Variability in the CO ro-vibrational lines from HD163296

4.6 MB·

by  Rosina P. Hein Bertelsen

MNRAS000,000–000(2015) Preprint19January2016 CompiledusingMNRASLATEXstylefilev3.0
Variability in the CO ro-vibrational lines from HD163296 (cid:63)
Rosina P. Hein Bertelsen1†, I. Kamp1, G. van der Plas2, M.E. van den Ancker3,
L.B.F.M.Waters4,5, W.-F. Thi6, P. Woitke7
1Kapteyn Astronomical Institute, Rijks-universiteit Groningen (RuG), Landleven 12, Groningen 9747, Netherlands
2Departamento de Astronom´ıa, Universidad de Chile, Casilla 36-D, Santiago, Chile
3European Southern Observatory, Karl-Schwarzschild-Str.2, D 85748 Garching bei Mu¨nchen, Germany
4Anton Pannekoek Astronomical Institute, University of Amsterdam, PO Box 94249, 1090 GE Amsterdam, The Netherlands
5SRON Netherlands Institute for Space Research, Sorbonnelaan 2, 3584 CA Utrecht, The Netherlands
6 6Max-Planck-Institut fu¨r extraterrestrische Physik, Giessenbachstrasse 1, 85748 Garching, Germany
1 7SUPA, School of Physics & Astronomy, University of St. Andrews, North Haugh, St. Andrews KY16 9SS, UK
0
2

n
a AcceptedtoMNRASJanuary2016
J

6
1 ABSTRACT
Wepresentforthefirsttimeadirectcomparisonofmulti-epoch(2001–2002and2012)

] CO ro-vibrational emission lines from HD163296. We find that both the line shapes
R
and the FWHM (Full Width Half Maximum) differ between these two epochs. The
S FWHMofthemedianobservedlineprofilesare10–25km/slargerintheearlierepoch,
. and confirmed double peaks are only present in high J lines from 2001–2002. The line
h
p wings of individual transitions are similar in the two epochs making an additional
- central component in the later epoch a likely explanation for the single peaks and
o the lower FWHM. Variations in NIR brightness have been reported and could be
r
linked to the observed variations. Additionally, we use the thermo chemical disc code
t
s ProDiMotocompareforthefirsttimethelineshapes,peakseparations,FWHM,and
a
line fluxes, to those observed. The ProDiMo model reproduces the peak separations,
[
andlowandmidJ linefluxeswell.TheFWHMhowever,areoverpredictedandhighJ

1 line fluxes are under predicted. We propose that a variable non-Keplerian component
v oftheCOro-vibrationalemission,suchasadiscwindoranepisodicaccretionfunnel,
5
is causing the difference between the two data sets collected at different epochs, and
2
between model and observations. Additional CO ro-vibrational line detections (with
2
CRIRES/VLTorNIRSPEC/Keck)or[Neii]lineobservationswithVISIR/VLTcould
4
help to clarify the cause of the variability.
0
.
1 Key words: protoplanetary discs – line: profiles – stars: variables: T Tauri, Herbig
0 Ae/Be – infrared: ISM – (stars:) circumstellar matter
6
1
:
v
i
X 1 INTRODUCTION systemhasbeenestimatedtoM∗∼2.3M(cid:12)(Montesinosetal.
2009).Usingspatiallyresolvedsub-mmdata,theinclination
r Herbig Ae/Be (HAeBe) stars are intermediate mass pre
a of the disc has been estimated to 46◦±4◦ and the position
main sequence stars that are surrounded by discs. CO ro-
angleto128◦±4◦(Isellaetal.2007).Thedischasbeenstud-
vibrational emission is frequently observed from these discs
ied through interferometric observations of both continuum
(Blake & Boogert 2004; Brittain et al. 2007, 2009; van der
and CO emission lines (0.87–7.0mm) (Isella et al. 2007).
Plas et al. 2015) and traces the inner regions close to the
Fromthelineemissiontheauthorsfoundthe12COand13CO
star, where planet formation is expected to occur.
to be optically thick and derive an outer radius of 550±50
HD163296 is a well studied Herbig Ae star of spectral
au.Thecontinuumdustemissionimpliesanouterradiusof
type A3Ve (Gray & Corbally 1998) located at a distance of
200±15au. Thus, a sharp drop in continuum emission of a
118.6pc (Hipparcos). It has been labeled a group II source
factor>30ataradiusof200au,wasinvokedtoexplainthe
(flatdisc)byMeeusetal.(2001)andthestellarmassofthe
lack of continuum emission from radii larger than 200 au
(Isella et al. 2007). Meanwhile, from a non-detection of the
(cid:63) Based on observations made with ESO Telescopes at the La red-shiftedjetbeyond500auradius(Gradyetal.2000)and
theextentofthescatteredlightemission(Wisniewskietal.
SillaParanalObservatoryunderprogrammeID088.C-0898A.
† [email protected] 2008)theradiusofthedustdisccanbeestimatedto500au.
