ebook img

Relating jet structure to photometric variability: the Herbig Ae star HD 163296 PDF

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

Preview Relating jet structure to photometric variability: the Herbig Ae star HD 163296

Astronomy&Astrophysicsmanuscriptno.HD163296˙140115˙arxiv (cid:13)c ESO2014 January17,2014 Relating jet structure to photometric variability: the Herbig Ae star HD 163296(cid:63) L.E.Ellerbroek1,L.Podio2,3,C.Dougados4,2,S.Cabrit5,2,M.L.Sitko6,7,24,H.Sana8,L.Kaper1,A.deKoter1,9, P.D.Klaassen10,G.D.Mulders11,I.Mendigut´ıa12,C.A.Grady13,14,K.Grankin15,H.vanWinckel9,F.Bacciotti3, R.W.Russell16,24,D.K.Lynch16,17,24,H.B.Hammel7,18,24,L.C.Beerman6,19,24,A.N.Day6,20,24, D.M.Huelsman6,21,24,C.Werren6,24,A.Henden22,andJ.Grindlay23 1 AstronomicalInstitute“AntonPannekoek”,UniversityofAmsterdam,SciencePark904,1098XHAmsterdam,TheNetherlands e-mail:[email protected] 2 InstitutdePlane´tologieetd’AstrophysiquedeGrenoble,414,RuedelaPiscine,38400St-Martind’He`res,France 4 3 INAF-OsservatorioAstrofisicodiArcetri,LargoEnricoFermi5,50125,Florence,Italy 1 4 CNRS/UniversidaddeChile,LaboratoireFranco-Chiliend’Astronomie(LFCA),UMI3386,Santiago,Chile 0 5 LERMA,ObservatoiredeParis,UMR8112duCNRS,61Av.del’Observatoire,75014,Paris,France 2 6 DepartmentofPhysics,UniversityofCincinnati,CincinnatiOH45221,USA 7 SpaceScienceInstitute,4750WalnutStreet,Boulder,CO80303,USA n 8 SpaceTelescopeScienceInstitute,3700SanMartinDrive,Baltimore,MD21218,USA a 9 InstituutvoorSterrenkunde,KULeuven,Celestijnenlaan200B,3001Leuven,Belgium J 10 LeidenObservatory,LeidenUniversity,POBox9513,2300RALeiden,TheNetherlands 5 11 LunarandPlanetaryLaboratory,TheUniversityofArizona,Tucson,AZ85721,USA 1 12 DepartmentofPhysicsandAstronomy,ClemsonUniversity,Clemson,SC29634-0978,USA 13 EurekaScientific,Inc.,Oakland,CA94602,USA ] 14 ExoplanetsandStellarAstrophysicsLaboratory,Code667,GoddardSpaceFlightCenter,Greenbelt,MD20771,USA R 15 CrimeanAstrophysicalObservatory,ScientificResearchinstitute,98409,Crimea,Nauchny,Ukraine S 16 TheAerospaceCorporation,LosAngeles,CA90009,USA . 17 ThuleScientific,Topanga,CA90290,USA h 18 AssociatedUniversitiesforResearchinAstronomy,Inc.,1212NewYorkAve.NW,Washington,DC20005,USA p 19 DepartmentofAstronomy,UniversityofWashington,Seattle,WA98105,USA - o 20 DepartmentofPhysics,MiamiUniversity,Oxford,OH45056,USA r 21 DepartmentofManagementScienceandEngineering,StanfordUniversity,Stanford,CA94305,USA st 22 AmericanAssociationofVariableStarObservers,49BayStateRoad,Cambridge,MA02138,USA a 23 Harvard-SmithsonianCenterforAstrophysics,60GardenStreet,Cambridge,MA02138,USA [ 24 VisitingAstronomer,InfraredTelescopeFacility,operatedbytheUniversityofHawaiiunderCooperativeAgreementno.NNX- 08AE38AwiththeNationalAeronauticsandSpaceAdministration,ScienceMissionDirectorate,PlanetaryAstronomyProgram. 1 v Received;accepted 4 4 Abstract 7 3 Herbig Ae/Be stars are intermediate-mass pre-main sequence stars surrounded by circumstellar dust disks. Some are observed to . produce jets, whose appearance as a sequence of shock fronts (knots) suggests a past episodic outflow variability. This “jet fossil 1 record”canbeusedtoreconstructtheoutflowhistory.Wepresentthefirstopticaltonear-infrared(NIR)spectraofthejetfromthe 0 HerbigAestarHD163296,obtainedwithVLT/X-shooter.Weaccuratelydeterminephysicalconditionsintheknots,aswellastheir 4 kinematic“launchepochs”.Knotsareformedsimultaneouslyoneithersideofthedisk,witharegularintervalof∼16yr.Thevelocity 1 dispersionrelativetothejetvelocity,aswellastheenergyinput,iscomparableinbothlobes.However,themasslossrate,velocity, : v andshockconditionsareasymmetric.Wefind M˙ /M˙ ∼ 0.01−0.1,whichisconsistentwithmagneto-centrifugaljetlaunching jet acc i models.Noevidencefordustisfoundinthehigh-velocityjet,suggestingalaunchregionwithinthesublimationradius(<0.5au).The X jetinclinationmeasuredfrompropermotionsandradialvelocitiesconfirmsitisperpendiculartothedisk.Atentativerelationisfound r betweenthestructureofthejetandthephotometricvariabilityofthecentralsource.EpisodesofNIRbrighteningwerepreviously a detectedandattributedtoadustydiskwind.Wereportforthefirsttimesignificantopticalfadingslastingfromafewdaysuptoayear, coincidingwiththeNIRbrighteningepochs.Thesearelikelycausedbydustliftedhighabovethediskplane;thissupportsthedisk windscenario.Thediskwindislaunchedatalargerradiusthanthehigh-velocityatomicjet,althoughtheiroutflowvariabilitymay haveacommonorigin.Nosignificantrelationbetweenoutflowandaccretionvariabilitycouldbeestablished.Ourfindingsconfirm thatthissourceundergoesperiodicejectionevents,whichmaybecoupledwithdustejectionsabovethediskplane. Keywords.Stars:formation–Stars:circumstellarmatter–Stars:variables:TTauri,HerbigAe/Be–ISM:jetsandoutflows–ISM: Herbig-Haroobjects–Stars:individualobjects:HD163296 (cid:63) BasedonobservationsperformedwithX-shooter(program089.C- 0874) mounted on the ESO Very Large Telescope on Cerro Paranal, Chile 1 L.E.Ellerbroeketal.:Relatingjetstructurewithphotometricvariability:HD163296 1. Introduction continuum emission is observed up to ∼ 240 au; the outer gas disk is detected in CO lines with a Keplerian rotation profile HerbigAe/Bestars(HAeBe)areintermediate-mass(2−10M ) (cid:12) (Rosenfeldetal.,2013;deGregorio-Monsalvoetal.,2013). pre-main sequence stars. Like their low-mass equivalent, the Thenear-infrared(NIR)excesshasacomponentthatpeaks classicalTTauristars(CTTS),HAeBestarsareassociatedwith at 3 µm and is well fitted by a blackbody of 1500 K. This accretion disks and, in a few cases, jets (e.g. Corcoran & Ray, suggests emission by dust at the evaporation temperature. 