Table Of Content

Mon.Not.R.Astron.Soc.000,000–000(0000) Printed5February2008 (MNLATEXstylefilev2.2) A search for strong, ordered magnetic fields in Herbig Ae/Be stars⋆ G.A. Wade1, S. Bagnulo2, D. Drouin1, J.D. Landstreet3, D. Monin4 1DepartmentofPhysics,RoyalMilitaryCollegeofCanada,POBox17000,Station’Forces’,Kingston,Ontario,CanadaK7K4B4 2EuropeanSouthernObservatory,Casilla19001,Santiago19,Chile 3DepartmentofPhysics&Astronomy,TheUniversityofWesternOntario,London,Ontario,Canada,N6A3K7 7 4NationalResearchCouncilofCanada,HerzbergInstituteofAstrophysics,5071WestSaanichRd.,Victoria,BC,Canada,V9E2E7 0 0 2 5February2008 n a J ABSTRACT 2 The origin of magnetic fields in intermediate-mass and high-mass stars is fundamentally a 1 mystery. Clues toward solving this basic astrophysical problem can likely be found at the pre-main sequence (PMS) evolutionary stage. With this work, we perform the largest and 1 mostsensitivesearchformagneticfieldsinpre-mainsequenceHerbigAe/Be(HAeBe)stars. v We seek to determine whether strong, orderedmagnetic fields, similar to those of main se- 7 quenceAp/Bpstars, can bedetectedin these objects,andif so,to determinethe intensities, 8 3 geometricalcharacteristics, and statistical incidence of such fields. Sixty-eightobservations 1 of 50 HAeBe stars have been obtained in circularly polarised light using the FORS1 spec- 0 tropolarimeter at the ESO VLT. An analysis of both Balmer and metallic lines reveals the 7 possiblepresenceofweak longitudinalmagneticfields inphotosphericlinesoftwo HAeBe 0 stars,HD101412andBFOri.Resultsfortwoadditionalstars,CPD-53295andHD36112, / aresuggestiveofthepresenceofmagneticfields,butnofirmconclusionscanbedrawnbased h p ontheavailabledata.TheintensityofthelongitudinalfieldsdetectedinHD101412andBF - Orisuggestthattheycorrespondtoglobally-orderedmagneticfieldswithsurfaceintensities o oforder1kG.Ontheotherhand,nomagneticfieldisdetectedin4otherHAeBestarsinour r t sampleinwhichmagneticfieldshadpreviouslybeenconfirmed.MonteCarlosimulationsof s thelongitudinalfieldmeasurementsoftheundetectedstarsallowustoplaceanupperlimits a : of about 300 G on the general presence of aligned magnetic dipole magnetic fields, and of v about500Gonperpendiculardipolefields.Takingintoaccounttheresultsofoursurveyand i X otherpublishedresults,wefindthattheobservedbulkincidenceofmagneticHAeBestarsin oursampleis8-12%,ingoodagreementwiththatofmagneticmainsequencestarsofsimilar r a masses.Wealsofindthatthermslongitudinalfieldintensityofmagnetically-detectedHAeBe starsisabout200G,similar tothatofApstarsandconsistentwith magneticfluxconserva- tion during stellar evolution.These results are all in agreementwith the hypothesisthat the magneticfieldsofmainsequenceAp/Bpstarsarefossils,whichalreadyexistwithinthestars at the pre-main sequence stage. Finally, we explore the ability of our new magnetic data to constrain magnetospheric accretion in Herbig Ae/Be stars, showing that our magnetic data arenotconsistentwiththegeneraloccurrenceinHAeBestarsofmagnetosphericaccretionas describedbythetheoriesofKo¨nigl(1991)andShuetal.(1994). Keywords: stars:magneticfields–polarisation 1 INTRODUCTION suppressing large-scalemixingandleadingtoanamazingarrayof atmosphericpeculiaritieswhicheffectivelydefinethesestarsspec- About 5-10% (e.g. Wolff 1968) of all intermediate mass (B and troscopically(e.g.Adelman1993). A) main sequence stars exhibit organised magnetic fields with strengthsrangingbetweenafewhundred andafew tensof thou- Their magnetic fieldshaveclearly defined observational and sandsofgauss.Thepresenceofthesefieldshasimportantconse- physicalpropertieswhicharetotallydifferentfromthoseofthesun quencesforthestructureoftheatmospheresoftheseAp/Bpstars, andotherlate-typestars.Thefieldsarestrong,withimportantdipo- larcomponentsandfillingfactorsnear100%;thefieldstructureis staticintherotatingstellarreferenceframe,andappearstobesta- ⋆ BasedonobservationsfromtheESOtelescopesattheLaSillaParanal bleontimescalesofdecades;andthefieldsshowonlyweakorno Observatory underprogrammeID072.C-0447,DDT-272.C-5063,074.C- correlationwithprojectedrotational velocity, mainsequence age, 0442. andotherphysicalparameters. 2 G.A. Wade,S. Bagnulo,D. Drouin,J.D. Landstreet,D. Monin Because these characteristicscontrast sostrongly withthose of active cool stars(theonly other broad classof non-degenerate starsinwhich magnetic fieldscan be studied systematically), the magneticAp/Bpstarsrepresentanimportantlaboratoryforinves- tigating the fundamental physics associated with stellar and cos- micmagnetism. Inparticular,becausestarsinthismassrangeare progenitorsofbothwhitedwarfsandneutronstars,theirmagnetic 4 characteristicsmayprovideimportantcluesneededinordertoun- derstandmagnetisminthesedegenerateobjects. lutionReomf tahreksaeblmy,avgenreytilcittfileelidssk.nRoewcnenatbionuvtetshteigfaotriomnastiboynaBnadgnevuolo- 6.0Birthline (10-5 Msun/yr) et al. (2003, 2004, 2006) and Po¨hnl et al. (2003, 2005), showing tvheartymyoaugnngetiocpaenndclcuhsetemrsic,asltlryonpgelcyulsiuaprpAo-rttytpheesvtiaerwsathreatptrheesefinetldins L)sun 5.0 L/* arepresentinthesestarsatthetimetheyreachthemainsequence. g( 4.0 o Thissuggeststhatthemagneticfieldsobserved inmainsequence l 3.5 Ap/Bp stars arepresent, and possibly produced, within their pre- 2 mainsequence (PMS)progenitors. Thisideaisalsosupported by 3.0 the theoretical work of Ste¸pien´ (2000) and of Ste¸pien´ and Land- 2.5 street(2002),whichsuggestthatstrongmagneticfieldsarerequired atthepre-mainsequencephaseinordertoproducetheslowrota- 2.0 tionwhichiscommonlyobservedintheAp/Bpstars.Theprimary aimofthisstudyistotestthesepropositionsbysearchingfordirect 1.5 evidence of magnetic fields in the photospheres of a statistically usefulsampleofPMSstarsofintermediatemass. ZA A second motivation of this study is to explore the gen- M S eral role of magnetic fields in the late stages of formation of 0 4.4 4.2 4 3.8 intermediate-massstars.Accretionontolower-massPMSTTauri log(T ) eff starsisnowgenerallysupposedtobemediatedbythepresenceof strong,large-scalemagneticfields(e.g.discussionbyJohns-Krull Figure1.HRdiagramofsamplestarswithknownTeff andlogL/L⊙,de- rivedfromTable1.Thedataareinterpretedusingthemodelevolutionary etal.(1999)andreferencestherein).Someauthors(e.g.Vinketal. calculations ofPalla&Stahler(1993),whichassumeanaccretionrateof 2002,Muzzeroleetal.2004)havesuggestedthatsimilar“magne- 10−5 M⊙/yrbeforethebirthline. Eachevolutionary trackislabelled with tosphericaccretion”mayoccurinintermediate-massPMSstarsas thecorresponding stellarmass.Notably,amajorityofthesampleiscon- well.Theobservationspresentedhereprovideauniqueopportunity centratedbetween1.5-3 M⊙. Starsinwhichmagneticfieldsarepossibly toexplorethisproposal. detectedinthisstudy(HD101412,BFOri,andHD36112)areindicated Many intermediate-mass PMS stars areidentifiable observa- byfilledsymbols(theluminosityofCPD-53295isnotknown,andthuswe tionally as Herbig Ae/Be (HAeBe) stars (Herbig 1960; Cohen & areunabletoplaceitontheHRdiagram). Kuhi 1979; Finkenzeller & Mundt 1984; Hillebrand et al. 1992). HAeBe stars are characterised by spectral types A and B with strong, often ubiquitous emission lines. They are distinguished 1kG.AsurveyundertakenbyGlagolevski&Chountonov(1998), from the classical Ae/Be stars by their IR colours and frequent including a larger sample of stars, also found no fields, but with presence within dust-obscured regions and association with neb- relativelypoor precision. On theother hand, Donati et al. (1997) ulae(Waters&Waelkens1998).Accordingtostandardstellarevo- andDonati(2000)reportedtheprobableanddefinite(respectively) lution theory, HAeBe stars should not posses deep outer convec- detection of a circular polarisation signature in metallic lines of tionzoneswhichgenerateimportantquantitiesofoutward-flowing the Herbig Aestar HD 104237, providing the first detection of a mechanicalenergy.Rather,thesestarsareexpectedtohaveconvec- magneticfieldinaHAeBestar.Morerecently,Hubrigetal.(2004, tivecoressurroundedby primarilyradiativesub-photosphericen- 2006a)haveclaimedthedetectionofstrongmagneticfieldsinsev- velopes(Iben1965;Gilliland1986).However,since1980,repeated eralHAeBestars.Clearly,indirectanddirectevidenceforthepres- observationsofthesestars(e.g.Praderieetal.1982;Felenboketal. enceofmagneticfieldsinsomeHAeBestarsinincreasing, butit 1983;Catalaetal.1986;Hamann&Persson1992)haveshownthat remainsverysparse. manyareintenselyactive.Inparticular,theydisplaycharacteristics In this study, we use the low-resolution spectropolarimeter oftenassociated withmagneticactivityand thepresence of chro- FORS1attheESOVLTtoconductasearchformagneticfieldsin mospheres or coronae (e.g.Skinner &Yamauchi 1996; Swartzet about50HAeBestars,thelargestsuchsampleeverstudied.First al. 2005). These properties havebeen proposed asindicators that resultsofthisstudyhavealreadybeenreportedinashortpaperby at least some of these stars or their circumstellar envelopes are Wadeetal.(2005). intenselymagneticallyactive. Previous attempts to detect magnetic fields in HAeBe stars haveproducedvariedresults.Catalaetal.(1993)attemptedtode- 2 SAMPLESELECTION tectZeemancircularpolarisationintheFeλ5018andHeλ5876 linesinthespectrumoftheprototypicalHAeBestarABAur.High Our study required the selection of a relatively large number of resolution,highsignal-to-noiseratiospectraofABAurresultedin HAeBestarstoallowustoderivestatisticallymeaningfulconclu- nodetectionofamagneticfield,withupperlimitsontheorderof sionsaboutthepresenceofmagneticfieldsinthesestars,andthe A searchforstrong,ordered magneticfieldsin HerbigAe/Bestars 3 relationshipofmagneticHAeBestarstotheAp/Bpstars.Various yieldsadispersionof1.2Åperpixel.Spectrawereobtainedwitha literaturesourceswereusedfortargetselection,primarilythecata- slitwidthof0.5”or0.8”,yieldingaresolvingpowerof1560or975, loguesofHAeBestarsandHAeBecandidatesbyThe´etal.(1994) respectively.Inafewcases, grism1200g(nowdecommissioned) andbyVieiraetal.(2003). was used, which covers the spectral range from 4310 to 5490 Å ThecatalogueofThe´ etal.(1994)containssixcategoriesof (whichincludesHβandHγonly).Itsdispersionis0.6Åperpixel, stars;thestarsselectedforourstudywereobtainedonlyfromthe andwithaslitwidthof0.5”or0.8”,thisgrismyieldsaresolving first category, which contains stars historically known as HAeBe powerof2800or1800,respectively. stars,orstrongcandidatesofthegroup.Accordingtotheauthors, Sixty-eightobservationsofthe50targetswereobtainedincir- allofthesestarspossessnear-orfar-infraredexcessandemission cularlypolarisedlightfollowingtheprocedure describedbyBag- lines,associatedwiththepresenceofcircumstellardust, discsand nulo et al. (2006). The light gathered by the telescope is sent energeticouflowswhichareusuallyfoundinHAeBestellarenvi- through the Longitudinal Atmospheric Dispersion Corrector and ronments.Ontheotherhand,Vieiraetal.