIRAScatalogueandFouqueetal.(1992). 4 Klotzetal.:ConstrainingstellarparametersforTXPsc opticallythinandcontainsalmostnodust. In Sect.4.2.1 and 4.2.2 low-resolution spectroscopic observa- tionsarecomparedwithsyntheticspectraofhydrostaticmodels tofixthefundamentalstellarparametersC/OratioandT .The eff overallenergydistributionaswellasthebandsofthemolecules that are present from 2.3-6µm put strong constraints on these parameters(Jørgensenetal.2000;Loidletal.2001;P11). Low resolution spectroscopy does not allow to ascertain mass and logg (Fig.6-10 in P11). To determine these parameters Sect.4.2.3 follows the approach described in P11: spectro- interferometric observations are compared to models of fixed T andC/Oratiobutvaryingloggandmass. eff 4.2.1. C/Oratio Jørgensenetal.(2000)foundtheratiobetweenthe3µmfeature (HCN and C H ) and the 5.1µm feature (CO and C ) to be a 2 2 3 Fig.4. Rosselandradiusofthesyntheticmodelsversusthede- sensitivemeasureoftheC/Oratio.Inordertobeindependentof rivedmid-IRUD-radiusofthesamemodels(crosses).Theblue distance each model spectrum is normalized to the ISO flux at lineisalinearfittothedata. 2.9µm (local minimum of molecular absorption, Aringeretal. 2009).Aχ2 testisappliedbetween2.9-6.0µmtocomparethe Table 4. Stellar parametersderivedfromobservations(middle observed ISO spectra (1996 Nov 26 and 1997 Dec 11) to the block)andcalculated(rightblock)usingapproach1. models. For both ISO spectra we find that the best solution is obtainedwithaC/O ratioof1.05.Consideringallsolutionsly- d θUD θRoss Teff R L ing within the 68% confidence level a C/O ratio of 1.1 can be [pc] [mas] [mas] [K] [R⊙] [L⊙] defined as an upper limit. Due to the coarse grid spacing for C/Othisvalueisnota strictupperlimitandnolowerlimitcan 275 10.51±0.70 9.99 3127+192 294 7406 −173 beassigned.Figs.5and6showtheISO/SWSspectrumplotted together with models of different temperatures and C/O ratios. 280 10.51±0.70 9.99 3127+192 299 7678 −173 The upper panels draw the region around the 3µm and 5.1µm 315 10.51±0.70 9.99 3127+192 337 9717 featuretoalargerscale.Theseplotsdemonstratethatasynthetic −173 spectrum with a C/O ratio largeror equal1.4 is not able to re- producetheobservations,becauseitover-evaluatesthedepthof the5.1µmfeature.ThisfindingisconsistentwiththeC/Oratios giveninliterature(seeTable3). This implies that for a hydrostatic C-star the Rosseland radius canbeapproximatedbythemid-IRUD-radiusifthederivedcor- 4.2.2. Effectivetemperature rectionfactorof0.95isapplied. Applyingthiscorrectionfactoryieldsan angularRosselanddi- We expectthe 3µm feature to be a goodtemperatureindicator ameter θ = 9.99mas. Together with the bolometric magni- forhydrostaticstars(P11).Consequently,inordertofindthebest Ross tude the distance-independenteffective temperature can be de- temperatureforTXPsc,aχ2testisusedtocomparetheobserved rived. Using the three different distance estimates d that are ISO spectra and the modelspectra between 2.9 - 3.6µm. Only availableforTXPsc(seeTable3)alinearradiusandluminosity model spectra lying within the confidence level of C/O (1.05, canbecalculated.Table4liststhederivedandcalculatedstellar 1.1;seeSect.4.2.1)areusedforthistest. parametersof approach1. Errors on the temperatureare deter- The large grid of models allows to determine Teff very pre- minedbyusingtheerrorsonθ andbyassuminganarbitrary cisely.Allsolutionslyingwithinthe68%confidencelevelhave UD errorof±0.1form thataccountsforthestellarvariabilityand a temperature of 3080+70 / 3170+70K for the visual mini- bol −60 −80 thefittingerror.Thesevaluesareinagreementwiththosegiven mum/maximum,respectively.TheupperleftpanelofFigs.5and inliterature(seeTable3). 