Millan-Gabet Et Al

Very close environments of young stars

Fabien Malbet

Institut de Planétologie et d’Astrophysique de Grenoble University of Grenoble / CNRS

CNRS Summer School - Roscoff 4 April 2010

VLTI Summer school 2010 - Inner regions of YSOs revealed by interferometry - F. Malbet Inner regions of Young Stellar Objects

From Isella et al. (2007)

VLTI Summer school 2010 - Inner regions of YSOs revealed by interferometry - F. Malbet Outline

• Introduction – Physical conditions in the inner regions of YSOs – Need for very high angular resolution – Physical processes

• Infrared interferometry – Principles and observables – Instruments available for inner regions studies – Elements of bibliography

• Inner disk physics – Morphology of circumstellar structures – Constraints on disk structure (T, z,…) – Dust mineralogy – Gas/dust connection – Transition disks

• Other AU-scale phenomena – Hydrogen line emission : Outflows / winds or Magnetosphere? – Importance of Binaries and multiple systems

• Future prospects

VLTI Summer school 2010 - Inner regions of YSOs revealed by interferometry - F. Malbet INTRODUCTION

– Formation of stars, disks and planets – Physical conditions in the inner regions of YSOs – Need for very high angular resolution – Physical processes

VLTI Summer school 2010 - Inner regions of YSOs revealed by interferometry - F. Malbet Formation of stars, disks and planets

NuageMolecular moléculaire clouds EffondrementCollapse FragmentationFragmentation

10 pc ! 4° 1 pc ! 20’ 0.1 pc ! 2’ (Taurus) Accretion disk and bipolar ejection Debris disk and protoplanets Planetary system Bouvier Bouvier & Malbet (2001, DPLS, (2001, 30, 84)

150 AU ! 1’’ 300 AU ! = 2’’1,3” = 1,300 mas30 AU ! 0.2’’ 5 Formation of stars, disks and planets

NuageMolecular moléculaire clouds EffondrementCollapse FragmentationFragmentation

10 pc ! 4° 1 pc ! 20’ 0.1 pc ! 2’ (Taurus) Accretion disk and bipolar ejection Debris disk and protoplanets Planetary system Bouvier Bouvier & Malbet (2001, DPLS, (2001, 30, 84)

150 AU ! 1’’ 300 AU ! = 2’’1,3” = 1,300 mas30 AU ! 0.2’’ 5 Formation of stars, disks and planets

NuageMolecular moléculaire clouds EffondrementCollapse FragmentationFragmentation

10 pc ! 4° 1 pc ! 20’ 0.1 pc ! 2’ (Taurus) Accretion disk and bipolar ejection Debris disk and protoplanets Planetary system

Bouvier Bouvier & Malbet a few AUs (2001, DPLS, (2001, 30, 84)

150 AU ! 1’’ 300 AU ! = 2’’1,3” = 1,300 mas30 AU ! 0.2’’ 5 Physical conditions in the close environment of young stellar objects

Dust

Magnetosphere

Accretion disk Planets Gas

• Physical phenomena • Physical conditions – Keplerian accretion disk: gas + dust – Radius ranging from 0.1 AU to 10 AU – Stars from K to B spectral types (4000K to – Temperature ranging from 150 K to 4000 K 10000K) – Velocity ranging from 10 km/s to few 100 km/s – Strong outflowing wind – Companions – Magnetophere – Protoplanets ! At 150 pc (Taurus), this corresponds to : 1µm ! " ! 20µm and spatial scales between 0.5 et 70 mas

VLTI Summer school 2010 - Inner regions of YSOs revealed by interferometry - F. Malbet 34 The inner regions of protoplanetary disks

!"#$%&'( Close environment of young stars

Fig. 1. This figure (Dullemond & Monnier 2010) sketches the different gas and dust regions of the disk surrounding a typical young star. With the angular resolution and spectral coverage of the Keck Interferometer, the inner disk regions (sub-AU to few-AU) can be directly probed.

Dullemond & Monnier (2010, ARAA) Figure 1: Pictogram of the structure and spatial scales of a protoplanetary disk. Note that the radialVLTI scale Summer school on 2010 the - Inner x-axis regions of in YSOs not revealed linear. by interferometry Above - F. Malbet t he pictogram it 0.18 AB Aur 3.5µm Visibility 1.0 is shown which techniquesAB Aur 3.5 µm model can image spatially resolve which scales. Below it is shown which kind1.0 of emission arises from which parts of the disk. Fig. 2. Model images (left panels) 0.15

0.8 and corresponding visibility curves 0.5 (right panels) at 3.5 µm and 10.7 µm 0.13 for one of our primary targets, the 0.6

0.0 Visibility 0.10 well known Herbig Ae object AB Surface Brightness Aur. We note that for ease of

-0.5 0.4 0.08 comparison, we have converted the observable measured by the KI

0.05 -1.0 0.2 Nuller (null depth) into its -1.0 -0.5 0.0 0.5 1.0 0 20 40 60 80 AU Baseline (m) equivalent visibility modulus. These models are predictions from the 0.21 AB Aur 10.7 micron model image AB Aur 10.7µm Visibility solution developed in Tannirkulam 1.0 100 Keck Segment-Tilting Data 2008 for this object. Note an 0.18 10 alternative way to visualize the data 0.8 50 that can be obtained: at each value 0.15 5

milliarcsec of the projected baseline, we will

0 0 0.6 AU have 382 spectrally dispersed 0.12 Visibility visibility points across the K, L and -5 (Surface Brightness (erg/s/cm^2) )^0.1 -50 0.09 0.4 N bands (330 in K, 10 in L and 42 in -10 N).

0.06 -100 0.2 0 20 40 60 80 -10 -5 0 5 10 Baseline (m) AU

Fig. 3. Representative KI N-band/Nulling KI N-band/Nulling examples of calibrated data 0.8 0.8 obtinaed thus far. All ry_tau ab_aur objects appear clearly 0.6 0.6 resolved. Note the large null N-band leakages (20- mwc275 0.4 0.4 80%). Note also the clearly higher V2 in the Br! line, Calibrated Null Leak Calibrated Null Leak v1057cyg 0.2 v1685cygmwc_480 0.2 showing that hot gas su_aur v1295_aql emission originates inside 0.0 0.0 dg_tau the dust sublimation radius. 8 9 10 11 12 13 14 8 9 10 11 12 13 14 wavelength (microns) ( wavelength (microns)

Millan-Gabet et al. NASA Keck Proposal 2011B 6 Instrumental requirements

• Wavelength domain Temperature ranges ! " ~ 1 to 20 µm: • Angular resolution Spatial scales

1.22 "/D 0.1 AU 1AU 5AU 10AU

75pc 1.5mas 15mas 70mas 150mas

150pc 0.7mas 7mas 30mas 70mas

450pc 0.2mas 2mas 10mas 20mas

VLTI Summer school 2010 - Inner regions of YSOs revealed by interferometry - F. Malbet Instrumental requirements

• Wavelength domain Temperature ranges ! " ~ 1 to 20 µm: • Angular resolution Spatial scales Spatial telescope 1.22 "/D 0.1 AU 1AU 5AU 10AU

75pc 1.5mas 15mas 70masAdaptive150mas optics Interferometry 150pc 0.7mas 7mas 30mas 70mas

450pc 0.2mas 2mas 10mas 20mas

VLTI Summer school 2010 - Inner regions of YSOs revealed by interferometry - F. Malbet Mm interferometry

Ground-based telescopes

Infrared Interferometry ALMA

Dust Wind sublimation

VLTI Summer school 2010 - Inner regions of YSOs revealed by interferometry - F. Malbet Mm interferometry

Ground-based telescopes

Infrared Interferometry ALMA

Dust Wind sublimation

Optical Interferometry

VLTI Summer school 2010 - Inner regions of YSOs revealed by interferometry - F. Malbet Infrared and visible Interferometry

– Principles and observables – Instruments available for YSO studies – Elements of bibliography on YSO science results

VLTI Summer school 2010 - Inner regions of YSOs revealed by interferometry - F. Malbet Basics of optical interferometry

Visibility ampltitude V(u,v)

Visibility phase !(u,v)

VLTI Summer school 2010 - Inner regions of YSOs revealed by interferometry - F. Malbet Spatial coherence

Interferogram 1 Reduced contrast Shifted phase Interferogram 2

Zernicke-van Cittert theorem Visibility = Fourier transform of the brightness spatial distribution

VLTI Summer school 2010 - Inner regions of YSOs revealed by interferometry - F. Malbet Visibilities

Uniform disk Binary with unresolved Binary with resolved components component

VLTI Summer school 2010 - Inner regions of YSOs revealed by interferometry - F. Malbet Interferometers involved in YSO research

Tel. Wavelength Baseline Facility Instrument # tel. Diam. (microns) (m) (m) Existing facilities PTI V" H, K 3 0.4 80-110 IOTA V", CP H, K 3 0.4 5-38 ISI Heterodyne 11 2/3 1.65 4-70 V" K, L / spectral KI 2 10 80 nulling N AMBER: V", CP 1-2.5 / spectral 3 8.2 40-130 VLTI 4 (8) MIDI: V", V 8-13 / spectral 2 1.8 8-200 CHARA V", CP, Imaging 1-2.5 (spectral) 2/4 (6) 1 50-350 Future facilities LBT V", nulling 1-12 µm 2 8.4 6-23 MROI V", CP, imaging V, NIR 3/6 (10) 1.4 7.5-340

VLTI Summer school 2010 - Inner regions of YSOs revealed by interferometry - F. Malbet Interferometers involved in YSO research

Tel. Wavelength Baseline Facility Instrument # tel. Diam. (microns) (m) (m) Existing facilities PTI V" H, K 3 0.4 80-110 Small apertures IOTA V", CP H, K 3 0.4 5-38 ISI Heterodyne 11 2/3 1.65 4-70 V" K, L / spectral KI 2 10 80 nulling N AMBER: V", CP 1-2.5 / spectral 3 8.2 40-130 VLTI 4 (8) MIDI: V", V 8-13 / spectral 2 1.8 8-200 CHARA V", CP, Imaging 1-2.5 (spectral) 2/4 (6) 1 50-350 Future facilities LBT V", nulling 1-12 µm 2 8.4 6-23 MROI V", CP, imaging V, NIR 3/6 (10) 1.4 7.5-340

VLTI Summer school 2010 - Inner regions of YSOs revealed by interferometry - F. Malbet Interferometers involved in YSO research

Tel. Wavelength Baseline Facility Instrument # tel. Diam. (microns) (m) (m) Existing facilities PTI V" H, K 3 0.4 80-110 IOTA V", CP H, K 3 0.4 5-38 ISI Heterodyne 11 2/3 1.65 4-70 V" K, L / spectral KI 2 10 80 nulling N Large apertures AMBER: V", CP 1-2.5 / spectral 3 8.2 40-130 VLTI 4 (8) MIDI: V", V 8-13 / spectral 2 1.8 8-200 CHARA V", CP, Imaging 1-2.5 (spectral) 2/4 (6) 1 50-350 Future facilities LBT V", nulling 1-12 µm 2 8.4 6-23 MROI V", CP, imaging V, NIR 3/6 (10) 1.4 7.5-340

VLTI Summer school 2010 - Inner regions of YSOs revealed by interferometry - F. Malbet Interferometers involved in YSO research

Tel. Wavelength Baseline Facility Instrument # tel. Diam. (microns) (m) (m) Existing facilities PTI V" H, K 3 0.4 80-110 IOTA V", CP H, K 3 0.4 5-38 ISI Heterodyne 11 2/3 1.65 4-70 V" K, L / spectral KI 2 10 80 nulling N Spectral AMBER: V", CP 1-2.5resolution / spectral 3 8.2 40-130 VLTI 4 (8) MIDI: V", V 8-13 / spectral 2 1.8 8-200 CHARA V", CP, Imaging 1-2.5 (spectral) 2/4 (6) 1 50-350 Future facilities LBT V", nulling 1-12 µm 2 8.4 6-23 MROI V", CP, imaging V, NIR 3/6 (10) 1.4 7.5-340

VLTI Summer school 2010 - Inner regions of YSOs revealed by interferometry - F. Malbet Interferometers involved in YSO research

CHARA V", CP, Imaging 1-2.5 (spectral) 2/4 (6)Imaging 1 50-350 Future facilities LBT V", nulling 1-12 µm 2 8.4 6-23 MROI V", CP, imaging V, NIR 3/6 (10) 1.4 7.5-340

VLTI Summer school 2010 - Inner regions of YSOs revealed by interferometry - F. Malbet Interferometry publication OLBIN database

http://olbin.jpl.nasa.gov

# 71 astrophysical results published to date, ~100 objects observed.

VLTI Summer school 2010 - Inner regions of YSOs revealed by interferometry - F. Malbet YSOs + Debris disks observed (1998-2009)

Refereed papers

$ !" !" !" # % ! $ ! ! " &''% &''' ! ! " # "!!! "!!& "!!" "!!( ! "!!# "!!) "!!$ "!!* Interferometers "!!% "!!' IOTA 14% Number of observations 40 PTI KI 15% 36% 30

20 ISI 10 1% CHARA 0 8% !& "& "' ! ! #" 1998 1999 # !" "# #( 2000 ! "& 2001 2002 2003 2004 2005 2006 2007 VLTI 2008 2009 26%

VLTI Summer school 2010 - Inner regions of YSOs revealed by interferometry - F. Malbet INNER DISK PHYSICS

– Disk structures – Constraints on disk structure (T, z,…) – Dust mineralogy – Gas/dust connection

VLTI Summer school 2010 - Inner regions of YSOs revealed by interferometry - F. Malbet UV and IR excesses in Spectral Energy Distribution

Geometrically thin optically thick accretion disk + irradiation = « classical » accretion disk model

Malbet & Bertout (1995, A&AS )

VLTI Summer school 2010 - Inner regions of YSOs revealed by interferometry - F. Malbet 18 Original disk concept

e.g. Malbet & Bertout (1995, A&AS 113, 369)

• Optically thick disk both for inner gas and outer dust • Simple power-law temperature distribution (T " r-0.75, T " r-0.5) • Oblique disk heating ! fits rather well spectral energy distributions (SEDs)

VLTI Summer school 2010 - Inner regions of YSOs revealed by interferometry - F. Malbet Millan-Gabet et al. (2001, ApJ 546, 358) Monnier & Millan-Gabet (2002, ApJ 579, 694 ) Eisner et al. (2003, ApJ 588, 360) Wilkin & Akeson (2003, Ap&SS 286, 145) Eisner et al. (2004, ApJ 613, 1049) Akeson et al. (2004, ApJ 622, 440) Eisner et al. (2005, ApJ 623, 952) Monnier et al. (2005, ApJ 624, 832) Akeson et al. (2005, ApJ 635, 1173)

VLTI Summer school 2010 - Inner regions of YSOs revealed by interferometry - F. Malbet Millan-Gabet et al. (2001, ApJ 546, 358) Monnier & Millan-Gabet (2002, ApJ 579, 694 ) e.g. DullemondEisner et al. et (2003, al. (2001, ApJ ApJ588, 560,360) 957) Wilkin & Akeson (2003, Ap&SS 286, 145) Eisner et al. (2004, ApJ 613, 1049) Akeson et al. (2004, ApJ 622, 440) Eisner et al. (2005, ApJ 623, 952) Monnier et al. (2005, ApJ 624, 832) Akeson et al. (2005, ApJ 635, 1173)

VLTI Summer school 2010 - Inner regions of YSOs revealed by interferometry - F. Malbet Herbigs Ae Be

Millan-Gabet et al. (2001, ApJ 546, 358) Monnier & Millan-Gabet (2002, ApJ 579, 694 ) Eisner et al. (2003, ApJ 588, 360) Wilkin & Akeson (2003, Ap&SS 286, 145) Eisner et al. (2004, ApJ 613, 1049) Akeson et al. (2004, ApJ 622, 440) Eisner et al. (2005, ApJ 623, 952) Monnier et al. (2005, ApJ 624, 832) Akeson et al. (2005, ApJ 635, 1173)

VLTI Summer school 2010 - Inner regions of YSOs revealed by interferometry - F. Malbet Herbigs Ae Be T Tauris

VLTI Summer school 2010 - Inner regions of YSOs revealed by interferometry - F. Malbet Herbigs Ae Be T Tauris Must include ACCRETION luminosity onto star Muzerolle et al. (2003, ApJ 597, 149)

VLTI Summer school 2010 - Inner regions of YSOs revealed by interferometry - F. Malbet Inner region discussion

• Inner rim shapes: how sharp is it? Dust sublimation Isella & Natta (2005, A&A 438, 899)

vs Dust settling & grain growth Tannirkulam et al. (2007, ApJ 661, 374)