(2003)producedacata- theinstrumentcollimator,thenpassesthroughasingleslitintothe logueofHAeBestarsandstrongcandidatesfromaninitialsearch topsectionoftheFORS1instrument.There,thestellarlightpasses fornewTTauristars(pre-mainsequencestarsoflowermass)using througharotatablequarter-waveplate(locatedinthefilter/camera theInfraredAstronomicalSatellite(IRAS)pointsourcecatalogue1. section of FORS1) oriented at ±45◦ with respect to the fast axis Becausetheinitialsearchwasbasedoncircumstellardustproper- of the analyser, which converts the circularly polarised light into ties, itincluded HAeBestarsalong withtheT Tauri stars. Viera linearly polarized light. The orthogonal linearly polarized beams etal.extractedtheHAeBestarsbyfilteringthedatausingspecific arethenseparatedbytheWollastonprismbeforebeingdispersed requirements such as a spectral type earlier than F5, emission at byagrism.Twospectra(each corresponding tooneof theoppo- Hα, and a minimum level of infrared emission. The majority of sitelypolarisedbeams)arefinallyrecordedsimultaneouslyonthe thestarswereassociatedbytheauthorswithastarformingregion CCD.Bychanging theorientationof thequarter-waveplatewith (SFR).Basedonthequalityofthesetwoliteraturesourcesandthe respect to the Wollaston fast axis (alternately +45◦ or −45◦) and arguments presented by theirauthors, weconclude that all of the obtaining an additional exposure, the positions of the right and starsinoursamplearebonafideHAeBestars.Intotal,50HAeBe left circularly polarised beams are exchanged throughout the in- starshavebeenselected,spanninginaspectraltyperangefromF2 strument, and in particular their respective positions on the CCD throughB0. areinterchanged.AsdescribedintheFORS1manual(VLT-MAN- TheselectedstarsarelistedinTable1.Thespectraltypesof ESO-13100-1543), this procedure allows for the removal of sys- the stars have been obtained from the bibliographic service SIM- tematicerrorsduetoCCDinhomogeneitiesanddifferentialwave- BAD (unless indicated otherwise). Table 1 also contains physical lengthcalibrationerrors. properties(whereavailable)suchasmass,radius,effectivetemper- Inordertoobtainahighsignal-to-noiseratio,multipleexpo- ature,luminosity, vsini,ageanddistance, found intheliterature. suresofeachstarwereobtained.Anevennumberofexposures(be- Wheneverthesourcespermitted,theuncertaintiesassociatedwith tween2and30, depending ontheapparent magnitude ofthestar theseparameterswereincluded. andtheseeing)wasobtained,eachpairofexposurescorresponding AscanbeseeninTable1, 7objectshavenowell-determined toaquarter-waveplateorientationofeither+45oor−45o.Thedata physical parametersintheliterature.Inaddition, thephysical pa- werereducedusingtheproceduredescribedindetailbyBagnuloet rameters (and in particular the luminosities) of many objects are al.(2006). highlyuncertain. Notwithstandingthislimitation,wehaveplaced Thejournalofobservations(Table2)reportsourobservations those stars with known temperatures and luminosities on the HR ofbothstandardandprogrammestars.Itshouldbenotedthatgrism diagraminFig.1,alongwiththemodelpre-mainsequenceevolu- 600Bwasusedforallobservationsunlessindicatedotherwise(in tionarytracksofPalla&Stahler(1993).BasedonFig.1,mostof whichcasegrism1200gwasused).Table2includestheJulianDate thestarsintheobservedsamplehavemassesabove2M ,andfully of mid-observation, the slit width, exposure time and number of ⊙ 18ofthe37starsonthediagram(49%)havemassesbetween2and exposures,andtheresultant peaksignal-to-noiseratio perÅ.The 4M .Eightstarshavemassesbelow2M ,and9starshavemasses tablealsocontainstheresultsofthemagneticfielddiagnosis,which ⊙ ⊙ largerthan4M . arediscussedinSect.4. ⊙ 3 OBSERVATIONSANDDATAREDUCTION 4 RESULTS ThesamplestarswereobservedattheEuropeanSouthernObser- vatory (ESO) Very Large Telescope (VLT) with the first of the Thelongitudinal magnetic fieldhasbeen inferredfromthepolar- two FOcal Reducer and low dispersion Spectrographs (FORS1). isationspectrain themanner described by Bagnulo et al.(2006). FORS1 was mounted at the Cassegrain focus of one of the tele- Briefly, the field diagnosis is obtained using a Least-Squares fit scope unitsand wasused inspectropolarimetricfast mode. Ato- based on the predicted circularly polarised flux V/I in the weak- tal of about five nights were allocated for our study, distributed fieldregime,givenby overthreeobservingruns(visitormoderunsinFebruary(P72)and V 1 dI November2004(P74),andoneservicemoderunofDirector’sDis- I =−geffCzλ2 I dλ hBzi, (1) cretionaryTimeinApril2004). Thelargemajorityofourspectrawereobtainedwiththegrism wheregeff istheeffectiveLande´factor,Iistheusual(unpolarised) 600B,whichcoversthespectralrangefrom3450to5900Å,and intensity,λisthewavelengthexpressedinÅ,hBziisthelongitudi- nalfieldexpressedingauss,and e C = (≃4.67×10−13Å−1G−1) 1 http://irsa.ipac.caltech.edu/ z 4πm c2 e 4 G.A. Wade,S. Bagnulo,D. Drouin,J.D. Landstreet,D. Monin Name Sp.Type Mass R∗ log(Teff) log(L∗/L⊙) vsini Age Distance (M⊙) (R⊙) (K) (km/s) (Myr) (pc) CPD-53295b F2 - - b3.83 - b175 - - HD17081d B7IV ∗4.6 ∗4.9 o4.09 o2.70 n18 o0.1 o135 HD278937a A3 g2.0 ∗1.9 g3.92 g1.21 - g8.0 g350 HD275877a A2IIvar+ g1.9 ∗1.7 g3.92 g1.12 i217±13 g11.0 g120 HD31293a A0Vpe f2.4 b2.7 f3.98±0.05 f1.68+0.13 f80 f2.0 b160 −0.11 HD31648a A3pshe+ f2.2 ∗2.5 f3.94±0.05 f1.51+0.15 i102±5 f2.5 p131±20 −0.13 HD293782a A3e c3.3 c3.2 c3.96±0.05 c1.79 i215±15 g4.0 c460 HD34282a A0e+sh j1.39 q0.6 j3.94+0.02 j1.13+0.06 j110±10 j6.4 j348 −0.01 −0.05 HD35187b A2e r2.2 ∗2.3 r3.96 r1.53 h105±9 r5.0 r150 HD287841a A5III:e m1.55±0.15 ∗1.4 m3.87±0.01 m0.74±0.05 - ∗20.0 - HD35929a A5 s3.6±0.2 ∗6.6 s3.85±0.01 s2.00+0.06 f150±30 f3.2 s390±30 −0.07 HD36112a A8e cc1.8±0.2 ∗2.3 q3.89±0.02 ∗1.24 cc60±6 q6.0 ∗200+63 −40 HD244604a A3 - - - - - - - HD245185a A5 g2.2 ∗1.8 g3.97 g1.36 - g8.0 g400 TOria A3V c3.6 c2.8 c4.03±0.05 c1.92 c100 g1.7 c460 V380Oria A0 aa2.8±0.3 aa2.7 c4.03±0.05 c1.93 - ∗1.4 c460 HD37258a A2V h3.0 - h3.96±0.02 t0.59 h160±13 - - BFOria A5II-IIIevar c1.4 c1.3 c3.83±0.05 c0.49 i37±2 ∗25.0 c460 HD37357a A0e - - b3.96 - - u66.0 u480 HD37411a B9V - - h4.03 - - - v510 HD37490a BIIIe c16.9 c11.4 c4.31±0.05 c4.30 c160 - c360 HD37806a A0 h3.0 ∗2.4 f3.95±0.05 f>1.51 f120±30 ∗2.6 - HD250550a B7e c4.8 c3.5 c4.09±0.05 c2.38 c110 g1.4 c700 HD259431a B6pe c12.2 c6.4 c4.31±0.05 c3.80 c100 g0.1 c800 HD52721a B2Vne c24.1 c17.4 c4.35±0.05 c4.83 c400 - c1150 HD53179a Bpe w16 - l3.90 l3.85 - w0.3 l1150 HD53367a B0IV:e e13±3 c17.4 e4.50±0.05 e4.04+0.34 c30 - e250+120 −0.26 −60 NXPupa A0 c3.0 c2.4 c3.99±0.05 c1.65 f120±10 ∗2.3 c450 HD68695b A0V - - - - - - - HD72106Ab A0IV aa2.4±0.4 - aa4.04 aa1.3 aa45 - b290 HD72106Bb A0IV aa1.75±0.25 - aa3.90 aa1.0 10 - b290 HD76534a B2Vne c11.4 c7.5 c4.25±0.05 c3.70 c110 - c870 HD85567a B5Vne - - x4.28 x4.0±0.3 x70 - x1500±500 HD87403b A1III - - - - - - - HD94509a A0Ib - - - - - - - HD95881a A1/A2III/IV - - - - y150 - - HD96042b B1V:ne - - - - b132 - - HD97048a A0pshe c3.4 c2.5 f4.00±0.05 f1.61+0.13 f140±20 f>6.3 c215 −0.10 HD98922a B9Ve ∗5.0 ∗9.2 f4.02±0.05 f>2.96 - ∗0.1 - HD100453b A9Ve k1.7 ∗1.7 k3.86 k0.95+0.18 - ∗10.0 k114 −0.30 HD100546a B9Vne e2.4 ∗1.7 e4.02±0.05 e1.51+0.06 - e∼10 e103+7 −0.05 −6 HD101412a B9.5V aa2.6±0.3 - b3.98 1.74 dd7 2 bb500−700 HD104237a A:pe f2.3 ∗2.7 f3.93±0.05 f1.55+0.06 - f6.3 - −0.05 HD132947a A0 - - - - - - - HD141569a B9.5e j2.0 ∗1.7 j4.00+0.01 j1.41±0.05 j236±15 j4.71 j108 −0.01 HD142666a A8Ve k1.8 ∗1.5 k3.93 k1.04+0.18 i72±2 ∗10.0 k116 −0.30 HD144432a A9/F0V k2.2 ∗3.0 k3.90 k1.51+0.18 i85±4 ∗2.0 k200 −0.30 HD144668a A7IVe e3.1±0.5 c3.9 e3.88±0.05 e1.89+0.17 h180±50 e0.6±0.4 e210+50 −0.15 −30 TyCraa B8e c1.9 c1.5 c3.92±0.05 c0.97 c88 ∗11.5 c130 HD190073a A2IVpe ∗2.7 ∗3.3 f3.95±0.05 f>1.80 z20 ∗1.3 f>290 Table 1. HAeBe programme stars ordered by right ascension, including spectral type, mass, radius, effective temeperature, luminosity, vsini, age and heliocentricdistance,whereavailable.Reference:aThe´etal.(1994),bVieiraetal.(2003),cHillenbrandetal.(1992),dMalfaitetal.(1998),evandenAncker etal.(1997),fvandenAnckeretal.(1998),gHerna´ndezetal.(2004),hBo¨hm&Catala(1995),iMoraetal.(2001),jMer´ıetal.(2004),kDominiketal.(2003), lFuenteetal.(2002),mPinheiroetal.(2003),nAcke&Waelkens(2004),oHabartetal.(2003), pGalazutdinov etal.(2003),qMannings&Sargent(2000), rNattaetal.(2004),sMarconietal.(2000),tKovalchuketal.(1997),uHolmesetal.(2003),vHovhannessianetal.(2001),wvandenAnckeretal.(2004), xMiroshnichenkoetal.(2001),yGradyetal.(1996),zCarporon&Lagrange(1999),aaWadeetal.(2005),bbCorradietal.(1997),ccBeskrovnayaetal.(1999), ddGuimaraesetal.(2006),∗Thiswork. A searchforstrong,ordered magneticfieldsin HerbigAe/Bestars 5 [ht] x 10−3 x 10−3 x 10−3 x 10−3 6 5 5 6 4 4 2 2 V/I 0 V/I 0 V/I 0 V/I 0 −2 −2 −4 −4 −5 −5 3850 3900 3950 4000 4060 4080 4100 4120 4140 4300 4320 4340 4360 4380 4800 4850 4900 4950 Wavelength (angstrom) Wavelength (angstrom) Wavelength (angstrom) Wavelength (angstrom) x 106 x 106 x 106 x 106 3 3 4 4 3 2 2 3 U) U) U) U) D D D2 D A A A A I (1 I (1 I ( I (2 1 0 0 0 1 3850 3900 3950 4000 4060 4080 4100 4120 4140 4300 4320 4340 4360 4380 4800 4850 4900 4950 Wavelength (angstrom) Wavelength (angstrom) Wavelength (angstrom) Wavelength (angstrom) Figure2.StokesIandVprofilesintheFORS1spectrumofHD94660.Fromlefttoright,individualframesshowH8,CaHandKandHǫ;Hδ;Hγ;and Hβ.Thinlinesrepresenttheobservedspectrum,whilethicklinesrepresentthecomputedStokesVspectrumdeterminedusingEq.(1).Notethestrong,clear signaturesassociatedwithallBalmerlines. whereeistheelectroncharge,m theelectronmass,andcthespeed This procedure has been repeated four different ways: using e oflight.FortheatmosphereofamainsequenceAorBtypestar, the whole spectrum, using only metal (and He) lines, using all theweak-fieldapproximationholds,forHlines,uptoaround10- Balmerlines,andusingallBalmerlinesexceptHβ.(Inourgrism 20kG,and,formetallines,uptoabout1kG. 600Bobservations,Hβtypicallyexhibitsthestrongestemissionof Toderivethelongitudinalfield,weminimisetheexpression anyoftheBalmerlines,withemissionintensitydecreasingstrongly for higher series members. Measurement of the field using only (y −hBix −b)2 BalmerlinesHγandhigherisanattempttomitigatetheeffectsof χ2 =X i σz2 i (2) thisemission.)Aflag”n”,”d”,or”D”,asdefinedabove,wasas- i i signed to each kind of analysis. It isclear that a series ”DDDD” denotesacertaindetection,a”nnnn”seriesdenotesanulldetecion, where, for each pixel i in the reduced spectrum, y = (V/I), i i whereasintermediatecases(includingpossibly”d”flags)indicate x =−g C λ2(1/I×dI/dλ),andbisaconstantterm.Uncertainties i eff z i i theneed of further observations tofully asseswhether thestar is σ associatedwitheachreducedspectralpixelareobtainedasde- i magneticornot.Theresultsofthefourdiagnosesofthemagnetic scribedbyBagnuloetal.(2006).Thelongitudinalfielduncertainty fieldaresummarisedincolumns6-9ofTable2,withtheirsignifi- isobtainedfromtheformaluncertainlyofthelinearregression,and cancesummarisedincolumn10. itsderivationisdescribedbyBagnuloetal.(2002). AsisdiscussedlaterinthisSection,weencounteredanumber ofcasesforwhichthefieldwasdetectedataboutthe3-σlevel,and 4.1 Standardstars in which minor changes in the data reduction would transform a marginaldetectioninanullorintoadefinitedetection.Although Inorder toevaluate the nominal operation of theinstrument, and thesecases should certainly beinvestigated further via additional theaccuracyofthereductionandanalysistechniques,weobserved observations,wetriedtoextractfurtherinformationfromtheavail- bothmagneticandnon-magneticstandardstars.ThemagneticAp ablespectra, toformulateamorerobust andreliablecriterionfor star HD 94660 (spectral type A0p) was observed 4 times: twice fielddetection. (consecutively) during the P72 run with grism 600B, and twice Eachobservationblockconsistedofseveralpairsofobserva- during the P74 run with both grism 600B and grism 1200g. The tionsobtainedwiththeretarderwaveplateat−45◦and+45◦.Asex- non-magneticA1VstarHD96441wasobservedduringtheDDT plainedbyBagnuloetal.(2006),therearetwowaystodetermine run.Themagneticdiagnosisforthesestandardstarsisincludedin hBi. (i) One can combine all observations to obtain a final, high Table2. z signaltonoiseratioStokesV profile,fromwhichisdetermineda InP72,thefieldofHD94660measuredinBalmerlineswas uniquevalueofhBi.(ii)Alternatively,onecancombineindividual −2534±63 G and −2429±70 G, and thefieldsmeasured in all z pairsofframesobtainedat−45◦ and+45◦ toobtainseverallower Balmerlinesexcept Hβwere−2363±78G and−2535±71 G. signal-to-noiseratioStokesVprofiles,whichyieldseveralhBi de- The consistency of these results is expected for a main sequence z i terminations.ThefinalhBivalueiscomputedfromtheaverageof ApstarwithnoBalmerlineemission.Ontheotherhand,thefields z thesehBi values.Ifbothmethodsyieldadetectionatthe3σlevel measuredinmetallinesweresignificantlyweaker, −1617±94G z i orhigher,weflagtheobservationwith”D”(implyingadefinitede- and−2026±54G.AsdiscussedbyBagnuloetal.