6showthatmodelswithhigher/lowertemperaturesdonotrepro- ducethedepthofthe3µmfeature. Thetemperaturederivedwithapproach1inSect.4.1.1iswithin 4.2.Approach2:Hydrostaticmodels the errors of the temperature that is derivedhere for the visual minimumandmaximum. The short wavelength part of the ISO spectrum of TXPsc is We confirm the finding of Jørgensenetal. (2000) that the dif- dominated by the 3µm feature which has contributions from ference in the two ISO spectra can be explainedby a tempera- HCN and C H . The 5µm feature, on the other hand, is due 2 2 ture change. The lightcurvesuggests that this difference is due to C and the fundamental band of CO. The region from 7 3 to variability effects, but time-series spectroscopy is needed to to 8µm is dominated by HCN and C H and according to 2 2 confirmthisfinding. Gautschy-Loidletal. (2004) the region from 8 to 9µm is af- fectedbyCS opacity.Thelowvariabilityinthe V-band(∆V ∼ 0.4mag) justifies the use of hydrostatic models in the near- as 4.2.3. Surfacegravityandmass wellasmid-IR.Additionally,hydrostaticmodelsareabletore- producelargepartsoftheoverallISOaswellasMIDIspectra. Interferometric observations are compared to synthetic visi- This indicates that the circumstellar environment of TXPsc is bilities of hydrostatic models of varying logg and mass us- 5 Klotzetal.:ConstrainingstellarparametersforTXPsc Fig.5. ISO/SWSspectrumofTXPsc atvisualminimumfrom1997Dec11(blackline)plottedwithhydrostaticmodels(colored lines)ofdifferenttemperatureandC/Oratio.Thebestfittingmodelisplottedingreen.Modelspectraarenormalizedtothefluxof thecorrespondingISOspectrumat2.9µm. Fig.6. SameasFig.5,butforatvisualmaximumfrom1996Nov21. ing a χ2 test. All synthetic visibilities are computed for the there is also a degeneracy between distance and logg. From best-fitting values of T and C/O-ratio from Sects.4.2.1 and current TP-AGB evolutionary calculations (e.g. Karakasetal. eff 4.2.2. Synthetic visibilities are derived at 3 different distances: 2002;Marigo&Girardi2007,Marigoetal.inprep.)weexpect 275pc, 280pc, 315pc (vanLeeuwen 2007; Claussenetal. that a 1M TP-AGB star with solar metallicity does not make ⊙ 1987; Bergeat&Chevallier 2005, respectively). In Fig.7 the thetransitiontotheC-richdomain.Thissuggeststhatthelogg wavelength-dispersedcalibratedvisibilitiesare plottedtogether valuesinTable5foundforM =2M arethemorereliableones. ⊙ withthesyntheticvisibilitiesofthebest-fittingmodels.There- But,consideringthedegeneracyandthelimitedmasssampling gion between 8 - 9µm is not considered in the fitting, as the (M =1,2M )inthegrid,alsomodelswithhighermasseswould ⊙ hydrostaticmodelsdonotincludetheCSopacitydata. reproduce the observed visibilities. To support this statement, The middle block of Table5 gives the best-fitting