• Inner hole? but –optically thick disk beyond the dust sublimation barrier e.g. TTS Akeson et al. (2006, ApJ 635, 1173) –disk halo with 0.15-0.8 optical depths

VLTI Summer school 2010 - Inner regions of YSOs revealed by interferometry - F. Malbet Inner region discussion

• Inner rim shapes: how sharp is it? Dust sublimation Isella & Natta (2005, A&A 438, 899)

vs Dust settling & grain growth Tannirkulam et al. (2007, ApJ 661, 374)

• Inner hole? but –optically thick disk beyond the dust sublimation barrier e.g. TTS Akeson et al. (2006, ApJ 635, 1173) –disk halo with 0.15-0.8 optical depths Vinkovic & Jurkic (2007, ApJ 658..462)

VLTI Summer school 2010 - Inner regions of YSOs revealed by interferometry - F. Malbet The geometry of the inner rim

VLTI Summer school 2010 - Inner regions of YSOs revealed by interferometry - F. Malbet bg= NIR TheThe structure geometry of the inner of rimthe inner rim interferometry How sharp is the inner rim? E. Tatulli Depends on dust grains size and distribution Principles Recombination • If inclined disk: asymmetries (skewness) depending on dust Atmospheric If inclined disk: asymmetries (skewness) depending on dust turbulence characteristicscharacteristics Tannirkulam et al. (2007, ApJ 661, 374) The observables Interferometry Closure phase is a powerful observable to probe such asymmetries & YSOs Context NIR continuum [Tannirkulam et al. 2007] emission Emission lines Imaging State of the art Next steps Perspectives

VLTI Summer school 2010 - Inner regions of YSOs revealed by interferometry - F. Malbet bg= bg= NIR The structure of the inner rim interferometryNIR The structure of the inner rim interferometry How sharpTheis thegeometry inner rim? of the inner rim E. Tatulli How sharp is the inner rim? E. Tatulli Principles Depends on dust grains size and distribution Principles Depends on dust grains size and distribution Recombination AtmosphericRecombination • If inclinedIf inclined disk: disk: asymmetriesasymmetries (skewness) (skewness) depending depending on dust on dust turbulenceAtmospheric If inclined disk: asymmetries (skewness) depending on dust turbulence The observables characteristicscharacteristicscharacteristics Tannirkulam et al. (2007, ApJ 661, 374) The observables InterferometryInterferometry • ClosureClosure phaseClosure phase phaseis aapowerful powerfulis a powerfulobservable observableobservable to probe to probeto such probe such asymmetries asymmetries such & YSOs& YSOs ContextContext Monnier et al. (2006, ApJ 646, 444) [Monnierasymmetries et al. 2006] NIRNIR continuum continuum [Tannirkulam et al. 2007] emissionemission EmissionEmission lines lines ImagingImaging StateState of the of the art art NextNext steps steps PerspectivesPerspectives

VLTI Summer school 2010 - Inner regions of YSOs revealed by interferometry - F. Malbet FU Orionis

Malbet et al. (2005, A&A, 437, 627)

VLTI Summer school 2010 - Inner regions of YSOs revealed by interferometry - F. Malbet FU Orionis

Malbet et al. (2005, A&A, 437, 627)

VLTI Summer school 2010 - Inner regions of YSOs revealed by interferometry - F. Malbet FU Orionis

• The 4 brightest FUors have been observed

• FU Ori well constrained Quanz et al. (2006, ApJ 648, 472) • Others like Z CMa appear more extended: background emission or companion? Millan-Gabet et al. (2006, ApJ 645, L77)

Malbet et al. (2005, A&A, 437, 627) • Recent FUOr: V1647 Ori Ábrahám, Mosoni, Henning et al. (2006, A&A 449, L13)

VLTI Summer school 2010 - Inner regions of YSOs revealed by interferometry - F. Malbet Radial temperature distribution of disks

Irradiated disks (King et al. 1998) T # r -1/2 T # r -3/4 Viscous disk (Lynden-Bell & Pringle 1974) Diameter [mas] Diameter

Wavelength [µm]

Commonly used analytic temperature-power-law disk models (T # r -1/2 or T # r -3/4) cannot describe the measured wavelength-dependence of the apparent size ! Detailed physical modeling required

Kraus et al. (2007, ApJ 676, 490)

VLTI Summer school 2010 - Inner regions of YSOs revealed by interferometry - F. Malbet MWC 147: a full disk model to understand NIR and MIR measurements

-6 Inclination: 60º, Ṁacc = 9$10 Mʘ/yr SED NIR NIR visibilities MIR visibilities

Kraus et al. (2007, ApJ 676, 490)

VLTI Summer school 2010 - Inner regions of YSOs revealed by interferometry - F. Malbet Effect of extended scattered light

• Ring radius fitting can lead to overestimated sizes • Careful modeling must be performed including all sources of radiation

Pinte et al. (2007, ApJ 673, L63)

VLTI Summer school 2010 - Inner regions of YSOs revealed by interferometry - F. Malbet Vertical structure @ 10 microns

Leinert et al. (2004, A&A, 423, 537)

VLTI Summer school 2010 - Inner regions of YSOs revealed by interferometry - F. Malbet Vertical structure @ 10 microns

Flaring

Self-shadowed

Leinert et al. (2004, A&A, 423, 537)

VLTI Summer school 2010 - Inner regions of YSOs revealed by interferometry - F. Malbet Vertical structure @ 10 microns

Flaring

Sizes consistent with flat self-shadowed / flaring disk model SED classification

Self-shadowed

Leinert et al. (2004, A&A, 423, 537)

VLTI Summer school 2010 - Inner regions of YSOs revealed by interferometry - F. Malbet Dust mineralogy in HAeBe

Whole Disk

Van Boekel et al. (2004, Nature, 432, 479)

VLTI Summer school 2010 - Inner regions of YSOs revealed by interferometry - F. Malbet Dust mineralogy in HAeBe

Whole Disk

Inner disk

Van Boekel et al. (2004, Nature, 432, 479)

VLTI Summer school 2010 - Inner regions of YSOs revealed by interferometry - F. Malbet Dust mineralogy in HAeBe

Whole Disk

Inner disk Outer disk

Van Boekel et al. (2004, Nature, 432, 479)

VLTI Summer school 2010 - Inner regions of YSOs revealed by interferometry - F. Malbet … also in T Tauri disks!

Whole disk TW Hya Inner region

Ratzka et al. (2007, A&A 471, 173) ! Inner disks (< 2 AU) have: –larger silicate grains –higher fraction of silicates is crystalline (40-100%)

VLTI Summer school 2010 - Inner regions of YSOs revealed by interferometry - F. Malbet 51 Oph: NIR CO overtone emission Averaged data in NIR

Tatulli, et al. (2008, A&A 489, 1151)

VLTI Summer school 2010 - Inner regions of YSOs revealed by interferometry - F. Malbet 51 Oph: NIR CO overtone emission Averaged data in NIR

All observations fitted by the standard disk model! but it seems not to be physically possible

Tatulli, et al. (2008, A&A 489, 1151)

VLTI Summer school 2010 - Inner regions of YSOs revealed by interferometry - F. Malbet Region of CO emission in 51 Oph

Visibility amplitudes

NIR excess continuum

Tatulli, et al. (2008, A&A 489, 1151)

VLTI Summer school 2010 - Inner regions of YSOs revealed by interferometry - F. Malbet Region of CO emission in 51 Oph

Visibility amplitudes

NIR excess continuum

CO emission 0.3AU size

Tatulli, et al. (2008, A&A 489, 1151)

VLTI Summer school 2010 - Inner regions of YSOs revealed by interferometry - F. Malbet Region of CO emission in 51 Oph

Visibility amplitudes

NIR excess continuum

CO emission 0.3AU size

Visibility phases

Tatulli, et al. (2008, A&A 489, 1151)

VLTI Summer school 2010 - Inner regions of YSOs revealed by interferometry - F. Malbet Region of CO emission in 51 Oph

Visibility amplitudes

NIR excess continuum

CO emission 0.3AU size

Visibility phases

Comprehension of the dust vs gas morphology will increase with better coverage.

Tatulli, et al. (2008, A&A 489, 1151)

VLTI Summer school 2010 - Inner regions of YSOs revealed by interferometry - F. Malbet Transition disks: the case of T Cha Olofsson et al.: VLTI/AMBER observations of the T Tauri2Olofssonetal.:VLTI star T Cha 3 /AMBER observations of the T Tauri star T Cha input variables of this model are the position angle of the disk Photosphere 1.0 30 (PA,0◦,meaningthemajorsemi-axisisorientednorth-south), Scattered stellar light 100 Thermal emission 0.8 ◦ -12 the inclination on the line of sight (i,0 is face-on), and the in- 10 20 band 0.6 K

ner diameter (θ). We ﬁxed the star-to-ring ﬂux ratio to account ) - o 50 2 10 V for H and K band excess derived using a Nextgen synthetic spec- ) 0.4 -2 trum of the star (see Sect. 3.2). The width W of the uniform ring 10-14 0.2