(2006),incon- tectionaccordingtobothcriteri).Ifa3σdetectionisobtainedwith sistency between metal line and Balmer line longitudinal fields onlyoneofthetwomethods,weflagtheobservationwith”d”.Fi- is common in stars with strong fields, and likely results from a nally, if both methods yieldsa null detection (< 3σ) weflagthe breakdown of the weak-field assumption (in the case of metallic observationwith”n”. lines),possiblycoupled withtheinfluenceofnonuniform surface 6 G.A. Wade,S. Bagnulo,D. Drouin,J.D. Landstreet,D. Monin 15 25 30 12 nts nts 20 Balmer nts 10 Metal nts 25 Fspuellctrum e e e e urem 10 urem 15 urem 8 urem 20 as as as 6 as 15 e e 10 e e m 5 m m 4 m 10 # of # of 5 # of 2 # of 5 0 0 0 0 0 2000 4000 6000 0 100 200 300 0 100 200 300 0 100 200 300 S/N σ(1) σ(2) σ(3) B B B Figure3.Histogramsofthesignal-to-noiseratio(S/N,column5ofTable2)andformaluncertaintiesofthelongitudinalfieldmeasurements(columns6-8of Table2). 25 25 ments20 Balmer ments20 Metal measure15 measure15 Number of 105 Number of 105 0 0 0 1 2 3|<B>/σ|4 5 6 7 0 1 2 3|<B>/σ|4 5 6 7 z B z B 25 25 ments20 Fspuellctrum ments20 Bwa/olm Heβr measure15 measure15 Number of 105 Number of 105 0 0 0 1 2 3|<B>/σ|4 5 6 7 0 1 2 3|<B>/σ|4 5 6 7 z B z B Figure4.Histogramsofthedetectionsignificancez=hBzi/σBofmagneticfieldmeasurementsofprogrammestars.Eachpanelcorrespondstooneofthe4 measurementsreportedincolumns6-9ofTable2.Thedashedlineindicatesthe3σdetectionthreshold. distributions of metal abundances. Moreover, these two measure- ble (i.e. a 0.5 arcsec slit) to get the most out of the metallic line mentswereobtainedusingdifferentslitwidths(0.5vs.1arcsec). spectrum,andthatprobablythedifferencebetweenthetwometal- Toexplorethepossibleeffectsoftheslitsizedifference,wehave liclinemeasurementsofHD94660isduetosomecombinationof performedafewnumericalexperiments,computingpolarisedsyn- theseeingandtheslitsize. theticspectraanddegradingtheresolutionofbothStokesI andV toanadjustablevalue. Wecomputed thebinned andsmoothed V InP74,thefieldmeasuredinBalmerlinesfromboththegrism anddI/dλ,andcalculatedtheslopeandinferredlongitudinalfield 600B and grism 1200g observations are consistent, and equal to fromtheregression.Theinterestingresultisthatastheresolving −2437±41G(600B)and−2337±40G(1200g).Thefieldsmea- powerdegradesbelowabout2000,thededucedfieldfrommetallic suredinmetallines(−2074±44Gand−2144±16G)arealsoin linesstartstochange alot. It generally getssmaller, but it varies reasonablygoodagreement. inasomewhatunpredictablemanner.Thisappearstobebasically TheresultsforHD94660arefullyconsistentwiththosede- because,aswereducetheresolution,wearegoingfromtheregime scribedbyBagnuloetal.(2006).Inparticular,thefieldsmeasured whereeachdepressionintheobservedspectrumisroughlyoneline during P74 agree to within 1σ with the fields reported by them. totheregimewhereeachdepressionisoneblendofnearbystrong StokesIandVprofilesofHD94660areillustratedinFig.2. lines, and treating this as a single line is an increasingly bad ap- proximation.Ontheotherhand,changingtheresolvingpowerwill ThesingleobservationofHD96441,obtainedduringtheDDT havenegligibleeffectonBalmerlines(asobserved).Fromthiswe run,wasconsistentwithzerofield(−108±43GinBalmerlines, conclude that it is best to use the highest resolving power possi- and78±65Ginmetallines). A searchforstrong,ordered magneticfieldsin HerbigAe/Bestars 7 x 10−3 x 10−3 x 10−3 x 10−3 4 4 4 4 2 2 2 2 V/I 0 V/I 0 V/I 0 V/I 0 −2 −2 −2 −2 −4 −4 −4 −4 3850 3900 3950 4000 4060 4080 4100 4120 4140 4300 4320 4340 4360 4380 4800 4850 4900 Wavelength (angstrom) Wavelength (angstrom) Wavelength (angstrom) Wavelength (angstrom) x 105 x 105 x 105 x 105 15 15 15 15 10 10 10 10 U) U) U) U) D D D D A A A A I ( 5 I ( 5 I ( 5 I ( 5 0 0 0 0 3850 3900 3950 4000 4060 4080 4100 4120 4140 4300 4320 4340 4360 4380 4800 4850 4900 Wavelength (angstrom) Wavelength (angstrom) Wavelength (angstrom) Wavelength (angstrom) Figure5.StokesIandVprofilesinthespectrumofHD101412. 4.2 Programmestars incolumn10ofTable2,onlyHD101412 andCPD-53295indi- cateasignificant fielddetectedaccording tobothcriteriaofBag- 4.2.1 Dataquality nuloetal.(2006)(andarethereforedenotedwiththesymbol“D”). Noother observation of anyprogramme starresultedinanyfield Twohundred and seventy-two magnetic fieldmeasurements have detectedabovethe3σlevel. beenobtainedfromthe68observationsofthe50programmestars. Themean peakS/N (perangstrom)ofthespectrawas2990:1,but varied significantly from 1340:1-5720:1 due to seeing and trans- 4.2.2 Detectedtargets parencyconditions,andapparentbrightnessofthetarget.Therange ofS/Nresultsinarelativelylargerangeoflongitudinalfielduncer- tainties (all uncertainties in this paper are at 1σ confidence). For HD101412 measurements obtained from Balmer lines only, the mean uncer- taintywas66G,rangingfrom19to162G.Formetallinesonly, HD101412(He3-692,PDS057)wasselectedfromthework themeanuncertaintywas105G,rangingfrom21to291G.Forthe ofThe´ etal.(1994) whereitisnotedasanhistoricallyknown or fullspectrum,themeanuncertaintywas48G,rangingfrom16to strongcandidateHAeBestar.ItwasfirstclassifiedasaHAeBestar 102G.ThesedistributionsareillustratedinFig.3(thedistribution byHuandZhou(1990) basedonitsIRASinfraredexcess.Itex- for all Balmer lines excluding Hβ was very similar to that of all hibitsdouble-peaked Hαemission,aswellasO[]λλ6300,6364 Balmerlines,andsoisnotillustratedhere). emission(e.g.Vieiraetal.2003),andmayexhibitpolarisationvari- Thedistributionoferrorbarsforthemetallinemeasurements abilityontimescalesofminutes(Yudin&Evans1998).Although issomewhatbroaderthanfortheBalmerlineorfullspectrumre- HD101412isnotassociatedwithareflectionnebula(Ray&Eis- gressions.AspointedoutbyBagnuloetal.(2006),thisisduetothe loeffel 1994), Vieira et al. (2003) were able to associated it with factthatBalmerlineshavesimilarstrengthinmostofthesestars, theDC295.0+1.3starformingregion(SFR)intheCentaurusand whereasmetallineschangegreatlyfromstartostar.Althoughitis Cruxcomplex.ItthereforeappearsthatHD101412isabonafide truethatthediversityofBalmerlineprofilesinspectraofHAeBe HAeBestarandthereforeapre-mainsequencestarofintermediate stars is somewhat greater than for main sequence A and B stars, mass. the responsible emission effects are mostly visible in Hα, and to ThesurfacetemperatureofHD101412reportedbyVieiraet alesserdegreeinHβ.AttheresolutionofFORS1,theprofilesof al.(2003)isabout9500K,toocoolforHD101412tobeamisclas- highermembersoftheseriesareverysimilartothoseofmainse- sifiedBestar.Noparallax dataisavailablefor thisstar, although quencestars. basedonthereporteddistancetotheDC295.0+1.3SFR(500-700 The magnetic field diagnoses obtained for the programme pc;Corradietal.1997),Wadeetal.(2005)estimateditspositionon starsaresummarisedinFig.4,whereweshow histogramsofthe theHRdiagram.Thisposition,interpretedusingthemodelevolu- detection significance z = |hBi|/σ . Each frame in Fig. 4 corre- tionarytracksofPalla&Stahler(1993),indicatesthatHD101412 z B spondstooneofthe4measurementsreportedinTable2.Only12 hasamassof2.6±0.3M ,andappearedatthebirthlineonlyabout ⊙ of the272 programme star measurements (corresponding to only 2Myrago.Thissuggeststhatthisstarisstillaccretingmaterialand 4 stars) indicate a magnetic field detection significant at greater burning deuterium as its main source of energy, and its structure thanorequaltothe3σlevel.ThesestarsareBFOri(whichcon- shouldbesignificantlydifferentthanthatofmainsequenceAstars. tributesdetectionsbetween3.2σand6.9σtoall4frames,“dddd”), HD101412wasobservedduringtheP72observingrun.Un- HD 101412 (which contributes detections from 4.0σ to 4.6σ to fortunately, observations of this star were obtained while clouds 3frames,“DddD”),CPD-53295(whichcontributesdetectionsof were passing, and only 4 exposures were obtained, resulting in 4.0σto2frames,“nDDn”),andHD36112(whichcontributesde- some frames with low S/N and others that were partially satu- tectionsfrom3.0σ-3.6σto3frames,“nddn”).Accordingtotheflag rated. Thereduced spectrum(which haslower (1350:1) S/N,and 8 G.A. Wade,S. Bagnulo,D. Drouin,J.D. Landstreet,D. Monin is illustrated in Fig. 5) shows strong Balmer lines which in gen- Klines,consistentwiththelatespectraltype.Neitherthegrism600 eralshow littleevidenceof emission.Hβshowsanasymmetryin northe1200gspectrumStokesVspectrumshowsanysignatures. the inner blue wing which is presumably due to emission. The Thelongitudinalmagneticfieldhasbeendiagnosedusingboth Balmer lines are much broader than those of HD 94660 (which Balmerlinesandmetalliclines(flags“nnnn”and“nDDn”,thelat- has T = 10750 K according to uvbyβ photometry), supporting terwithgrism1200g). Accordingtobothofthedetectioncriteria eff the cooler temperature derived by Vieira et al. (2003). If we as- ofBagnuloetal.(2006),nofieldatallisdetectedinthe600Bob- sume this temperature, the wings of the Hγ and Hδ profiles are servation.Ontheotherhand,afieldisdetectedinthemetallines well-fitusingATLAS9Balmerprofilescomputedforlogg = 4.0. andfullspectruminthe1200gobservation,bothatthe4.0σlevel The metallic-line spectrum is substantially weaker than that of (139±35G,129±32G).Whenthisspectrumisexaminedbyeye, HD94660, although theCaKline(firstpanel ofFig.5) isrel- nosignificantsignaturesareevident. ativelystrong.TheMgλ4481lineandFelinesofmultiplet42 Although we are unable to simply discard the apparent (λλλ4923,5018,5169)arealsoprominent. metallic-linedetectioninthe1200gspectrum,giventhatbothob- Thelongitudinalmagneticfieldhasbeendiagnosedusingboth servations of thisstar were obtained within about 1 hour of each Balmerlinesandmetalliclines(flag“DddD”).Accordingtoboth other,andthattheBalmerlinediagnosisofthe1200gobservation ofthedetectioncriteriaofBagnuloetal.(2006),afieldisdetected isinconsistentwiththemetalliclinediagnosis,wedonotconsider intheBalmerlines:atthe4.6σlevel(512±111G)fromtheBalmer thedetectionofCPD-53295tobeconvincingevidenceofthepres- lineregressionbasedonthefinalreducedStokesI andV spectra, enceofamagneticfield. andthe3.2σlevelbasedontheweightedaverageofregressionre- sultsobtainedfromthe4pairsofStokesI/Vspectra(572±181).In BFORI addition,eachofthe4pairsofspectraproduceBalmerlineregres- sionresultsthatareconsistenttowithin1σwiththeresultreported BFOri(BD-06◦1259)wasselectedfromthecatalogueofThe´ in Table 2. Although no clear Stokes V signatures are evident in etal.(1994). Thisobjectisawell-knownandextensivelystudied BalmerlinesshowninFig.5,theredoesappeartobeareasonably pre-mainsequencestar,exhibitinglargeinfraredexcess,Hαemis- convincingtendancyforthecircularpolarisationtobe positivein sion,andpolarimetricvariability.Hillenbrandetal.(1992) derive thebluewing andnegativeintheredwingoftheline(consistent withapositivelongitudinalmagneticfield)2. an effective temperature of 6750 K, and a mass of 1.4 M⊙, sub- stantiallylowerthanindicatedbytheSIMBADspectraltype(A5 Thefullspectrumregressionalsoindicatesalongitudinalfield II-IIIevar)andderivedbyGrininetal.