stellar pa- oneadditionalmodelwith3M iscalculatedandoverplottedin ⊙ rameters for a given distance that were determined using ap- Fig.7asdashedline.Themodelsarealmostindistinguishable. proach 2. It is clear from the table and from Fig.7 that, given TherightpartofTable5givesstellarparametersthatarecalcu- the error barson the visibilities, there is a degeneracybetween latedfromthederivedlogg,massandeffectivetemperature.The logg and mass. As the distance defines the level of visibility, luminositiesaresignificantlylargerthanthe L = 5200L used ⊙ 6 Klotzetal.:ConstrainingstellarparametersforTXPsc Fig.7. Wavelength-dispersedcalibratedvisibilitiespluserrors(darkgreyshadedarea)plottedwiththebest-fittingsyntheticvisibil- itiesofhydrostaticmodels(fulllines).Toshowthedegeneracybetweenmassandloggonemodelwith3M isoverplotted(dotted ⊙ line). byGautschy-Loidletal.(2004),butcomparabletoL=7700L ⊙ derivedbyClaussenetal.(1987).Theluminositiesandradiiare also in perfect agreement with the ones determined with ap- proach1(seeSect.4.1). 4.2.4. Photometricconstraintsonthebestmodel The best fitting hydrostatic models of visual mini- mum/maximum (model with lowest χ2 in Table5) are logg,M overplotted to the ISO and MIDI spectra as well as to pho- tometric measurements in Fig.8. Photometric measurements from Johnsonetal. (1966), MendozaV.&Johnson (1965) and Catchpoleetal. (1979) were observed with the Johnson filter system. Zero points to convert these measurements from mag- nitudestoJanskyaretakenfromCox(2000).Thesezeropoints arealsousedtoconvertobservationsfromBergeatetal.(1976) andBergeat&Lunel(1980)astheauthorsclaimthattheirfilter system is similar to the Johnson filter system. Zero points for Fig.8. Best fitting hydrostaticmodelsfor the visualminimum 2MASSphotometry(Cutrietal.2003)aregiveninCohenetal. (blackline)andmaximum(greyline).SuperimposedaretheISO (2003). Olofssonetal. (1993) and Kerschbaumetal. (1996b) spectra of visual minimum (green line) and visual maximum usedtheESOfiltersystem.ZeropointsaretakenfromLeBertre (blue line), MIDI spectra (violet lines) and photometric mea- (1988)andWamsteker (1981).Noguchietal.(1981) usedtheir surements of different works (orange symbols). Model spectra own filter system and corresponding zero points are given in arenormalizedtothefluxofthecorrespondingISOspectrumat theirpaper. 2.9µm. 5. Comparisonwithevolutionarytracks with solar-metallicity and C/O within a narrow interval (i.e. WefollowtheapproachdescribedinP11andcomparethestel- 1<C/O≤1.1). As we see in Fig. 9, the TP-AGB evolutionary larparameterswithnewevolutionarytracksofthermallypulsing tracksintheC-richregimeextendtomuchlowereffectivetem- AGBstars(Marigoetal.,inprep). peratures than the derived values for TX Psc. This cooling is Figure9 depictsevolutionarytracksin the regionofAGB stars mainlydrivenbytheincreaseoftheC/Oratioaftereachdredge- intheH-Rdiagram.Overplottedarethedeterminedluminosity up episode, as well as by the progressive strengthening of the and temperature for TXPsc for the two approaches. Only the mass-loss efficiency.The relativelywarm effective temperature best-fitting luminosity (see Table 5) at d=280pc is plotted for