(cid:13)c 2015TheAuthors
2 R. P. Hein Bertelsen et al.
Tilling et al. (2012) presented Herschel/PACS obser- 10”, seen from ALMA observations of the CO J=2–1 and
vations of the disc, and used the line detections from this J=3–2 lines.
dataset together with additional line and continuum data Using high spectral resolution (R ∼12000) VLTI-
tofitaProDiMo(Woitke etal. 2009)model,inanattempt AMBER observations Garcia Lopez et al. (2015) studied
to determine disc properties. A continuous disk with inner the Brγ line of HD163296 and attempted to constrain the
and outer radii of 0.45 and 700au provided a reasonable physicaloriginofthisline.Theycomparedtheobservations
fit to the optical to millimetre continuum spectral energy topredictionsfromalineradiativetransferdiscwindmodel
distribution, the low an mid J CO line intensities, and the andfoundthattheobservationsandthemodellingbothsug-
spectralprofilesfromground-basedinterferometricobserva- gestthattheBrγemissioninthisdisccomesfromadiscwind
tions. ALMA observations (band 7, ∼850µm) have put the with launching radii from ∼0.02au to ∼0.04au, while the
outerdustdiscradiusat240auandtheCOouterradiusat entire disc wind emitting region stretches out to ∼0.16au.
575 au (de Gregorio-Monsalvo et al. 2013; Rosenfeld et al. In this paper we will study the variability in the CO
2013). A well resolved dusty disc was seen with no indica- ro-vibrational lines over a period of ten years (2001–2002
tions of gaps or holes beyond 25au. de Gregorio-Monsalvo to 2012) and investigate the connection with variability ob-
et al. (2013) conclude that a standard tapered-edge model servedintheopticalandNIRcontinuum.Firstly,wepresent
withoneuniquedensityprofilecannotmatchbothdatasets thethreesetsofobservationaldata,andthelineprofilesand
at once. The CO channel maps require a thicker gas disc, FWHM derived from them in a homogeneous way (Sect.
while the mid- and far-infrared SED require a flatter dust 2 and 3). Hereafter, we compare with modelled CO ro-
disc than presented in Tilling et al. (2012). vibrationallinesproducedusingapreviouslypublisheddisc
CO ro-vibrational v=1–0 emission lines from this disc structure(Tillingetal.2012),inordertounderstandtheob-
have been studied on several occasions (Blake & Boogert served CO ro-vibrational line emission (Sect. 4). This also
2004;Salyketal.2011;Brittainetal.2007,Bertelsenetal. tests the predicting power of the model, keeping in mind
submitted, from hereon B15). In Salyk et al. (2011), using that it has been fitted to a wide variety of observational
NIRSPECdatafromtheKecktelescope,theauthorsidenti- data, but never before to the CO ro-vibrational lines them-
fieddoublepeakedprofilesandFWHM∼83km/s(forJ>25) selves.Wefinishthepaperwithadiscussionofhowtheline
indicatingthattheoriginoftheemissionisclosetothestar. variability can be interpreted in the context of both model
AssumingKeplerianrotation,thepeakseparationmeasured and observations (Sect. 5) and present the final conclusions
from the high J profiles would suggest an outer radius for in Sect. 6.
the emission of 2–3au. Recently, we have collected spectra
containing CO ro-vibrational emission lines with CRIRES
at the VLT, and found single peaked profiles for low J lines
2 OBSERVATIONAL DATA
while high J lines displayed asymmetric flat topped profiles
(B15). We use high resolution spectra (R∼100,000) of HD163296
Incoronagraphicimagesabipolarjet(HH409)hasbeen collected in March 2012 with the VLT cryogenic high-
seen(Gradyetal.2000),withradialvelocitiesforthegasin resolution infrared echelle spectrograph (CRIRES Kaeufl
thejetof200–300km/s.HD163296hasshownvariationsin et al. 2004). Frome hereon we refer to this dataset as C12.
NIR (Near Infra Red) brightness (over 30% in J-band, 20% These observations cover a wavelength range from 4.5µm
at H-band, and 15% at K-band, de Winter et al. 2001) on to 5µm with 6 different grating settings (4.6575, 4.7363,
timescales of years (Sitko et al. 2008). This was seen as an 4.9948, 4.6376, 4.8219, and 5.0087µm). The details of the
outburstinthe1–5µmwavelengthregionduring2002.This data reduction and flux calibration can be found in B15.
wavelength region corresponds to a bump frequently seen From C12, we include lines from J=0 up to J=36.
in SEDs of many HAeBes and is thought to be related to For comparison we use high resolution spectra
the inner wall of the dust disc (Natta et al. 2001; Dominik (R∼25000)ofHD163296collectedwithNIRSPEC(McLean
etal.2003).Variabilityinthisregioncouldberelatedtoan et al. 1998) on the Keck II telescope. These observations
increaseofthesurfaceareaoftheinnerwall(puffedupinner arecollectedduringtheperiod2001–2002.Thespectrafrom
wall, Sitko et al. 2008). Sitko et al. (2008) found no SED differentepochswerecombined,resultinginonesinglespec-
model (based on data from one epoch) that could explain trum per wavelength region. This stacking of the lines is
the full nature of this source. necessary in order to improve the S/N on the line profiles.