[m] 0 0 is given by W/(θ/2) = 0.18. We performed a grid exploration (W m

v 1.0 ! 2 F for these three parameters, with a χr minimization. The bottom ! 2 0.8 three panels of Fig. A.1 show the reduced χ maps for the three -16 -10 r 10 -50 Closure Phase ( band pairs of free parameters. Only the diameter is determined well H 0.6 - with this approach, while the position angle and inclinationcan- 2 -20 -100 V 0.4 not be strongly constrained, because of the almost unidirectional -18 2 10 0.2 -30 alignment of the (u, v)points.Thefinalbestfit(χr = 0.76) re- ◦+91 ◦+32 +0.52 1 10100 50100 0 1000-50 -100 0 20 40 60 80 36 38 40 42 44 46 48 50 = = = -6 -1 -6 -1 turns PA 67 −56, i 42 −40,andθ 1.82−0.38 mas. Therefore, ! (µm) u [m] B/! (10 rad ) B /! (10 rad ) adustringmodelatatenthofAUorsofromthestariscon- Huélamoet al.: A companion candidate in the gap ofMax the T Cha disk. Fig. 2. Observed and modeled SED of TCha. Photometric mea- sistent with the AMBER visibilities and with the photometric Fig. 1. Left panel:(u, v)coverageoftheobservations.ObservedsquaredvisibilitOlofsson et al. (2011, A&A 528, L6)ies (middle)andCP(right). H and K-bands data surements were gathered from the literature (Brown et al. 2007, excesses in the H-andK-bands. Nonetheless, the above model are represented by squares and circles symbols, respectively. Overplotted are the predictions from our best disk model (empty Lommen et al. 2010, Alcaláet al. 1993 and reference therein). 500 2981 is an oversimplified description of a disk and ignores the disk’s symbols). Dashed and dotted lines in the middle panels are a 2nd-order polynomial fit the modeled visibilities for the extreme values The gray area represents the modeled SED for the extreme val- asymmetry, as well as the SED1992; at other wavelengths. Alcala We develo etp al. 1993; Schisano et al. 2009). If the line 2714 ues on R ,withintheuncertainties(seeSect.3).on Rin, inner disk within the uncertainties (see Sect. 3). 1 arefinedmodelinthenextsection. in, inner• diskTransition disks are disks with lack of MIR is related to accretion episodes, then the average accretion 2448 3. Radiative transfer modeling Table 1. Parametersemission2. of Results the disk model andcompared first analysis to FIR (discovered byWe do not observe strong discrepancies between H and K bands ˙ 9 visibilities, suggesting that the same structure is responsible for 2181 Hipparcos has estimated a 66+rate19 pc distance is forM T Cha,=4 but, as 10Parameter− M /yr. Inner disk Outer disk −12 2.1. VLTI/AMBER observations most of the emission in H and K bands. Finally, the CP are explained in Schisano et al. (2009), proper motions measure- Size distributionSpitzer)⊙ index p -3.5 -3.5 1914 × µ ments by Frink et al. (1998) and TerranegraThe et al. age (1999) of sug- theamin source–amax [ m] TChawasobservedattheVLTI,usingtheAMBERinstrument is variously0.1 – 10 0.01 – estimated 3000 to be be-close to zero, indicative of a centro-symmetric structure atthe M [M ] 1 ×10−9 1.75 ±0.25 × 10−4 0 1648 0 gest that this distance is underestimated. T Cha may in fact dust,disk & that allows the simultaneous combination of three beams in the given spatial resolution of the observations. We first investigate Rin [AU]• Believe 0.13to± 0have.03 disk 7.5±0.5 holes and therefore belong to a star association clustertween located 2–10 at 100 pc, Myr which according to different methods (Fernándezthe possibilitydec [mas] of a companion around T Cha, and then examine 1381 Rout [AU] near-IR with0.17 spatial±0.03 filtering 300 (Petrov et al. 2007). The instrument is the distance that we adopt in this study. The disk inclina- Flaring indexhaving (β) starting1.11 planetary 1.11 formation whether an inner dusty disk as proposed by Brown et al. (2007) ◦ delivers spectrally dispersed interferometric visibilities, closure 1114 Phase (degrees) tion is also uncertain. An inclinationet al. of 75 2008).is commonly used A completeH/R (at 1 AU) phases study (CP0.106 hereafter), of andthe 0.15 differential Cha phases atassociation spectral resolu- can reproduce the data. in the literature. This value comes from v sin i measurements Surface density index (α) -1 -1 −1 • Discoverytions up to 12 000.of Indisk the following, gap between we present K -and0.17H-bands and 847 (54 km.s ,Alcaláetal.1993)indicativeofahighlyinclinedby Torres et al. (2008) provides an average age of 6 Myr,2.2. Geometric models −1 (around 2.2 and 1.65 µm, respectively) observations taken in the object. Schisano et al. (2009) measured v sin i of 37 ±2km.s , 7.5 AU (Olofsson et al. 2011) 581 −1 which could suggest a slightlywhile smaller inclination. da Silva However, etremain al. unchanged (2009) oncelow the spectralUBVRI estimate resolutionbands were mode reproduced a (LR; slightlyR well∼35) with the older 8.2 m Unit ageTo see whether our interferometric datapoints are compatible µ enough. In the end,Telescopes the star is defined (UTs), with coupled the following with the para usem- of adaptiveoptics. T Cha with a companion−500 around TCha, we used a toy model that aims 314 TChadisplaysCOinabsorptioninthe4.7 mfundamentalband 500 0 −500 −5 0 5 (Brown et al. in prep), which requires(between a high inclination 5-10 (∼60– Myr)eters: Teff based=• 5400Discovery K,was log( on observedg) = the3.5, R withof∗ = Lithiuma1 5. 3 diplanetRff&,erentM∗ = baselines1 .by5 M content& ,Huelamo of two VLTI of configu-et the al. to describe a binary system where both stars are individuallyun- ◦ and Av = 1.5. 70 ). In the following, the inclination is therefore consideredas rations (UT1-2-4 and UT1-3-4), during June 11 and 12, 2009. resolved (e.g., Lawson 1999). To findRA a[mas] model that best repro- Baseline (meter) afreeparameterandisfurtherdiscussedinSect.3.2. The disk is known(2011)The to longest have at a depletedbaseline62mas zone length (gap)+/- is ∼ according7mas,130 m corresponding i.e. 6.7AU to a maxi- duces both AMBER visibilities and CP, we performed a χ2 min- Cha members. Finally,to SED modeling a (Brown dynamical et al. 2007). Therefore, evolution two separate study of the Huélamo et al. (2011, A&A 528, L7) mum angular resolution of 1.3 milli-arcseconds (mas). The inte- imization exploration through a 3-dimensional grid: flux ratio 3.1. Model setup zones were defined for the circumstellar material: an inner and η Cha cluster, whichan outer dusty probably structure.gration Given time belongsthe per high frame angular was 50 resolution to ms. Inthe additionof toCha T Cha, a associ- calibra- between the two components (F2/F1), separation (ρ), and po- To model the disk emission and to reproduce simultaneously, the VLTI Summer school 2010 - Inner regions of YSOs revealed by interferometry - 2F. Malbet our observations, thetor outer (HD disk 107145, is overresolved, F5V) was so observed changing toits correct for instrumental sition angle (PA). The reduced χr values are then transformed squared visibilities and CP in Hation,& K bands, as provides well as the SED, ancharacteristics age willofe onlyff 6.7ects. affect Observations Myr the SED. (Ortega For were the performed inner disk et we without al. fringe-trackin 2009).g. Forinto normalized probabilities to perform a statistical (Bayesian) we used the MCFOST radiative transfer code (Pinte et al. 2006). use a mixture of astro silicates (Draine & Lee 1984) and amor- analysis. The top three panels of Fig. A.1 display the reduced χ2 MCFOST calculates syntheticthe observables, purpose such as SEDs of and this paper,Data we was adopt reduced anfollowing age the of standard 7 Myr. procedures de- r phous carbon (Zubkoscribed et al. 1996)in Tatulli with et a al. mass (2007) ratio and of 4:1 Chelli be- et al. (2009), using the maps for the three pairs of free parameters. We derived uncer- monochromatic raytraced images. The computation is done by tween silicate and carbon. We choose to use this composition, propagating packets of photons insideThe the disk. Light distance scattering, to the source,amdlib package, based release on 2.99, the and the Hipparcosyorick interface pro- par-tainties from the probability distributions. The binary parame- because these grains will produce a very weak emission fea- ters for the best model (χ2 = 0.29) are the following: flux ratio absorption, and re-emission processes are included. The geom- µ vided by the Jean-Marie Mariotti Center (http://www.jmmc.fr). r ture at 10 m, which has not been detected in the IRS spectrum of F /F = 0.95+0.05,separationofρ = 1.15+0.57 mas, and po- etry of the disk is defined by severalallax, parameters: is 66the inner pc and 15(Brown pc. et al. 2007).ARaw more From spectral previous reliable visibilities tests with di andfferent CPvalue composi- were extracted of 100 for all the pc frames was 2 1 −0.85 −0.49 ◦+25 outer radii (Rin and Rout,respectively),theindexα for the sur- of each observing file. A selection of 80% of the highest qual- sition angle on the sky of PA = 68 east of north. The un- α ± tions (e.g., less carbon), such features, even a small one, will fill −68 face density (Σ(r) = Σ(r/r0) ),obtained and a disk scale height using H(r) as- properthe dip in the 7-15 motionµmrangeandwillnotproduceagoodfittotheity frames was studies made to minimize (Frink the effect et of al. instrumental 1998;certainty on the flux ratio is rather large, because the squared − 2 2 suming a vertical Gaussian distribution (exp[ z /2H(r) ]), ger- SED. We try using largerjitter grains and non-optimal (≥ 10 µm) to avoid light any injection emission into the instrument.The visibilities (36 values) can also be reproduced for a small-flux mane to a disk at the hydrostatic equilibrium. The disk’s flaring Terranegra et al.features, 1999). without Torres anyAMBER satisfying+VLTI et results al. instrumental on both (2008) visibilit transferies provided and function was calibrated a kine-ratio. However, in that case, CP (12 values) are not close to zero is described by a powerlaw determining the scale height as a −9 β SED. In our model,theusing dust measurements mass of the inner of HD disk 107145, is 10 M after&, correcting for its di- and CP are only matched for higher flux ratio, close to unity. The function of the radius (H(r) = H0(r/r0) ). The dust content is matical distanceand of falls 109 in the optically pcameter for thick (0.3 Tregime.± 0. Cha,1mas,whichwasobtainedusingthegetCalsoft- The near-IR and excess an can- average valuelarge uncertainties on F /F therefore reflect the larger amount described by a differential powerlaw for the grain size distribu- not be reproduced by an optically thin inner disk. However, as 2 1 ∝ −p ware, http://nexsciweb.ipac.caltech.edu/gcWeb/gcWeb.jsp). The of visibility with respect to the number of CP. That the flux ratio tion (dn(a) a da), betweenof the minimum 108 (amin9)andmaximal pc for thethe AMBER whole observations association. only trace the τ = 1surface,further We have adopted the (amax)grainsizes.Theradiativetransfercodecanhandlespa- constraints on the innerH-band disk mass measurements cannot be established. have a lower Forthe signal-to-noise ratio since is about F2/F1 ∼ 1impliesthattheemissionprofileofthecom- tially separated disk zones. ± they have been obtained close to the instrument-limiting mag- panion, if any, on detector (deg) Orientation Companion would be similar to the one of the primary and not latter value for thisouter disk, paper. we use grain opacities of astro silicates, and theto- Time (Hours LST) tal dust mass of thenitude. outer disk, With for about grain sizes 20 spectral up to 3 mm, channels of through the H and colder, as expected in order to match the near-IR excess probed 3.2. Results and discussion −4 Finally, we note[1.75 ± 0.25] that× 10 KMbands, previous& is constrained the total data well works set by (after the millimeter processing) based is made on of 585 radial and from SED modeling (Brown et al. 2007). Besides this consider- To reproduce the stellar photosphere, we adopt a NextGen stel- observations. We choose to use an inclination of 60◦,while75◦ 195 spectrally dispersed visibilities and CP, respectively(aver- ation, SchisanoFig. 1.et al. (2009)Results conclude from from radial the velocityLmea- SAM observations: a companion lar atmosphere model (Allard et al. 1997). All the stellar pa- is often used in the literature. With an inclination higher than surements, that even a low-mass object (0.05 M%)atadistanceof velocity (RV) and◦ directaged imaging for each baseline and and baseline coronographic triplet to provide 36 and tech-12 rameters, except the inclination i,weretakenfromBrownetal. 60 ,theouterdiskhidestheinnerdisk,whichwouldaccountbroad-band visibilities and CP, respectively). 0.1 AUcandidate would have been detected. is detected Additionally, at NaCo 61 narrow-.8 7.4 mas from the central source (2007). The two parameters T ff and R∗ were then varied only to neither for the NIR excess nor for the interferometric measure- eniques have not reported theFigure1 presence presents the final of reducedany (stellar data for TCha: or the ( veryu, v) band imaging observations of T Cha at 2.12 and 1.75 µmruleout± 2 coverage (left panel), the squared visibilities in H and K broad any stellarwith companion a flux with a ratio mass higher of than 0.92 0.15 M% 0.20between % in L’. Upper Left panel: χ as low-mass) companion around T Cha (Schisano et al. 2009;∼ bands as a function of the spatial frequencies (gray and black, re- 2and10AU(Vicenteetal.submitted).Aclosecompanionina function of the position± of companion candidate (degrees of Chauvin et al. 2010, Vicentespectively, et middle al. panel), (submitted)). and the CP (right panel). SAM The squared ob-the first tenths of AU around T Cha is therefore not a satisfying visibilities show that the circumstellar emission around T Cha solution.freedom = 314). The black circle correspond to the imaging res- servations allow to fill theis marginally gap resolved: between the squared between visibilities decrease RVwith and in- We now consider the case of a uniform ring aimed at repre- creasing spatial frequency, but remain at a high level (V2 ≥ 0.3). sentingolution the inner disk of edge the (e.g., telescope Eisner et al. 2004). (1 The.22 thλree/D). Upper Right panel: best fit of direct imaging observations. the model of the companion (solid line) overplotted on the deconvolved phase. Lower panel: Orientation of companion detections made using each of the 9 individual data files 3. Observations and Data Reduction plotted in raw detector coordinates. Spurious structures The observations presented here were obtained with NAOS- should appear fixed, while real features will rotate with CONICA (NACO), the AO system at the Very Large the sky, as illustrated by the overplotted solid line de- Telescope (VLT), and SAM (Tuthill et al. 2010) in two dif- picting the expected orientation for an object with sky position angle of 78 degrees. ferent campaigns. L’ observations were obtained on March 2010 under excellent atmospheric conditions (average co- herent time of τ0=8 ms, and average seeing of 0. 6), while Ks band data were obtained on July 2010 under moderate 4. Main Results atmospheric conditions (τ0=4 ms, and seeing of 1. 0). On March 2010, T Cha was observed with the 4.1. L detection L27 objective, the 7-hole mask and the L filter (λc= 3.80µm, ∆λ= 0.62 µm). The target and a calibrator star The three free parameters of the fit are: the flux ratio, the (HD 102260) were observed during 10 minutes each. We re- separation and the position angle of the companion candi- peated the sequence star+calibrator 9 times, integrating a date. The upper left panel of Fig. 1 depicts the minimum total of 48 minutes on-source. The observational procedure χ2 as a function of position angle and separation. For an included a dithering pattern that placed the target in the arbitrary fit, χ2 is high (reduced χ2 of 9), however the ≈ four quadrants of the detector. We acquired datacubes of map shows a clear minima for a companion to the west of 100 frames of 0.4 sec integration time in each offset posi- the star. The phase corresponding to the best fit model is tion. The plate scale, 27.10 0.10 mas/pix, and True shown on the right panel of the same figure. It consists of a North orientation of the± detector, -0.48 0.25 de- sinusoidal curve with a specific angular direction, and a pe- grees, were derived using the astrometric± calibrator riod of half the resolution of the 8.2 meter telescope. On the θ Ori1 C observed on April 2010. same figure are plotted phases deconvolved from the mea- In the case of the Ks-band data, we used the S27 ob- sured closure phases and projected onto the orientation of jective and the same strategy but integrating in datacubes the best-fit binary. of 100 frames with 0.5 sec of individual exposure time. We The best-fitting companion parameters for the L-band spent a total of 20 minutes on-source, and we used two data are separation of 61.8 7.4 mas, position angle 78.5 stars, HD 102260 and HD 101251, as calibrators. 1.2 degrees and a fractional± flux relative to the central ob-± All data were reduced using a custom pipeline detailed ject of 0.92 0.20%. To confirm the validity of the de- in Lacour et al. (in preparation). In brief, each frame is tection, each± one of the nine star+calibrator data pairs flat-fielded, dark-subtracted, and bad-pixel corrected. The was also analyzed separately. As observations were taken complex amplitudes of each one of the 7 6/2 = 21 fringe in ‘pupil tracking mode’, all optical and electronic aberra- spatial frequencies are then used to calculate× the bispec- tions should remain at frozen orientation on the detector, trum, from which the argument is taken to derive the clo- while real structure on the sky will rotate with the azimuth sure phase. Lastly, the closure phases are fit to a model of pointing of the telescope (close to the sidereal rate). This a binary source. expected rotation of the detection is illustrated in

2 HYDROGEN LINE EMISSION

– Outflowing winds? – Magnetospheres?

VLTI Summer school 2010 - Inner regions of YSOs revealed by interferometry - F. Malbet Disk and wind spatially are spectrally resolved in MWC 297

Br%

Continuum

VLTI Summer school 2010 - Inner regions of YSOs revealed by interferometry - F. Malbet Disk and wind spatially are spectrally resolved in MWC 297

Br%

Continuum

Star hidden

hidden

Malbet et al. (2007, A&A 464, 43)

VLTI Summer school 2010 - Inner regions of YSOs revealed by interferometry - F. Malbet Disk and wind spatially are spectrally resolved in MWC 297

Br%

Continuum

Disk Disk

Star hidden

hidden

Disk Disk Malbet et al. (2007, A&A 464, 43)

VLTI Summer school 2010 - Inner regions of YSOs revealed by interferometry - F. Malbet Disk and wind spatially are spectrally resolved in MWC 297

Br% Outflowing wind

Continuum

Disk Disk

Star hidden

hidden

Wind Outflowing wind

Disk Disk Malbet et al. (2007, A&A 464, 43)

VLTI Summer school 2010 - Inner regions of YSOs revealed by interferometry - F. Malbet Disk and wind spatially are spectrally resolved in MWC 297

Br% Outflowing wind

Continuum

Disk Disk

Star hidden

hidden

Wind Outflowing wind

Disk Disk Malbet et al. (2007, A&A 464, 43) # Investigate the link between disk, star and wind

VLTI Summer school 2010 - Inner regions of YSOs revealed by interferometry - F. Malbet G. Weigelt et al.: Spectro-interferometry of MWC 297 with VLTI/AMBER G. Weigelt et al.: Spectro-interferometry of MWC 297 with VLTI/AMBER

0 0 -20 25 -20 25 0 0 Y / R Y / R -25 G. Weigelt et al.: Spectro-interferometry ofG. MWC Weigelt-25 297 et with al.: VL Spectro-interferometryTI/AMBER of MWC 297 with VLTI/AMBER

-25 0 25 -25 0 25 0 0 -10 25 10 25 -10 10 -20 25 -20 25 0 0 Y / R 0 Y / R 0 Y / R / Y Y / R / Y -25 -25 -25 G. Weigelt et al.: Spectro-interferometryG. WeigeltG. et Weigelt al.: Spectro-interferometry of MWCet al.:-25 Spectro-interferometry 297G. with Weigelt VLTI/AMBER of et MWC al.: Spectro-interferometry 297 of MWC with VL 297TI/AMBER with VL ofTI/AMBER MWC 297 with VLTI/AMBER

-25 0 25 G. Weigelt et al.:-25 Spectro-interferometry 0 25 of MWC 297 with VLTI/AMBER 0 0 0 0 TableFig. 2. Geometric 6. Normalized model fit radii of MWC 297 Br in theγ continuumline profiles and the center of of the Br theγ emission disk-wind line. model Fig. 6. Normalized Brγ line profiles of the disk-wind model -20 -20 25-10-20 20 25 1025 -10 25 -20 10 20 -20 25 25 -20 25 25 ◦ ◦ 5 fromWavelength Table componentC1 A.1 at thediameterofC1 inclination componentC2 anglesdiameterofC2i fluxratio=10 (dotted 5 from Table A.1 at the inclination angles i =10 (dotted 0 0 0 range◦ (dominant compact core)◦ (halo)◦ (C1/C2) ◦ ◦ ◦ line), 20a (solid),0 40 ±(dashed), and 60c (dashed-dotted; line), 20 (solid), 40 (dashed), and 60 (dashed-dotted;

Y / R > 0 Y / R Continuum Gaussian0 4.63 0.21 mas Gaussian ∼ 22 mas 4.6 Y / R 0 Y / R 0 a b ± > c

Continuum Ring (20%) R / Y 4.47 0.20 mas Gaussian ∼ 19 mas 3.6 Y / R / Y Y / R / Y

Y / R / Y d -25 -25 highBrγ center spectral Gaussian resolution, 12.6 ± 0.75 not mas degraded-- to the - spectral reso- high spectral resolution, not degraded to the spectral reso- -25 G. Weigelt et al.: Spectro-interferometryG. Weigelt-25-25 et al.: Spectro-interferometry of MWC 297 with VL ofTI/AMBER MWC 297 with VLTI/AMBER -25 -25 G. Weigelt et al.: Spectro-interferometrya G. Weigelt of MWC et al.: 297 Spectro-interferometry withb VLTI/AMBER of MWC 297 with VLTI/AMBER Noteslution – -25average ofover continuum AMBER). wavelengths between 2.147 and 2.194 µm(exceptthelineregion). Inner radius of a ring lution of AMBER). c model with a ring width of 20% of the inner ring radius. Due to the lack of data at small spatial frequencies, the size of the extended halo component is not well constrained by our measurements. Therefore, we can only give a rough estimate: d -25 022(19) 25-25± 5masorlarger,butsmallerthantheAMBERATFOVof250mas. 0 25 Continuum-compensated diameter (see text). 12000 G. Weigelt et al.: Spectro-interferometry of MWC-25 297 with VL 0TI/AMBER 25 12000 -25 0 25 0 0 0 0 Table 3. RangeFig. of10 parameter 6. Normalized variations Brγ line for our proﬁles continuum-30 of theFig. disk-wind 6. Normalized modelTable Br 3.γ lineRange proﬁles of parameter of the disk-wind variations model for our continuum- -30 -10 -20 3025 25-10 10 -1020 The25 continuum-compensated-202510-10 visibilities required-30 for the 20 10 25-20-20velocity (LSR) [km/s] velocity (LSR) [km/s] -20 -20velocity (LSR) [km/s] 25 size determination of the line-emitting region25 discussed ◦ ◦ -100 0 100 25 25 -100 0 100 25-100 0 100 25 5.0 5.0 5.0 abovedisk were calculated plus25 in disk-wind the following5 way. from Within model Table the calculations A.1 at the inclination angles5 fromi =10 Table(dotted A.1disk at plus the inclination disk-wind model angles calculationsi =10 (dotted PA 65 deg PA 300 deg 4.0 4.0 4.0 0 Magneto-wavelength0 region of Brγ, the measured visibility has◦ ◦ ◦ ◦ ◦ ◦ 0 3.0 3.0 3.0 0 10000 two constituents, a pure0 line-emittingline), component0 20 and a (solid), 40 (dashed), and 60line),(dashed-dotted;10000 20 (solid), 40 (dashed), and 60 (dashed-dotted; Y / R Y / R Flux Flux 0 Flux 0 0 0 Y / R 0 Y / R continuum-emitting one, the second of which includes emis- 2.0 2.0 2.0 0 Y / R / Y Y / R / Y Y / R / Y

Y / R / Y G. Weigelt et al.: Spectro-interferometry of MWC 297 with VLTI/AMBER Y / R / Y Y / R / Y