ofTXPscsuggeststhatthisstarisobservedcloseafterthetran- TXPsc.Errorsfortheluminosityareassumedtolieintheorder sitionintotheC-stardomain,inanearly‘quiet’stageinwhich of∼40%(uncertaintyonthegivendistancemeasurement). thestrongwindhasnotyetdeveloped.Thispictureisnicelysup- We note an encouraging agreement between the observed lo- portedbytheobservationalfindingsalreadydiscussedinthepre- cation of TX Psc in the H-R diagram and the predicted ranges vioussections. of luminosity and effective temperature for a carbon-rich star Itisvisible fromthe evolutionarytracksthatan AGBstar with 7 Klotzetal.:ConstrainingstellarparametersforTXPsc Table5. Stellarparametersderivedfromobservations(middleblock)andcalculated(rightblock)usingapproach2. d T T C/O logg M χ2 R L L eff,min eff,max logg,M min max [pc] [K] [K] [M ] [R ] [L ] [L ] ⊙ ⊙ ⊙ ⊙ 275 3080+70 3170+70 1.05 -0.5 1 0.50 295 7019 7876 −60 −80 -0.2 2 0.50 295 7019 7876 280 3080+70 3170+70 1.05 -0.5 1 0.48 295 7019 7876 −60 −80 -0.2 2 0.53 295 7019 7876 315 3080+70 3170+70 1.05 -0.6 1 0.49 331 8836 9915 −60 −80 -0.3 2 0.53 331 8836 9915 solar metallicity will turn into a carbon-richAGB star only for Acknowledgements. TheauthorsthankAngelaBaierforfruitfuldiscussionson masses around 2M and higher. Also, the position of TXPsc ISO spectra and Walter Nowotny for helpful discussions on SEDs and pho- ⊙ in Fig.9 suggests that the mass lies between 2 and 3M . This tometric filter systems.This workis supported bytheAustrian Science Fund ⊙ FWFunderprojectnumberAP23006.BAacknowledgessupportfromAustrian is not in agreement with the best fitting models having 1M ⊙ Science Fund (FWF) Projects AP23006 & AP23586 and from contract ASI- (Sect.4.2.3), but in good agreement with the models having INAFI/009/10/0.ThisresearchhasmadeuseoftheSIMBADdatabase, oper- 2M⊙. atedatCDS,Strasbourg,France.Weacknowledgethevariablestarobservations fromtheAAVSOInternationalDatabasethatwereusedinthisresearch. 6. Conclusion References Aringer,B.,Girardi,L.,Nowotny,W.,Marigo,P.,&Lederer,M.T.2009,A&A, In this work we determined stellar parameters for TXPsc by 503,913 comparing observations to geometric models (Klotzetal. Baschek,B.,Scholz,M.,&Wehrse,R.1991,A&A,246,374 2012b), state-of-the-art hydrostatic model atmospheres Bergeat,J.&Chevallier,L.2005,A&A,429,235 (Aringeretal. 2009) and evolutionary models (Marigo et Bergeat,J.,Knapik,A.,&Rutily,B.2001,A&A,369,178 Bergeat,J.&Lunel,M.1980,A&A,87,139 al. in prep.). Two different approaches were used to fix the Bergeat,J.,Sibille,F.,Lunel,M.,&Lefevre,J.1976,A&A,52,227 parameters: Bressan,A.,Marigo,P.,Girardi,L.,etal.2012,ArXive-prints A1. Spectro-interferometric observations were used to de- Catchpole,R.M.,Robertson,B.S.C.,Lloyd-Evans,T.H.H.,etal.1979,South termine a wavelength-dispersed uniform disk diameter. A AfricanAstronomicalObservatoryCircular,1,61 Chesneau,O.2007,NewARev.,51,666 correction factor for hydrostatic C-stars was derived from Claussen,M.J.,Kleinmann,S.G.,Joyce,R.R.,&Jura,M.1987,ApJS,65,385 hydrostaticmodelstoconverttheUDdiametertotheRosseland Cohen,M.,Wheaton,W.A.,&Megeath,S.T.2003,AJ,126,1090 radius,whichwasthenusedtodeterminetheeffectivetempera- Cox,A.2000,Allen’sAstrophysicalQuantities(Springer) tureT . Cruzale`bes,P.,Lopez,B.,Bester,M.,Gendron,E.,&Sams,B.1998,A&A,338, eff