(2001: 8750K).Thechal- detection, but only according to one of the criteria. On the other lengeofclassifyingthisstarresults,atleastinpart,fromthepres- hand,theregressionusingonlymetalliclines,althoughcorrespond- enceofcomplexnon-photosphericabsorptionlines(Hernandezet ingtoasingle-criteriondetection,showssimilarmagnitudebutop- al.2004).BFOriiswithin1′ofanX-raysourceidentifiedbyDami- positesign.Althoughthiscouldbeattributedtothelargererrorbar ani et al. (1994). According to Boehm & Catala (1994), thisstar resultingfromtheweakandnoisymetallic-linespectrum,anddif- doesnotexhibitO[]λλ6300,6364emission. fersfromtheBalmer-lineresultbyonly2.3σ,itmotivatesustobe BFOriwasobservedduringtheP72observingrun.Eightex- more cautious about claiming theexistence of afieldin thisstar. posureswereobtained,yieldingaS/Nof2900:1.Thespectrum(il- Forthemoment,weconcludetentativelythatHD101412hostsa lustratedinFig.6)ischaracterisedbystrong,asymmetricBalmer photospheric magnetic fieldwithalongitudinal intensityof order lines (a systematic asymmetry can be followed down to at least 500G,butfurtherobservationsareclearlynecessary. H10,andpossiblyH11),andarich,strongmetallic-linespectrum. TheFelinesofmultiplet42areespeciallyprominent,andappear CPD-53295 toshowweakemissioninthewings.TheCaKlineisverystrong. Althoughitisattheveryedgeofourspectraldomain,weseequite CPD-53295(PDS002)wasselectedfromtheworkofVieira clearlyanabsorptionprofileofHeλ58763. etal.(2003),whowereunabletoassociateitwithaspecificstar- Themagneticdiagnosisforthisstarissummarisedbytheflag formingregion.SIMBADindicatesaspectraltypeofF2,andVieira “dddd”, indicatingadetectionaccordingtooneofthecriteriafor et al. report an effective temperature of 6750 K and a rotational each of the four measurements (−180±38 (4.7σ) from Balmer velocityof175kms−1.Otherwise,thisstarisverypoorlystudied, lines, −95±30 G (3.2σ) from metal lines, −144±21 G (6.9σ) appearinginonly5articlesintheADSsince1970. fromthefullspectrum, and−158±45G(3.5σ)fromallBalmer CPD-53 295 was observed on a single night during the P74 linesexcludingHβ.InFig.6,noisybutapprentlysignificantcircu- observing run withboth grisms,obtaining S/Nsof 2925:1 (grism larpolarisationispresentinallBalmerlines.Basedonthedetection 600B)and1800:1(grism1200g).Thespectraarerichinmetallic ofconsistentlongitudinalmagneticfieldsinbothBalmerlinesand lines,showBalmerlinesinabsorption,andverystrongCaHand metallines,wetentativelyconcludethatBFOrihostsamagnetic fieldwithalongitudinal intensityoforder 150G.However, asin thecaseofHD101412, furtherobservationsareclearlynecessar- ilyinordertounambiguouslyestablishthepresenceofamagnetic 2 AsexplainedattheendofSect.8andinFig.8ofBagnuloetal.(2006),in field. theirsurveyofclustermagneticApstarstheyencounterednumerouscases where a simple inspection of the V/I profiles did not show any obvious polarization signal, but in which the magnetic field was detected by the linearcorrelationbetweenV/Iandthequantity1/I ×dI/dλ.Thereliability of this result is supported by the histograms of Fig. 7 of Bagnulo et al. 3 Thedetection ofthis strongabsorption is confirmedfromexamination (2006),showingthatvirtuallynodetectionsarefoundin”normal”Astars, ofhigh-resolution spectra inourpossession. Thedetection ofH λ5876 wherewedonotexpect tofindanymagnetic field.Magnetic fieldswere is remarkable given the assumed temperature of this star, and it will be detectedonlyinApstars,i.e.,wheretheyareexpectedtobefound. exploredindetailinafuturepaper. A searchforstrong,ordered magneticfieldsin HerbigAe/Bestars 9 x 10−3 x 10−3 x 10−3 x 10−3 5 5 5 5 V/I 0 V/I 0 V/I 0 V/I 0 −5 −5 −5 −5 3850 3900 3950 4000 4060 4080 4100 4120 4140 4300 4320 4340 4360 4380 4800 4850 4900 Wavelength (angstrom) Wavelength (angstrom) Wavelength (angstrom) Wavelength (angstrom) x 106 x 106 x 106 x 106 4 6 6 5 3 4 4 4 U) U) U) U) D2 D D D3 A A A A I ( I (2 I (2 I ( 1 2 0 0 0 1 3850 3900 3950 4000 4060 4080 4100 4120 4140 4300 4320 4340 4360 4380 4800 4850 4900 Wavelength (angstrom) Wavelength (angstrom) Wavelength (angstrom) Wavelength (angstrom) Figure6.StokesIandVprofilesinthespectrumofBFOri. HD36112 current,standardisedreductionandanalysispresentedinthispaper (according to the methodology reported by Bagnulo et al. 2006), which we believe to be more accurate and robust and to provide HD 36112 (PDS 183, MWC 758) is a well-studied HAeBe morerealisticerrorbars,doesnotresultinafielddetectionforthis star,selectedfromthecatalogueofThe´ etal.(1994).Vieiraetal. (2003) reportHαandO[]forbiddenlineemission.Thisstarhas star(longitudinalfieldsof166±70GinBalmerlinesand−11±91G in metal lines). Nor is a magnetic field detected in HD 72106B beenstudiedindetailbyBeskrovnayaetal.(1999),whodiscussthe (detection flag“nnnn”, hBi = 76±67 G fromthefull spectrum spectroscopic,photometricandpolarimetricproperties.Theypoint z analysis).ThedifferenceintheresultsforHD72106Acanbeex- outinparticularthestrongnear-IRexcessduetowarmcircumstel- plainedasanaturalresultofthesomewhatlargererrorbarsderived lardust,andpeculiar,rapidvariabilityofcircumstellarlineprofiles, here,coupledwiththeexpectedstatisticalfluctuationsinmeasure- whichtheyspeculatemaybeconnectedwithcircumstellarjet-like mentsobtainedfromindependentreductionsofthedata.However, inhomogeneities. TheHipparcosparallaxof4.89±1.16masyieldsadistance themagneticfieldofHD72106AwasdetectedrepeatedlybyWade of205+63 pc.IncombinationwiththeeffectivetemperatureT = etal.(2005)andFolsometal.(2006)usingtheESPaDOnShigh- 7700±−24000K(Beskrovnayaetal.1999)andthemodelevolutioneffary resolutionspectropolarimeter, withclearStokesV signaturesand tracksofPalla&Stahler(1993),thisyieldsamassof1.8±0.2M , longitudinalfielderrorbarsofabout60G:HD72106Aistherefore ⊙ aconfirmedmagneticobject. andanageof5-10Myr(consistentwiththeresultsofMannings& Sargent(2000),whofind6Myr).Theprojectedrotationalvelocity HD 72106B (PDS031S - thefainter southerncomponent of ofHD36112is60±6km/s(Beskrovnayaetal.1999). this0.8′′doublestar;Harkopfetal.1996;ESA1997)wasselected HD 36112 was observed during P74. Eight exposures were from the investigation of HAeBe group members and candidate obtained,foranetS/Nof3725:1.Thespectrum,illustratedinFig. starsbyVieiraetal.(2003). According tothoseauthors, thisstar 7,isrichinmetallinesandshowsclearcoreemissionatHβ,and exhibitsweakHαemissionandsignificantIRexcess,andmaybe emission at He  λ5876. The Ca  K and Mg  λ4481 lines are anevolvedHAeBeobject,possiblynearingtheendofitsPMSevo- prominent.Themagneticdiagnosisofthisstar(“nddn”)yieldsno lution.Bothstarsexhibitlargepropermotionsandthereisnode- detectioninBalmerlines,butmarginaldetectionsaccordingtoone tectablerelativemotionofthepair(Harkopfetal.1996).According ofthecriteriaofBagnuloetal.(2006)inthemetalliclines(−166± toFolsometal.(2006),theradialvelocitiesareidenticalwithinthe 56G;3.0σ)andthefullspectrum(−149±41G;3.6σ).Although measurementsuncertainties.Thissuggeststhattheyarephysically thisdoesnotconstitutestrongevidenceforamagneticfield,given associated,whichimpliesthattheyarelikelyco-eval.Weobserved theactivitydiscussedbyBeskrovnayaetal.(1999),thisstarshould bothstarsindividually,underconditionsofexcellentseeing,using bereobserved. aslitsizeof0.5′′. Although no temperatures are available in the literature for the individual stars, based on the B−V colours of Fabricius & 4.2.3 Othertargetsofnote Marakov(2000)bothcomponentsappeartotheA-typestars.Our twoFORS1spectraconfirmthis,withwell-developedBalmerlines and similar metallic-line spectra. The Balmer lines have similar HD72106A depth in spectra of both stars, but they are considerably broader in the B component. The Mg  λ4481 lines are well-developed Wade et al. (2005) reported the detection of a weak lon- (strongerinthespectrumoftheAcomponent),asistheCaKline gitudinal magnetic field in HD 72106A using both FORS1 and (substantiallystrongerintheBcomponent).Overall,theAcompo- ESPaDOnS. The FORS1 detection, which corresponded to 5.2σ nent’sspectrumissimilartothatofHD94660-themetallicspec- (Balmerlines) and5.8σ (fullspectrum) detections, wasbased on trum is weaker, and the Balmer lines marginally broader. On the anearlierreductionandanalysisofthesedata(Drouin2005).The otherhand,theBcomponent’sspectrumissuggestiveofacooler 10 G.A. Wade,S. Bagnulo,D. Drouin,J.D.Landstreet, D.Monin x 10−3 x 10−3 x 10−3 x 10−3 3 3 3 3 2 2 2 2 1 1 1 1 V/I 0 V/I 0 V/I 0 V/I 0 −1 −1 −1 −1 −2 −2 −2 −2 −3 −3 −3 −3 3850 3900 3950 4000 4060 4080 4100 4120 4140 4300 4320 4340 4360 4380 4800 4850 4900 Wavelength (angstrom) Wavelength (angstrom) Wavelength (angstrom) Wavelength (angstrom) x 106 x 106 x 106 x 106 6 6 7 8 5 6 7 4 U) U) U) U) D D4 D5 D6 A A A A I (2 I ( I ( I ( 3 4 5 0 2 3 4 3850 3900 3950 4000 4060 4080 4100 4120 4140 4300 4320 4340 4360 4380 4800 4850 4900 Wavelength (angstrom) Wavelength (angstrom) Wavelength (angstrom) Wavelength (angstrom) Figure7.StokesIandVprofilesinthespectrumofHD36112. HD144432,HD31648andHD144668 1.1 1 HD 144432, HD 31648 and HD 144668 are 3 HAeBe stars 0.9 that have been discussed by Hubrig et al. (2004, 2006a). These 0.8 authorsalsousedFORS1tosearchformagneticfieldsinBalmer malised flux00..67 l2tie2ncetGiso,(nr4ei.n0pσoHr)tDirneg1s4pm4ec6at6rigv8ie.nlOyal)fdipneatHreticDctuio1lan4rs4i4(n31t2e1r9aens±tdi3Hs8tDhGe3r1(e36p.41o8σr,t)abnayndHdnuo8b7drei±g- Nor0.5 etal.(2006b)ofstrongStokesVsignaturesinsomespectrallinesof 0.4 thesestars,whichtheyinterprettobedueto“circumstellar”mag- neticfields. 0.3 WehaveobservedHD144432,HD31468andHD144668as 0.2 partofthisprogramme.Noneofthesestarsyieldsanysignificant 0.1 4260 4280 4300 4320 4340 4360 4380 4400 4420 magneticfielddetectionintheirmetallicorBalmerlines,withun- Wavelength (angstrom) certainties comparable to those of Hubrig et al. (2004). We have alsocarefully examined our reduced spectra, and wefindno evi- Figure 8. Observed (symbols) and computed (lines) Hγ profiles of HD dence of strong Stokes V signatures in our observations of these 72106A (brighter northern component - Top) and HD 72106B (fainter stars. southerncomponent-Bottom). HD104237 star.WehaveemployedtheFORS1spectraofthetwocomponents, withtheirrobustcontinuaandBalmerlineprofiles,incombination withATLAS9Balmerlineprofiles(convolvedtothesameresolu- HD104237isacoolHerbigAestarwithasharpmetallic-line tionastheFORS1spectraassumingagaussianinstrumental pro- spectrum in which Donati et al. (1997) claimed the marginal de- file),toconstrainthetemperaturesofthetwocomponents.Wede- tectionofamagneticfieldusinghigh-resolutionspectropolarime- rivefromthecomparisonofobservedandtheoreticalprofilesthat tryandLeast-SquaresDeconvolution.Thisresultwassubsequently HD 72106A has T = 11000±1000 K, 3.5 6 logg 6 4.5 and confirmedbyDonati(2000),whodetectedStokesVsignaturesand eff HD72106BhasT =8000±500K,4.06logg64.5.Thebest- derived a longitudinal field of about 50 G (Donati, priv. comm.) eff fit results (for T = 11000 K, logg = 4.0 and T = 8000 K, using new data and similar techniques. Acke & Waelkens (2004) eff eff logg=4.5)areshowninFig.8. found approximately solar photospheric abundances based on an Vieira et al. (2003) associated HD 72106 with the Vela and equivalentwidthabundancesanalysis. GumNebulaSFRs.ThesystemhasacombinedHipparcosparal- HD 104237 was observed during both P72 and P74. Three lax of 3.47±1.43 mas, yielding a distance of 288+202 pc, which observations (corresponding to12, 10and 8 exposures) were ob- −84 isconsistentwiththatoftheSFR.Wadeetal.(2005)estimatedits tained,fornetS/Nsof2375:1,2800:1and2400:1. Thespectrum, position on the HR diagram. This position, interpreted using the illustrated in Fig. 10, exhibits significant emission at Hβ, which modelevolutionarytracksofPalla&Stahler(1993),indicatesthat showsastrongandvariableemissioncore,aswellasemissionin thecomponents of HD72106 havemassesof 1.75±0.25 M (B the multiplet 42 Fe  lines. The magnetic diagnoses of this star ⊙ component)and2.4±0.4M (Acomponent),andaresituatedclose yieldnodetectionineitherBalmerormetallinesaccordingtoany ⊙ totheZAMS.Althoughuncertain,theseH-Rdiagrampositionsare ofthecriteriaofBagnuloetal.(2006),with besterrorbarsof64G consistentwithaco-evalsystem. (Balmerlines)and35G(metallines).Theuncertaintiesobtained

ebook img

A search for strong, ordered magnetic fields in Herbig Ae/Be stars PDF

0.35 MB·English
by  G. A. Wade
#additional_collections #journals #arxiv
Save to my drive
Quick download
Download
Upgrade Premium
Most books are stored in the elastic cloud where traffic is expensive. For this reason, we have a limit on daily download.