Y / R / Y sion of both the circumstellar environment and the unre- 1.0 1.0 1.0 ParametersR / Y high spectralRange resolution, not degradedModel 5 tohigh the spectral resolution,reso-Parameters not degraded toRange the spectral reso- Model 5 -25 solved-25 star. The emission line visibility VBrγ can be written 1.0 1.0 1.0 -25 -25 -25 centrifugally driven-25 0.9 -25 -250.9 0.9 -25 -25 as (WeigeltG.G.-25 Weigeltet Weigelt al. 2007) et et al.: al.: Spectro-interferometry Spectro-interferometry of of MWC MWC 297 297 with with VL VLTI/AMBERTI/AMBER 0.8 0.8 0.8 Disk:-25 lution of AMBER). lution of AMBER).Disk: 0.7 0.7 0.7 G. Weigelt et al.: Spectro-interferometry of MWC 297 with VLTI/AMBER 0.6 0.6 0.6 Fig. 4. Sketch of the MWC&'() 297 disk-wind model adopted 0.5 0.5 0.5 Rin 0.25–3 AU (8.8–105 R∗) 0.3 AU (10.5 R∗) Rin 0.25–3 AU (8.8–105 R∗) 0.3 AU (10.5 R∗) Visibility 0.4 Visibility 0.4 Visibility 0.4 FBrγVBrγ = in this paper. 0.3 0.3 0.3 -258000 0 25 12000-25 0-25 25 0 25 * 12000 8000 0.2 0.2 0.2 -25 0 25 disk wind2 2 = |F V | + |F V | − 2 F V F V · cosΦ, (1) 0.1 0.1 0.1 tot tot outc c tot tot c c ∗ ∗ Rout 1–5 AU (35–175 R∗) 3AU(105R∗) R 00 1–5 AU (35–175 R ) 3AU(105R ) 0.0 0.0 0.0 160 o 160 160 ! PBL 14.0m, PA 68 Fig. 6. Fig. 6. Normalized Brγ line profiles of the disk-wind model o 40 Fig. 6.Table-40 3. Range of parameterNormalized variations40 Brγ lineTable for our profiles 3. continuum- of the disk-wind model PBL 28.0m, PA 68 Range of parameter variations for our continuum- -40120 120 120 Normalized Brγ line profiles of the disk-windFig. model 6. o 30 Normalized Brγ line profiles of the disk-wind model -20 -20 where V 10(F ) denotes020 the measured total20 visibility (flux)10 30 %# PBL 42.1m, PA 68 -20 -30 -10 20 -10 10tot tot -30s 10 s ∗ ∗ 80 80 -10 80 2525 -10R -20 0.85–1.25 AU (30–44 R∗)20 0.9 AU (31.5 R∗) R 0.85–1.25 AU (30–44 R ) 0.9 AU (31.5 R ) 25 -20-20 25 - ◦ ◦ ◦ 40 25 2540 25 40 25 25 25 in the Br25γ line, Vc (25Fc) is the measured visibility (flux) in topology of the magnetic field in the disk: magnetic field ◦ 0 0 0 25 5 from Table A.1 at the inclination angles i =10 (dotted 25 25 disk plus"# disk-wind5 from model Table calculations A.1 at thedisk inclination plus disk-wind angles i =10 model(dotted calculations the continuum, and Φ describes5 from the measured Table wavelength-− A.1− at the inclination%$ angles− i =105(dotted from Table A.1 at the inclination− angles− i =10 (dotted− -40 -20-40 -40 lines in the wind-launching region are inclined with re- 1 1 "$ α 0.4– 0.75 ◦ 0.5 α 0.4– 0.75 0.5 Diff. Phase [deg] -80 Diff. Phase [deg] -80 Diff. Phase [deg] -80 differential phase within the Brγ line. ¿From◦ the AMBER spect to the◦ disk plane by less◦ than 60 (Blandford◦ &◦ ◦ ◦ ◦ ◦ ◦◦ ◦ -120 -120 -120 00 spectrum (see Fig. 1), we find Ftot =4.5andFBrγ =3.5×Fc line), 20 line),(solid), 20 40(solid),(dashed), 40 (dashed), and 60 and(dashed-dotted; 60 (dashed-dotted; 0 − − − 0 -160 -1600 -160 0 60000 100000 2 line), 20 (solid),− Payne− 401982). The(dashed), radiation pressure of and10000 hot stars− proba-60 (dashed-dotted;6000line), 20 (solid),α2 40 (dashed),0 and.34– 600.4(dashed-dotted; 0.33 0 Y / R Y / R α 0.34– 0.4 0.33 160 160 0 160 0 at the emission line center.0 Y / R / Y Y / R / Y 0 0 0 bly changes the trajectories+ of the gas streams, leading to Y / R / Y Y / R / Y Y / R / Y

120 120 120 R / Y Y / R / Y Y / R / Y R / Y We derivedR / Y a large continuum-compensated radius 80 80 80 R / Y Y / R / Y Y / R in high spectralParameters resolution,a flatter diskhigh wind (de not Kool spectral & degradedRange Begelmanhigh 1995; resolution, spectral Drew to et al. the resolution, spectral notParameters degradedModel reso- not 5 degraded to theT spectralin toRange the reso- spectral1400–2000 reso- Model K 5 1800 K 40 40 40 T 1400–2000 K 1800 K high spectral resolution, not degraded to the spectral reso- -25-25 (HWHM) of 6.3 ± 0.4mas(1.6 ± 0.1 AU) for the line- 1998; Everett et al., 2001). Figure 4 shows a sketch of the -25 0 -250 0 -25 -25 emitting region at the peak wavelength of the-25 emission line, -* -40 -25 -40 -25 -40 -25 -25 -25 -25 disk-wind model that we consider.lution It assumes that of both AMBER). an -80 -80 -25 -80 -25 Disk: lution of AMBER). Disk: Closure Phase [deg] Closure Phase [deg] Closure Phase [deg] which is ∼3 times larger thanlution the 0.56 AU of HWHM AMBER). radius of lution of AMBER). -120 -120 -120 accretion disk and a disk wind contribute to the observed -160 -160 -160 the compact continuum-emitting Gaussian component.in The emission. in 2.163 2.164 2.165 2.166 2.167 2.168 2.169 2.163 2.164 2.165 2.166 2.167 2.168 2.169 2.163 2.164 2.165 2.166 2.167 2.168 2.169 smallerDisk line-continuum wind: size ratio obtained by KrausR et al. 0.25–3 AU (8.8–105 R∗) 0.3 AUR (10.5 R∗0.25–3)Disk wind:AU (8.8–105 R∗) 0.3 AU (10.5 R∗) Wavelength [µm] Wavelength [µm] Wavelength [µm] 12000 12000 4000 4000 -25-2512000(2008a)Fig. 0 can0 A.1.8000 be 25explained 25Sketch by of the the lower geometry spectral resolution of the of disk-wind model adopted. 8000 12000 R = 1500, whichω leads1 to an averaging0.1–3 over differentR AUout diam- (3.5–1054.1. Geometry and1–5R kinematics∗) AU of the (35–175 disk0.5 wind AUR (17.5∗) R∗) 3AU(105Rout R∗) 1–5ω AU1 (35–1750.1–3R∗) AU (3.5–1053AU(105R∗)R∗) 0.5 AU (17.5 R∗) Fig.-50 7. ComparisonWeigelt of the observationset al. (left(2011, panel; see Fig. 1) with the corresponding model50 quantities of model 5 for -25 0 25 -50 Table 3.50 two different PAs of the model image (middle and right-40 panel; the model quantities are calculated for AMBER’s-30 spectral -3040 eters corresponding-40 to30 diffTableerent wavelengths. 3.30 FigureRange 3Fig. also Fig. of 6. parameter 6.Table40Normalized 3. variationsRangeFig. Brγ ofline 6. forRangeparameterFig.Normalized our profiles of continuum- 6. parameterTable variations of the Br 3.γ disk-windline variationsRange for profiles our of model continuum- parameter for of theour disk-wind continuum- variations model for our continuum- resolution of 12 000, as discussed in-30 Sect. A.3 and Fig. A.3). The middle and right panels-20 show the dependence of the 30-10-10 -20 20 -30 1010 20 Normalized30 Brγ line profiles ofNormalized the disk-wind Br modelγ line profiles of the disk-wind model 25 A&A 527, A103)25 -20 25 25 shows20 that-20 theω sizeN of the line-emitting0.5–5.7 region25 is smallerR AUs in (17.5–200To model20 the0.85–1.25 extendedR disk∗) wind, AU we follow1AU(35 (30–44 an approachR∗R)∗) 0.9 AURs (31.5 R0.85–1.25∗) ◦ ωN AU (30–440.5–5.7R∗ AU) ◦ (17.5–2000.9 AU (31.5R∗)R∗) 1AU(35R∗) interferometric observables25 (spectrum, visibilities, wavelength-differential phases,25 and closure phases) of our best-fit disk- 25 25 25 ! used by Shlosman &! Vitello (1993) in their studies of disk ◦ ◦ wind model 5 (disk-wind emitting region, continuum accretion disk,25 plus central star; Tables 3 and A.1) on the wavelength2525 -10 the25 wings of the Brγ emission line than10 at the line center. 5 fromdisk Table plus A.1disk disk-wind at5 plus the from disk-windinclination model Table calculations A.1 model angles at the calculationsi inclination=10 (dotted angles i =10 (dotted ◦ disk plus disk-wind model− calculations− disk− plus disk-wind model calculations across the Brγ line for an inclination angle of i =20, clock-wise motion of the disk wind, and for two different PAs of This wavelength dependence is one of several observationalα51 − fromwinds in Table cataclysmic variables.MWC A.10.4– This at approach297 the0.75 provides inclinationat a 5 from angles Table1 0i.5 A.1=10 at(dotted the− inclination− angles−i =10−(dotted ◦ ◦ 25 γ 1– 5 2 α γ0.4– 0.75 1– 5 0.5 2 the projected disk polar axis on the sky: 65 (middle) and 300 (right). Several other PAs that are also approximately results that have to be explained by models, such as the simple parametrization◦ ◦ of a disk wind, which◦ has◦ geomet- ◦ ◦ ◦ ◦ ◦ ◦ ◦ 0in agreement with the observations are discussed0 in the text and Figs. A.4 and A.5. The detailed0 dependence of the 0 100000 line),rical and10000 kinematical 20 properties(solid),− similar− to 40line), those used(dashed), by 20 (solid), and2000− 6040 (dashed-dotted;(dashed), and 60 (dashed-dotted; interferometric observables on0 the PA is presented in Figs. A.4 and A.5. 0 2000 0 10000disk-wind model60000 presented0 in the next section. αline),2 20 (solid),0.34– 40 60000(dashed),.4 10000line), and 20 602(solid),0.(dashed-dotted;33 40 (dashed),−f − and 60 0.5–3(dashed-dotted;− 0.5–3 f R / Y 0.5–3 0.5–3 α 0.34– 0.4 0.33 Y / R / Y 0 00 0 Y / R / Y Y / R / Y R~120000 Y / R / Y Y / R / Y Kurosawa et al. (2006) for T Tauri stars (see Fig. A.1 in Y / R / Y Y / R / Y R / Y Y / R / Y Y / R / Y Y / R / Y R / Y Y / R / Y the Appendix).Parameters Here, we use a similarParameters description of the Range Range Model 5 Model 5 What type of matter radiates in this compact inner 2008). The formation of this gap could be a result of bina-0 ParametersTinhighhigh spectral spectralRange resolution,1400–2000 resolution,high K not notspectral degradedhigh degradedModel spectral resolution,T to1800Parameters 5in the to resolution,K the spectral not spectral degraded reso-1400–2000β not reso- degraded toRange the K spectral to the0.3–2 spectral reso-1800Model Kreso- 5 1 disk inside the dust sublimation radius? We assume that rity (Artymowicz & Lubow 1994). Another possibility for β 0.3–2 1 -25 -25 -25 R / Y -25 4. Comparison-25 of the observations with-25 a model of kinematical parameters of the wind where according to the ◦ ◦ ◦ it is a mixture of the warm,-25 mostly neutral molecular gas gap-25 formation-25 is the interaction of stellar radiation-25-25 and -25 -25 -25 ◦conservation◦ of angularDisk: momentum, the tangentialDisk: velocity◦ plus refractory (e.g. graphite) grains. Similar properties of wind with the inner disk. The strong radiation pressure a magneto-centrifugallyθ1 drivenDisk: disk wind lution10lution–80 of AMBER). of AMBER). lution80 oflution AMBER). of AMBER).Disk: θ1 10 –80 80 the inner compact material were found by Benisty et al. and the stellar wind can blow away the disk atmosphere-25 u decreases along a streamline from the Keplerian velocity (2010) for the HAeBe star HD 163296. Interaction of the (Drew et al. 1998). As a result, the disk can partially or We now present a disk-wind model andin compare−9 the ob- −at6 the base of streamline.−1 R Thein radial velocity−R7v0.25–3inincreases AU−1 0.25–3 (8.8–105 AUR (8.8–105∗) 0.3R AU∗) (10.50.3R AU∗) (10.5−9 R∗−) 6 −1 −7 −1 dust grains with the stellar radiation possibly plays an im- fully dissipate in the vicinity of the star. ˙ R Disk wind:0.25–3 AU (8.8–105 R∗) 0.3 AUDisk (10.5 wind:RR∗)in 0.25–3˙ w AU (8.8–105 R∗) 0.3 AU (10.5 R∗) served spectrum,M visibilities,w and phases of10 MWC 297–10 with alongM8000 the$ streamlineyr from the initial value10v0 toMv∞ $(seeyr M 10 –10 M$yr 10 M$ yr portant role in the acceleration of dust and gas and the 0 8000 4000 1200012000 8000 12000 120004000 8000 0 formation of the disk wind. 4) The critical rotation velocity v = (2GM∗/(3R∗) the model predictions. In this disk-wind model, mass ejec- Eqs. A.1 and A.2 and Fig. A.2 in Appendix A),out where v crit Rout 1 1–5 AU (35–175Rout R∗R) 1–50∗ AU3AU(105 (35–1751–5 AU1RR∗ (35–175)∗)out ∗ R3AU(105∗) 3AU(105R∗) R∗) ∗ (Maeder&Meynet 2000) of MWC 297! would be tion is driven by magneto-centrifugal forces. The theoryω of and v∞ are the0.1–3 model parameters. AU (3.5–105 R ) 0.5 AUω (17.5R R 0.1–3) AU1–5 (3.5–105 AU (35–175R∗)R ) 0.5 AU3AU(105 (17.5 R∗)R ) -60 -50 −1 50 -60 50 ∗ $ ∗ $ 40 60 As shown above, an inner gap in-40 the disk of MWC 297 450 km s for R =6R and60M =10-40M ,asadopted 40 -40 these winds30 was-50 originally40 proposed by Blandford & Payne30TableTable 3. 3.RangeRangeFig.40Fig. of parameter of 6. 6. parameterTableNormalizedNormalized 3. variationsTableRange variations Br Br 3.γγ oflineline forRange parameter for our profiles profiles our continuum-of parameter continuum- of of variations the the disk-wind disk-wind variations for our model model continuum- for our continuum- is needed to explain the observations. This gap region may in our paper.-30 For our-30 determined inclination angle i = -20-20 -30 30 -30 -40 2020 The local30 mass-loss rates per unit area of theR disk,s m ˙ w, 0.85–1.25 AU∗ (30–44 R∗) 0.9 AU∗ (31.5 R∗) 25 25 ◦ 25 25 (1982) and25 has been elaboratedR bys many25 ω authors,N 0.85–1.250.5–5.7 AUR AU (30–44 (17.5–200R0.85–1.25∗) R0.9∗) AU AU1AU(35 (31.5 (30–44N R∗R)sR)∗) 0.9 AU (31.5 R ) ∗ ∗ ∗ ∗ be filled with material that25 is either fully transparent or 20 ± 10 , the25 maximum value of vrot sin i would be 150 ± 25 25 Fig. 6. ω R 0.5–5.70.85–1.25 AU (17.5–200 AU◦◦ (30–44R ) R )1AU(350.9 AUR (31.5) R ) − is a function of the cylindric radius ω. To describeNormalizedm ˙ w,we Brγ line profiles of the disk-wind model 251 2525 25 semi-transparent in the infrared (e.g., Tannirkulam et al. 80 km s for vrot