1998;Gradyetal.,2000,2004).Keyquestionsarehowexactly Interferometricobservations(Tannirkulametal.,2008a;Benisty these jets are launched and how this process relates to disk ac- et al., 2010a) show that a major fraction of this emission orig- cretion(Ferreiraetal.,2006;Bai&Stone,2013).InHerbigsys- inates from within the theoretical value of the dust sublima- tems,thejet-diskcouplingmaybedifferentthaninCTTS.The tion radius (R ∼ 0.5 au for HD 163296, eq. 1 in Dullemond sub relativelyhighstellarluminositycausesgrainparticlestorapidly & Monnier 2010). Also, the emission profile is smooth, i.e. evaporate at distances within ∼ 1 au, creating a dust-free zone does not originate from a sharply contrasted inner disk rim. (Kamaetal.,2009).AccretionisobservedinHAeBestars(e.g. HD 163296 also shows significant variations in its NIR bright- Muzerolleetal.,2004;Mendigut´ıaetal.,2011),soaninnergas nessontimescalesofyears(Sitkoetal.,2008). disk likely exists. Since jets are expected to originate (at least A number of scenarios have been put forward to explain in part) from this region (Blandford & Payne, 1982), they may theseobservations.Hydrostaticdiskmodelsarenotpreferred,as help to improve our understanding of the coupling of accretion theNIRemissionismuchstrongerthanpredictedbyhydrostatic andoutflowintheinnerdisk. equilibrium;also,thesescenariosdonotexplainthephotometric We may constrain the launching process by observing jet variability. Alternative scenarios include a dust halo (Vinkovic´ structureandmotion.Jetsfromyoungstarsareusuallyobserved et al., 2006) and a dusty disk wind (Bans & Ko¨nigl, 2012). In as a sequence of shock fronts or “knots”. These are likely the both cases, dust would not exist within R , but at larger dis- sub result of a variable outflow velocity (Rees, 1978; Raga et al., tances and above the disk plane, making its observation within 1990). A significant asymmetry in velocity and shock condi- R a projection effect. Vinkovic´ & Jurkic´ (2007) suggest that sub tions is often observed between the two lobes of jets (Hirth dustejections,possiblyrelatedtojetlaunching,areresponsible et al., 1994; Ray et al., 2007). By tracing the trajectories of forNIRexcessandvariability. the knots through space, we are able to reconstruct the epochs The “fossil record” contained in the jet as described above whentheywereformedandtheir(quasi-)periodicoccurence,if mayhelpconstrainthephysicalpropertiesandvariabilityofthe any (e.g., Ellerbroek et al., 2013). On-source photometric and innerregionsofthedisk.InthispaperwepresentopticaltoNIR spectroscopicobservationsthatweremadeduringthese“launch spectra of the jet and central source. These were obtained with epochs” (when available) may shed light on what happens in X-shooterontheESOVeryLargeTelescope(VLT).Wecombine the disk whenever a knot is created. These observations may the spectra and archival images to reveal the outflow history of also clarify the origin of the asymmetry between lobes, which thesystem.Wethencomparethiswiththephotometricvariabil- isobservedinmanyjetsystems.Inthispaperwecombinetime- ity of the central object. In Sect. 2 we describe the newly ob- resolved jet and disk observations and diagnostics to constrain tainedandarchivalobservationaldata.InSect.3wepresentour the properties of the launch mechanism, the structure of the analysisofthejetkinematicsandphysicalconditions.InSect.4, launchregion,theoriginofjetasymmetry,andtherelationwith we present a multi-band lightcurve of the central object com- diskaccretion. piledfromarchivalandpreviouslyunpublisheddata.Weanalyze The Herbig Ae star HD 163296 is a promising test case thevariabilityandcolorsofthesource,andestimateitshistoric for this observing strategy. Its disk is well-studied and associ- accretionrate.InSect.5wediscusstheconstraintsputonthejet ated with a jet, and a copious amount of time-resolved imag- launching by our observations. We also propose a possible ex- ing and spectroscopic data are available. Located at a distance planationforthesourcevariabilityandcommentonitsrelation of 119±11 pc (van Leeuwen, 2007), the system does not ap- withjetstructure.WepresentourconclusionsinSect.6. pear to be associated with a star-forming cluster or dark cloud (Finkenzeller & Mundt, 1984). Meeus et al. (2012) do find ex- tended[Cii]emissionthatmayoriginatefromabackgroundor 2. Observations,datareductionandarchivaldata surroundingmolecularcloud. The bipolar jet HH 409 was discovered on coronagraphic Inthissectionwegiveanoverviewofthespectroscopicandpho- images (and later confirmed with long slit data) of the Space tometricdatapresentedinthispaper.Animageofthejetin[Sii] Telescope Imaging Spectrograph (STIS) on the Hubble Space and the definition of the knots can be found in Wassell et al. Telescope (HST Grady et al., 2000; Devine et al., 2000). One (2006,W06;theirFig.1). of the knots (A) has also been associated with X-ray emission (Swartz et al., 2005). The high-velocity gas in the jet has ra- 2.1. VLT/X-shooterspectroscopy dial velocities of 200 − 300 km s−1. A blue-shifted molecular outflow (up to 13(cid:48)(cid:48) from the source) was found in CO 2–1 and Spectra of HD 163296 and its jet were obtained with VLT/X- 3–2emissionwiththeAtacamaLargeMillimiterArray(ALMA) shooter (Vernet et al., 2011), which covers the optical to NIR byKlaassenetal.