Preview A search for strong, ordered magnetic fields in Herbig Ae/Be stars

Mon.Not.R.Astron.Soc.000,000–000(0000) Printed5February2008 (MNLATEXstylefilev2.2) A search for strong, ordered magnetic fields in Herbig Ae/Be stars⋆ G.A. Wade1, S. Bagnulo2, D. Drouin1, J.D. Landstreet3, D. Monin4 1DepartmentofPhysics,RoyalMilitaryCollegeofCanada,POBox17000,Station’Forces’,Kingston,Ontario,CanadaK7K4B4 2EuropeanSouthernObservatory,Casilla19001,Santiago19,Chile 3DepartmentofPhysics&Astronomy,TheUniversityofWesternOntario,London,Ontario,Canada,N6A3K7 7 4NationalResearchCouncilofCanada,HerzbergInstituteofAstrophysics,5071WestSaanichRd.,Victoria,BC,Canada,V9E2E7 0 0 2 5February2008 n a J ABSTRACT 2 The origin of magnetic fields in intermediate-mass and high-mass stars is fundamentally a 1 mystery. Clues toward solving this basic astrophysical problem can likely be found at the pre-main sequence (PMS) evolutionary stage. With this work, we perform the largest and 1 mostsensitivesearchformagneticfieldsinpre-mainsequenceHerbigAe/Be(HAeBe)stars. v We seek to determine whether strong, orderedmagnetic fields, similar to those of main se- 7 quenceAp/Bpstars, can bedetectedin these objects,andif so,to determinethe intensities, 8 3 geometricalcharacteristics, and statistical incidence of such fields. Sixty-eightobservations 1 of 50 HAeBe stars have been obtained in circularly polarised light using the FORS1 spec- 0 tropolarimeter at the ESO VLT. An analysis of both Balmer and metallic lines reveals the 7 possiblepresenceofweak longitudinalmagneticfields inphotosphericlinesoftwo HAeBe 0 stars,HD101412andBFOri.Resultsfortwoadditionalstars,CPD-53295andHD36112, / aresuggestiveofthepresenceofmagneticfields,butnofirmconclusionscanbedrawnbased h p ontheavailabledata.TheintensityofthelongitudinalfieldsdetectedinHD101412andBF - Orisuggestthattheycorrespondtoglobally-orderedmagneticfieldswithsurfaceintensities o oforder1kG.Ontheotherhand,nomagneticfieldisdetectedin4otherHAeBestarsinour r t sampleinwhichmagneticfieldshadpreviouslybeenconfirmed.MonteCarlosimulationsof s thelongitudinalfieldmeasurementsoftheundetectedstarsallowustoplaceanupperlimits a : of about 300 G on the general presence of aligned magnetic dipole magnetic fields, and of v about500Gonperpendiculardipolefields.Takingintoaccounttheresultsofoursurveyand i X otherpublishedresults,wefindthattheobservedbulkincidenceofmagneticHAeBestarsin oursampleis8-12%,ingoodagreementwiththatofmagneticmainsequencestarsofsimilar r a masses.Wealsofindthatthermslongitudinalfieldintensityofmagnetically-detectedHAeBe starsisabout200G,similar tothatofApstarsandconsistentwith magneticfluxconserva- tion during stellar evolution.These results are all in agreementwith the hypothesisthat the magneticfieldsofmainsequenceAp/Bpstarsarefossils,whichalreadyexistwithinthestars at the pre-main sequence stage. Finally, we explore the ability of our new magnetic data to constrain magnetospheric accretion in Herbig Ae/Be stars, showing that our magnetic data arenotconsistentwiththegeneraloccurrenceinHAeBestarsofmagnetosphericaccretionas describedbythetheoriesofKo¨nigl(1991)andShuetal.(1994). Keywords: stars:magneticfields–polarisation 1 INTRODUCTION suppressing large-scalemixingandleadingtoanamazingarrayof atmosphericpeculiaritieswhicheffectivelydefinethesestarsspec- About 5-10% (e.g. Wolff 1968) of all intermediate mass (B and troscopically(e.g.Adelman1993). A) main sequence stars exhibit organised magnetic fields with strengthsrangingbetweenafewhundred andafew tensof thou- Their magnetic fieldshaveclearly defined observational and sandsofgauss.Thepresenceofthesefieldshasimportantconse- physicalpropertieswhicharetotallydifferentfromthoseofthesun quencesforthestructureoftheatmospheresoftheseAp/Bpstars, andotherlate-typestars.Thefieldsarestrong,withimportantdipo- larcomponentsandfillingfactorsnear100%;thefieldstructureis staticintherotatingstellarreferenceframe,andappearstobesta- ⋆ BasedonobservationsfromtheESOtelescopesattheLaSillaParanal bleontimescalesofdecades;andthefieldsshowonlyweakorno Observatory underprogrammeID072.C-0447,DDT-272.C-5063,074.C- correlationwithprojectedrotational velocity, mainsequence age, 0442. andotherphysicalparameters. 2 G.A. Wade,S. Bagnulo,D. Drouin,J.D. Landstreet,D. Monin Because these characteristicscontrast sostrongly withthose of active cool stars(theonly other broad classof non-degenerate starsinwhich magnetic fieldscan be studied systematically), the magneticAp/Bpstarsrepresentanimportantlaboratoryforinves- tigating the fundamental physics associated with stellar and cos- micmagnetism. Inparticular,becausestarsinthismassrangeare progenitorsofbothwhitedwarfsandneutronstars,theirmagnetic 4 characteristicsmayprovideimportantcluesneededinordertoun- derstandmagnetisminthesedegenerateobjects. lutionReomf tahreksaeblmy,avgenreytilcittfileelidssk.nRoewcnenatbionuvtetshteigfaotriomnastiboynaBnadgnevuolo- 6.0Birthline (10-5 Msun/yr) et al. (2003, 2004, 2006) and Po¨hnl et al. (2003, 2005), showing tvheartymyoaugnngetiocpaenndclcuhsetemrsic,asltlryonpgelcyulsiuaprpAo-rttytpheesvtiaerwsathreatptrheesefinetldins L)sun 5.0 L/* arepresentinthesestarsatthetimetheyreachthemainsequence. g( 4.0 o Thissuggeststhatthemagneticfieldsobserved inmainsequence l 3.5 Ap/Bp stars arepresent, and possibly produced, within their pre- 2 mainsequence (PMS)progenitors. Thisideaisalsosupported by 3.0 the theoretical work of Ste¸pien´ (2000) and of Ste¸pien´ and Land- 2.5 street(2002),whichsuggestthatstrongmagneticfieldsarerequired atthepre-mainsequencephaseinordertoproducetheslowrota- 2.0 tionwhichiscommonlyobservedintheAp/Bpstars.Theprimary aimofthisstudyistotestthesepropositionsbysearchingfordirect 1.5 evidence of magnetic fields in the photospheres of a statistically usefulsampleofPMSstarsofintermediatemass. ZA A second motivation of this study is to explore the gen- M S eral role of magnetic fields in the late stages of formation of 0 4.4 4.2 4 3.8 intermediate-massstars.Accretionontolower-massPMSTTauri log(T ) eff starsisnowgenerallysupposedtobemediatedbythepresenceof strong,large-scalemagneticfields(e.g.discussionbyJohns-Krull Figure1.HRdiagramofsamplestarswithknownTeff andlogL/L⊙,de- rivedfromTable1.Thedataareinterpretedusingthemodelevolutionary etal.(1999)andreferencestherein).Someauthors(e.g.Vinketal. calculations ofPalla&Stahler(1993),whichassumeanaccretionrateof 2002,Muzzeroleetal.2004)havesuggestedthatsimilar“magne- 10−5 M⊙/yrbeforethebirthline. Eachevolutionary trackislabelled with tosphericaccretion”mayoccurinintermediate-massPMSstarsas thecorresponding stellarmass.Notably,amajorityofthesampleiscon- well.Theobservationspresentedhereprovideauniqueopportunity centratedbetween1.5-3 M⊙. Starsinwhichmagneticfieldsarepossibly toexplorethisproposal. detectedinthisstudy(HD101412,BFOri,andHD36112)areindicated Many intermediate-mass PMS stars areidentifiable observa- byfilledsymbols(theluminosityofCPD-53295isnotknown,andthuswe tionally as Herbig Ae/Be (HAeBe) stars (Herbig 1960; Cohen & areunabletoplaceitontheHRdiagram). Kuhi 1979; Finkenzeller & Mundt 1984; Hillebrand et al. 1992). HAeBe stars are characterised by spectral types A and B with strong, often ubiquitous emission lines. They are distinguished 1kG.AsurveyundertakenbyGlagolevski&Chountonov(1998), from the classical Ae/Be stars by their IR colours and frequent including a larger sample of stars, also found no fields, but with presence within dust-obscured regions and association with neb- relativelypoor precision. On theother hand, Donati et al. (1997) ulae(Waters&Waelkens1998).Accordingtostandardstellarevo- andDonati(2000)reportedtheprobableanddefinite(respectively) lution theory, HAeBe stars should not posses deep outer convec- detection of a circular polarisation signature in metallic lines of tionzoneswhichgenerateimportantquantitiesofoutward-flowing the Herbig Aestar HD 104237, providing the first detection of a mechanicalenergy.Rather,thesestarsareexpectedtohaveconvec- magneticfieldinaHAeBestar.Morerecently,Hubrigetal.(2004, tivecoressurroundedby primarilyradiativesub-photosphericen- 2006a)haveclaimedthedetectionofstrongmagneticfieldsinsev- velopes(Iben1965;Gilliland1986).However,since1980,repeated eralHAeBestars.Clearly,indirectanddirectevidenceforthepres- observationsofthesestars(e.g.Praderieetal.1982;Felenboketal. enceofmagneticfieldsinsomeHAeBestarsinincreasing, butit 1983;Catalaetal.1986;Hamann&Persson1992)haveshownthat remainsverysparse. manyareintenselyactive.Inparticular,theydisplaycharacteristics In this study, we use the low-resolution spectropolarimeter oftenassociated withmagneticactivityand thepresence of chro- FORS1attheESOVLTtoconductasearchformagneticfieldsin mospheres or coronae (e.g.Skinner &Yamauchi 1996; Swartzet about50HAeBestars,thelargestsuchsampleeverstudied.First al. 2005). These properties havebeen proposed asindicators that resultsofthisstudyhavealreadybeenreportedinashortpaperby at least some of these stars or their circumstellar envelopes are Wadeetal.(2005). intenselymagneticallyactive. Previous attempts to detect magnetic fields in HAeBe stars haveproducedvariedresults.Catalaetal.(1993)attemptedtode- 2 SAMPLESELECTION tectZeemancircularpolarisationintheFeλ5018andHeλ5876 linesinthespectrumoftheprototypicalHAeBestarABAur.High Our study required the selection of a relatively large number of resolution,highsignal-to-noiseratiospectraofABAurresultedin HAeBestarstoallowustoderivestatisticallymeaningfulconclu- nodetectionofamagneticfield,withupperlimitsontheorderof sionsaboutthepresenceofmagneticfieldsinthesestars,andthe A searchforstrong,ordered magneticfieldsin HerbigAe/Bestars 3 relationshipofmagneticHAeBestarstotheAp/Bpstars.Various yieldsadispersionof1.2Åperpixel.Spectrawereobtainedwitha literaturesourceswereusedfortargetselection,primarilythecata- slitwidthof0.5”or0.8”,yieldingaresolvingpowerof1560or975, loguesofHAeBestarsandHAeBecandidatesbyThe´etal.(1994) respectively.Inafewcases, grism1200g(nowdecommissioned) andbyVieiraetal.(2003). was used, which covers the spectral range from 4310 to 5490 Å ThecatalogueofThe´ etal.(1994)containssixcategoriesof (whichincludesHβandHγonly).Itsdispersionis0.6Åperpixel, stars;thestarsselectedforourstudywereobtainedonlyfromthe andwithaslitwidthof0.5”or0.8”,thisgrismyieldsaresolving first category, which contains stars historically known as HAeBe powerof2800or1800,respectively. stars,orstrongcandidatesofthegroup.Accordingtotheauthors, Sixty-eightobservationsofthe50targetswereobtainedincir- allofthesestarspossessnear-orfar-infraredexcessandemission cularlypolarisedlightfollowingtheprocedure describedbyBag- lines,associatedwiththepresenceofcircumstellardust, discsand nulo et al. (2006). The light gathered by the telescope is sent energeticouflowswhichareusuallyfoundinHAeBestellarenvi- through the Longitudinal Atmospheric Dispersion Corrector and ronments.Ontheotherhand,Vieiraetal.