A2. Spectroscopic measurements were compared to synthetic 132 Cutri, R. M., Skrutskie, M. F.,van Dyk, S., et al. 2003, VizieR Online Data spectra from hydrostatic models to determine T and C/O eff Catalog,2246,0 ratio. The mass and logg were constrained by comparing deGraauw,T.,Haser,L.N.,Beintema,D.A.,etal.1996,A&A,315,L49 spectro-interferometric observations with synthetic visibility deVegt,C.1974,A&A,34,457 profilesfromhydrostaticmodels. Dunham,D.W.,Evans,D.S.,Silverberg,E.C.,&Wiant,J.R.1975,MNRAS, 173,61P The main advantage of approach1 is the distance-independent Dyck,H.M.,vanBelle,G.T.,&Benson,J.A.1996,AJ,112,294 determination of T . On the other hand, conversion of the eff Fouque,P.,LeBertre,T.,Epchtein,N.,Guglielmo,F.,&Kerschbaum,F.1992, UD-radius to the Rosseland radius and the use of the apparent A&AS,93,151 bolometric magnitude introduces uncertainties. Approach 2 Gautschy-Loidl,R.,Ho¨fner,S.,Jørgensen,U.G.,&Hron,J.2004,A&A,422, allows to constrain not only T , but also C/O, logg and M. 289 eff Goorvitch,D.&Chackerian,Jr.,C.1994,ApJS,91,483 Oneofthedisadvantagesofthistechniqueistheunknownerror Harris,G.J.,Pavlenko,Y.V.,Jones,H.R.A.,&Tennyson,J.2003,MNRAS, that is introduced by the model. Additionally, the uncertainty 344,1107 in distance, that is needed to constrain logg and M, and the Harris,G.J.,Tennyson,J.,Kaminsky,B.M.,Pavlenko,Y.V.,&Jones,H.R.A. degeneracybetweenthesetwoparameters,limitstheaccuracyof 2006,MNRAS,367,400 Hauschildt,P.H.,Allard,F.,Ferguson,J.,Baron,E.,&Alexander,D.R.1999, theparameterdetermination.Thissuggeststhathigh-resolution ApJ,525,871 spectroscopyis neededto fullydiscriminatebetween massand Ho¨fner,S.,Gautschy-Loidl, R.,Aringer, B.,&Jørgensen, U.G.2003,A&A, logg. 399,589 There is a very good agreement between the best-fitting hy- Iben,Jr.,I.&Renzini,A.1983,ARA&A,21,271 drostatic model atmosphere and observations (interferometry, Ireland,M.J.,Scholz,M.,&Wood,P.R.2008,MNRAS,391,1994 Ireland,M.J.,Scholz,M.,&Wood,P.R.2011,MNRAS,418,114 spectroscopyandphotometry). Jaffe,W.J.2004,inSocietyofPhoto-OpticalInstrumentationEngineers(SPIE) Our spectro-interferometric results are also an important tool Conference Series, Vol. 5491, Society of Photo-Optical Instrumentation to constrain and validate stellar AGB models, that are still Engineers(SPIE)ConferenceSeries,ed.W.A.Traub,715 subjectto severe uncertainties.We foundthatpresentTP-AGB Johnson, H. L., Mitchell, R. I., Iriarte, B., & Wisniewski, W. Z. 1966, CommunicationsoftheLunarandPlanetaryLaboratory,4,99 tracks with a detailed treatment of molecular opacities nicely Jørgensen, U. G. 1997, in IAU Symposium, Vol. 178, IAU Symposium, ed. reproducethederivedT ,L,C/OvaluesforTXPsc. eff E.F.vanDishoeck,441–456 8 Klotzetal.:ConstrainingstellarparametersforTXPsc Fig.9. ZoomintotheAGBregionoftheH-Rdiagram.Linesdenotesolarmetallicityevolutionarytracks(Marigoetal.,inprep.) andnumbersindicatethemassontheearly-AGB.Yellow/greylinesmarktheregionofcarbon-richAGBstarswithC/O>1.0.Black linesmarktheregionofoxygen-richAGBstars(C/O≤1).Forbettervisibility,thetrackofthe2 M modelisplottedwithadotted ⊙ line. Different colored symbols refer to the luminosity and effective temperature