(2013)andalsorecoveredonSubMillimeter spectralregioninthreeseparatearms:UVB(290–590nm),VIS ArrayCO2–1images(C.Qi,privatecommunication).Thema- (550–1010nm),andNIR(1000-2480nm).Table1liststheset- terialinthemolecularoutflowpropagatesanorderofmagnitude tingsandcharacteristicsoftheobservations.Toobservethejet, slower than the fast-moving jet, peaking at −18 km s−1 in the multiple overlapping pointings of the 11(cid:48)(cid:48) slit were performed. systemicrestframe. On6July2012,oneshorton-sourceexposurewasfollowedby TheradiusofthedustdiskaroundHD163296isestimated severaloffsets,coveringbothlobesoftheHH409jetupto25(cid:48)(cid:48). ataround500au,basedonthenon-detectionofthered-shifted A narrow-slit deeper exposure covering the red lobe up to 40(cid:48)(cid:48) jetuptothisdistance(Gradyetal.,2000)andtheextentofthe was taken on 7 July 2012. Sky frames were obtained before or scattered light emission (Wisniewski et al., 2008). The 850 µm aftereveryexposuretocorrectfortelluricemissionlines.Since 2 L.E.Ellerbroeketal.:Relatingjetstructurewithphotometricvariability:HD163296 Table1.JournaloftheX-shooterobservations. Target HD163296 HH409 Date 6Jul2012 6Jul2012 7Jul2012 2Jul2013 14Jul2013 UT(startobs.) 02:36 02:42 01:34 06:18 01:33 HJD-2400000(startobs.) 56114.608 56114.613 56115.565 56475.763 56487.565 Slitpositionangle(NthroughE) 42.5◦ 42.5◦ 42.5◦ 42.5◦ 42.5◦ Sectioncovered((cid:48)(cid:48)) 0 5−25 10−40 2−13 2−13 (measuredfromsource) (NEandSW) (NE) (NEandSW) (NEandSW) Exposuretime(s) 2 240(5−15(cid:48)(cid:48)), 300 330 330 440(15−25(cid:48)(cid:48)) Slitwidth,UVB/VIS/NIR((cid:48)(cid:48)) 0.5/0.4/0.4 1.0/0.9/0.6 0.8/0.7/0.4 1.0/0.9/0.6 1.0/0.9/0.6 Resolution,∆(cid:51)(kms−1) 33/17/27 59/34/37 48/27/27 59/34/37 59/34/37 V-bandseeing((cid:48)(cid:48)) 0.8−0.9 0.8−0.9 0.6−0.8 0.6−0.8 0.6−0.8 thesourceisverybright,wehaveacquiredthejetspectrabyoff- 2.2. Archivaldata setting the slit with respect to the source. As a result, the inner jet region (< 5(cid:48)(cid:48)) is not covered in the 2012 observations. On Measurements of the positions of knots A, B, and C were made on archived HST/STIS imaging and spectroscopy, 2and14July2013,follow-upobservationswerecarriedoutin two offsets up to 2(cid:48)(cid:48) from the source, in order to constrain the HST/Advanced Camera for Surveys (ACS) imaging, and GoddardFabry-Pe´rot(GFP)imaging(Devineetal.2000;Grady positionsoftheinnerknots. etal.2000;W06;Gu¨ntheretal.2013,G13).Thepositionofthe TheframeswerereducedusingtheX-shooterpipeline(ver- knotswasmeasuredonthe[Sii]673nmlineintheSTISG750L sion1.5.0,Modiglianietal.,2010),employingthestandardsteps spectra,andontheLyαlineintheSTISG140Mspectra. ofdatareduction,i.e.biassubtraction,orderextraction,flatfield- Multi-wavelength photometric data of HD 163296 were ing,wavelengthcalibration,andskysubtraction,toproducetwo- taken from various papers and data catalogs, as well as previ- dimensional spectra. The wavelength calibration was verified ouslyunpublisheddata.Foranoverviewoftheseresources,see by fitting selected OH lines in the sky spectrum, resulting in TableA.1.Overtheperiod1979–2012thesourcehas beenob- a calibration accuracy of a few km s−1. Flux-calibration was served with a regularity that varies per band and epoch. The performedusingspectraofthespectrophotometricstandardstar source is best covered in the optical bands. We retrieved data GD153(aDAwhitedwarf).Thewidthoftheknotsperpendic- frommonitoringcampaignsconductedatMadainakObservatory ulartothejetaxiswasassumedtobenarrowerthantheseeing, (see also Grankin et al., 2007) and at La Silla with the sothattheslitlosseswereestimatedfrommeasuringtheseeing Swiss Telescope in the period 1983–2000. Observations from FWHMfromthespatialprofileofapointsourceonthe2Dframe Las Campanas observatory cover the V-band from 2001–2009 (on the first night, HD163296, on the second night, a telluric (Pojmanski&Maciejewski,2004). standard star). These estimates were refined by comparing the Most of the NIR observations were taken at the Infrared obtainedSEDtotheaveragedphotometry(seeSect.4.1).They Telescope Facilty (IRTF) and at Mount Lemmon Observing were subsequently corrected for the slit widths used in the jet Facility(MLOF)intheperiod1996–2009.Partofthesedataare observations. This procedure results in an uncertainty of about alsopresentedinSitkoetal.(2008).SomeoftheLand M band 10%ontheabsolutefluxcalibration;therelativefluxcalibration data were extracted from spectra obtained with The Aerospace isaccuratetowithin3%. Corporations Broad-band Array Spectrograph System (BASS), The wavelengths and velocities used throughout this paper which covers the 3–13 µm wavelength region. BASS is de- are expressed in the systemic rest frame, for which we adopt scribed more fully in Sitko et al. (2008). Additional JHKLM (cid:51) = 5.8±0.2 km s−1 with respect to the Local Standard of photometric data were obtained using the SpeX spectrograph sys Rest,determinedbyQietal.