Ellerbroek et al. (2014) matched transient optical fad- Furthermore, there is no sign of variability between epochs
ing with the enhanced NIR excess (also noticed by Sitko beyond the noise in the spectra (Salyk private communica-
et al. 2008) and relate these to the jet (HH 409). The au- tion),hencethecombiningofthespectraisjustified.Frome
thorssuggestascenariowheredustclouds,thatarelaunched hereonwerefertothisdatasetasN01/02.Duetolowtrans-
above the disc plane to heights where they cross the ob- mission (60–80%) certain regions were left out of this final
servers line of sight to the star, are responsible for the op- spectrum, causing several transitions to have gaps in the
tical and NIR variability. Ellerbroek et al. (2014) derive a centreoftheirlineprofiles(seemiddleframeofFig.1).The
period of 16.0±0.7years for the increased outflow activity observations span a wavelength range from ∼4.65–5.15µm
fromthejet.Duetoepoch-sparsityofthephotometricdata, and include lines from J=1 up to J=37. Details on data ac-
it could not be confirmed whether the jet period matches quisition and reduction for N01/02 can be found in Salyk
the NIR and optical variations, and thereby that a single et al. (2011, 2009) and Blake & Boogert (2004).
mechanismisinfactresponsibleforbothjetandphotomet- WealsoincludedadditionalspectrafromNIRSPECon
ric variability. Klaassen et al. (2013) reported the detection theKeckIItelescope,obtainedinMarch2002.Fromehereon
ofarotatingmoleculardiscwindthatextendstomorethan we refer to this dataset as N02b. The observations span
MNRAS000,000–000(2015)
Variable CO ro-vibrational lines in HD163296 3
J, while mid and high J lines show either double peaked or
Table 1. Observing details, FWHM and errors from Gaussian
fits. N01/02= NIRSPEC 2001/2002 (Salyk et al. 2011), N02b= flattoppedprofiles.However,thecentraldipsinthedouble
NIRSPEC2002(Brittainetal.2007),C12=CRIRES2012(Hein peakedprofilesarecomparabletothenoise,andmanydou-
Bertelsenetal.submitted). ble peaked profiles have central gaps in the spectra, where
low transmission regions due to strong telluric absorption
occurred(Fig.1).Evenlineprofilesthatdisplaywhatlooks
Dataset N01/02 N02b C12 like the normal central dip of a Keplerian profile could in
fact be affected by telluric residuals that mimic the typi-
Instrument NIRSPEC NIRSPEC CRIRES
cal shape. Hence, the double peaked nature is uncertain for
Observingdates 06-08-2001 23-03-2002 06-03-2012
low and mid J lines. High J lines however, are unlikely to
08-08-2001
21-04-2002 be affected heavily by telluric absorption (lines of J∼25–
22-07-2002 30 or higher show no significant telluric absorption, se e.g.
R 25.000 25.000 100.000 theC12spectra),andarethereforemorereliable.TheN02b
Fcont at4.770 µm 12.1Jy — 16.3Jy spectrum shows a general mix between single peaked and
Observedlines J=1–37 J=1–33 J=0–36 double peaked emission lines with no clear connection with
highorlowJ.Also,severallinessufferfromtelluricovercor-
FWHM [km/s] [km/s] [km/s] rectiongivingirregularlines.Again,thecentraldipismostly
comparable to the noise and a single peaked nature cannot
allJ 73.9±1.4 60.0±1.4 55.0±2.5 be ruled out from this dataset either. CRIRES has higher
J<10 66.9±1.8 45.8±1.6 54.5±3.9
spectral resolution than NIRSPEC, and the C12 data have
10<J<20 74.3±3.8 — 50.9±4.9
manyhighsignaltonoiselines.Hence,ifpresentatthetime
J>20 85.0±2.8 78.9±2.2 59.2±4.3
of observation, the double peak should have been visible in
this spectrum.
a wavelength region from ∼4.64–5.02µm and include lines However, the C12 dataset has been collected a decade
fromJ=1uptoJ=33.However,thisdatasethasseverallines later than the other two datasets, thus it is possible that
distortedby telluricover-correction, due tovariable telluric the variability noted for this source in the near-IR con-
water lines during the night of observations. The data re- tinuum and in the optical (Sitko et al. 2008; Ellerbroek
duction for N02b can be found in Brittain et al. (2007). et al. 2014) plays a role. For these studies observations col-
ForC12,wefoundacontinuumfluxat4.770µmof16.3 lectedatseveraldifferentepochswereuseddatingfrom1979
Jy (B15). For typical CRIRES spectra, we can expect the until 2012. During this time, the NIR brightness was in-
accuracyofthefluxcalibrationtobearound∼30%(Brown creased by more than 10% (in H-, K-, L-, and M-band) in
et al. 2013), due to differences in width of the PSF (Point 3 epochs: 1986, 2001–2002, and 2011–2012. Each of the CO
Spread Function) for the science versus the telluric stan- ro-vibrational datasets are collected during or shortly after
dard spectra (these differences are expected to be due to one of these epochs (N01/02 in 2001–2002, N02b in 2002,
variations in the performance of the AO system). From the and C12 in 2012). Hence the observed variations of the CO
N01/02 spectrum a continuum flux at 4.770µm of 12.1 Jy ro-vibrational line profiles are not directly linked to the in-
was found. An error of 10–20% is expected. The continuum creased NIR brightness epochs. Meanwhile, this does not
fluxes measured for C12 and N01/02 are consistent within exclude that one underlying mechanism could in fact be
theuncertainties.ForBN02b,wedonothavefluxcalibrated driving both the NIR variability and the CO ro-vibrational
spectra. variability in different ways.