(2003)producedacata- theinstrumentcollimator,thenpassesthroughasingleslitintothe logueofHAeBestarsandstrongcandidatesfromaninitialsearch topsectionoftheFORS1instrument.There,thestellarlightpasses fornewTTauristars(pre-mainsequencestarsoflowermass)using througharotatablequarter-waveplate(locatedinthefilter/camera theInfraredAstronomicalSatellite(IRAS)pointsourcecatalogue1. section of FORS1) oriented at ±45◦ with respect to the fast axis Becausetheinitialsearchwasbasedoncircumstellardustproper- of the analyser, which converts the circularly polarised light into ties, itincluded HAeBestarsalong withtheT Tauri stars. Viera linearly polarized light. The orthogonal linearly polarized beams etal.extractedtheHAeBestarsbyfilteringthedatausingspecific arethenseparatedbytheWollastonprismbeforebeingdispersed requirements such as a spectral type earlier than F5, emission at byagrism.Twospectra(each corresponding tooneof theoppo- Hα, and a minimum level of infrared emission. The majority of sitelypolarisedbeams)arefinallyrecordedsimultaneouslyonthe thestarswereassociatedbytheauthorswithastarformingregion CCD.Bychanging theorientationof thequarter-waveplatewith (SFR).Basedonthequalityofthesetwoliteraturesourcesandthe respect to the Wollaston fast axis (alternately +45◦ or −45◦) and arguments presented by theirauthors, weconclude that all of the obtaining an additional exposure, the positions of the right and starsinoursamplearebonafideHAeBestars.Intotal,50HAeBe left circularly polarised beams are exchanged throughout the in- starshavebeenselected,spanninginaspectraltyperangefromF2 strument, and in particular their respective positions on the CCD throughB0. areinterchanged.AsdescribedintheFORS1manual(VLT-MAN- TheselectedstarsarelistedinTable1.Thespectraltypesof ESO-13100-1543), this procedure allows for the removal of sys- the stars have been obtained from the bibliographic service SIM- tematicerrorsduetoCCDinhomogeneitiesanddifferentialwave- BAD (unless indicated otherwise). Table 1 also contains physical lengthcalibrationerrors. properties(whereavailable)suchasmass,radius,effectivetemper- Inordertoobtainahighsignal-to-noiseratio,multipleexpo- ature,luminosity, vsini,ageanddistance, found intheliterature. suresofeachstarwereobtained.Anevennumberofexposures(be- Wheneverthesourcespermitted,theuncertaintiesassociatedwith tween2and30, depending ontheapparent magnitude ofthestar theseparameterswereincluded. andtheseeing)wasobtained,eachpairofexposurescorresponding AscanbeseeninTable1, 7objectshavenowell-determined toaquarter-waveplateorientationofeither+45oor−45o.Thedata physical parametersintheliterature.Inaddition, thephysical pa- werereducedusingtheproceduredescribedindetailbyBagnuloet rameters (and in particular the luminosities) of many objects are al.(2006). highlyuncertain. Notwithstandingthislimitation,wehaveplaced Thejournalofobservations(Table2)reportsourobservations those stars with known temperatures and luminosities on the HR ofbothstandardandprogrammestars.Itshouldbenotedthatgrism diagraminFig.1,alongwiththemodelpre-mainsequenceevolu- 600Bwasusedforallobservationsunlessindicatedotherwise(in tionarytracksofPalla&Stahler(1993).BasedonFig.1,mostof whichcasegrism1200gwasused).Table2includestheJulianDate thestarsintheobservedsamplehavemassesabove2M ,andfully of mid-observation, the slit width, exposure time and number of ⊙ 18ofthe37starsonthediagram(49%)havemassesbetween2and exposures,andtheresultant peaksignal-to-noiseratio perÅ.The 4M .Eightstarshavemassesbelow2M ,and9starshavemasses tablealsocontainstheresultsofthemagneticfielddiagnosis,which ⊙ ⊙ largerthan4M . arediscussedinSect.4. ⊙ 3 OBSERVATIONSANDDATAREDUCTION 4 RESULTS ThesamplestarswereobservedattheEuropeanSouthernObser- vatory (ESO) Very Large Telescope (VLT) with the first of the Thelongitudinal magnetic fieldhasbeen inferredfromthepolar- two FOcal Reducer and low dispersion Spectrographs (FORS1). isationspectrain themanner described by Bagnulo et al.(2006). FORS1 was mounted at the Cassegrain focus of one of the tele- Briefly, the field diagnosis is obtained using a Least-Squares fit scope unitsand wasused inspectropolarimetricfast mode. Ato- based on the predicted circularly polarised flux V/I in the weak- tal of about five nights were allocated for our study, distributed fieldregime,givenby overthreeobservingruns(visitormoderunsinFebruary(P72)and V 1 dI November2004(P74),andoneservicemoderunofDirector’sDis- I =−geffCzλ2 I dλ hBzi, (1) cretionaryTimeinApril2004). Thelargemajorityofourspectrawereobtainedwiththegrism wheregeff istheeffectiveLande´factor,Iistheusual(unpolarised) 600B,whichcoversthespectralrangefrom3450to5900Å,and intensity,λisthewavelengthexpressedinÅ,hBziisthelongitudi- nalfieldexpressedingauss,and e C = (≃4.67×10−13Å−1G−1) 1 http://irsa.ipac.caltech.edu/ z 4πm c2 e 4 G.A. Wade,S. Bagnulo,D. Drouin,J.D. Landstreet,D. Monin Name Sp.Type Mass R∗ log(Teff) log(L∗/L⊙) vsini Age Distance (M⊙) (R⊙) (K) (km/s) (Myr) (pc) CPD-53295b F2 - - b3.83 - b175 - - HD17081d B7IV ∗4.6 ∗4.9 o4.09 o2.70 n18 o0.1 o135 HD278937a A3 g2.0 ∗1.9 g3.92 g1.21 - g8.0 g350 HD275877a A2IIvar+ g1.9 ∗1.7 g3.92 g1.12 i217±13 g11.0 g120 HD31293a A0Vpe f2.4 b2.7 f3.98±0.05 f1.68+0.13 f80 f2.0 b160 −0.11 HD31648a A3pshe+ f2.2 ∗2.5 f3.94±0.05 f1.51+0.15 i102±5 f2.5 p131±20 −0.13 HD293782a A3e c3.3 c3.2 c3.96±0.05 c1.79 i215±15 g4.0 c460 HD34282a A0e+sh j1.39 q0.6 j3.94+0.02 j1.13+0.06 j110±10 j6.4 j348 −0.01 −0.05 HD35187b A2e r2.2 ∗2.3 r3.96 r1.53 h105±9 r5.0 r150 HD287841a A5III:e m1.55±0.15 ∗1.4 m3.87±0.01 m0.74±0.05 - ∗20.0 - HD35929a A5 s3.6±0.2 ∗6.6 s3.85±0.01 s2.00+0.06 f150±30 f3.2 s390±30 −0.07 HD36112a A8e cc1.8±0.2 ∗2.3 q3.89±0.02 ∗1.24 cc60±6 q6.0 ∗200+63 −40 HD244604a A3 - - - - - - - HD245185a A5 g2.2 ∗1.8 g3.97 g1.36 - g8.0 g400 TOria A3V c3.6 c2.8 c4.03±0.05 c1.92 c100 g1.7 c460 V380Oria A0 aa2.8±0.3 aa2.7 c4.03±0.05 c1.93 - ∗1.4 c460 HD37258a A2V h3.0 - h3.96±0.02 t0.59 h160±13 - - BFOria A5II-IIIevar c1.4 c1.3 c3.83±0.05 c0.49 i37±2 ∗25.0 c460 HD37357a A0e - - b3.96 - - u66.0 u480 HD37411a B9V - - h4.03 - - - v510 HD37490a BIIIe c16.9 c11.4 c4.31±0.05 c4.30 c160 - c360 HD37806a A0 h3.0 ∗2.4 f3.95±0.05 f>1.51 f120±30 ∗2.6 - HD250550a B7e c4.8 c3.5 c4.09±0.05 c2.38 c110 g1.4 c700 HD259431a B6pe c12.2 c6.4 c4.31±0.05 c3.80 c100 g0.1 c800 HD52721a B2Vne c24.1 c17.4 c4.35±0.05 c4.83 c400 - c1150 HD53179a Bpe w16 - l3.90 l3.85 - w0.3 l1150 HD53367a B0IV:e e13±3 c17.4 e4.50±0.05 e4.04+0.34 c30 - e250+120 −0.26 −60 NXPupa A0 c3.0 c2.4 c3.99±0.05 c1.65 f120±10 ∗2.3 c450 HD68695b A0V - - - - - - - HD72106Ab A0IV aa2.4±0.4 - aa4.04 aa1.3 aa45 - b290 HD72106Bb A0IV aa1.75±0.25 - aa3.90 aa1.0 10 - b290 HD76534a B2Vne c11.4 c7.5 c4.25±0.05 c3.70 c110 - c870 HD85567a B5Vne - - x4.28 x4.0±0.3 x70 - x1500±500 HD87403b A1III - - - - - - - HD94509a A0Ib - - - - - - - HD95881a A1/A2III/IV - - - - y150 - - HD96042b B1V:ne - - - - b132 - - HD97048a A0pshe c3.4 c2.5 f4.00±0.05 f1.61+0.13 f140±20 f>6.3 c215 −0.10 HD98922a B9Ve ∗5.0 ∗9.2 f4.02±0.05 f>2.96 - ∗0.1 - HD100453b A9Ve k1.7 ∗1.7 k3.86 k0.95+0.18 - ∗10.0 k114 −0.30 HD100546a B9Vne e2.4 ∗1.7 e4.02±0.05 e1.51+0.06 - e∼10 e103+7 −0.05 −6 HD101412a B9.5V aa2.6±0.3 - b3.98 1.74 dd7 2 bb500−700 HD104237a A:pe f2.3 ∗2.7 f3.93±0.05 f1.55+0.06 - f6.3 - −0.05 HD132947a A0 - - - - - - - HD141569a B9.5e j2.0 ∗1.7 j4.00+0.01 j1.41±0.05 j236±15 j4.71 j108 −0.01 HD142666a A8Ve k1.8 ∗1.5 k3.93 k1.04+0.18 i72±2 ∗10.0 k116 −0.30 HD144432a A9/F0V k2.2 ∗3.0 k3.90 k1.51+0.18 i85±4 ∗2.0 k200 −0.30 HD144668a A7IVe e3.1±0.5 c3.9 e3.88±0.05 e1.89+0.17 h180±50 e0.6±0.4 e210+50 −0.15 −30 TyCraa B8e c1.9 c1.5 c3.92±0.05 c0.97 c88 ∗11.5 c130 HD190073a A2IVpe ∗2.7 ∗3.3 f3.95±0.05 f>1.80 z20 ∗1.3 f>290 Table 1. HAeBe programme stars ordered by right ascension, including spectral type, mass, radius, effective temeperature, luminosity, vsini, age and heliocentricdistance,whereavailable.Reference:aThe´etal.(1994),bVieiraetal.(2003),cHillenbrandetal.(1992),dMalfaitetal.(1998),evandenAncker etal.(1997),fvandenAnckeretal.(1998),gHerna´ndezetal.(2004),hBo¨hm&Catala(1995),iMoraetal.(2001),jMer´ıetal.(2004),kDominiketal.(2003), lFuenteetal.(2002),mPinheiroetal.(2003),nAcke&Waelkens(2004),oHabartetal.(2003), pGalazutdinov etal.(2003),qMannings&Sargent(2000), rNattaetal.(2004),sMarconietal.(2000),tKovalchuketal.(1997),uHolmesetal.(2003),vHovhannessianetal.(2001),wvandenAnckeretal.(2004), xMiroshnichenkoetal.(2001),yGradyetal.(1996),zCarporon&Lagrange(1999),aaWadeetal.(2005),bbCorradietal.(1997),ccBeskrovnayaetal.(1999), ddGuimaraesetal.(2006),∗Thiswork. A searchforstrong,ordered magneticfieldsin HerbigAe/Bestars 5 [ht] x 10−3 x 10−3 x 10−3 x 10−3 6 5 5 6 4 4 2 2 V/I 0 V/I 0 V/I 0 V/I 0 −2 −2 −4 −4 −5 −5 3850 3900 3950 4000 4060 4080 4100 4120 4140 4300 4320 4340 4360 4380 4800 4850 4900 4950 Wavelength (angstrom) Wavelength (angstrom) Wavelength (angstrom) Wavelength (angstrom) x 106 x 106 x 106 x 106 3 3 4 4 3 2 2 3 U) U) U) U) D D D2 D A A A A I (1 I (1 I ( I (2 1 0 0 0 1 3850 3900 3950 4000 4060 4080 4100 4120 4140 4300 4320 4340 4360 4380 4800 4850 4900 4950 Wavelength (angstrom) Wavelength (angstrom) Wavelength (angstrom) Wavelength (angstrom) Figure2.StokesIandVprofilesintheFORS1spectrumofHD94660.Fromlefttoright,individualframesshowH8,CaHandKandHǫ;Hδ;Hγ;and Hβ.Thinlinesrepresenttheobservedspectrum,whilethicklinesrepresentthecomputedStokesVspectrumdeterminedusingEq.(1).Notethestrong,clear signaturesassociatedwithallBalmerlines. whereeistheelectroncharge,m theelectronmass,andcthespeed This procedure has been repeated four different ways: using e oflight.FortheatmosphereofamainsequenceAorBtypestar, the whole spectrum, using only metal (and He) lines, using all theweak-fieldapproximationholds,forHlines,uptoaround10- Balmerlines,andusingallBalmerlinesexceptHβ.(Inourgrism 20kG,and,formetallines,uptoabout1kG. 600Bobservations,Hβtypicallyexhibitsthestrongestemissionof Toderivethelongitudinalfield,weminimisetheexpression anyoftheBalmerlines,withemissionintensitydecreasingstrongly for higher series members. Measurement of the field using only (y −hBix −b)2 BalmerlinesHγandhigherisanattempttomitigatetheeffectsof χ2 =X i σz2 i (2) thisemission.)Aflag”n”,”d”,or”D”,asdefinedabove,wasas- i i signed to each kind of analysis. It isclear that a series ”DDDD” denotesacertaindetection,a”nnnn”seriesdenotesanulldetecion, where, for each pixel i in the reduced spectrum, y = (V/I), i i whereasintermediatecases(includingpossibly”d”flags)indicate x =−g C λ2(1/I×dI/dλ),andbisaconstantterm.Uncertainties i eff z i i theneed of further observations tofully asseswhether thestar is σ associatedwitheachreducedspectralpixelareobtainedasde- i magneticornot.Theresultsofthefourdiagnosesofthemagnetic scribedbyBagnuloetal.(2006).Thelongitudinalfielduncertainty fieldaresummarisedincolumns6-9ofTable2,withtheirsignifi- isobtainedfromtheformaluncertainlyofthelinearregression,and cancesummarisedincolumn10. itsderivationisdescribedbyBagnuloetal.(2002). AsisdiscussedlaterinthisSection,weencounteredanumber ofcasesforwhichthefieldwasdetectedataboutthe3-σlevel,and 4.1 Standardstars in which minor changes in the data reduction would transform a marginaldetectioninanullorintoadefinitedetection.Although Inorder toevaluate the nominal operation of theinstrument, and thesecases should certainly beinvestigated further via additional theaccuracyofthereductionandanalysistechniques,weobserved observations,wetriedtoextractfurtherinformationfromtheavail- bothmagneticandnon-magneticstandardstars.ThemagneticAp ablespectra, toformulateamorerobust andreliablecriterionfor star HD 94660 (spectral type A0p) was observed 4 times: twice fielddetection. (consecutively) during the P72 run with grism 600B, and twice Eachobservationblockconsistedofseveralpairsofobserva- during the P74 run with both grism 600B and grism 1200g. The tionsobtainedwiththeretarderwaveplateat−45◦and+45◦.Asex- non-magneticA1VstarHD96441wasobservedduringtheDDT plainedbyBagnuloetal.(2006),therearetwowaystodetermine run.Themagneticdiagnosisforthesestandardstarsisincludedin hBi. (i) One can combine all observations to obtain a final, high Table2. z signaltonoiseratioStokesV profile,fromwhichisdetermineda InP72,thefieldofHD94660measuredinBalmerlineswas uniquevalueofhBi.(ii)Alternatively,onecancombineindividual −2534±63 G and −2429±70 G, and thefieldsmeasured in all z pairsofframesobtainedat−45◦ and+45◦ toobtainseverallower Balmerlinesexcept Hβwere−2363±78G and−2535±71 G. signal-to-noiseratioStokesVprofiles,whichyieldseveralhBi de- The consistency of these results is expected for a main sequence z i terminations.ThefinalhBivalueiscomputedfromtheaverageof ApstarwithnoBalmerlineemission.Ontheotherhand,thefields z thesehBi values.Ifbothmethodsyieldadetectionatthe3σlevel measuredinmetallinesweresignificantlyweaker, −1617±94G z i orhigher,weflagtheobservationwith”D”(implyingadefinitede- and−2026±54G.AsdiscussedbyBagnuloetal.