(2011)basedonsub-mmemission (Rayner et al., 2003). 0.8–5 µm spectra were obtained using lines from the outer disk. This value agrees with the observed a slit width of 0.8(cid:48)(cid:48) and the echelle diffraction gratings. Zero- radialvelocityofphotosphericabsorptionlinesintheX-shooter point corrections were applied by also observing the star with spectrumofHD163296,whichgive7.2±2.0kms−1. theprism(0.7-2.4µm)andaslitwidthof3.0(cid:48)(cid:48).Thismethodhas beenshowntoprovideresultscomparabletophotometricimag- Afterreduction,theframesweremerged,averagingtheover- ing(within5%)whentheseeingis1.0arcsecorbetterandthe lappingregionsbetweenobservations.Intheobservationsclos- airmassdoesnotexceed2.0(seeSitkoetal.,2012). est to the source, a reflection (“ghost spectrum”) of the central The magnitudes are defined in the Johnson (UBVRI), source was present at the edges of the slit. We subtracted this 2MASS (JHK ), and ESO (LM) filter systems. The measure- s contributioninthefollowingway.Inaregionaroundeveryemis- mentshaveatypicaluncertaintyof0.01magintheopticaland sionlineofinterest,ateveryspatialrowinthe2Dframes,wedi- 0.05magintheNIR. videdthespectralprofilebytheon-sourcespectrum.Wethenfit- tedazero-orderpolynomialtotheresidual.Finally,theoriginal spectral profile was divided by the source spectrum multiplied 3. Results:thejet by this fitted value. This results in a position–velocity diagram freeofsourcecontamination(Fig.1).Faint,residualcontinuum W06presentanoverviewofHSTdataofthejet.Ithasaposition emission is seen at +6−7(cid:48)(cid:48), which likely results from a back- angleof42.5±3.5◦ (norththrougheast)andextendsupto30(cid:48)(cid:48) groundcontinuumsource(alsovisibleatthislocationintheim- in both the south-west (blue-shifted, or “blue lobe”) and north- agesofGradyetal.2000). east (“red lobe”) direction. This section contains an analysis of 3 L.E.Ellerbroeketal.:Relatingjetstructurewithphotometricvariability:HD163296 Figure1. Position-velocity diagram of Hα, [S ii] λ673 nm, and [Fe ii] λ1643 nm in July 2012, and [S ii] λ673 nm in July 2013. The y-axis denotes the position along the slit; y > 0 corresponds to the NE lobe, while y < 0 corresponds to the SW lobe. Due to the brightness of the source, the central region was not observed. Black contours correspond to logF /[erg s−1 cm−2 Å−1] λ = (−17.5,−17,−16.5,−16,−15.5). The blue contours indicate the 3σ detection level of the knots. The black crosses denote the derivedpositionsandvelocitiesoftheknots,andtheir1σuncertainties.NotethebowshockfeaturesinknotAandCinHα. theX-shooterspectraofthejet.Itisdividedintotwoparts:the strained,aresimilartotheotherknots.Assumingconstantmo- kinematicsandthephysicalconditions. tion for all knots allows us to make an estimate of their launch epochs(t ).Theseepochsshouldthenbeviewedastheinter- launch valsduringwhichthebrightknotshaveformedinthejet,close 3.1. Kinematics:propermotionsvs.radialvelocities tothedrivingsource. Since the knots are regularly spaced, a better constraint on ThebipolarHH409jetconsistsofasequenceoflocalemission their trajectories is obtained by making a global fit to the data line maxima, or “knots” (Fig. 1). These knots are detected in presented in Fig. 2a. We assumed that each knot was created 40 emission lines of various allowed and forbidden transitions: H i, [C i], [N ii], [O i], [O ii], [S ii], Ca ii, [Ca ii], [Fe ii], and simultaneously with its counterpart, that knot creation is peri- [Ni ii]. No H emission is observed along the jet. Fig. 1 dis- odic, and that knots propagate uniformly on either side of the plays the X-s2hooter spectrum of the Hα, [S ii] λ673 nm, and system.WeomittedknotA2fromthefittingprocedure,asithas [Fe ii] λ1643 nm emission lines. Knots A, A2, B, C, D, E, F, no counterpart. The best-fit model (χ2red = 0.82) has a period G, and the onset of H, defined by W06, are indicated in this 16.0±0.7yrandpropermotionsvt,red = 0.28±0.01(cid:48)(cid:48) yr−1 and 2Dspectrum.Theknotpositionsareidentifiedbyco-addingthe vt,blue =0.49±0.01(cid:48)(cid:48)yr−1.TheglobalfitisdisplayedinFig.A.1. position-velocity diagrams of a selection of the strongest emis- An independent estimate of the launch epoch estimates is sionlines.Twomorerecentlyejectedknots:A3(identifiedinthis achievedbyusingX-shooterradialvelocities((cid:51))andassuming r work)andB2(alsoreportedbyG13)areonlypartlycoveredby thatthejetisperpendiculartothediskequatorialplane.Existing the X-shooter slit in July 2012. The movements of these knots estimates of the disk inclination are based on interferometry of aredeterminedfromtheJuly2013observationsandtheobserva- the outer disk (i = 46 ± 4◦, Isella et al. 2007; i = 44◦, disk disk tionsoftheLyαcounterpartofknotB2(G13)andthemolecular Rosenfeld et al. 2013) and the inner disk (i = 48 ± 2◦, disk counterpartofA3(Klaassenetal.,2013).Thejetlobesareasym- Tannirkulametal.2008a).Inthispaperweadoptthelattervalue. metric,thebluelobebeingfaster(220−300kms−1)thanthered From the co-added position-velocity diagrams (Fig. 1) of the lobe(130−200kms−1).Bow-shockfeaturesinknotAandC, brightest lines, a spectral profile is extracted at every pixel of whicharespatiallyresolvedintheHSTimages(W06),showup width= 0(cid:48).