From all three datasets, we find broad emission
lines (FWHM∼50–80km/s) suggesting that the CO ro-
vibrational lines are emitted from the inner radius of the
3 OBSERVATIONAL RESULTS AND
disc.However,theFWHMmeasuredfromthethreesetsare
ANALYSIS
not in agreement. In Fig. 3, we show the FWHM versus J
up
The individual lines collected from the three observational for lines collected from all three datasets in the same man-
datasetsareshowninFig.1.FortheCRIRESdatathespec- ner.TheFWHMvariationscanalsobeseenfromthemedi-
tral resolution is R=100000, corresponding to ∆v=3km/s, anscollectedforhigh,midandlowJseparately,displayedin
while for NIRSPEC it is R = 25000, corresponding to Fig. 2 and listed with FWHM in Table 1. We see that lines
∆v=12km/s. The lines displayed in Fig. 1 have smaller from N01/02 are overall wider than those from C12. Low J
spacing between velocity channels than the ∆v=3km/s and lines from N02b are narrower than both N01/02b and C12,
∆v=12km/smentionedabove(theC12datahas∼1.5km/s, but high J lines from N02b are comparable in FWHM to
the N01/02 data has ∼4km/s, and the N02b data has N01/02. As discussed above, the N02b dataset has several
∼5km/s)duetofinesamplingofthespectralresolution.In irregular lines, which may be due to imperfect telluric cor-
Fig. 2, we show normalised line profiles compiled from co- rections. Thus for the further comparison we focus on the
addingeitheralllines,alllowJ lines(J<10),allmidJ lines C12 and N01/02 datasets.
(10<J<20), or all high J lines (J>20), in separate medians To explore in more details the differences between C12
for C12, N01/02 and N02b. andN01/02,wealsocomparefluxcalibratedlineprofilesfor
The C12 spectrum shows single peaked emission lines individual transitions that were present in both data sets
for low and mid J values while flat topped asymmetric pro- (Fig. 4). This shows that the general width and line wings
files with shoulders seem to be present at higher J. The of the individual line profiles are the same in the two data
N01/02spectrumshowssinglepeakedemissionlinesforlow sets, while (in the case of high and medium J lines) the in-
MNRAS000,000–000(2015)
4 R. P. Hein Bertelsen et al.




Figure 1.Individuallineprofilesandthemedianatthebottomfromallthreedatasets.Left:LineprofilesfromN01/02.Middle:Line
profilesfromN02b.Right:LineprofilesfromC12.Lineidentificationsareindicatedontheplots.SeverallinesfromtheN01/02dataset
have’gaps’atthelinecentres,thustheregionmarkedinblueontheN01/02median,correspondingtothesemissingregions,shouldbe
deemedlessreliable.SomelinesintheC12datasetaredetectedtwiceduetooverlappingwavelengthsettings.
Figure 2.ComparisonofthemedianlineprofilesfromtheC12data(blue),theN01/02data(red),andtheN02bdata(black).Ranges
ofJ-levelsusedforthemedians(co-addition)areindicatedontheplots.
MNRAS000,000–000(2015)
Variable CO ro-vibrational lines in HD163296 5
Figure 3. FWHM versus Jup for all three samples. Lines collected from C12 are shown in blue, lines collected from N01/02 are shown
inred,andlinesfromN02bareshowninblack.
Figure 4. Individual line profile comparisons of flux calibrated lines, that were present in both the C12 data (blue) and the N01/02
data(red).Lineidentificationsarenotedontheplots.Awideplottingrangeischosentoshowthetypicalnoiseonthecontinuum.
MNRAS000,000–000(2015)
6 R. P. Hein Bertelsen et al.
tensity and the shape of the central part of the lines differ.
Table 2. {Key model parameters adapted or fitted in T12
ForlowJ (theP4line)theshapeiswellmatchedinthetwo or adapted from other work: B10=Benisty et al. (2010),
datasets. An interpretation could be that the single peaked B16=this work, E09=Eisner et al. (2009), G00=Grady et al.
lineprofiles(i.e.linesfromtheCRIRESdataandlowJlines (2000), H08=Hughes et al. (2008), I07=Isella et al. (2007),
from the NIRSPEC data) are composites of multiple emit- L07=van Leeuwen (2007), M09=Montesinos et al. (2009),
ting regions, where the main (Keplerian) components are R10=Renard et al. (2010), T08a=Tannirkulam et al. (2008b),
the same in the two epochs and the additional component, T08b=Tannirkulam et al. (2008a), T12=Tilling et al. (2012),
W06=Wasselletal.(2006).FormoredetailsseeT12.