Y / R / Y α 0.34– 0.4 0.33 Y / R / Y f 0.5–3 0.5–3 Y / R / Y Y / R / Y Y / R / Y R / Y the disk. R / Y Y / R / Y Y / R / Y R / Y Y / R / Y 0 the observations: line), 20 (solid), 40 (dashed), andthe observations: 60 (dashed-dotted; Tin β ParametersParameters1400–2000Thighinhigh0.3–2 K spectral spectralTRangeParametersin Range resolution, resolution,1400–2000Parameters18001400–2000 K K not not1T degraded degradedinModelRangeModel K 5 to toRange1800 5 the the1400–2000 K spectral spectral1800Model K reso- reso- 5 Model 5 1800 K -25 -25 -25 R / Y -25 -25 5 β 0.3–2 1 -25 -25 -25 -25-25 -25 -25 -25 -25 high◦ spectral◦ resolution, not◦ degraded1) to To the◦ select spectral◦ the reso-most appropriate◦ the disk parameters, where1)d is To the distance select between the point mostθ1S and appropriate theDisk: star.Disk: The κ islutionlution the the10 integrated–80 disk of of AMBER). parameters,AMBER). lineDisk: opacity in theDisk: consideredθ180 spectral 10 –80 80 -25 total mass-loss rate is then line,9 τ(ν, r)6 is the line optical1 depth at point r7and for the 1 we calculate the continuum˙ in intensityRinDisk− lution wind: of0.25–3−Disk the of AMBER). wind: disk, AU−R (8.8–105inId,atinR0.25–3−∗) 0.3 AU− AU (8.8–105we (10.5 calculate−9 RR∗∗−))6 the0.3∗− continuum AU1 (10.5 R∗− intensity)7 ∗−1 of the disk, Id,at N Disk wind: w R 0.25–3 AU (8.8–105 RR∗) ˙Diskw0.3 wind: AU0.25–3 (10.5 AUR∗ (8.8–105) R ) 0.3 AU (10.5 R ) 4000 0 θ 80008000 4000M 12000120004000 800010 frequency–108000ν inM0 the$yr direction4000 to an observer,10M andMn$i isyr the 10 –10 M$yr 10 M$ yr ˙ ˙ out 1 out -100 100 wavelengthsMw = M50w(θ) sin nearθdθ.ω150 Brγ-100and120000.1–3R compareoutR(A.4) AU (3.5–105numberω it1 1–5 with density1–5 AUR the of∗ω AU) atoms0.1–3 (35–175100 stellarR (35–175 in0.5 the AUi-th AUR in-0.1–3 (3.5–105∗ stateR)R (17.5out∗) AU1–5R3AU(105R∗ (3.5–105ω AU)∗1)3AU(105 (35–1751–5wavelengths0.5 AURR∗ AU0.1–3)∗ (35–175)RR∗∗ (17.5))0.5 AU nearR AUR (3.5–105∗∗3AU(105)) (17.5 Brγ andR3AU(105∗)R compare∗) 0.5 AUR∗ it) (17.5 withR∗ the) stellar in- -60 -50 50 -50 ! 1 50 -50 -40-40 60 -40 40 40-40θ -60 -50 3030 40 40 60 TableTable 3. 3. RangeRange of of parameter parameter variations variations for for our our continuum- continuum- 25 25 25 -3025-30 25 N 25 s Rs ωN 0.85–1.25zmax ω0.5–5.7N AURs AU (30–44"0.5–5.7 (17.5–200s0.85–1.25R∗ AU) (17.5–200R0.9∗) AU AU (30–44 (31.51AU(35R∗)RR∗∗))1AU(35R∗)0.9 AUR (31.5∗) R∗) 25 25 25 tensity25 I25∗. A goodω solution0.5–5.7R has AU to reproduce (17.5–2000.85–1.25Table 3.R the AU"∗) observed (30–44v1AU(35z(r ) R" R∗) R0.9∗)ωN AU0.85–1.25 (31.5tensity0.5–5.7R AU∗)I∗ (30–44. AU A good (17.5–200R∗) solution0.9R∗ AU) has (31.51AU(35 to reproduceR∗) R∗) the observed 25 2525 -30 In this case, we can write30 the continuity equation in its τ(ν, r)= κ(r )Rangeφ(ν − ν of)dz parameter, (A.8) variations for our continuum- 25 usual form α1 − diskdisk plus plus− −γ disk-wind disk-wind0.−4–α10−0c .−75 model model− calculations calculations−1–0 5.4–− −−0−0..755 − − 2 −0.5 − ratio Id/I∗,theobservedγ Kα–band1 1– flux,γdisk 5 ! andz plus0 the.4– disk-wind observed0.75 1– modelα2 51 calculationsγ 0ratio.5 0.I4–d/I2∗0,theobserved.75 1– 5 K–band0.5 flux,2 and the observed 0 0 0 0 0 2 ˙ proach0 to find2000 models that can approximatelyproach to reproduce find models that can approximately reproduce 0 0 0 00 0 20004πρ0 (l, θ)v(l, θ)l0 = Mw6000(θ). 2000 1000010000 (A.5)2 6000 6000− −f 2000−2 − 0.5–3− " − −−− − 0.5–3− − Y / R / Y 6000 α f 0.34–α 00.5–3.4 2 0.34– 0.033.4 0.5–3 0.33 Y / R / Y Y / R / Y 2 Y / R / Y Y / / R Y α r 0.34– 0.4 0.33 Y / / R Y f 0.5–3where φ is the profile function normalized0.5–3 to unity, is

Y / R / Y α 0.34– 0.4 0.33 Y / R / Y R / Y f 0.5–3 0.5–3 Y / R / Y

Y / R / Y visibilities. Y / R / Y R / Y Y / R / Y visibilities. 0 the observations:10000 the" observations:" Using this equation, one can calculate the distributioninTin of theβ vectorParametersParameters with coordinates1400–2000β Tin (x, y, z0.3–2 K), andinz RangeRangechanges0.3–21400–2000 from 1800 K K 1 ModelModel1 5 51800 K Y / R / Y -25 β T 0.3–2 1400–2000 K T1 β 1800 K1400–2000 K0.3–2 1800 K 1 -25 -25 -25 -25 -25 -25 -25-25 -25 -25the-25 number2) We-25 density determine at each point-25 in the the disk velocity wind. andthe currentParameters density value z distributionin Eq. A.6 to the◦ outer◦ boundaryRange◦ value◦ 2) We◦ determineModel 5◦ the velocity and density distribution 1) To select the◦ most1 ◦ appropriateDisk:Disk:θ1 the disk1)◦ To parameters,10 select–80 the most appropriate◦ 80◦ the disk parameters,◦ -25 θ1 10 z–80maxθ . 10 –8080 θ1 80 10 –80 80 along each streamline corresponding9 6 toThe the opticalDisk: wind depth1 of thegeometry− disk9 wind−7 in6 the− line9 frequencies−11−6 along−1 − each7 streamline−−17 corresponding−1 to the wind geometry VLTI Summer school 2010 - Inner regions ofA.2. YSOs Calculation revealed of theby interferometry modelwe intensity calculate distributions- F. MalbetDisk the− wind: continuum− ˙ w RR−inin intensity˙Diskw wind:0.25–30.25–3 of−we the calculateAU AU disk, (8.8–105− (8.8–105Id the,atRR∗∗ continuum)) 0.30.3−9 AU AU intensity− (10.5 (10.56 RR∗∗−))1 of the disk,−7 Id,at−1 4000M˙ w 0 Disk100 wind:–104000MM$yr M 10 10–10DiskM10M wind:$$yryr–10r M˙ wM$yr 10 10M$ yr–1010 MM$$yryr 10 M$ yr -25 0 25 -25 0 25 0 and kinematics4000 parameters-2580008000 0 (see 25 Sect.in the 4-25 direction and4000R 0in Appendix toward 250 the star0.25–3 is A).τ∗(ν)= AUτ(ν, (8.8–105=0).We andR∗) kinematics0.3 AU (10.5 parametersR∗) (see Sect. 4 and Appendix A). wavelengths nearω1 Brassumeγ and that0.1–3RR compare thereoutout AUis a completeω (3.5–1051 it1–51–5 with redistributionwavelengths AU AUR1 the∗0.1–3) (35–175 (35–175 of stellar the AU0.5 line near AUR (3.5–105R in-∗∗)) (17.5 Brγ andRR3AU(1053AU(105∗∗)) compare0.5 AURR∗∗ it)) (17.5 withR∗ the) stellar in- -60 -100 -50-50 -60 60 -60100-50 50For50-50 the-100 calculations60-60 of4040 the60 2-D intensity50 8000 distributionsω150100 600.1–3 AU (3.5–105 R∗ω) 0.5 AU0.1–3 (17.5 AUR∗ (3.5–105) R∗) 0.5 AU (17.5 R∗) X / R X / R 25 25-40-40 (maps)3) of the Using emitting the region, velocitywe use the coordinateX and/ R sys- densityfrequencies distribution,RX in out/ the sRs reference frame1–5 and of the AU atom (35–175 and use the R∗) 3) Using3AU(105 the velocityR∗) and density distribution, and 25 25 25 25 2525 25 25 25 25 40tensity I∗. AωN goodωN solution0.5–5.70.5–5.7RR has AU AU to(17.5–200ω reproduce(17.5–200N0.85–1.250.85–1.25tensityRω∗N0.5–5.7)R the AU∗ AU)I∗1AU(35. (30–44observed (30–44 AU A0.5–5.71AU(35 good (17.5–200RR∗∗R AU) solution)∗)R (17.5–2000.9∗0.9)R∗ AU AU) has (31.5 (31.5R1AU(35 to∗) reproduceRR∗∗)) 1AU(35R∗) theR observed∗) -40 tem (x, y, z)centeredonthestar.The(x, y) plane coincides Doppler profilesφ in the calculations of the Brγ and the 25 R11 − 0.85–1.25−− AU−− (30–44− choosingR∗) 0.9 the AU−− temperature (31.5 R∗) law along each streamline (we choosingwith the sky plane, the the x temperatureaxisratio is the intersectionI /I∗,theobserved law ofγ theγ diskalongVoigt each profileKαα for streamline–band the−H1–α line 51–γ flux, (with 5 (we theandγ same00.. the4–4– constants observed∗00..7575 as 1–2 52 −1– 5 00..55 2 2 0 0 proach to findd modelsproach thatin Kurosawa tocanproach find et approximately1 al. 2006).models to find thatratio models− reproduce canId/I that− approximately,theobserved can approximately reproduceK–band− reproduce flux, and the observed 0 0 0 0 00 0 plane0 0 with the0 skyFig. plane,2000 5. and the z axis is6000 parallel6000 to the2000 α 22 −−0proach.4– −−0.75 to find models that−−0.5 can approximately reproduce Fig. 5. / R Y 2000 Intensity distributions of our / R Y best-fit disk-wind 2000 Intensity distributionsf α ofα our0.5–3 best-fitf disk-wind00..34–34– 00..440.5–3use0.5–3 a constant00. electron.3333 0.5–3 temperature of 8000 K), we solve Y / R / Y R / Y Y / R / Y use a constant electron temperature of 8000 K), we solve Y / R / Y Y / / R Y