(cid:48)2alongtheslit.AGaussianfunctionisfittedtothis as sharp velocity drops in the X-shooter Hα position-velocity profile; its centroid velocity is converted to a launch epochs as diagram. t = x/((cid:51) tani ). The fluxes and launch epochs of these launch t r disk Fig. 2a shows the measured positions (x) and lengths (es- individualcomponentsaredisplayedinFig.2b.The16-yearpe- t riodicityinthejetstructure,aswellastheconcurrenceofknots timated as the FWHM of the spatial profiles along the jet axis) inbothlobesareevidentfromthisfigure. oftheknotsobservedoverthe1998–2013epoch.Thetrajecto- riesofknotsA,A2,B,andCarewellfittedbyuniformmotion. Fig.2a(arrows)and2cillustratetheagreementinthelaunch ThetrajectoriesofknotsA3,B2,andD–G,thoughpoorlycon- epoch estimates between the proper motion and radial veloc- 4 L.E.Ellerbroeketal.:Relatingjetstructurewithphotometricvariability:HD163296 Figure2.(a)PositionsoftheknotsofHH409overtime;verticalbarscorrespondtotheFWHMofthespatialprofile.Theshaded bandsanddashedlinesarelinearfitstotheirtrajectories.Thearrowsrepresentthevelocitiesontheskycalculatedfromtheradial velocitiesanddiskinclination.References:(1)Devineetal.(2000);(2)Gradyetal.(2000);(3)Wasselletal.(2006);(4)Gu¨nther etal.(2013);(5)Klaassenetal.(2013).(b)LaunchepochsfromradialvelocitiesintheX-shooterspectra.Everysymbolrepresents theflux(onalogarithmicscale)andt ofonerowofthetwo-dimensionalspectrum,obtainedthroughtheproceduredescribed launch inthetext.Onlypeaksabovethe3σbackgroundlevelaredisplayed.Thedatashowaperiodicityof16yr,whichwasalsoobtained fromtheglobalfittothejetpropermotions(dottedlines).Thelaunchepochsalsoagreewellbetweentheblueandredlobes.(c) Launch epochs of knots in the blue (open squares) and red (closed squares) lobes, derived from proper motions and from radial velocities.Thedottedhorizontallinesareplottedona16-yearinterval;notethelogarithmicscale. 5 L.E.Ellerbroeketal.:Relatingjetstructurewithphotometricvariability:HD163296 ity methods. The latter figure also demonstrates the periodic- ity and simultaneity (in both lobes) of the launch events. The jetinclinationanglecalculatedfromtheaveragejetvelocitiesis i =tan−1(cid:104)v(cid:105)/(cid:104)v(cid:105)=47±2◦.Thered-shiftedknotsB2,B,andC jet t r arelaunchedsimultaneouslywiththeblue-shiftedknotsA3,A, andH,respectively.ForknotH,t isestimatedfromitsposi- launch tionin2004andtheaverage(cid:51) inthebluelobe.Duetotheirhigh r velocities,thecounterpartsofthered-shiftedknotsD,E,F,and GarelikelylocatedtensofarcsecondsbeyondknotHintheblue lobe and are not covered by our observations. The most recent launchepochtracedbyourobservationspeakedin2001–2004, markingtheformationofknotsA3andB2.Intheremainderof this paper, we adopt the t estimates from radial velocities, launch asthesehavethesmallestmeasurementerror. 3.2. Physicalconditions The line emission along the jet is a result of collisional excita- tioninshockfronts.Conditionsmayvarysignificantlybetween the regions up- and downstream of the shock front (“pre”- and “post”-shock, respectively; Hartigan et al. 1994). The physical conditions of the shocked gas in the knots (post-shock elec- tron and total density, n , n ; temperature, T ; and ion- e,post H,post e ization, X )canbeestimatedbycomparingobservedlineratios e withthosepredictedbycollisionalexcitationmodels(Bacciotti & Eislo¨ffel 1999; for an application to X-shooter spectra see e.g. Ellerbroek et al. 2013). Similarly, the pre-shock parame- ters (shock velocity, (cid:51) ; pre-shock density n ; and mag- shock H,pre neticfield,B)canbeobtainedbycomparinglineratiostoshock models(e.g.,Hartiganetal.,1994). The results presented in this section are summarized in Table2.Weconsidertheemission> 3σabovethebackground noise.ThisisthecaseforknotsA,A2,andA3inthebluelobe and B, C, and D in the red lobe. The line flux per knot is ob- tained by integrating over the blue contours seen in Fig. 1. We consideronlythehigh-velocityemission.Wedidnotcorrectfor thespatialextentoftheknotsbeyondtheslitaperture,whichwe assume to be low-velocity emission. The line fluxes predicted by models and used in diagnostics in this section are corrected to match the solar abundances of Asplund et al. (2005), which resultsinascalingoftypically0.1dex. 3.2.1. Extinction Extinctionbydustaffectstheobservedlineratiosandhencethe estimated physical and shock parameters. The value of A can V beestimatedbyconsideringforbiddenlinesfromthesameupper levelwhicharefarapartinthespectrum.[Feii]infraredlinesare well suited for this purpose, although large uncertainties in the Einsteincoefficientsaffecttheoutcome(seeNisinietal.,2005, for a discussion). Alternatively H i lines can be used, but their ratiosdependalsoonothermodelparameters,liketheshockve- locity(Hartiganetal.,1994). Figure3. Emission line ratios for six knots in the blue (open Wefindthatwithintheerrorsthe[Feii]1643/1321nmand symbols) and red (filled symbols) lobes. Contours indicate the 1643/1256 nm line ratios are consistent with theoretical ones predictedratiosbythemodelsofHartiganetal.(1994)foragrid computed for A = 0. Similarly, the Hα/Hβ line ratios follow V of pre-shock densities, magnetic field strengths, and shock ve- thepredictionsfromtheHartiganetal.