present in lines from C12 and in low J lines from N01/02,
couldberelatedtotheNIRvariability.Thisadditionalcom-
ponentwouldthen’drive’theFWHMinC12tolowervalues, Quantity Symbol Value Reference
since the half maximum location is shifted upwards. innerradius R [au] 0.45 E09,T08b,
in
R10,B10
outerradius Rout [au] 700.0 B16
4 COMPARISON WITH A DISC MODEL tapering-offradius Rtaper [au] 125. H08
stellarluminosity L∗/L(cid:12) 37.7 M09
WeuseapreviouslypublisheddiscstructureforHD163296 stellarmass M∗/M(cid:12) 2.47 M09
producedwithProDiMo(Model3fromTillingetal.2012), effectivetemperature Teff[K] 9250.0 M09
to model for the first time a sample of CO ro-vibrational discgasmass Mdisc/M(cid:12) 0.071 FitinT12
lines.Inthefollowingsections(Sections4.1and4.2)wefirst dusttogasmassratio dust-to-gas 0.01 FitinT12
dustgrainmaterial-
describe the model and then present how the model results
compare to the observed lines. massdensity ρgrain[g/cm3] 3.36 T12
minimumgrainsize amin[µm] 0.01 FitinT12
maximumgrainsize amax[µm] 2041 FitinT12
dustsizedistribution-
4.1 Model description
powerlawindex p 3.68 FitinT12
ProDiMo is a radiation thermo-chemical disc code (Woitke flaringindex β 1.066 FitinT12
et al. 2009; Kamp et al. 2010), which solves the radiative distance d [pc] 118.6 L07
transfer, the chemical network and the gas heating/cooling PAHabundance-
balance.TheTillingetal.(2012)modelassumesaparame- relativetoISM fPAH 0.0068 FitinT12
inclination i 50.0◦ G00,W06,I07,
terised disc structure using power laws for the surface den-
T08a,B10
sity and the gas scale height. Level populations are calcu-
lated from statistical equilibrium and detailed line transfer
calculationsareperformedtopredictemissionlines.Inorder Table3.COlinefluxes.Themodellinefluxesarecomputedfrom
to model CO ro-vibrational emission lines, we use the large escapeprobabilityandhighJrotationallineswereobservedwith
Herschel/PACS (Tilling et al. 2012), low J rotational lines were
CO ro-vibrational model described in Thi et al. (2013). We
includefluorescencepumpingtotheA1Πelectronicleveland observed from the ground (Isella et al. 2007), and the observed
ro-vibrationallinefluxesarefromSalyketal.(2011).
use40rotationallevelswithin7vibrationallevelsofboththe
ground electronic state X1Σ+ and the excited state A1Π.
Mol. λ Obs. Model
The Tilling et al. (2012) model has been computed us-
[µm] [10−18W/m2] [10−18W/m2]
ing a Monte-Carlo evolutionary χ2-minimisation strategy
(Woitke et al. 2011), varying 11 parameters to find the 13COJ=1–0 2720.41 0.0124 0.0105
best fit to observed emission lines (not including the CO COJ=2–1 1300.4 0.379 0.40
ro-vibrational lines) and the dust spectral energy distribu- COJ=3–2 866.96 1.65 1.21
COJ=18–17 144.78 <13.1 2.69
tion (SED), from HD163296, simultaneously. The spatial
COJ=29–28 90.16 <11.1 0.624
CO interferometric data were not used as a constraint for
COv=1–0P37 5.053 60.3 122.
the modelling, however, the extracted line profiles for the
COv=1–0P14 4.793 86.5. 71.3
pure rotational CO lines (Isella et al. 2007) were used as
COv=1–0P4 4.700 158 70.8
a constraint for the radial extent and tapered edge. Key
parametersusedinthemodelaredisplayedinTable2.The
gasvolumedensitydistribution,thegastemperatureandthe resolutioncouldthusbegoodenoughtoseparatethepeaks,
COabundanceareshowninFig.5.InTable3,emissionline depending however on the size of the depression between
predictions from the model are compared to the observed the peaks. In the upper panels of Fig. 6, the peak separa-
lines from the literature. Even though this model has not tion, the FWHM, and the line fluxes are shown for selected
been made to fit CO ro-vibrational observations, we use it line transitions (the observed line sample and a few extra
heretoinvestigatetheemittingregionofCOro-vibrational linestofillingapsinJ coverage)fromthemodel.Thelower
transitions and to compare with observations. panelsshowthesesameplotswithalltransitionscalculated
in LTE. For comparison we also show the observed peak
separations, FWHM, and line flux. To obtain a more re-
4.2 Line profile comparison
alistic comparison between model and observation, we add
We derive line flux, FWHM and peak separation from the appropriate noise (S/N∼100 on the continuum) and con-
modelled CO ro-vibrational line profiles. Line shapes in the volve the model line profiles with the spectral resolution
disc model do not vary strongly with J. The peak separa- of the relevant instruments (CRIRES: 3km/s, NIRSPEC:
tionsinthemodelledlinesare∼50km/s,andtheNIRSPEC 12km/s) to create simulated CRIRES and NIRSPEC line
MNRAS000,000–000(2015)
Variable CO ro-vibrational lines in HD163296 7
0.30 0.1 0.30 3 0 0 1K00 K 40K 0.30
4 6 8 10 12 14 1.01.52.02.53.03.54.0 -12 -10 -8 -6 -4
0.25 0.1log n<H> [cm-3] 0.25 1000loKg Tgas [K] 0.25 log (cid:161) (CO)
5000KK 3 0 0 K 1 0 0 K
0.20 0.20 00 0.20
0
z / r 0.15 z / r 0.151 40 K z / r 0.15
=1 AV=10 AV=1 AV=10 AV=1 20 AV=10 AV=1
0.10A=1V,rad 0.10=1 K 0.10=1
V V V
0.05A 0.05A 0.05A
0.1 20K
0.00 0.00 0.00
1 10 100 1 10 100 1 10 100
r [au] r [au] r [au]
Figure 5. Left panel: Gas volume density, together with contours showing Av=1.0 and Av=10.0 (dashed white lines), Av,rad=0.1, and
Av,rad=1.0(dashedredlines).Middlepanel:Gastemperature,togetherwithcontoursshowingAv=1.0andAv=10.0(whitedashedlines),
andgastemperaturecontoursat20,40,100,300and1000K(blackdashedlines).Rightpanel:COabundance,togetherwithcontours
showingAv=1.0andAv=10.0(dashedwhitelines)andlog(χ/n)=-3(dashedblueline).