Y / / R Y f 0.5–3 0.5–3 Y / R / Y Y / R / Y R / Y f 0.5–3 0.5–3 Y / R / Y The calculations of the ionization state and the number 0 line-of-sight. In this case,visibilities. the intensity of6000 radiationthe at a observations:the2 observations:visibilities.− − − the observations: densities of theαinin atomic levels were performed0the.34– in a observations: cylinder0.4 0.33 Y / R / Y frequency ν within a spectral line is TT 1400–20001400–2000 K K 18001800 K K model 5 (i.e., intensity-25 -25 distribution-25 of-25 the continuum-25-25-25 -25 disk -25the-25 equations-25 model of the 5 (i.e., statistical intensityβ equilibriumβ distribution for the of0.3–2 the0.3–2 hydrogenβ continuumβ disk 0.3–21the1 equations0.3–2 of the statistical1 equilibrium1 for the hydrogen -25 zmax 2) We determine1)with the To a radius select velocityTin1)rc up To the◦ to 300and select◦ most◦R∗ and density◦ the appropriatea1400–2000 height2) most Weh distributionc =600 appropriate determineR K∗ the◦ ◦ disk◦◦ the◦ parameters, the velocity◦ disk1800 parameters, K and density◦ distribution◦ 1) Tovz select(r) −τ(ν,r the) mostθ1 appropriate the10 θ–801 disk parameters,1) To10 select–8080 the most appropriate80 the disk parameters, plus the disk wind; the central star is not-25 shown; see atomsIw(ν,x,y)= andplus computeS(r)φ( theν − ν0 disk the) e population wind;κ(θr)1dz, (A.6) the ofdivided central the into hydrogen grid star10 cells–80 in isl, θ atomic notcoordinates. shown;θ1 In each cell, see we 80atoms10 and–80 compute the population80 of the hydrogen atomic ! alongc each streamline corresponding9 6 to thealong wind1 each geometry9 streamline67 corresponding11 7 to the wind1 geometry zminwe calculate the continuum˙we calculatesolvedDiskDisk the intensitywe equations−9 wind:calculate wind: the− of− the6 continuum of−˙ statistical the the−1 equilibrium disk,− continuumwe intensityI ford calculate,at− the− 15-7 intensity− of− the9 the−1 disk,− continuum−6 of theId−,at disk,1 intensity− Id,at−−7 of the− disk,1 Id,at Tables 3 and A.1) at the center of the Brγ line (v = Tables0 3 and A.1)40004000M˙ wM atw0 the center10 10–10 of–10 theMw$M Bryr$γyrMline˙ w 10 (v 10=–1010M10levels$Myr$–10yryr andM the$yr ionization10 M$10 degreeyr M$ alongyr each streamline. -25 0 25 -25 0 25 levelswhere r is and a-25 vector,0 the 0|r| =( ionization 25x2 + y2 + z-252)1 degree/2, v 0(r) is 25 alongthe level each hydrogen0 streamline. atom + continuum, taking into account both − -100 and100−1 kinematicsz wavelengths parametersDiskwavelengths (see wind: near11 Sect. Brγ near 4and andand Br compare Appendix kinematicsγ and compare it A). with parameters∗∗ the it with stellar (see the in- Sect. stellar∗∗ 4 and in- Appendix A). 1 -100 -60 -100 100 -60 projection of100 thewavelengths-100 velocity5050 at point r on near the4000 line-of-sight, Brγ andcollision100 compare andω radiativeω it processes with0.1–30.1–3 of the excitation AUstellar AUwavelengths and (3.5–105 (3.5–105 ionization in- RR near)) Br0.50.5γ AU AUand (17.5 (17.5 compareRR )) it with the stellar in- 0kms ) and at 14 other velocities-60 (the labels-50-50 give the 60 60-604) We0kms then calculate) and60 the at 14 Br60γ otherline velocities profile1 and (the compare labels give the 4) We then calculate the Brγ line profile and compare − 25 25X /25 R 25 2525X25-50 / R 25 25and S is a source25 X function / R50 for a3) transition Using between−X / R the energytensity velocity(see GrininI∗tensity. Aωω &ω andN MitskevichN goodI density∗. solution 1990 A0.5–5.70.5–5.70.1–3 good for more distribution, AU3) AUhas AU solution details). (3.5–105Using (17.5–200 (17.5–200to We reproduce as- has the andR toR∗R)∗ velocity∗) reproduce) the0.51AU(351AU(35 observed AU and (17.5 the densityRR observedR∗∗))∗) distribution, and 1 tensity I∗. A good1 solutionsumed has that to the distribution reproduce of the atomic the sub-levels observedtensity followsI∗. A good solution has to reproduce the observed velocity in km s ). For the calculation of the25 model im- itlevels withi and j thevelocity observed in km line s profile.). For The the calculation solutionωN is of good0.5–5.7 the if model AU−− (17.5–200 im- itR with∗) the1AU(35 observedR∗) line profile. The solution is good if choosing the temperatureratiotheirId/I statisticalratio∗,theobserved lawγγ weights.Id along/I∗ Johnson,theobserved each collisionchoosingK streamline–band rates1–1– (Johnson the 5 5K flux,–band temperature (we and flux, the observed and law the22 along observed each streamline (we 0 0 3 ratio Id−1/I∗,theobservedproach to findK–band modelsproach flux, that and to can findthe approximately observedratio modelsId/I that∗,theobserved canreproduce approximatelyK–band reproduce flux, and the observed ages in this figure, a0 clockwise motion0 0 of the disk00 wind0 was the0 theoretical2hν 0nagesj (r) gi in line this profile figure,proach20002000 is a toin clockwise good find1972) models agreement were motion usedγ for that all of transitions with the canproach disk the approximately except wind 1-2; to−1– for find was the 5 lat- models reproducethe theoretical that can2 approximately line profile is in reproduce good agreement with the Y / R / Y Fig. 5. R / Y Fig. 5. r Intensity / R Y distributions of / R Y our best-fitS( )= disk-windIntensity distributions− 1use, a constant of our(A.7) electron best-fit temperature disk-windff of 8000 K),0.5–30.5–3 we solve 0.5–30.5–3 Y / R / Y R / Y Y / R / Y use a constant electron temperature of 8000 K), we solve Y / R / Y Y / / R Y Y / / R Y 0 2 visibilities. visibilities.ter, we usedvisibilities. those from Scholz et al. (1990). c " nassumed.i(r) gj # Therefore,the2000 observations: in the blue-shiftedf the observations: imagesthe observations: (leftvisibilities.0.5–3 pan- 0.5–3 assumed. Therefore, in the blue-shifted imagesR / Y (left pan- intensity and shape ofthe the observations: observed one. intensity and shape of the observed one. model 5 (i.e., intensity-25 -25 distribution-25-25-25 of-25 themodel continuum 5 (i.e., intensity disk the distribution equations of of the the statistical continuumβ2)β equilibrium We disk determinethe for equations the0.3–2 the0.3–2 hydrogen velocity of the and statistical density equilibrium11 distribution for the hydrogen -25 -25 -25 -25 2) We determine1) theTo2) select velocity We the determine and most1) density appropriate To the select velocity distribution the◦◦ the2) most and◦ We◦ disk density appropriate determine parameters, distribution the the velocity disk◦◦ parameters, and density distribution els), mainly the disk regions on the right hand-25 side of the 12 5) Forels), each mainly disk-continuum the1) disk To regions select and disk-wind the on most theβ11 right model appropriate hand (see1) side To the0.3–2 selectof disk the the parameters, most5) For appropriate each1 disk-continuum the disk parameters, and disk-wind model (see plus the disk wind; the central star isplus not the shown; disk wind; see atoms the central and compute star is the not populationalong shown;θθ each see of streamline theatoms hydrogen1010◦ and corresponding–80–80◦ compute atomic the to population the wind8080◦ geometry of the hydrogen atomic along each streamlinealong corresponding each streamlineθ1 to the corresponding wind−−99 geometry10along−−–8066 each to−− the streamline11 wind geometry− corresponding−7780 −−11 to the wind geometry star are bright. The radius of the inner edge of the disk Table 3), westar compute are bright.we thewe calculate two-dimensional The calculate radius the of the continuum the continuumintensity˙˙wwwe inner calculate edge intensity distri-we intensity calculate of the the of continuum disk ofthe the the disk, continuumTable disk, intensityId,at 3),Id,at we intensity compute of the disk, of the the two-dimensionalId disk,,at Id,at intensity distri- Tables-25 3 and 0 A.1) 25 at-25 the 0 center-25 25 0 of 25-25 theTables 0 Br-25 25γ 3line 0 and 25 (-25v A.1)=-25 0 atlevels 25 0 the 25 and center00 the-25 ionizationofand 0 the kinematics 25 Brγand degreeMMline kinematics parameters ( alongv = each1010levels parameters−9 (see streamline.–10–10 and− Sect.6 MM the$$ (seeyryr 4 ionization− and1 Sect. Appendix 4 and10 degree10−7 AppendixMM A).$$ alongyryr−1 each A). streamline. − and kinematicswavelengths parameters (see nearM Sect.˙ Brwwavelengthsγ 4and and compare Appendix10 near–10and it Br with kinematicsA).Mγ$andyr the compare stellar parameters10 init withM$ (seeyr the Sect. stellar 4 and in- Appendix A). wind ejection region0kms (i.e.,1) radius and at of-100 14 the-100 other inner velocities hole)-60-60 is ω (the1 -100= labels100bution−1001-100 give map thewind for6060 ejection several100wavelengths inclination region0 100 (i.e., angles. near radius Br We ofγ and the calculate inner comparewavelengths hole) all it is withω1 near= the Brbution stellarγ and in- map compare for several it with inclination the stellar angles. in- We calculate all X / R 25 X / R X / R 250kmsX / R X )/ R and atX / 143)R X other Using/ R4) We velocities the then velocityX calculate / (the R3) and Using labels the density3) give Brthe Usingγ velocityline the distribution, profilethe andvelocity4) and We density and compare then and calculate distribution, density the distribution, Br andγ line profile and and compare 17.5 R∗ (∼0.3 AU). The inclination25−1 angle (angle2525 -60 between 25 17.5−1 60R∗ (∼0.3tensity AU).tensityI The∗.I A∗. inclination good A good solution solutiontensity angle hasItensity (anglehas∗ to. A reproduceto good betweenIreproduce∗.3) A solution good Using the the solutionobserved has the observed to velocity reproduce has to andreproduce the density observed the distribution, observed and velocity in km s ). For the calculation25 velocity of the model in km sim-choosing). Forit with the thecalculation temperature observedchoosing of law line thechoosing the along profile. model temperature each im- the The streamline temperature solutionit with lawchoosing along the is (we good law observed each the along if temperature streamline line each profile. streamline (we law The along (we solution each streamline is good if (we the polar axis and the viewing direction)0 of the model0 is i the polar axisratio andratioI thed/IId∗ viewing/I,theobserved∗,theobservedproachproach direction)ratio to to find findKI–bandKd of/Iratio models models–band the∗,theobserved flux, modelId/I flux, that that∗ and,theobserved is andcan cani the approximately the approximatelyK observed–band observedK flux,–band and reproduce reproduce flux, the and observed the observed ◦ ages in this figure,Fig.0 a 5. clockwiseFig.Intensity 5. motion00 Intensity distributions ofages the distributions disk in0 this of wind our figure, was best-fit of a◦ clockwiseourthe disk-wind best-fit theoretical motion disk-winduse line of thea profile constant diskuse is wind a in constant electron good was agreement temperaturethe electron theoretical temperature with of the line 8000 profile K), of 8000 we is solve in K), good we solve agreement with the Fig. 5. R / Y R / Y Fig. 5. Intensity distributions of / R Y / R Y our best-fit disk-wind Intensityuse a constant distributions electron of temperature our best-fit of disk-wind 8000 K), we solve

Y / R / Y use a constant electron temperature of 8000 K), we solve =20 (i.e., almost pole-on). TheR / Y colors represent0 the inten- =20 (i.e., almostvisibilities.visibilities. pole-on).proachthethe The observations: observations: colorsvisibilities. to find representvisibilities. models7 the that inten- can approximately reproduce 7 modelassumed. 5 (i.e., intensity Therefore,model distribution in 5 the (i.e.,model / R Y blue-shifted intensity of 5 the(i.e., continuum intensity distributionassumed. images distribution (leftdisk Therefore, of pan- the continuum ofintensity in the the continuum blue-shifted− disk and− shapethe− disk equations images of− thethe observed (left equations of the pan- statistical one. ofintensity the statistical equilibrium and shape equilibrium for of the the hydrogen observedfor the hydrogen one. −1 −1 ˚−1 −2-25 -25 -25-25 -25 model-25 5 (i.e.,sitythe intensity equations in erg ster distribution of12) thes 2) We1 statisticalA˚ We determine1 cm of determinethe the2 equilibrium. observations: In continuum the these the velocity2) images, velocity We for disk determine the2) and AMBER’s We and hydrogenthe density determine densityequations the distribution velocity distribution ofthe the velocity and statistical density and equilibrium density distribution distribution for the hydrogen sity in erg ster els),s mainlyA cm the.plus In disk these the regionsplus images, disk on-25 the wind; the AMBER’s disk right the wind;els), centralhand mainly the side star central theof theis disk not star regions shown; is5) not For on see eachshown; the rightatoms disk-continuum see hand andatoms1)1) compute side To To and ofselect select and the computethe disk-wind the the population most most5) the For appropriate modelappropriate population each of the (see disk-continuum hydrogen the theof the disk disk hydrogen atomic parameters, parameters, and atomic disk-wind model (see plus the disk wind; the central star is not shown;plus seethe diskatoms wind; and thecomputealongalong central each the each streamline star population streamline is1) not To correspondingalong of shown; select corresponding the each hydrogenalong the see streamline most each toatoms atomic theto appropriate streamline the wind and corresponding wind compute geometry geometry corresponding the disk the to parameters, population the wind to the geometry of wind the geometry hydrogen atomic spectral resolutionstar of are 12 bright. 000 was The modeled, radiusTables as of described the 3 and inner A.1)star in edge are at of the bright. the center diskspectral The ofTable radius the resolution 3), Br ofγ we theline compute of inner 12 (v 000= edgewewe the was calculate calculate oftwo-dimensional modeled, the disk the the as continuumTable continuum described intensity 3), we in intensity intensity distri- compute of of the the the two-dimensional disk, disk, IIdd,at,at intensity distri- Tables 3 and A.1) atTables the-25-25 center 3 0 and 0 25 of A.1) 25 the-25 at Br-25 the 0γ -25line 0 25 centerTables 25 0 (v -25 25= of 3 the and 0levels 25-25 Br A.1)γ and 0line at-25 25the the(v ionization 0= center 25 levels of degreewe and thelevels calculate the alongBrγ ionization andline each the the (v streamline. ionization continuum= degreelevels along degree and intensity each the along streamline. ionization of each the streamline. disk, degreeI ,at along each streamline. Sect. A.3. − −1 −1 Sect. A.3. andand kinematics kinematicswavelengthswavelengths parameters parametersand kinematics (see near nearand (see Sect. Br Br kinematics Sect.γγ 4andand parameters and 4 compareand compare Appendix parameters Appendix (see it it with A).with Sect. A). (see the the 4 and Sect. stellar stellar Appendix 4d and inin- Appendix A). A). 0kmswind1) and ejection at 14 region0kms other (i.e., velocities)0kms radius and at (theof-100-100) 14 andthe other labels innerwind at 14 velocities0kms give hole) ejection other the is−1 velocitiesω) (theregion1 and= labels4)100100 at (i.e.,bution We (the 14 give then radius other labels map the calculate velocitiesof for give the several inner the4) the We(the inclination Br hole) thenγ labels4)line is We calculateω profile1 give angles.then= the andbution calculate the We compare Br calculate mapγ4)line the We for Brprofile then severalallγ line calculate and profile inclination compare the and Br compare angles.γ line We profile calculate and compare all X / RX / R 2525−-100 X /− RX / RX / R X / R 100X / R X / 3)R 3) Using Using thewavelengths the velocity velocity3) and near Using and density Br3) densityγ the Usingand distribution,velocity compare distribution, the velocity and it with and density and and the density distribution,stellar in- distribution, and and 17.5 R∗ (∼−0.31 AU). Thevelocity inclination1 in km angle s17.51). (angleR For∗ (∼ the between0.3 calculation AU).− The1 ofinclination the model angle im-tensitytensity (angle betweenII∗∗.. A A good good solution solution has has to to reproduce reproduce the the observed observed velocity in km s ). Forvelocity the calculation in km25 s ). of For the the modelvelocity calculation im- init km of with the s ). model the For observed the im- calculationit line with profile.tensity the ofit with the observed TheI∗. model Athe solution good observed line im- solution profile. isit good line with has The profile. if the to solution reproduce observed The is solution linegood the profile.observed if is good The if solution is good if choosingchoosing the the temperatureratioratio temperatureII choosing/I/I∗∗,theobserved,theobserved law lawchoosing along the along temperature each the eachKK streamline–band–band temperature streamline law flux, flux, along (we and and law (we eachthe the along observed observed streamline each streamline (we (we the polar axis andages the in thisviewingages figure, in0 direction)0 this a clockwise figure,the of a polar the clockwise motion model axis of and motion is thei the disk of viewing wind the disk was direction) windthe was theoretical of thethe modeld theoreticald line is profilei line is profile in good is agreement in good agreement with the with the ages in this◦ figure,Fig. a 5. clockwiseIntensity motion distributionsFig. of 5. theIntensity disk of windages◦ our distributions was in best-fit thisthe figure, disk-wind theoretical of a clockwise our best-fituse line a motion profile constant disk-windratio of is the in electronI good diskduse/I∗,theobserved wind agreement a temperature constant was withthe electron ofK theoretical the 8000–band temperature K), flux, we line and solve profile of the 8000 observed is in K), good we solve agreement with the Fig. 5. Intensity distributionsR / Y R / Y Fig. of 5. ourIntensity best-fit disk-wind distributionsuse of a our constant best-fit electron disk-wind temperatureuse a constant of 8000 K), electron we solve temperature of 8000 K), we solve =20 (i.e., almostassumed. pole-on).assumed. Therefore, The0 colors Therefore, in represent the=20 blue-shifted in(i.e., the almost inten- blue-shifted images pole-on). (left images The pan- colors (left pan- representvisibilities.visibilities.intensity the inten- and shape of the observed7 one. 7 assumed. Therefore,model in 5 (i.e., the blue-shifted intensityR / Y model distribution images 5 (i.e., (left intensityassumed. of pan-the continuum distribution Therefore,intensity disk and of in the the shape continuum blue-shifted ofintensity the observedvisibilities. disk images and shape one. (left of pan- the observedintensity one. and shape of the observed one. model 5−1 (i.e.,−1 ˚ intensity−1 −2-25-25 distributionmodel of 5 the (i.e., continuum intensity−1 −1 ˚ distribution disk−1 the−2 the equations of equations the continuum of the of2)2) the statistical We Wethe statistical disk determine determine equations equilibriumthe equilibrium equations ofthe the the velocity velocity for statistical of for the the the and hydrogen and statistical hydrogen equilibrium density density equilibrium distribution distribution for the hydrogen for the hydrogen els), mainlysity in erg the ster diskels), regionss mainlyA onels),cm the the mainly-25. right In disk these hand regions thesity images, disk side inels), on regions erg of the AMBER’s mainly the ster right ons the hand5)A rightdisk For sidecm regions each hand of. the disk-continuumIn side on these theof the5) images, right For hand each and AMBER’s5) For side disk-winddisk-continuum each of the disk-continuum model and5) (see For disk-wind each and disk-continuum disk-wind model (see model and (see disk-wind model (see plusplus the the disk disk wind; wind; theplus the central the centralplus disk star the star wind; is disk notis the notwind; shown; central shown; the see central star seeatoms isatoms star not and shown; is and compute not computealongalong shown; see2) the each We eachatoms the population see determine streamline streamline population andatoms compute of the and corresponding corresponding the of velocity the compute hydrogen the hydrogen population and the atomic to to density population atomic the the of wind wind the distribution geometry hydrogen geometry of the hydrogen atomic atomic star arespectral bright. resolution Thestar radius of are 12 of bright.star 000 the was are inner The bright.modeled, edge radiusspectral The of as ofstar the radius described the resolution disk are inner of bright. the in edge of inner The 12 of 000 the radius edge was disk of of modeled, the theTable disk inner 3), asTable describedwe edge compute of3), the we in compute thedisk two-dimensional the two-dimensional intensity intensity distri- distri- TablesTables 3 and 3 and A.1) A.1) atTables at the the-25 center-25 3Tables center and 0 0 of 25 25 A.1) 3 of the and the Brat-25-25 A.1)γ Br theline 0γ 0Table centeratline 25 25 (v the 3), (=v of we center=levels the computelevels Br of andγ the andline the the Br the two-dimensional ionization(alongγv ionizationline= each (levelsv degree= streamline degree andlevels intensity along the along and ionizationcorrespondingTable each distri- the each streamline. 3), ionization streamline. wedegree compute to the along degree wind the each along geometrytwo-dimensional streamline. each streamline. intensity distri- Sect. A.3. −1 −1 Sect. A.3. andand kinematics kinematics parameters parameters (see (see Sect. Sect. 4 4 and and Appendix Appendix A). A). wind ejection region−1 (i.e.,wind radius ejectionwind of region theejection inner-25 (i.e., region 0 hole) radius 25−wind1 (i.e.,is ω of1 ejection-25 radius the= inner0bution of 25 region the hole) map inner (i.e., is forω hole)1 radius several= isbution ofω inclination1 the=and map innerbution kinematics for hole) angles. several map is ω Wefor parameters1 inclination= several calculatebution inclination (see angles. all map Sect. for We 4angles. several and calculate Appendix We inclination calculate all A). angles. all We calculate all 0kms0kms) and) and at at 14 14 other0kms other velocities0kms velocitiesX)X // and RR at (the) and(the14 labels other at labelsXX // 14 RR give velocities other give the the velocities (the4) labels We4) (the We then give labels then calculate the3) calculate3) give Using Using the the4) the Wethe theBrγ Br then velocity velocity4)lineγ Weline calculate profile then profile and and and calculate density thedensity and compare Br compareγ the distribution, distribution,line Br profileγ line and profile and and compare and compare 17.5 R∗ (∼0.3 AU). The17.5 inclinationR∗−1(17.5∼−0.31 R AU).∗ angle(∼0.3 The (angle AU). inclination17.5 between− The1 R inclination∗ angle−(∼1 0.3 (angle AU). angle The between (angle inclination between angle (angle between velocityvelocity in kmin km s s). For).velocity For the the calculationvelocityX in calculation/ R km s in). of km For theof s theX the model/). R model For calculation im- the im- calculationit ofwithit thewith the model of the observed thechoosing observedchoosing im- model3) Using lineit im- the linethe with profile. the temperature temperature profile.it the with velocity The observed The the solution lawand solutionlaw observed line along along density is profile. good is line each each good distribution, ifprofile. The streamline streamline if solution The and (wesolution (we is good is if good if the polar axis andages the inthe this viewing polar figure,the axis direction) a polarages andclockwise in axisthe of this viewing the and motion figure, model thethe direction) viewing of a polar is clockwise thei disk axis direction) of wind and motion the the model was of of viewing the thethe is modeli disk theoretical direction) wind is ichoosing was lineof thethe profile the model theoretical temperature is is ini good line agreement law profile along is in with each good the streamline agreement (we with the ◦ ages in this figure,◦ Fig.Fig. a 5. 5. clockwise◦ IntensityIntensityages motion distributions distributions in this of the figure,◦ disk of of awind clockwise our our was best-fit best-fit motionthe disk-wind disk-wind theoretical of the diskuseuse line wind a a profile constant constant was is inthe electron electron good theoretical agreement temperature temperature line with profile of of the 8000 8000 is in K), K), good we we7 solve solveagreement7 with the =20 (i.e., almostassumed. pole-on).=20 Therefore,Fig.(i.e., The=20 5. almost colorsassumed. inIntensity(i.e., represent the pole-on). almost blue-shifted Therefore, distributions pole-on).the The=20 inten- colors inimages(i.e., The the representof almost colors ourblue-shifted (left best-fit pole-on). pan- represent the inten- imagesdisk-wind The the inten- colors (leftuse pan- represent a constant the inten- electron temperature7 of 8000 K), we solve 7 assumed. Therefore,modelmodel in 5 5− (i.e., the(i.e.,− blue-shifted intensity intensityassumed.−−1 −−1 distribution distribution− Therefore, images1 −2 (left of ofin pan-the the the continuum continuum blue-shiftedintensityintensity disk disk and images and shapethethe shape (left equations equations ofintensity the of pan- the observed of of observedintensity the the and statistical statistical one. shape one. and of shape equilibriumthe equilibrium observed of the observed for for one. the the hydrogen hydrogen one. sity in erg ster−1 s−1 A˚sity−1 cm in erg−2sity. ster In in these1 ergs images,1 sterA˚ 1 cms AMBER’s2A˚. Incm these. images, In− these1 − AMBER’s1 images,˚−1 − AMBER’s2 els),els), mainly mainly themodelplusplus the disk disk the the regions 5els), (i.e., regions disk disk mainly intensity on wind;els), wind; on the the mainlythe right the thesity distribution disk right central central inhand the regions erg hand disk side ster star star side of regionson of thes is the isthe of notcontinuum nottheA right on shown; shown; thecm5) hand right5) For. disk In Forsidesee see each thesehand each oftheatomsatoms disk-continuum the images, side equations disk-continuum andof and the AMBER’s5) compute compute of For the and each5) statistical andthe the For disk-wind disk-continuum population population disk-wind each equilibrium disk-continuum model modelof of the the (see and hydrogen for hydrogen (see disk-wind the and hydrogen atomic atomic disk-wind model (see model (see spectral resolution ofspectral 12 000 was resolutionspectral modeled, resolution of 12 as 000 described of was 12 modeled, 000 in was modeled, as described as described in in starstar are are bright. bright.plusTablesTables The The the radiusstar 3 3 radius disk and and are of A.1)wind;star A.1) bright.the of the are innerat at thespectral inner the Thebright. the central edge center center radius edge resolution The of star theof of of of radius the the the disk is ofdisk not Brinner Br of 12γγTable the shown; 000linelineTable edge inner was 3), ( (vv of 3),see we modeled,== edge the we computeatomslevelslevels diskcompute of the as and and theTable described disk the the computethe two-dimensional 3), two-dimensional ionization ionizationTable we in the compute 3), population wedegree degree intensity compute the intensity along along two-dimensional of distri- the the each each distri- two-dimensional hydrogen streamline. streamline. intensity atomic intensity distri- distri- Sect. A.3.Sect.−− A.3. Sect. A.3. windwind ejection ejection regionTables region (i.e., 3wind11 (i.e., and radius ejection radius A.1)wind of atthe of ejection regionSect. the the inner inner center A.3.(i.e., region hole) hole) radius of is (i.e.,ω the is1 ofω radius=1 Br the=γbution innerline ofbution the hole)( mapv inner map= for is hole)ωlevels for several1 = several is andbutionω inclination1 the= inclination ionization mapbution angles. for mapangles. several degree We for We inclination severalcalculate along calculate each inclination all angles. streamline. all We angles. calculate We calculate all all 0kms0kms−1 )) and and at at 14 14 other other velocities velocities (the (the labels labels give give the the 4)4) We We then then calculate calculate the the Br Brγγ lineline profile profile and and compare compare 17.517.5R∗ R(∼∗ 0.3(∼0.30kms AU). AU). The17.5 The) inclinationandR inclination∗17.5 at−(−∼11 140.3R other∗ angle AU).(∼ angle0.3 velocities (angle The AU). (angle inclination between The between (the inclination labels angle give (angle angle the between (angle4) between We then calculate the Brγ line profile and compare velocityvelocity in in km km s s−1 ).). For For the the calculation calculation of of the the model model im- im- itit with with the the observed observed line line profile. profile. The The solution solution is is good good if if thethe polar polar axis axis andvelocityagesages and the in in the thisthe this viewing in viewing km polar figure, figure, sthe direction) axis). a adirection) polar clockwiseFor clockwise and the axis the of calculation the of and motionviewing motion the model the model of viewing of direction) is the ofthei is the disk diski direction) model windof wind the was im- was model of theitthethe is with modeli theoretical theoretical the is i observed line line profile profile line profile. is is in in good good The agreement agreement solution is with with good the the if ◦ ◦ ◦ ◦ 7 7 =20=20(i.e.,(i.e., almost almostagesassumed.assumed. pole-on). in pole-on). this=20 Therefore, Therefore, figure, The(i.e.,=20 The colors a almost clockwise colors in in(i.e., represent the the pole-on). almost represent blue-shifted blue-shifted motion the pole-on). The the inten- of colors inten-the images images The disk represent colors (left wind(left pan- pan- represent was the inten-theintensityintensity the theoretical inten- and and shape shape line profile of of the the observed observed is in good one. one. agreement7 with the 7 − −−1 −1− −1 − −2 −1 −1 −−1 − −2− − sitysity in erg in erg ster sterassumed.els),els),1 s 1s mainly mainlyA˚sityA˚1 Therefore,cm in the thecm erg2sity. disk In disk ster. thesein In in regions regions erg these thes images, ster images,A˚ blue-shifted on on1 s thecm the AMBER’s1 A˚ AMBER’s right right.1 Incm images thesehand hand2. In images, side side (left these of of pan- the images,the AMBER’sintensity AMBER’s5)5) For For and each each shape disk-continuum disk-continuum of the observed and and one. disk-wind disk-wind model model (see (see spectralspectral resolution resolutionels),starstar of mainly are are 12 ofspectral bright. bright. 000 12 the 000 wasspectral resolution disk The wasThe modeled, regions modeled, radius radius resolution of as of ofon 12 asdescribed the the 000the described of inner inner right was 12 000 modeled, in edgehand edge inwas of ofside modeled, the the as of described disk disk the asTableTable described in5) For3), 3), we we each in compute compute disk-continuum the the two-dimensional two-dimensional and disk-wind intensity intensity model distri- distri- (see Sect.Sect. A.3. A.3. starwindwind are ejection ejectionSect. bright. A.3. region regionSect. The A.3.(i.e., (i.e., radius radius radius of the of of innerthe the inner inner edge hole) hole) of the is is ωω disk== Tablebutionbution 3), map map we for for compute several several the inclination inclination two-dimensional angles. angles. We We intensity calculate calculate distri- all all wind ejection region (i.e., radius of the inner hole) is ω 11 = 17.517.5 RR∗∗ ((∼∼0.30.3 AU). AU). The The inclination inclination angle angle (angle (angle between between1 bution map for several inclination angles. We calculate all 17.5thethe polar polarR∗ (∼ axis axis0.3 and andAU). the the The viewing viewing inclination direction) direction) angle of of (angle the the model model between is is ii the polar◦◦ axis and the viewing direction) of the model is i 77 =20=20◦ (i.e.,(i.e., almost almost pole-on). pole-on). The The colors colors represent represent the the inten- inten- =20 (i.e., almost−− −− pole-on).−− The−− colors represent the inten- 7 sitysity in in erg erg ster ster 11ss 11A˚A˚ 11cmcm 22.. In In these these images, images, AMBER’s AMBER’s −1 −1 ˚−1 −2 sityspectralspectral in erg resolution resolution ster s of ofA 12 12 000cm 000 was was. In modeled, modeled, these images, as as described described AMBER’s in in spectralSect.Sect. A.3. A.3. resolution of 12 000 was modeled, as described in Sect. A.3. Nature of Br% in the HAe star HD104237