(1994)shockmodelsfor locities.Thecoloredsymbolscorrespondtomodelswith(cid:51) = the observed range in (cid:51) (see Sect. 3.2.3). Thus, we adopt shock shock (15,20,25,30,35,40,50,60,70,80,90,100)kms−1. A = 0 in the knots along the jet. This is consistent (assuming V a normal dust-to-gas ratio in the ISM) with the absence of in- terstellar(gas)featuresinthespectrumofthecentralsourceand wellasthelowon-sourceextinctionmeasuredfromphotometry the jet (Finkenzeller & Mundt, 1984; Devine et al., 2000), as (seeSect.4.1). 6 L.E.Ellerbroeketal.:Relatingjetstructurewithphotometricvariability:HD163296 Table2.PhysicalparametersandmasslossrateofHH409. Knot x fromsource (cid:51) n X (cid:104)n (cid:105) T M˙ ,cr.sect. M˙ ,L M˙ ,L t r e,post e H e jet jet [SII] jet [OI] (July2012,(cid:48)(cid:48)) (102kms−1) (102cm−3) (102cm−3) (104K) (10−10M yr−1) (cid:12) Bluelobe A 14.1±0.6 2.4±0.2 4.6±1.0 0.71±0.02 1.8±1.0 1.3±0.3 2.5±1.9 4.2±1.2 5.1±1.6 A2 11.8±0.8 2.6±0.3 5.9±0.8 0.75±0.01 2.2±1.2 1.5±0.3 2.1±1.7 4.2±0.9 5.1±1.2 A3 <7.6 2.7±0.2 6.5±0.8 0.80±0.01 2.3±1.3 1.5±0.3 1.0±1.1 4.1±0.9 5.0±1.2 Redlobe B 9.2±0.7 1.7±0.2 6.7±0.3 0.32±0.01 5.8±3.2 1.0±0.1 2.5±2.3 11±1 13±2 C 13.7±1.0 1.7±0.2 3.4±0.2 0.43±0.03 2.2±1.2 1.1±0.1 1.9±1.4 8.6±1.6 10±1 D 17.9±0.9 1.5±0.1 2.4±0.3 0.21±0.01 3.2±1.8 0.9±0.1 4.4±3.0 6.6±1.3 8.0±1.6 3.2.2. Electrondensity,ionizationfraction,electron totalnumberdensityinthejetisoforder100cm−3.The[Sii]/Hα temperature ratiointheredlobeisunderpredictedbyallmodelsexceptthose with n = 1000cm−3 and B = 1000 µG.This may indicatethat Theelectrondensity,hydrogenionizationfraction,andelectron a strong magnetic field inhibits compression and enhances the temperatureoftheshockedgasintheknotsarecalculatedfrom emissionfromlowerionizationspecies. theBEmethod(Bacciotti&Eislo¨ffel,1999).Wefindanelectron densityof∼ 500cm−3 thatdecreasesalongthejet.Theioniza- tionfractionishigherinthebluelobe(0.7−0.8)thaninthered 3.2.4. Masslossrate lobe(0.2−0.4).Theelectrontemperaturedecreasesawayfrom Themasslossrate M˙ isanimportantparameterinjetphysics. thesourceandishigherinthebluelobe(T ∼ 1−1.5×104 K) jet e It determines the amount of energy and momentum injected thanintheredlobe(T ∼104K).Excitationconditionsarethus e in the ISM, and its ratio with the accretion rate reflects the higherinthebluelobethanintheredlobe. (in)efficiency of the accretion process. Various methods have These trends are similar to those found by W06 and G13, been developed to calculate M˙ from the physical parameters althoughthevaluesfor n foundbytheseauthorsare anor- jet e,post (for a review, see Dougados et al., 2010). We apply two of der of magnitude higher than our calculations. This can be due these to our dataset: (1) we calculate the mass loss rate from to the fact that the slit used in these studies (with a width of thejetcrosssection(πR2),totaldensity(cid:104)n (cid:105),andvelocity((cid:51) = 0(cid:48).(cid:48)2)onlysamplesthecentralregionofthejet,whichislikelyto J H J |(cid:51) |/cosi);and(2)fromthe[Sii]and[Oi]lineluminosities. haveahigherdensity(seee.g.,Bacciottietal.,2000;Hartigan& r Morse,2007).Thedifferencesintheseandotherparametersare consistent within the large uncertainties in the line fluxes mea- Cross section: This method (“BE” method, Bacciotti & suredbytheseauthors.Thetotaldensityinthepost-shockregion Eislo¨ffel, 1999) calculates the mass loss rate from the average isestimatedasnH,post =ne,post/Xe,undertheassumptionthathy- density,velocity,andcrosssectionofthejet: drogenatomsarethemaindonoroffreeelectrons. M˙ =µm (cid:104)n (cid:105)πR2(cid:51) , (1) jet H H J J 3.2.3. Shockvelocity,magneticfieldstrength,compression where we adopt µ = 1.24 for the mean molecular weight. We take the jet radius R to be equal to half of the jet FWHM at Fig. 3 displays the values of a selection of line ratios against J the selected knot position as measured from resolved HST ob- thosepredictedbyHartiganetal.(1994)foragridofshockmod- els with n = (102,103,104) cm−3, B = (10,102,103) µG, servationsinthe[Sii]lines(W06).Fortheknotsconsidered,RJ H,pre increasesfrom40auupto100auawayfromthesource. and (cid:51) = 10 − 100 km s−1. The observed ratios are best shock represented by the models with pre-shock electron densities n =100cm−3,whilethemagneticfieldstrengthBcanadopt Lineluminosities: Ifthephysicalconditionsareuniformwithin H,pre values of 10 − 100 µG. Finally, the observed line ratios indi- each knot, the mass loss rate is proportional to the number cate that in the blue lobe the shock velocity is higher ((cid:51) ∼ of emitting atoms in the observed volume (Podio et al., 2006, shock 80−100kms−1)thanintheredlobe((cid:51) ∼35−60kms−1). 2009): shock This is consistent with the higher excitation conditions in the bluelobe(seealsoW06andG13). (cid:32) Xi X(cid:33)−1 (cid:51) sini Thepre-shockdensityisoforder100cm−3,asindicatedby M˙jet =µmHLline hνAi fi X H J l , (2) t thecomparisonwithshockmodels.Fortheestimatedshockve- locities and a typical magnetic field strength of 30 µG, the whereL isthelineluminosity;ν,A,and f arethefrequency, line i i compression factor,C = n /n varies between 6 and 20 radiative rate, and upper level population fraction for the con- H,post H,pre (Fig. 17 of Hartigan et al., 1994). We adopt C = 13 ± 7 in sideredtransition,respectively. Xi/X isthefractionofatomsof all the knots. The average density is estimated as the geomet- the considered species in the Xi ionization state and X/H the ricmeanofpre-andpost-shockdensities.Thiscanbeexpressed species’ relative abundance with respect to hydrogen. The up- intermsofthepost-shockconditionsandthecompressionfactor per level population, f, is calculated from the statistical equi- i as(cid:104)n (cid:105)=n X−1C−1/2.