(2006),incon- tectionaccordingtobothcriteri).Ifa3σdetectionisobtainedwith sistency between metal line and Balmer line longitudinal fields onlyoneofthetwomethods,weflagtheobservationwith”d”.Fi- is common in stars with strong fields, and likely results from a nally, if both methods yieldsa null detection (< 3σ) weflagthe breakdown of the weak-field assumption (in the case of metallic observationwith”n”. lines),possiblycoupled withtheinfluenceofnonuniform surface 6 G.A. Wade,S. Bagnulo,D. Drouin,J.D. Landstreet,D. Monin 15 25 30 12 nts nts 20 Balmer nts 10 Metal nts 25 Fspuellctrum e e e e urem 10 urem 15 urem 8 urem 20 as as as 6 as 15 e e 10 e e m 5 m m 4 m 10 # of # of 5 # of 2 # of 5 0 0 0 0 0 2000 4000 6000 0 100 200 300 0 100 200 300 0 100 200 300 S/N σ(1) σ(2) σ(3) B B B Figure3.Histogramsofthesignal-to-noiseratio(S/N,column5ofTable2)andformaluncertaintiesofthelongitudinalfieldmeasurements(columns6-8of Table2). 25 25 ments20 Balmer ments20 Metal measure15 measure15 Number of 105 Number of 105 0 0 0 1 2 3|<B>/σ|4 5 6 7 0 1 2 3|<B>/σ|4 5 6 7 z B z B 25 25 ments20 Fspuellctrum ments20 Bwa/olm Heβr measure15 measure15 Number of 105 Number of 105 0 0 0 1 2 3|<B>/σ|4 5 6 7 0 1 2 3|<B>/σ|4 5 6 7 z B z B Figure4.Histogramsofthedetectionsignificancez=hBzi/σBofmagneticfieldmeasurementsofprogrammestars.Eachpanelcorrespondstooneofthe4 measurementsreportedincolumns6-9ofTable2.Thedashedlineindicatesthe3σdetectionthreshold. distributions of metal abundances. Moreover, these two measure- ble (i.e. a 0.5 arcsec slit) to get the most out of the metallic line mentswereobtainedusingdifferentslitwidths(0.5vs.1arcsec). spectrum,andthatprobablythedifferencebetweenthetwometal- Toexplorethepossibleeffectsoftheslitsizedifference,wehave liclinemeasurementsofHD94660isduetosomecombinationof performedafewnumericalexperiments,computingpolarisedsyn- theseeingandtheslitsize. theticspectraanddegradingtheresolutionofbothStokesI andV toanadjustablevalue. Wecomputed thebinned andsmoothed V InP74,thefieldmeasuredinBalmerlinesfromboththegrism anddI/dλ,andcalculatedtheslopeandinferredlongitudinalfield 600B and grism 1200g observations are consistent, and equal to fromtheregression.Theinterestingresultisthatastheresolving −2437±41G(600B)and−2337±40G(1200g).Thefieldsmea- powerdegradesbelowabout2000,thededucedfieldfrommetallic suredinmetallines(−2074±44Gand−2144±16G)arealsoin linesstartstochange alot. It generally getssmaller, but it varies reasonablygoodagreement. inasomewhatunpredictablemanner.Thisappearstobebasically TheresultsforHD94660arefullyconsistentwiththosede- because,aswereducetheresolution,wearegoingfromtheregime scribedbyBagnuloetal.(2006).Inparticular,thefieldsmeasured whereeachdepressionintheobservedspectrumisroughlyoneline during P74 agree to within 1σ with the fields reported by them. totheregimewhereeachdepressionisoneblendofnearbystrong StokesIandVprofilesofHD94660areillustratedinFig.2. lines, and treating this as a single line is an increasingly bad ap- proximation.Ontheotherhand,changingtheresolvingpowerwill ThesingleobservationofHD96441,obtainedduringtheDDT havenegligibleeffectonBalmerlines(asobserved).Fromthiswe run,wasconsistentwithzerofield(−108±43GinBalmerlines, conclude that it is best to use the highest resolving power possi- and78±65Ginmetallines). A searchforstrong,ordered magneticfieldsin HerbigAe/Bestars 7 x 10−3 x 10−3 x 10−3 x 10−3 4 4 4 4 2 2 2 2 V/I 0 V/I 0 V/I 0 V/I 0 −2 −2 −2 −2 −4 −4 −4 −4 3850 3900 3950 4000 4060 4080 4100 4120 4140 4300 4320 4340 4360 4380 4800 4850 4900 Wavelength (angstrom) Wavelength (angstrom) Wavelength (angstrom) Wavelength (angstrom) x 105 x 105 x 105 x 105 15 15 15 15 10 10 10 10 U) U) U) U) D D D D A A A A I ( 5 I ( 5 I ( 5 I ( 5 0 0 0 0 3850 3900 3950 4000 4060 4080 4100 4120 4140 4300 4320 4340 4360 4380 4800 4850 4900 Wavelength (angstrom) Wavelength (angstrom) Wavelength (angstrom) Wavelength (angstrom) Figure5.StokesIandVprofilesinthespectrumofHD101412. 4.2 Programmestars incolumn10ofTable2,onlyHD101412 andCPD-53295indi- cateasignificant fielddetectedaccording tobothcriteriaofBag- 4.2.1 Dataquality nuloetal.(2006)(andarethereforedenotedwiththesymbol“D”). Noother observation of anyprogramme starresultedinanyfield Twohundred and seventy-two magnetic fieldmeasurements have detectedabovethe3σlevel. beenobtainedfromthe68observationsofthe50programmestars. Themean peakS/N (perangstrom)ofthespectrawas2990:1,but varied significantly from 1340:1-5720:1 due to seeing and trans- 4.2.2 Detectedtargets parencyconditions,andapparentbrightnessofthetarget.Therange ofS/Nresultsinarelativelylargerangeoflongitudinalfielduncer- tainties (all uncertainties in this paper are at 1σ confidence). For HD101412 measurements obtained from Balmer lines only, the mean uncer- taintywas66G,rangingfrom19to162G.Formetallinesonly, HD101412(He3-692,PDS057)wasselectedfromthework themeanuncertaintywas105G,rangingfrom21to291G.Forthe ofThe´ etal.(1994) whereitisnotedasanhistoricallyknown or fullspectrum,themeanuncertaintywas48G,rangingfrom16to strongcandidateHAeBestar.ItwasfirstclassifiedasaHAeBestar 102G.ThesedistributionsareillustratedinFig.3(thedistribution byHuandZhou(1990) basedonitsIRASinfraredexcess.Itex- for all Balmer lines excluding Hβ was very similar to that of all hibitsdouble-peaked Hαemission,aswellasO[]λλ6300,6364 Balmerlines,andsoisnotillustratedhere). emission(e.g.Vieiraetal.2003),andmayexhibitpolarisationvari- Thedistributionoferrorbarsforthemetallinemeasurements abilityontimescalesofminutes(Yudin&Evans1998).Although issomewhatbroaderthanfortheBalmerlineorfullspectrumre- HD101412isnotassociatedwithareflectionnebula(Ray&Eis- gressions.AspointedoutbyBagnuloetal.(2006),thisisduetothe loeffel 1994), Vieira et al. (2003) were able to associated it with factthatBalmerlineshavesimilarstrengthinmostofthesestars, theDC295.0+1.3starformingregion(SFR)intheCentaurusand whereasmetallineschangegreatlyfromstartostar.Althoughitis Cruxcomplex.ItthereforeappearsthatHD101412isabonafide truethatthediversityofBalmerlineprofilesinspectraofHAeBe HAeBestarandthereforeapre-mainsequencestarofintermediate stars is somewhat greater than for main sequence A and B stars, mass. the responsible emission effects are mostly visible in Hα, and to ThesurfacetemperatureofHD101412reportedbyVieiraet alesserdegreeinHβ.AttheresolutionofFORS1,theprofilesof al.(2003)isabout9500K,toocoolforHD101412tobeamisclas- highermembersoftheseriesareverysimilartothoseofmainse- sifiedBestar.Noparallax dataisavailablefor thisstar, although quencestars. basedonthereporteddistancetotheDC295.0+1.3SFR(500-700 The magnetic field diagnoses obtained for the programme pc;Corradietal.1997),Wadeetal.(2005)estimateditspositionon starsaresummarisedinFig.4,whereweshow histogramsofthe theHRdiagram.Thisposition,interpretedusingthemodelevolu- detection significance z = |hBi|/σ . Each frame in Fig. 4 corre- tionarytracksofPalla&Stahler(1993),indicatesthatHD101412 z B spondstooneofthe4measurementsreportedinTable2.Only12 hasamassof2.6±0.3M ,andappearedatthebirthlineonlyabout ⊙ of the272 programme star measurements (corresponding to only 2Myrago.Thissuggeststhatthisstarisstillaccretingmaterialand 4 stars) indicate a magnetic field detection significant at greater burning deuterium as its main source of energy, and its structure thanorequaltothe3σlevel.ThesestarsareBFOri(whichcon- shouldbesignificantlydifferentthanthatofmainsequenceAstars. tributesdetectionsbetween3.2σand6.9σtoall4frames,“dddd”), HD101412wasobservedduringtheP72observingrun.Un- HD 101412 (which contributes detections from 4.0σ to 4.6σ to fortunately, observations of this star were obtained while clouds 3frames,“DddD”),CPD-53295(whichcontributesdetectionsof were passing, and only 4 exposures were obtained, resulting in 4.0σto2frames,“nDDn”),andHD36112(whichcontributesde- some frames with low S/N and others that were partially satu- tectionsfrom3.0σ-3.6σto3frames,“nddn”).Accordingtotheflag rated. Thereduced spectrum(which haslower (1350:1) S/N,and 8 G.A. Wade,S. Bagnulo,D. Drouin,J.D. Landstreet,D. Monin is illustrated in Fig. 5) shows strong Balmer lines which in gen- Klines,consistentwiththelatespectraltype.Neitherthegrism600 eralshow littleevidenceof emission.Hβshowsanasymmetryin northe1200gspectrumStokesVspectrumshowsanysignatures. the inner blue wing which is presumably due to emission. The Thelongitudinalmagneticfieldhasbeendiagnosedusingboth Balmer lines are much broader than those of HD 94660 (which Balmerlinesandmetalliclines(flags“nnnn”and“nDDn”,thelat- has T = 10750 K according to uvbyβ photometry), supporting terwithgrism1200g). Accordingtobothofthedetectioncriteria eff the cooler temperature derived by Vieira et al. (2003). If we as- ofBagnuloetal.(2006),nofieldatallisdetectedinthe600Bob- sume this temperature, the wings of the Hγ and Hδ profiles are servation.Ontheotherhand,afieldisdetectedinthemetallines well-fitusingATLAS9Balmerprofilescomputedforlogg = 4.0. andfullspectruminthe1200gobservation,bothatthe4.0σlevel The metallic-line spectrum is substantially weaker than that of (139±35G,129±32G).Whenthisspectrumisexaminedbyeye, HD94660, although theCaKline(firstpanel ofFig.5) isrel- nosignificantsignaturesareevident. ativelystrong.TheMgλ4481lineandFelinesofmultiplet42 Although we are unable to simply discard the apparent (λλλ4923,5018,5169)arealsoprominent. metallic-linedetectioninthe1200gspectrum,giventhatbothob- Thelongitudinalmagneticfieldhasbeendiagnosedusingboth servations of thisstar were obtained within about 1 hour of each Balmerlinesandmetalliclines(flag“DddD”).Accordingtoboth other,andthattheBalmerlinediagnosisofthe1200gobservation ofthedetectioncriteriaofBagnuloetal.(2006),afieldisdetected isinconsistentwiththemetalliclinediagnosis,wedonotconsider intheBalmerlines:atthe4.6σlevel(512±111G)fromtheBalmer thedetectionofCPD-53295tobeconvincingevidenceofthepres- lineregressionbasedonthefinalreducedStokesI andV spectra, enceofamagneticfield. andthe3.2σlevelbasedontheweightedaverageofregressionre- sultsobtainedfromthe4pairsofStokesI/Vspectra(572±181).In BFORI addition,eachofthe4pairsofspectraproduceBalmerlineregres- sionresultsthatareconsistenttowithin1σwiththeresultreported BFOri(BD-06◦1259)wasselectedfromthecatalogueofThe´ in Table 2. Although no clear Stokes V signatures are evident in etal.(1994). Thisobjectisawell-knownandextensivelystudied BalmerlinesshowninFig.5,theredoesappeartobeareasonably pre-mainsequencestar,exhibitinglargeinfraredexcess,Hαemis- convincingtendancyforthecircularpolarisationtobe positivein sion,andpolarimetricvariability.Hillenbrandetal.(1992) derive thebluewing andnegativeintheredwingoftheline(consistent withapositivelongitudinalmagneticfield)2. an effective temperature of 6750 K, and a mass of 1.4 M⊙, sub- stantiallylowerthanindicatedbytheSIMBADspectraltype(A5 Thefullspectrumregressionalsoindicatesalongitudinalfield II-IIIevar)andderivedbyGrininetal.(2001: 8750K).Thechal- detection, but only according to one of the criteria. On the other lengeofclassifyingthisstarresults,atleastinpart,fromthepres- hand,theregressionusingonlymetalliclines,althoughcorrespond- enceofcomplexnon-photosphericabsorptionlines(Hernandezet ingtoasingle-criteriondetection,showssimilarmagnitudebutop- al.2004).BFOriiswithin1′ofanX-raysourceidentifiedbyDami- positesign.Althoughthiscouldbeattributedtothelargererrorbar ani et al. (1994). According to Boehm & Catala (1994), thisstar resultingfromtheweakandnoisymetallic-linespectrum,anddif- doesnotexhibitO[]λλ6300,6364emission. fersfromtheBalmer-lineresultbyonly2.3σ,itmotivatesustobe BFOriwasobservedduringtheP72observingrun.Eightex- more cautious about claiming theexistence of afieldin thisstar. posureswereobtained,yieldingaS/Nof2900:1.Thespectrum(il- Forthemoment,weconcludetentativelythatHD101412hostsa lustratedinFig.6)ischaracterisedbystrong,asymmetricBalmer photospheric magnetic fieldwithalongitudinal intensityof order lines (a systematic asymmetry can be followed down to at least 500G,butfurtherobservationsareclearlynecessary. H10,andpossiblyH11),andarich,strongmetallic-linespectrum. TheFelinesofmultiplet42areespeciallyprominent,andappear CPD-53295 toshowweakemissioninthewings.TheCaKlineisverystrong. Althoughitisattheveryedgeofourspectraldomain,weseequite CPD-53295(PDS002)wasselectedfromtheworkofVieira clearlyanabsorptionprofileofHeλ58763. etal.