Disk truncated by magnetosphere

Gas within the disk

Outflowing wind

Tatulli et al. (2007, A&A 464, 55)

VLTI Summer school 2010 - Inner regions of YSOs revealed by interferometry - F. Malbet Disk/star interaction ?

VLTI Summer school 2010 - Inner regions of YSOs revealed by interferometry - F. Malbet A systematic study of the origin of the Brϒ

emissionS. Kraus etin al.: Origin Herbig of hydrogen line emission Ae/Be in Herbig Ae/Be stars stars 1169

with that of the Spitzer/IRS spectrum taken with a smaller aper- Dust sublimation radius (Tevp=2000K, ε=1) ture (see Fig. 7b) suggests that the SED at λ > 15 µmisdom- Dust sublimation radius (Tevp=1500K, ε=1) Dust sublimation radius (T =1000K, ε=1) ∼ evp 1000K inated by emission from this envelope. The SED model ﬁts by 10 Continuum-emitting region Borges Fernandes et al. (2007)suggestthatthesystemisseen Brγ line-emitting region 1500K 2000K under low to intermediate inclination (i < 70◦). The line proﬁles of the strong hydrogen∼ recombination lines were modeled as emission from a spherically symmetric gas en- 1 velope (Benedettini et al. 1998). Acke et al. (2005)interpreted

the proﬁle of the optical [O ]lineinthecontextofawind Ring radius [AU]

originating from the surface layer of a passive disk. 0.1 Stellar Surface Using the AMBER visibilities measured at continuum wavelengths, we compare the ring diameter measured towards various

PAs (covering a position angle range of 68◦)andﬁndnoindi- MWC480 HD163296 HD104237 HD98922 MWC297 V921Sco cations for an elongation of the continuum-emitting∼ region. This 0.01 is consistent with the measured small Brγ line closure phase (see Table 4), indicating that the brightness distribution of the 2 combined line- and continuum-emitting region is nearly centro- symmetric, as in the case when the line-emitting region is seen 1.5 nearly face-on. Furthermore, V921 Sco exhibits a continuum- cont / R γ

emitting region which is more compact than expected for an Br

R 1 irradiated dust disk, which might suggest that the NIR continuum emission is dominated by emission from an optically thick gaseous inner accretion disk (similar to MWC 297). 0.5 Within the Brγ line, we measure a slight increase in visibility (see Fig. 7). Furthermore, the continuum-corrected line- 0 10 100 1000 10000 visibilities VBrγ (ranging between 0.7 and 0.5) show that the Stellar Luminosity [Lsun] Brγ-emitting region is also spatially resolved and onlyKraus slightly et al. (2008, A&A 489, 1157) more compact than the continuum-emitting region (R /R Fig. 8. Top:thefittedringradiiforthecontinuum(R ,blackpoints) Brγ cont ≈ cont 0.7). Therefore, the Brγ region is too extended to be consis- and Brγ line (RBrγ,redpoints)plottedasafunctionofstellarluminos- tent with magnetosphericVLTI accretion Summer school or an 2010 X-wind - Inner as regions dominant of YSOs ity.revealed For theby interferometry Brγ line, we - showF. Malbet the ring radius, determined by fitting Brγ-emitting mechanism, which makes a strong stellar wind, a only the spectral channel with the line center. For comparison, we plot disk wind or a gaseous inner disk the most likely scenario. the spatial extension of the stellar surface R$ (grey area) and the dust sublimation radius corresponding to dust sublimation temperatures of 2000, 1500, and 1000 K (computed using Eq. (2)andassuming% = 1). Bottom:providinganaturalmeasureforthetemperaturedistribution 7. Discussion of general trends within the circumstellar disk, we normalized RBrγ by the size of the Since our sample covers a wide range of stellar parameters, continuum-emitting region Rcont.Inparticular,thisnormalizationseems important considering the large range of stellar luminosities covered by we investigated whether the measured size of the continuum our sample. and Brγ-emitting region correlates with the stellar parameters or the spectroscopic properties. Revealing such relationships could provide more insight into the involved physical mecha- 1000 K using the analytic expression for the dust sublimation nisms and is also essential to confirm the empirically found cor- radius Rsubl by Monnier & Millan-Gabet (2002) relations between the Brγ luminosity and other estimators for 2 the mass accretion rate (Calvet et al. 2004). In order to fur- 1500 K L$ 1 ther expand our sample, we include in this section not only Rsubl = 0.034 AU (2) Tsubl # L % · the five objects which we have investigated with AMBER, but ! " $ also the MWC 480 Keck-Interferometer measurement published In this relation, % denotes the ratio of the absorption efficiencies by Eisner (2007). For MWC 480, we assume L = 17 L , d = of the dust at the sublimation temperature Tsubl and at the stellar $ 140 pc, Teff = 8700 K, a P-Cygni Hα-line profile (Acke et al. effective temperature T$,i.e.% = κP(Tsubl)/κP(T$)withκP(T) = 2005), and a Brγ line luminosity log (L(Brγ)/L ) = 2.8(as Qabs(λ)Bλ(T)dλ/ Bλ(T)dλ.Weassume% = 1, corresponding $ − determined by Testi and Natta from unpublished TNG spectra; to rather large dust grains of several µminsizeand,therefore,to private communication). $the innermost region$ of the inner dust rim. We find that the measured continuum radii of three of the five objects (HD163296, HD104237, HD98922) follow the 7.1. Continuum emission R L1/2 law rather closely assuming dust sublimation tem- subl ∝ $ Earlier interferometric investigations of the continuum-emitting peratures Teff between 1300 K and 1500 K, which is in agree- region of HAeBe stars have established a relation between the ment with the results obtained in earlier NIR broadband interfer- size of the continuum-emitting region and the stellar luminosity, ometric survey observations (Monnier & Millan-Gabet 2002). suggesting that the continuum emission mainly traces hot dust For the Herbig Be stars MWC297 and V921Sco, the derived located at the dust sublimation radius. In order to check for sim- ring diameters are significantly more compact than the expected ilar trends, we plot the determined K-band continuum ring radii dust sublimation radii. A possible explanation for the small ap- as a function of the stellar luminosity L$ in Fig. 8.Forcompar- parent continuum sizes might be inclination effects, although, ison, we computed the predicted inner disk radii corresponding based on complementary information from the literature (see to dust sublimation temperatures Tsubl of 2000 K, 1500 K, and Sect. 6), we expect notable inclination only for HD 163296. A systematic study of the origin of the Brϒ

emissionS. Kraus etin al.: Origin Herbig of hydrogen line emission Ae/Be in Herbig Ae/Be stars stars 1169

with that of the Spitzer/IRS spectrum taken with a smaller aper- Dust sublimation radius (Tevp=2000K, ε=1) ture (see Fig. 7b) suggests that the SED at λ > 15 µmisdom- Dust sublimation radius (Tevp=1500K, ε=1) Dust sublimation radius (T =1000K, ε=1) ∼ evp 1000K inated by emission from this envelope. The SED model ﬁts by 10 Continuum-emitting region Borges Fernandes et al. (2007)suggestthatthesystemisseen Brγ line-emitting region 1500K < 2000K undermagnetospheric low to intermediate inclination accretion (i 70◦). • The line proﬁles of the strong hydrogen∼ recombination lines were modeled as emission from a spherically symmetric gas en- 1 •velopedisk (Benedettini wind et al. 1998). Acke et al. (2005)interpreted

the proﬁle of the optical [O ]lineinthecontextofawind Ring radius [AU]