Figure 6. Upper panels: The line fluxes (left), the peak separations (middle), and the FWHM (right), as a function of Jup for the
modelled sample of CO fundamental ro-vibrational (black squares). The green squares are the model convolved with the NIRSPEC
resolution (the model convolved with the CRIRES resolution corresponds very closely to the un-convolved model and is therefore not
shown).Observationsareshownforcomparisonasbluecrosses(C12)andredcrosses(N01/02).Inadditiontothedisplayederrorbars
fortheobservedlines,thereisalsoanadditionalsystematicerrorof50%fortheC12dataand20%fortheN01/02data.Thiserrorhas
asimilareffectonalllines,andthusdonotaffectlineratios.Bottompanels:ThesameplotswiththelinescalculatedinLTE.
profiles from the modelled lines. The simulated line profiles from the models have FWZI around 110–120km/s (mea-
madeforthreerepresentativeCOro-vibrationaltransitions, sured by eye), similar to the observed FWZI (The FWZI of
v(1–0)P04,v(1–0)P14,andv(1–0)P37,areshowntogether the lines is determined by the inner radius of the disc. In
with the original model line profiles in Fig. 7. These three the model that is set to 0.45 au, while values derived from
linesarechosensincetheyexemplifyalowJ,amidJ,anda observationslieintherange0.2–0.55au(Renardetal.2010;
highJline,andsincetheyaredetectedinbothobservational Eisner et al. 2009; Tannirkulam et al. 2008b; Benisty et al.
data sets, N01/02 and C12. 2010).). The FWHM values predicted by the model are a
factor1.3toowidecomparedtoobservedlinesfromN01/02
The simulated CRIRES and NIRSPEC line profiles
MNRAS000,000–000(2015)
8 R. P. Hein Bertelsen et al.
17.0 0.01" x 0.01" 0.6 CO @ 4.700 µm 0.01" x 0.01" 0.6 CO @ 4.793 µm 0.01" x 0.01" CO @ 5.053 µm
0.4 i = 50.0o d = 118 pc 0.4 i = 50.0o d = 118 pc 0.5 i = 50.0o d = 118 pc
16.8 0.2 FlinFe = 1=.4 17.E5-71E6+ W01/ mJy2 17.0 0.2 FlinFe = 1=.9 11.E5-71E6+ W01/ mJy2 17.0 FlinFe = 2=.0 16.E5-41E6+ W01/ mJy2
0.0 FWcHontM =104.8km/s 0.0 FWcHontM =105.0km/s 0.0 FWcHontM =103.1km/s
-0.2 (cid:54)v = 54.05 km/s -0.2 (cid:54)v = 54.05 km/s (cid:54)v = 52.90 km/s
16.6 -0.4 sep -0.4 sep -0.5 sep
-0.6 -0.6 16.5
F [Jy](cid:105)16.4 -00..60 E+I0-l00in.4e4 .[7EJ-+y00.62 k9m.40E./+0s06/a10r..4c2Es+0e7c0.124.]9E+007.6 F [Jy](cid:105)16.5 -0.06.0 E+I0-l00in.4e6 .[1EJ-+0y0.62 k1m.20E./+0s07/a1r0.8.c2Es+0e7c02.24.]4E+007.6 F [Jy](cid:105) 0.0 E+I0l-0i0n.e55 .[6EJ+y06 k1m.10E./+0s07/a1r.7cEs+0e70c.522.]2E+07
16.2
16.0
16.0 16.0
15.8 15.5
-150 -100 -50 0 50 -150 -100 -50 0 50 -150 -100 -50 0 50
v [km/s] v [km/s] v [km/s]
17.0 0.01" x 0.01" 0.6 CO @ 4.700 µm 0.01" x 0.01" 0.6 CO @ 4.793 µm 0.01" x 0.01" CO @ 5.053 µm
0.4 i = 50.0o d = 118 pc 0.4 i = 50.0o d = 118 pc 0.5 i = 50.0o d = 118 pc
16.8 0.2 FlinFe = 1=.4 17.E5-71E6+ W01/ mJy2 17.0 0.2 FlinFe = 1=.9 11.E5-71E6+ W01/ mJy2 17.0 FlinFe = 2=.0 16.E5-41E6+ W01/ mJy2
0.0 FWcHontM =104.8km/s 0.0 FWcHontM =105.0km/s 0.0 FWcHontM =103.1km/s
-0.2 (cid:54)v = 54.05 km/s -0.2 (cid:54)v = 54.05 km/s (cid:54)v = 52.90 km/s
16.6 -0.4 sep -0.4 sep -0.5 sep
-0.6 -0.6 16.5
F [Jy](cid:105)16.4 -00..60 E+I0-l00in.4e4 .[7EJ-+y00.62 k9m.40E./+0s06/a10r..4c2Es+0e7c0.124.]9E+007.6 F [Jy](cid:105)16.5 -0.06.0 E+I0-l00in.4e6 .[1EJ-+0y0.62 k1m.20E./+0s07/a1r0.8.c2Es+0e7c02.24.]4E+007.6 F [Jy](cid:105) 0.0 E+I0l-0i0n.e55 .[6EJ+y06 k1m.10E./+0s07/a1r.7cEs+0e70c.522.]2E+07
16.2
16.0
16.0 16.0
15.8 15.5
-150 -100 -50 0 50 -150 -100 -50 0 50 -150 -100 -50 0 50
v [km/s] v [km/s] v [km/s]
Figure7.Modelledlineprofileswithandwithoutconvolutionandnoise.Fromlefttoright:v(1–0)P04,v(1–0)P14,andv(1–0)P37.The