(2003),whowereunabletoassociateitwithaspecificstar- Themagneticdiagnosisforthisstarissummarisedbytheflag formingregion.SIMBADindicatesaspectraltypeofF2,andVieira “dddd”, indicatingadetectionaccordingtooneofthecriteriafor et al. report an effective temperature of 6750 K and a rotational each of the four measurements (−180±38 (4.7σ) from Balmer velocityof175kms−1.Otherwise,thisstarisverypoorlystudied, lines, −95±30 G (3.2σ) from metal lines, −144±21 G (6.9σ) appearinginonly5articlesintheADSsince1970. fromthefullspectrum, and−158±45G(3.5σ)fromallBalmer CPD-53 295 was observed on a single night during the P74 linesexcludingHβ.InFig.6,noisybutapprentlysignificantcircu- observing run withboth grisms,obtaining S/Nsof 2925:1 (grism larpolarisationispresentinallBalmerlines.Basedonthedetection 600B)and1800:1(grism1200g).Thespectraarerichinmetallic ofconsistentlongitudinalmagneticfieldsinbothBalmerlinesand lines,showBalmerlinesinabsorption,andverystrongCaHand metallines,wetentativelyconcludethatBFOrihostsamagnetic fieldwithalongitudinal intensityoforder 150G.However, asin thecaseofHD101412, furtherobservationsareclearlynecessar- ilyinordertounambiguouslyestablishthepresenceofamagnetic 2 AsexplainedattheendofSect.8andinFig.8ofBagnuloetal.(2006),in field. theirsurveyofclustermagneticApstarstheyencounterednumerouscases where a simple inspection of the V/I profiles did not show any obvious polarization signal, but in which the magnetic field was detected by the linearcorrelationbetweenV/Iandthequantity1/I ×dI/dλ.Thereliability of this result is supported by the histograms of Fig. 7 of Bagnulo et al. 3 Thedetection ofthis strongabsorption is confirmedfromexamination (2006),showingthatvirtuallynodetectionsarefoundin”normal”Astars, ofhigh-resolution spectra inourpossession. Thedetection ofH λ5876 wherewedonotexpect tofindanymagnetic field.Magnetic fieldswere is remarkable given the assumed temperature of this star, and it will be detectedonlyinApstars,i.e.,wheretheyareexpectedtobefound. exploredindetailinafuturepaper. A searchforstrong,ordered magneticfieldsin HerbigAe/Bestars 9 x 10−3 x 10−3 x 10−3 x 10−3 5 5 5 5 V/I 0 V/I 0 V/I 0 V/I 0 −5 −5 −5 −5 3850 3900 3950 4000 4060 4080 4100 4120 4140 4300 4320 4340 4360 4380 4800 4850 4900 Wavelength (angstrom) Wavelength (angstrom) Wavelength (angstrom) Wavelength (angstrom) x 106 x 106 x 106 x 106 4 6 6 5 3 4 4 4 U) U) U) U) D2 D D D3 A A A A I ( I (2 I (2 I ( 1 2 0 0 0 1 3850 3900 3950 4000 4060 4080 4100 4120 4140 4300 4320 4340 4360 4380 4800 4850 4900 Wavelength (angstrom) Wavelength (angstrom) Wavelength (angstrom) Wavelength (angstrom) Figure6.StokesIandVprofilesinthespectrumofBFOri. HD36112 current,standardisedreductionandanalysispresentedinthispaper (according to the methodology reported by Bagnulo et al. 2006), which we believe to be more accurate and robust and to provide HD 36112 (PDS 183, MWC 758) is a well-studied HAeBe morerealisticerrorbars,doesnotresultinafielddetectionforthis star,selectedfromthecatalogueofThe´ etal.(1994).Vieiraetal. (2003) reportHαandO[]forbiddenlineemission.Thisstarhas star(longitudinalfieldsof166±70GinBalmerlinesand−11±91G in metal lines). Nor is a magnetic field detected in HD 72106B beenstudiedindetailbyBeskrovnayaetal.(1999),whodiscussthe (detection flag“nnnn”, hBi = 76±67 G fromthefull spectrum spectroscopic,photometricandpolarimetricproperties.Theypoint z analysis).ThedifferenceintheresultsforHD72106Acanbeex- outinparticularthestrongnear-IRexcessduetowarmcircumstel- plainedasanaturalresultofthesomewhatlargererrorbarsderived lardust,andpeculiar,rapidvariabilityofcircumstellarlineprofiles, here,coupledwiththeexpectedstatisticalfluctuationsinmeasure- whichtheyspeculatemaybeconnectedwithcircumstellarjet-like mentsobtainedfromindependentreductionsofthedata.However, inhomogeneities. TheHipparcosparallaxof4.89±1.16masyieldsadistance themagneticfieldofHD72106AwasdetectedrepeatedlybyWade of205+63 pc.IncombinationwiththeeffectivetemperatureT = etal.(2005)andFolsometal.(2006)usingtheESPaDOnShigh- 7700±−24000K(Beskrovnayaetal.1999)andthemodelevolutioneffary resolutionspectropolarimeter, withclearStokesV signaturesand tracksofPalla&Stahler(1993),thisyieldsamassof1.8±0.2M , longitudinalfielderrorbarsofabout60G:HD72106Aistherefore ⊙ aconfirmedmagneticobject. andanageof5-10Myr(consistentwiththeresultsofMannings& Sargent(2000),whofind6Myr).Theprojectedrotationalvelocity HD 72106B (PDS031S - thefainter southerncomponent of ofHD36112is60±6km/s(Beskrovnayaetal.1999). this0.8′′doublestar;Harkopfetal.1996;ESA1997)wasselected HD 36112 was observed during P74. Eight exposures were from the investigation of HAeBe group members and candidate obtained,foranetS/Nof3725:1.Thespectrum,illustratedinFig. starsbyVieiraetal.(2003). According tothoseauthors, thisstar 7,isrichinmetallinesandshowsclearcoreemissionatHβ,and exhibitsweakHαemissionandsignificantIRexcess,andmaybe emission at He  λ5876. The Ca  K and Mg  λ4481 lines are anevolvedHAeBeobject,possiblynearingtheendofitsPMSevo- prominent.Themagneticdiagnosisofthisstar(“nddn”)yieldsno lution.Bothstarsexhibitlargepropermotionsandthereisnode- detectioninBalmerlines,butmarginaldetectionsaccordingtoone tectablerelativemotionofthepair(Harkopfetal.1996).According ofthecriteriaofBagnuloetal.(2006)inthemetalliclines(−166± toFolsometal.(2006),theradialvelocitiesareidenticalwithinthe 56G;3.0σ)andthefullspectrum(−149±41G;3.6σ).Although measurementsuncertainties.Thissuggeststhattheyarephysically thisdoesnotconstitutestrongevidenceforamagneticfield,given associated,whichimpliesthattheyarelikelyco-eval.Weobserved theactivitydiscussedbyBeskrovnayaetal.(1999),thisstarshould bothstarsindividually,underconditionsofexcellentseeing,using bereobserved. aslitsizeof0.5′′. Although no temperatures are available in the literature for the individual stars, based on the B−V colours of Fabricius & 4.2.3 Othertargetsofnote Marakov(2000)bothcomponentsappeartotheA-typestars.Our twoFORS1spectraconfirmthis,withwell-developedBalmerlines and similar metallic-line spectra. The Balmer lines have similar HD72106A depth in spectra of both stars, but they are considerably broader in the B component. The Mg  λ4481 lines are well-developed Wade et al. (2005) reported the detection of a weak lon- (strongerinthespectrumoftheAcomponent),asistheCaKline gitudinal magnetic field in HD 72106A using both FORS1 and (substantiallystrongerintheBcomponent).Overall,theAcompo- ESPaDOnS. The FORS1 detection, which corresponded to 5.2σ nent’sspectrumissimilartothatofHD94660-themetallicspec- (Balmerlines) and5.8σ (fullspectrum) detections, wasbased on trum is weaker, and the Balmer lines marginally broader. On the anearlierreductionandanalysisofthesedata(Drouin2005).The otherhand,theBcomponent’sspectrumissuggestiveofacooler 10 G.A. Wade,S. Bagnulo,D. Drouin,J.D.Landstreet, D.Monin x 10−3 x 10−3 x 10−3 x 10−3 3 3 3 3 2 2 2 2 1 1 1 1 V/I 0 V/I 0 V/I 0 V/I 0 −1 −1 −1 −1 −2 −2 −2 −2 −3 −3 −3 −3 3850 3900 3950 4000 4060 4080 4100 4120 4140 4300 4320 4340 4360 4380 4800 4850 4900 Wavelength (angstrom) Wavelength (angstrom) Wavelength (angstrom) Wavelength (angstrom) x 106 x 106 x 106 x 106 6 6 7 8 5 6 7 4 U) U) U) U) D D4 D5 D6 A A A A I (2 I ( I ( I ( 3 4 5 0 2 3 4 3850 3900 3950 4000 4060 4080 4100 4120 4140 4300 4320 4340 4360 4380 4800 4850 4900 Wavelength (angstrom) Wavelength (angstrom) Wavelength (angstrom) Wavelength (angstrom) Figure7.StokesIandVprofilesinthespectrumofHD36112. HD144432,HD31648andHD144668 1.1 1 HD 144432, HD 31648 and HD 144668 are 3 HAeBe stars 0.9 that have been discussed by Hubrig et al. (2004, 2006a). These 0.8 authorsalsousedFORS1tosearchformagneticfieldsinBalmer malised flux00..67 l2tie2ncetGiso,(nr4ei.n0pσoHr)tDirneg1s4pm4ec6at6rigv8ie.nlOyal)fdipneatHreticDctuio1lan4rs4i4(n31t2e1r9aens±tdi3Hs8tDhGe3r1(e36p.41o8σr,t)abnayndHdnuo8b7drei±g- Nor0.5 etal.(2006b)ofstrongStokesVsignaturesinsomespectrallinesof 0.4 thesestars,whichtheyinterprettobedueto“circumstellar”mag- neticfields. 0.3 WehaveobservedHD144432,HD31468andHD144668as 0.2 partofthisprogramme.Noneofthesestarsyieldsanysignificant 0.1 4260 4280 4300 4320 4340 4360 4380 4400 4420 magneticfielddetectionintheirmetallicorBalmerlines,withun- Wavelength (angstrom) certainties comparable to those of Hubrig et al. (2004). We have alsocarefully examined our reduced spectra, and wefindno evi- Figure 8. Observed (symbols) and computed (lines) Hγ profiles of HD dence of strong Stokes V signatures in our observations of these 72106A (brighter northern component - Top) and HD 72106B (fainter stars. southerncomponent-Bottom). HD104237 star.WehaveemployedtheFORS1spectraofthetwocomponents, withtheirrobustcontinuaandBalmerlineprofiles,incombination withATLAS9Balmerlineprofiles(convolvedtothesameresolu- HD104237isacoolHerbigAestarwithasharpmetallic-line tionastheFORS1spectraassumingagaussianinstrumental pro- spectrum in which Donati et al. (1997) claimed the marginal de- file),toconstrainthetemperaturesofthetwocomponents.Wede- tectionofamagneticfieldusinghigh-resolutionspectropolarime- rivefromthecomparisonofobservedandtheoreticalprofilesthat tryandLeast-SquaresDeconvolution.Thisresultwassubsequently HD 72106A has T = 11000±1000 K, 3.5 6 logg 6 4.5 and confirmedbyDonati(2000),whodetectedStokesVsignaturesand eff HD72106BhasT =8000±500K,4.06logg64.5.Thebest- derived a longitudinal field of about 50 G (Donati, priv. comm.) eff fit results (for T = 11000 K, logg = 4.0 and T = 8000 K, using new data and similar techniques. Acke & Waelkens (2004) eff eff logg=4.5)areshowninFig.8. found approximately solar photospheric abundances based on an Vieira et al. (2003) associated HD 72106 with the Vela and equivalentwidthabundancesanalysis. GumNebulaSFRs.ThesystemhasacombinedHipparcosparal- HD 104237 was observed during both P72 and P74. Three lax of 3.47±1.43 mas, yielding a distance of 288+202 pc, which observations (corresponding to12, 10and 8 exposures) were ob- −84 isconsistentwiththatoftheSFR.Wadeetal.(2005)estimatedits tained,fornetS/Nsof2375:1,2800:1and2400:1. Thespectrum, position on the HR diagram. This position, interpreted using the illustrated in Fig. 10, exhibits significant emission at Hβ, which modelevolutionarytracksofPalla&Stahler(1993),indicatesthat showsastrongandvariableemissioncore,aswellasemissionin thecomponents of HD72106 havemassesof 1.75±0.25 M (B the multiplet 42 Fe  lines. The magnetic diagnoses of this star ⊙ component)and2.4±0.4M (Acomponent),andaresituatedclose yieldnodetectionineitherBalmerormetallinesaccordingtoany ⊙ totheZAMS.Althoughuncertain,theseH-Rdiagrampositionsare ofthecriteriaofBagnuloetal.(2006),with besterrorbarsof64G consistentwithaco-evalsystem. (Balmerlines)and35G(metallines).Theuncertaintiesobtained

See more

The list of books you might like

Upgrade Premium
Most books are stored in the elastic cloud where traffic is expensive. For this reason, we have a limit on daily download.
DMCA & Copyright: Dear all, most of the website is community built, users are uploading hundred of books everyday, which makes really hard for us to identify copyrighted material, please contact us if you want any material removed.
Copyright @ PDFdrive.to, 2022 | We love our users