determined in this work (for the two different approachesA1andA2atvisualminimum/maximum). Jorgensen,U.G.,Almlo¨f,J.,&Siegbahn,P.E.M.1989,ApJ,343,554 Richichi,A.,Chandrasekhar,T.,Lisi,F.,etal.1995,A&A,301,439 Jørgensen,U.G.,Hron,J.,&Loidl,R.2000,A&A,356,253 Sacuto,S.,Aringer,B.,Hron,J.,etal.2011a,A&A,525,A42 Jorissen,A.,Mayer,A.,vanEck,S.,etal.2011,A&A,532,A135 Sacuto, S.,Jorissen, A.,Cruzale`bes, P.,etal. 2011b,inAstronomical Society Karakas,A.I.,Lattanzio,J.C.,&Pols,O.R.2002,PASA,19,515 ofthePacificConferenceSeries,Vol.445,WhyGalaxies CareaboutAGB Kerschbaum,F.&Hron,J.1996a,A&A,308,489 Stars II: Shining Examples and Common Inhabitants, ed. F. Kerschbaum, Kerschbaum,F.,Lazaro,C.,&Habison,P.1996b,A&AS,118,397 T.Lebzelter,&R.F.Wing,171 Klotz,D.,Sacuto,S.,Kerschbaum,F.,etal.2012a,A&A,541,A164 Samus,N.N.,Durlevich,O.V.,&etal.2009,VizieROnlineDataCatalog,1, Klotz,D.,Sacuto,S.,Paladini,C.,Hron,J.,&Wachter,G.2012b,ArXive-prints 2025 Kurucz, R. 1993, Limbdarkening for 2 km/s grid (No. 13): [+0.0] to Scholz,M.2003,inSocietyofPhoto-OpticalInstrumentationEngineers(SPIE) [-5.0]. Kurucz CD-ROM No. 17. Cambridge, Mass.: Smithsonian Conference Series, Vol. 4838, Society of Photo-Optical Instrumentation AstrophysicalObservatory,1993.,17 Engineers(SPIE)ConferenceSeries,ed.W.A.Traub,163–171 Lambert,D.L.,Gustafsson,B.,Eriksson,K.,&Hinkle,K.H.1986,ApJS,62, Sloan,G.C.,Kraemer,K.E.,Price,S.D.,&Shipman,R.F.2003,ApJS,147, 373 379 Lasker,B.M.,Bracker,S.B.,&Kunkel,W.E.1973,PASP,85,109 vanLeeuwen,F.2007,A&A,474,653 LeBertre,T.1988,A&A,190,79 Wamsteker,W.1981,A&A,97,329 Leinert,C.,Graser,U.,Przygodda,F.,etal.2003,Ap&SS,286,73 Wittkowski,M.,Boboltz,D.A.,Driebe,T.,etal.2008,A&A,479,L21 Leinert,C.,vanBoekel,R.,Waters,L.B.F.M.,etal.2004,A&A,423,537 Wittkowski,M.,Boboltz,D.A.,Ireland,M.,etal.2011,A&A,532,L7 Loidl,R.,Lanc¸on,A.,&Jørgensen,U.G.2001,A&A,371,1065 Wittkowski,M.,Hummel,C.A.,Johnston,K.J.,etal.2001,A&A,377,981 Marigo,P.&Aringer,B.2009,A&A,508,1539 Marigo,P.&Girardi,L.2007,A&A,469,239 Mart´ı-Vidal,I.,Marcaide,J.M.,Quirrenbach,A.,etal.2011,A&A,529,A115 MendozaV.,E.E.&Johnson,H.L.1965,ApJ,141,161 Neilson,H.R.&Lester,J.B.2008,A&A,490,807 Noguchi,K.,Kawara,K.,Kobayashi,Y.,etal.1981,PASJ,33,373 Ohnaka,K.,Tsuji,T.,&Aoki,W.2000,A&A,353,528 Olofsson,H.,Eriksson,K.,Gustafsson,B.,&Carlstroem,U.1993,ApJS,87, 305 Paladini,C.,Aringer,B.,Hron,J.,etal.2009,A&A,501,1073 Paladini,C.,vanBelle,G.T.,Aringer,B.,etal.2011,A&A,533,A27(P11) Querci,F.,Querci,M.,&Tsuji,T.1974,A&A,31,265 Quirrenbach, A., Mozurkewich, D., Hummel, C. A., Buscher, D. F., & Armstrong,J.T.1994,A&A,285,541 Ragland,S.,Traub,W.A.,Berger,J.-P.,etal.2006,ApJ,652,650 Ratzka, T.2005,PhDthesis,Max-Planck-Institute forAstronomy,Ko¨nigstuhl 17,69117Heidelberg,Germany 9 Klotzetal.:ConstrainingstellarparametersforTXPsc,OnlineMaterialp1 Table2.Propertiesofthecalibratortargets. HD Name Sp.T.a F a θb 12 [Jy] [mas] HD48915 Sirius A1 143.1±3 6.08±0.03 HD20720 τ04Eri M3/M4 162.7±6 10.14±0.04 HD224935 YYPsc M3 86.9±5 7.25±0.03 HD49161 17Mon K4 10.4±5 2.44±0.01 HD18884 αCet M1.5 234.7±3 12.28±0.05 HD45348 Canopus F0 154.8±3 6.87±0.03 Notes. (a)http://www.eso.org/observing/dfo/quality/MIDI/qc/ calibrators obs.html (b) http://simbad.u-strasbg.fr/simbad/