black lines in the upper three panels show the model line profiles convolved with the CRIRES resolution (R=100000) and an applied
signaltonoiseratioof100onthecontinuum,whiletheun-convolvedmodellineprofilesareshowninred.Theblacklinesinthelower
threepanelsshowthemodellineprofilesconvolvedwiththeNIRSPECresolution(R=25000)andanappliedsignaltonoiseratioof100
onthecontinuum.Thelinewavelength,flux,FWHMandpeakseparationofthe(un-convolved)modelarereportedontheplot.
CO 4.700µm CO 4.793µm CO 5.053µm
106 106
(cid:111) 110024 (cid:111)cont (cid:111)line (cid:111) 110024 (cid:111)cont (cid:111)line (cid:111) 110024 (cid:111)cont
100 100 100
10-2 10-2 10-2 (cid:111)
mulative F [%]line12468000000 Fline = 7.31E-17 W/m2 mulative F [%]line12468000000 Fline = 7.38E-17 W/m2 mulative F [%]line12468000000 Fline = 1.23E-16 W/m2 line
cu000000......22322305005000 -4 -2100 00Klog2 nCO 4[cm306-30]K81001K0 100K cu000000......22322300050050 -4 -2100 00Klog2 nCO 4[cm306-30]K81001K0 100K cu000000......22322300050050 -4 -2100 00Klog2 nCO 4[cm306-30]K81001K0 100K
z/rz/r000000......0110115055051000K 300K 100K z/rz/r000000......0110115055051000K 300K 100K z/rz/r000000......0110115055051000K 300K 100K
00..0000 00..0000 00..0000
11 1100 110000 11 1100 110000 11 1100 110000
rr [[AAUU]] rr [[AAUU]] rr [[AAUU]]
Figure8.Lineandcontinuumopticaldepth,thecumulativelineflux,andtheCOdensity,asafunctionofradius,forthethreemodelled
linesv(1–0)P04,v(1–0)P14,andv(1–0)P37.Thelinewavelengthsandfluxesfromsimpleverticalescapeprobabilityareindicatedonthe
plot. Vertical dashed lines indicate the radii within which 15% and 85% of the total flux is emitted, the black box indicates the radial
andverticalregionfromwhich50%ofthefluxoriginates,andtheredcontourlinesrepresentgastemperaturesof100Kand300K.
MNRAS000,000–000(2015)
Variable CO ro-vibrational lines in HD163296 9
(Fig. 6). The line fluxes predicted by the model match the 5 DISCUSSION
observedlowJ line fluxes (N01/02and C12) butthe highJ
5.1 Observed line profile comparison
line fluxes are up to a factor four too faint and the overall
shape of the line flux versus J curve is different than what N01/02 and C12 show clear differences in the line profile
is observed. The LTE model captures the line flux versus shapes that instrumental differences alone cannot explain.
J curve shape much better but the line fluxes are almost a ThesinglepeakednatureoftheC12linespointsatvariabil-
factor ten too high. ity in the lines as an explanation. Furthermore, a detailed
Themodelpredictscleardoublepeaksforalllines(sig- thermo-chemicaldiscmodelpredictsdoublepeakedlinepro-
nificantcentraldepressionbetweenpeaks),whichisnotseen fileswhichareonlyseenforhighJlineprofilesfromN01/02.
foralllinesintheobservations.SincetheCOro-vibrational A possible explanation for single peaks could be that an
lines are not spatially resolved in observations the observed additional non-Keplerian CO ro-vibrational emission com-
single peaks cannot be caused by emission at larger radii ponent is present in both N01/02 and C12, but that this
in the disc. Thus, the observed single peaks in both the componentisweakerinN01/02,andhencemainlyaffecting
CRIRES lines and the low J NIRSPEC lines could indi- lowJ lineshere.TheobservedCOro-vibrationalemissionis
cate that lines from both N01/02 and C12 have a second not spatially extended, hence, a Keplerian emission compo-
non-Kepleriancomponent,althoughmuchweakerinN01/02 nentfromlargeradiiinthedisccannotbeinvokedtoexplain
than in C12. We do see indications of double peaks present the single peaked profiles.