(Hartiganetal.,1994).Theresulting librium equations using the values of n and T calculated H e,post e e,post e 7 L.E.Ellerbroeketal.:Relatingjetstructurewithphotometricvariability:HD163296 dustgrains(Savage&Sembach,1996).Whendustgrainsevap- orate in the launching region or are sputtered in shocks along thejet,thesespeciesarereleasedintothegas-phase(Jonesetal., 1994;Jones,2000;Mayetal.,2000;Draine,2004;Guilletetal., 2009, 2011). Gas phase depletion of refractory elements has been observed by Nisini et al. (2005) and Podio et al. (2006, 2009, 2011) in HH jets from Class I sources at large distances fromtheirdrivingsource.Ithasalsobeenobservedintheinner 100auofaCTTSjet(Agra-Amboageetal.,2011). In order to check for the presence of dust grains in the jet launch zone, we estimate the Ca and Fe gas-phase abundance following the procedure illustrated in e.g., Nisini et al. (2005) andPodioetal.(2006).Wecompareobservedlineratiosofre- fractory(Ca,Fe)andnon-refractory(S)specieswithratioscom- putedthroughtheestimatedparameters(n ,X ,andT ).We e,post e e use the flux ratios of [Ca ii] λ729 nm, [Fe ii] λ1256 nm, and [Feii]λ1643nmto[Sii]λ673nm(Fig.4). WeassumethatFeandSaresinglyionizedandusethe16- levels model presented in Nisini et al. (2002, 2005) with colli- sional coefficients by Nussbaumer & Storey (1988) for Fe and by Keenan et al. (1996) for S. For the [Ca ii]/[S ii] ratio, it is assumed that no neutral calcium is present, since its ionisation potentialisverylow,6.1eV.Fortheionizationbalancewecon- siderthreelimitingcases: (1) AllcalciumisinCa+.Thisislikelyanoverestimate,assome Camaybedoublyionized. (2) The Ca++/Ca+ fraction is equal to the hydrogen ionisation fraction,X .AsexplainedinPodioetal.(2009),thisisjusti- e fiedbecausetheionizationpotentialofCa+ issimilartothat of hydrogen (11.9 eV and 13.6 eV, respectively). This also Figure4. Top two panels: Predicted versus observed [Fe ii] / goesfortherecombinationandcollisionalionizationcoeffi- [Sii]ratios,basedonthephysicalconditionsderivedinSect.3.2. cientsfortemperatureslowerthan3×104K. Bottompanel:Sameasabove,for[Caii]/[Sii].Predictionsare (3) The Ca++/Ca+ fraction is calculated by assuming coronal made for the three limiting cases discussed in the text. No evi- equilibrium at the estimated knot temperature T. Upward denceisfoundfordepletionofrefractoryelementsinthejet. transitions are assumed to be due to electron collisions and downward transitions occur by spontaneous emission. This underestimatestheCa+ fraction,astherecanbenoequilib- from the line diagnostics. The knot length lt is measured along riuminthejetwhilethegasismoving. theslit.Thismethodisappliedtotwolines:[Sii]λ673nmand [O i] λ630 nm. We assume that all sulphur is singly ionized. Fig.4showstheobservedandpredictedlineratios.Nosig- Tocomputetheionizationfractionofoxygen,weconsidercol- nificantdepletionofrefractoryspeciesisfoundinthejetknots. lisional ionization, simple and dielectronic recombination, and Somecareshouldbetakenwheninterpretingthisresult.Thecol- chargeexchangewithH. lisionalcoefficientsandatomicparametersfor[Feii]transitions The mass loss rates calculated from the line luminosities are affected by large uncertainties (Giannini et al., 2008). The (Table2)agreewitheachotherwithintheuncertainties.Theval- [Fe ii] / [S ii] ratio may be overpredicted if foreground dust is uesfromthecrosssectionmethodaresignificantlyandsystem- present, which we do not consider likely. The [Fe ii] and [S ii] atically lower than these. This discrepancy may be caused by linesmayhavedifferentfillingfactors;theFelinescouldorigi- an underestimated magnetic field strength and hence, an over- natefromadenserorlargerregion(seeNisinietal.,2005;Podio estimated compression factor. Additionally, upon comparison etal.,2006).Whenthisistakenintoaccount,oneshouldtakethe with the Hartigan et al. (1994) models (Fig. ??), the observed estimatedFegasabundanceasanupperlimit.Despitetheseun- [Nii]λ683nm/[Oi]λ630nmratiossuggestalowerXethande- certainties,ouranalysisgivesnoindicationofdustgrainsinthe rivedfromtheBEmethod,resultinginahigher M˙jet.Themass launchregionofthehigh-velocityjet. lossrateisconstantacrosseachlobe;thosefoundfortheredlobe areafactor2higherthaninthebluelobe.Theaveragemassloss rateis(cid:104)M˙ (cid:105)=5±2×10−10M yr−1. jet (cid:12) 4. Results:variabilityofthecentralsource Inthissectionwedescribethehistoriclightcurveofthesource, 3.2.5. Dustcontent aswellasthephotometricandspectroscopicaccretiondiagnos- The line ratios observed in the jet may also be used to test for tics.WeadoptthestellarparametersdeterminedbyMontesinos depletion of refractory species, such as Ca and Fe. This indi- et al. (2009), which we find to be consistent with the 2012 X- catesthepresenceofdustinthejetlaunchregion.Thegas-phase shooter spectrum. These parameters are the effective tempera- abundanceofthesespeciesisstronglydepletedintheISMwith tureTeff = 9250±200K,stellarradiusR∗ = 2.3±0.2R(cid:12),and respecttosolarabundancesbecausetheiratomsarelockedonto massM =2.2±0.1M . ∗ (cid:12) 8 L.E.Ellerbroeketal.:Relatingjetstructurewithphotometricvariability:HD163296 Figure5.Top:LightcurveofHD163296inV(0.55µm)andL(3.76µm).Thedottedlineandthegrayshadedareascorrespondtothe meanvalueand1σspread.ThetimeintervalsdenotedabovetheV-bandlightcurveareexpandedinthebottompanel.Horizontal bars indicate the jet launch epochs estimated from radial velocities (Sect. 3.1). Bottom: V-band lightcurve for selected intervals, exhibitingtheshape,duration,andfrequencyofthefadingevents.Thecalendardatecorrespondingtotheminimumvalueonthe x-axisisdisplayedinthebottomleftcornerofeachgraph.Referencesforplotsymbols:(1)deWinteretal.