in some observed lines (mostly high and mid J lines from Comparinglinefluxes,theoverallshapeorcurvaturein
N01/02) and the peak separations measured from these are line flux versus J diagram agrees well between N01/02 and
similar to the modelled lines. However, the absence of a C12, and the line flux values are also similar. Hence, the
strong double peak could also be due to many lines in the line fluxes alone do not indicate any kind of variability in
N01/02samplesufferingfromlowtransmissionatthecentre the lines. However, the flux calibration, is rather uncertain,
of the line, due to strong telluric absorption lines (see Sect. around30%forC12and20%forN01/02,andthusthesim-
3). The high J transitions are less likely to be significantly ilarity in line fluxes does not necessarily exclude variability.
affected by telluric absorption and lend themselves better Spectra from multiple epochs were combined to obtain
for acomparisontothemodels.InFig.9we showfluxnor- theN01/02datasetsincenoindicationofvariabilitybetween
malised medians from the simulated CRIRES line profiles line profiles was detected beyond the noise. No indepen-
and NIRSPEC line profiles from the model together with dentfluxcalibrationexistsbetweenthemultipleepochs;the
thecorrespondingobserved(N01/02,C12)medianlinepro- N01/02 data were neither obtained nor analysed with the
files.WeshowbothmediansmadefromallJtransitionsand intention to study variability. The same is true for the C12
medians made from only high or low J lines in the sample. data. Hence, there is now a clear need for new systemati-
We can confirm that the high J median from the N01/02 cally collected data to better understand the variability of
samplecompareswelltothemodel.Lookinginparticularat this source (we discuss this further in Sect. 5.3).
thelinewings,themediancreatedfromalllinesfromN01/02
andthemediancreatedfromhighJ linesfromN01/02look
5.2 Model comparison
similar. However, comparison of the medians from C12 and
thelowJmediansfromN01/02revealssignificantdifferences Consideringthefactthatthismodelwasnotsetuptofitthe
in the line wings. In these cases, the line profiles might be CO ro-vibrational lines initially, this comparison provides a
multiple component profiles with a non-Keplerian, central strong test. For high J lines from N01/02, the peak separa-
component. Thus, the direct comparison of medians from tions compare well and the FWHM are near the observed
modelandfromobservations,bothnormalisedtounity,can values. If all other observed lines are in fact affected by the
be misleading, since the model contains only a disc com- presenceofanadditionalnon-Kepleriancomponent(thatis
ponent. We can obtain a match between the line wings of not included in the model), the FWHM or peak separation
the observed C12 lines and the line wings of the simulated oftheremainingobservedlinescannotbeexpectedtomatch
CRIRES line profiles from the model by re-normalising the the Keplerian lines of the model.
C12 median line profile to the wings of the model (multi- The individual line fluxes predicted by the model are
plying the normalised C12, low J median by 3.5, shown in similar to the observed values up to J∼26. The modelled
blue). A similar match could also be achieved for the low high J lines fluxes deviate by up to a factor 4 from the ob-
J median from N01/02 (multiplying the normalised N01/02 servedones.TheoverallshapeofthefluxversusJdiagramis
median by 2.2, shown in blue). not well fitted by the model. The modelled lines calculated
The shape of a line profile is determined by how the inLTEgiveamuchbettershapepredictionforthefluxver-
flux builds up as a function of radius. In Fig. 8 we show sus J diagram, but the line fluxes are in this case almost a
a radial profile of the cumulative line flux from the model. factor of ten too high,a difference that cannot be explained
FromthepeakseparationofthehighJlinesfromN01/02,we by flux calibration uncertainties (C12, 30% systematic flux
inferanouterradiusfortheemittingregionof∼2.1±2.0au, calibration uncertainty, N01/02, 20% systematic flux cali-
using the assumption of Keplerian rotation. In the model, brationuncertainty).Thedifferenceinlinestrengthbetween
85% of the emission builds up within ∼1.5 au. Thus, the the (non-LTE) model and observations may point to a dif-
model provides a good prediction of the emitting region. ferent vertical structure in the inner disk model and/or the
In summary, provided there is a non-Keplerian additional excitation of the lines (thermal and fluorescent). However,
componentcausingacentralpeakinthelinesfromC12and the difference in flux versus J curvature could arise from
in the low J lines from N01/02, the model compares well to an additional non-Keplerian component that is more pro-
the observed CO ro-vibrational lines of HD163296. nouncedatlowJ comparedtohighJ lines(alsoresponsible
MNRAS000,000–000(2015)
10 R. P. Hein Bertelsen et al.
Figure9.Upperpanels:Individualnormalisedmodelledlineprofiles(dottedblacklines)andtheirmedians(thickblack)withanapplied
signal to noise ratio of 100 (on the continuum) and the spectral resolution of CRIRES (left) or NIRSPEC (right). For comparison, the
observed line profile median is overplotted in thick red (left: C12, right: N01/02). Medians created using only the low J lines (J<10)
showninthemiddlepanels,andmedianscreatedusingonlythehighJlines(J>20)showninthelowerpanels.FortheC12mediansand
the low J N01/02 median, we additionally overplot the observed medians re-normalised instead to the line wings of the model median
(thickblue)toevaluatethepresenceofanadditionalnon-Kepleriancomponentcausingthecentralpeaks.
MNRAS000,000–000(2015)