(2001);(2)Manfroid etal.(1991);(3)Mt.Maidanak(Grankinetal.,inprep.);(4)Swiss(thiswork);(5)Perrymanetal.(1997);(6)Hillenbrandetal. (1992);(7)Eiroaetal.(2001)(8)Pojmanski&Maciejewski(2004);(9)AAVSO(thiswork);(10)Tannirkulametal.(2008b);(11) Mendigut´ıa et al. (2013); (12) Sitko et al. (2008); (13) de Winter et al. (2001); (14) Berrilli et al. (1992); (15) BASS; (16) SpeX (Sitkoetal.,2008,thiswork).SeeTableA.1formoredetails. 9 L.E.Ellerbroeketal.:Relatingjetstructurewithphotometricvariability:HD163296 Figure6. Average SED of HD 163296 (open symbols), fitted witha9250KKuruczmodelreddenedwith A = 0.5;theNIR Figure7. (B−V,V) color-magnitude diagram for HD 163296. V Datapointsspantheperiod1983–2012;seeFig.5forthecover- excessisfittedwithablackbodyat1500K(thecombinedmodel ageofthisepoch.Onlypointswitherrors<0.1magareplotted. spectrum is shown as a solid gray line). Filled symbols indi- Fadingeventsin1988,1991,1993,and2000(letterscorrespond cate measurements during NIR brightening and optical fading to Fig. 5) are highlighted with colored symbols. During these epochs; the number of observations are displayed in brackets. events, the colors change along the extinction vector (Cardelli Dash-dotted,dashed,anddottedgraylinesrepresentmodelsred- denedwithA =0.7,0.9,and1.3,respectively. et al., 1989), which is plotted for RV = 3.1 and RV = 5.0. The V originofthisvector,indicatedwitharedcross,representsthein- trinsiccolorsandmagnitudeofanA1Vstarat119pc(Kenyon 4.1. Photometricvariability &Hartmann,1995). A lightcurve of HD 163296 is constructed from the collected data described in Sect. 2.2 and summarized in Table A.1. The toppanelofFig.5displaystheV-andL-banddataforthe1978– The 2002 event marked a 30% increase in flux in the L-band. 2013epoch.Fig.A.2displaysallthephotometricdata. Note,however,thatboththemeasurementerrorandtheintrinsic The optical brightness of the source over the period 1980– scatter of the NIR brightness are large. A denser coverage of 2012fluctuatesaroundasteadylevel((cid:104)V(cid:105) = 6.93±0.14mag). thelightcurvewouldbeneededtoevaluatethesignificanceand Overthisperiod,severalfadingeventsareseen,duringwhichthe uniquenessofsuchevents. opticalbrightnessdecreasessignificantly.Themostprominentof Thephotometryduringfadingeventsisfurtherillustratedin thesestartedinMarch2001andlastedforatleast6months(up Fig.6.TheaverageSEDwascomputedbytakingthemeanofthe toatmost1.1years).Withinonemonth,Vincreasedupto0.71± photometry, omitting measurements that deviate more than 2σ 0.15 mag above its average value. The bottom panel of Fig. 5 fromthemeanvalue.Itisfittedwitha9250KKuruczmodeland displaysa“zoomin”onthelightcurveoverthisandseveralother ablackbodyat1500K.Intheliterature,estimatesfortheextinc- fading events. Fadings of order 0.1 mag occur predominantly tiondeviatefrom E(B−V) = 0.015(fromtheabsenceofinter- in the 2001–2006 epoch with durations from hours to several stellargasabsorption; Devineetal.,2000)uptoE(B−V)=0.15 weeks. (fromSEDfitting; Tillingetal.,2012).Theaveragereddening Optical photometric observations dating back as far as the in our photometric data is E(B−V) = 0.065±0.016 (Fig. 7). 1890s were obtained from the DASCH catalog (see Grindlay FortheplotinFig.6weadopt E(B−V) = 0.15andRV = 3.1, et al., 2012). Since the target was either saturated or in the hence AV = 0.5(Cardellietal.,1989),whichbestfitstheSED non-lineardomainonmostofthephotographicplates,wewere intheUBVRI-bands.Thedisagreementbetweentheextinction only able to retrieve lower limits on the photometric measure- estimatesmaybeexplainedbyacircumstellarextinctioncarrier ments. However, prolonged fadings of more than 3σ are ob- withalowergas-to-dustratiothanintheISM,variabilityofits servedthroughoutthelightcurve,indicatingthatthesearerecur- column density, or both. The optical colors during the fading ringevents. events in 1988, 1991, 1993, and 2000 are well fitted by an en- In the NIR, the time coverage is much sparser. The mean hanceddustextinctionof∆AV ∼0.2−0.4magandRV =3.1−5.0 brightness in the L-band over the 1978–2012 period is (cid:104)L(cid:105) = (Fig.7).Themajorfadingeventin2001,forwhichcolorsarenot 3.36±0.10mag.ThemeanbrightnesslevelintheH,K,L,and available,impliesanenhancementofAV by0.7–0.8mag. M bands is exceeded by 10% in three epochs: in 1986 (N = The2002NIRbrighteningmayhaveacommonoriginwith obs 2, see Fig. A.2), 2001–2 (N = 3), and 2011–12 (N = 4). the optical fading in 2001. This is further reinforced by the si- obs obs The first two epochs were also reported by Sitko et al. (2008). multaneousmeasurementofopticalfadingandNIRbrightening 10

See more

The list of books you might like

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