•originatingX-wind from or the surfacedisk layer wind of a passive ? disk. 0.1 Stellar Surface Using the AMBER visibilities measured at continuum wavelengths, we compare the ring diameter measured towards various

emitting region which is more compact than expected for an Br

emissionS. Kraus etin al.: Origin Herbig of hydrogen line emission Ae/Be in Herbig Ae/Be stars stars 1169

with that of the Spitzer/IRS spectrum taken with a smaller aper- Dust sublimation radius (Tevp=2000K, ε=1) ture (see Fig. 7b) suggests that the SED at λ > 15 µmisdom- Dust sublimation radius (Tevp=1500K, ε=1) Dust sublimation radius (T =1000K, ε=1) ∼ evp 1000K inated by emission from this envelope. The SED model ﬁts by 10 Continuum-emitting region Borges Fernandes et al. (2007)suggestthatthesystemisseen Brγ line-emitting region 1500K < 2000K undermagnetospheric low to intermediate inclination accretion (i 70◦). • The line proﬁles of the strong hydrogen∼ recombination lines were modeled as emission from a spherically symmetric gas en- 1 •velopedisk (Benedettini wind et al. 1998). Acke et al. (2005)interpreted

the proﬁle of the optical [O ]lineinthecontextofawind Ring radius [AU]

PAs (covering a position angle range of 68◦)andﬁndnoindi- MWC480 HD163296 HD104237 HD98922 MWC297 V921Sco ! Nocations correlation for an elongation with ofL the∗ as continuum-emitting suggested∼ region.by Eisner This 0.01 is consistent with the measured small Brγ line closure phase et al.(see 2007 Table 4), indicating that the brightness distribution of the 2 combined line- and continuum-emitting region is nearly centro- symmetric, as in the case when the line-emitting region is seen 1.5 ! Wenearly are face-on. probing Furthermore, mostly V921 outﬂows Sco exhibits phenomena: a continuum- cont / R γ

emitting region which is more compact than expected for an Br

Brγ indirect accretion tracer through accretion-driven mass loss? R 1 irradiated dust disk, which might suggest that the NIR continuum emission is dominated by emission from an optically thick gaseous inner accretion disk (similar to MWC 297). 0.5 Within the Brγ line, we measure a slight increase in visibility (see Fig. 7). Furthermore, the continuum-corrected line- 0 10 100 1000 10000 visibilities VBrγ (ranging between 0.7 and 0.5) show that the Stellar Luminosity [Lsun] Brγ-emitting region is also spatially resolved and onlyKraus slightly et al. (2008, A&A 489, 1157) more compact than the continuum-emitting region (R /R Fig. 8. Top:thefittedringradiiforthecontinuum(R ,blackpoints) Brγ cont ≈ cont 0.7). Therefore, the Brγ region is too extended to be consis- and Brγ line (RBrγ,redpoints)plottedasafunctionofstellarluminos- tent with magnetosphericVLTI accretion Summer school or an 2010 X-wind - Inner as regions dominant of YSOs ity.revealed For theby interferometry Brγ line, we - showF. Malbet the ring radius, determined by fitting Brγ-emitting mechanism, which makes a strong stellar wind, a only the spectral channel with the line center. For comparison, we plot disk wind or a gaseous inner disk the most likely scenario. the spatial extension of the stellar surface R$ (grey area) and the dust sublimation radius corresponding to dust sublimation temperatures of 2000, 1500, and 1000 K (computed using Eq. (2)andassuming% = 1). Bottom:providinganaturalmeasureforthetemperaturedistribution 7. Discussion of general trends within the circumstellar disk, we normalized RBrγ by the size of the Since our sample covers a wide range of stellar parameters, continuum-emitting region Rcont.Inparticular,thisnormalizationseems important considering the large range of stellar luminosities covered by we investigated whether the measured size of the continuum our sample. and Brγ-emitting region correlates with the stellar parameters or the spectroscopic properties. Revealing such relationships could provide more insight into the involved physical mecha- 1000 K using the analytic expression for the dust sublimation nisms and is also essential to confirm the empirically found cor- radius Rsubl by Monnier & Millan-Gabet (2002) relations between the Brγ luminosity and other estimators for 2 the mass accretion rate (Calvet et al. 2004). In order to fur- 1500 K L$ 1 ther expand our sample, we include in this section not only Rsubl = 0.034 AU (2) Tsubl # L % · the five objects which we have investigated with AMBER, but ! " $ also the MWC 480 Keck-Interferometer measurement published In this relation, % denotes the ratio of the absorption efficiencies by Eisner (2007). For MWC 480, we assume L = 17 L , d = of the dust at the sublimation temperature Tsubl and at the stellar $ 140 pc, Teff = 8700 K, a P-Cygni Hα-line profile (Acke et al. effective temperature T$,i.e.% = κP(Tsubl)/κP(T$)withκP(T) = 2005), and a Brγ line luminosity log (L(Brγ)/L ) = 2.8(as Qabs(λ)Bλ(T)dλ/ Bλ(T)dλ.Weassume% = 1, corresponding $ − determined by Testi and Natta from unpublished TNG spectra; to rather large dust grains of several µminsizeand,therefore,to private communication). $the innermost region$ of the inner dust rim. We find that the measured continuum radii of three of the five objects (HD163296, HD104237, HD98922) follow the 7.1. Continuum emission R L1/2 law rather closely assuming dust sublimation tem- subl ∝ $ Earlier interferometric investigations of the continuum-emitting peratures Teff between 1300 K and 1500 K, which is in agree- region of HAeBe stars have established a relation between the ment with the results obtained in earlier NIR broadband interfer- size of the continuum-emitting region and the stellar luminosity, ometric survey observations (Monnier & Millan-Gabet 2002). suggesting that the continuum emission mainly traces hot dust For the Herbig Be stars MWC297 and V921Sco, the derived located at the dust sublimation radius. In order to check for sim- ring diameters are significantly more compact than the expected ilar trends, we plot the determined K-band continuum ring radii dust sublimation radii. A possible explanation for the small ap- as a function of the stellar luminosity L$ in Fig. 8.Forcompar- parent continuum sizes might be inclination effects, although, ison, we computed the predicted inner disk radii corresponding based on complementary information from the literature (see to dust sublimation temperatures Tsubl of 2000 K, 1500 K, and Sect. 6), we expect notable inclination only for HD 163296. Spatially and Spectrally Resolved Hydrogen Gas within 0.1 AU of T Tauri and Herbig Ae/Be Stars

–46– –45– –47–

6 1.0 Disk RWAur 0.9 Disk RYTau 11 4 Infall/Outflow 4 Disk MWC275 Infall/Outflow 0.25 0.9 2 4.0 0.8 Infall/Outflow 0.25 2 10 2 0.8 1

3.5 2 0.7 9

0.20 V 0 2 0.7 0.20 2 V 0 (degrees)

Flux (Jy) 8 V 0 (degrees) "#

Flux (Jy) 0.6 -2

3.0 (degrees)

0.6 Flux (Jy) "# -2 0.15 7 0.15 "# -4 -1 0.5 0.5 2.5 6 -4 -6 0.10 -2 0.10 2.160 2.165 2.170 2.175 2.160 2.165 2.170 2.175 2.160 2.165 2.170 2.175 5 2.160 2.165 2.170 2.175 2.160 2.165 2.170 2.175 2.160 2.165 2.170 2.175 !!"#m) !!"#m) !!"#m) !!"#m) !!"#m) !!"#m) 2.160 2.165 2.170 2.175 2.160 2.165 2.170 2.175 2.160 2.165 2.170 2.175 !!"#m) !!"#m) !!"#m) 2 6 Disk MWC758 0.40 1.6 Disk DGTau 0.40 Infall/Outflow 0.5 Disk AS353 Infall/Outflow 4 3.0 0.35 1 1.0 10 1.4 Infall/Outflow 0.35 2 5 1.2 2 0.30 0.4 0.8

V 0 2 0.30 2.5 V 0 2 (degrees) Flux (Jy)

1.0 V 0 (degrees) "#

Flux (Jy) 0.25 0.25 (degrees) Flux (Jy) "# -2 -1 0.3 0.6

0.8 "# 2.0 -5 0.20 0.20 0.6 -4 -2 0.4 -10 0.4 0.15 -6 2.160 2.165 2.170 2.175 2.160 2.165 2.170 2.175 2.160 2.165 2.170 2.175 0.2 2.160 2.165 2.170 2.175 2.160 2.165 2.170 2.175 2.160 2.165 2.170 2.175 !!"#m) !!"#m) !!"#m) !!"#m) !!"#m) !!"#m) 2.160 2.165 2.170 2.175 2.160 2.165 2.170 2.175 2.160 2.165 2.170 2.175 !!"#m) !!"#m) !!"#m)

Disk AS205A Disk DKTau 0.7 0.9 0.8 4 Infall/Outflow Disk V1057Cyg Infall/Outflow 1 7 0.9 1.0 2.5 0.8 Infall/Outflow 0.7 2 0.6 0.5

0.7 2 6 0.8

V 0 2 V

0.6 0 2 2.0 (degrees) Flux (Jy)

V 0.0

0.6 (degrees) "# Flux (Jy) 0.5 0.7 (degrees) Flux (Jy) "# 5

0.5 -2 -1 "# 0.5 -0.5 1.5 0.6 0.4 0.4 0.4 -4 4 -1.0 2.160 2.165 2.170 2.175 2.160 2.165 2.170 2.175 2.160 2.165 2.170 2.175 0.5 2.160 2.165 2.170 2.175 2.160 2.165 2.170 2.175 2.160 2.165 2.170 2.175 !!"#m) !!"#m) !!"#m) !!"#m) !!"#m) !!"#m) 2.160 2.165 2.170 2.175 2.160 2.165 2.170 2.175 2.160 2.165 2.170 2.175 !!"#m) !!"#m) !!"#m) 4 3.5 1.4 Disk MWC863 Disk DRTau Infall/Outflow 0.20 Disk V1331Cyg Infall/Outflow 0.7 2 0.45 1.1 2 2 3.0 1.2 Infall/Outflow 0.40 1.0 0.15 1

0.6 2

2.5 V 0

2 0.35 1.0

V 0 (degrees) 2 0.9 Flux (Jy)

V 0 (degrees) "#

Flux (Jy) 0.30 0.5 0.10 (degrees) Flux (Jy) "# 2.0 -2 0.8 0.8 "# -1 -2 0.25 0.4 0.7 1.5 0.20 0.6 0.05 -4 -2 2.160 2.165 2.170 2.175 2.160 2.165 2.170 2.175 2.160 2.165 2.170 2.175 0.15 0.6 2.160 2.165 2.170 2.175 2.160 2.165 2.170 2.175 2.160 2.165 2.170 2.175 !!"#m) !!"#m) !!"#m) !!"#m) !!"#m) !!"#m) 2.160 2.165 2.170 2.175 2.160 2.165 2.170 2.175 2.160 2.165 2.170 2.175 !!"#m) !!"#m) !!"#m)

0.30 0.35 6 0.8 Disk V2508Oph 15 4.0 Disk MWC480 Infall/Outflow 12 3 0.25 10 Disk MWC1080 0.40 Infall/Outflow 0.30 4 0.7 Infall/Outflow 2 3.5 0.20 5 0.25 2 0.35 0.6 2 10 1

V 0 3.0 2 0.15

V 0 (degrees) 2 0.20 Flux (Jy) 0.30 0.5 V 0

(degrees) -5 "# Flux (Jy) 0.10 (degrees) "# 2.5 -2 Flux (Jy) 8 0.15 0.4 -10 0.25 "# -1 0.05 -4 -15 -2 2.0 0.20 0.10 0.3 0.00 6 -6 2.160 2.165 2.170 2.175 2.160 2.165 2.170 2.175 2.160 2.165 2.170 2.175 -3 2.160 2.165 2.170 2.175 2.160 2.165 2.170 2.175 2.160 2.165 2.170 2.175 !!"#m) !!"#m) !!"#m) !!"#m) !!"#m) !!"#m) 2.160 2.165 2.170 2.175 2.160 2.165 2.170 2.175 2.160 2.165 2.170 2.175 !!"#m) !!"#m) !!"#m)

2 Fig. 14.— Figure 13 continued. Fig. 13.— Observed fluxes, V ,and∆φ values (gray regions indicate 1-σ confidence inter- Fig. 15.— Figure 13 continued. vals), and synthetic data for best-fit disk (dotted curves) and infall/outflow (dashed curves) models. Eisner et al. (2010, ApJ 718, 774)

VLTI Summer school 2010 - Inner regions of YSOs revealed by interferometry - F. Malbet V2129 Oph magnetic field

2.1kG octupole + 0.9kG dipole, both tilted at ~20 deg onto the rot. axis

2009 field strength *1.5-3 stronger than in June 2005 : Cold spot Hot spot non-stationary dynamo

Magnetic map consistent with hot/cold spot location (as in AA Tau) Donati et al. 2010

VLTI Summer school 2010 - Inner regions of YSOs revealed by interferometry - F. Malbet V2129 Oph MHD simulations

Tilted 1.2kG octupole + 0.35kG dipole

Disc truncation by the dipole (dominates at large distances) @ rm=6.2R*

Accretion flow redistributed into the octupole close to the star

Dipole + Dipole octupole alone

Romanova et al. 2010

VLTI Summer school 2010 - Inner regions of YSOs revealed by interferometry - F. Malbet Spectro-astrometry of Z CMa

Br only Br! + Cont. ! 0.2 2 10 10 Disk? Disk 0 0 (deg) !" 0.1 Binary 1 −10 −10

10 10 >) (mas) >) (mas) 0 0.0 −− 0 0 −− (deg) !"

−10 −10 Jet?Jet V (North V (North −0.1 −1 10 10

0 0 (deg) !"

−10 −10 −0.2 −2 0.2 0.1 0.0 −0.1 −0.2 2 1 0 −1 −2 −1000 −500 0 500 1000−1000 −500 0 500 1000

v (km/s) v (km/s) U (<−− East) (mas) U (<−− East) (mas) 6 differential-phase signals Br# + continuum Br# only –16–

0.6 (a) continuum (b) -600 to -500 kms-1 (c) -450 to -300 kms-1

0.4 N E

Δφ " photocenter displacement " spectro-astrometry K1

0.2 Offset (arcsec)

0.0 H H F F H Clear asymmetric displacements up to ~150 μas F -0.2 0.6 at red- and blue-shifted velocities. (d) -250 to -100 kms-1 (e) -100 to 50 kms-1 (f) SINFONI, -250 to -100 kms-1 0.4

0.2 Benisty et al. (2010, A&A 517, L3) Offset (arcsec)

+ 0.0 H H F F H F OSIRIS OSIRIS @ Keck

-0.2 -0.2 -0.1 0.0 0.1 0.2 -0.2 -0.1 0.0 0.1 0.2 -0.2 -0.1 0.0 0.1 0.2

Offset (arcsec) Whelan, Dougados et al. (2010)

Fig. 1.— Continuum subtracted velocity channel maps of the Z CMa [FeII] 1.257 µm jet emission with the dashed lines delinating the PA of the jet structures. Contour levels start at 1%-2% of the maximum and increase by factors of √2. The emission below S/N = 3 is

masked. Jet A (PA 245◦) is revealed, the origin of which is the Herbig Be star (H). This jet ∼ shows signs of possible precession (also see Figures 2 and 3) which explains why the chosen PA does not correspond exactly to the jet axis (on the smallest scales). A lower velocity jet,

Jet B at a PA of 235◦ is also seen. This jet extends to 0.4 and is clearly driven by the ∼ ∼ FUOR (F). Note the position of knot K1 and the lower velocity emission associated with both jets. Panel (f) is a channel map extracted from a SINFONI observation obtained in December 2008. Due to the smaller pixel scale and shorter integration time only Jet B was detected. It is included here as it conﬁrms that the FUOR is the driver of Jet B. K1 is also detected in the SINFONI data. MULTIPLE SYSTEMS

VLTI Summer school 2010 - Inner regions of YSOs revealed by interferometry - F. Malbet bg= 1 NIR ApertureShort historic synthesis of the resultsimaging of the θ Orionis C interferometry Aperture synthesissystem imaging of with the θ1 OrionisIOTA C system with IOTA E. Tatulli [Kraus et al. 2007] Principles Recombination Atmospheric turbulence The observables Interferometry & YSOs Context NIR continuum emission Emission lines Imaging State of the art Next steps Perspectives

Kraus et al. (2007, A&A 466, 649)

VLTI Summer school 2010 - Inner regions of YSOs revealed by interferometry - F. Malbet HD 98800B: orbit and masses

Boden et al. (2005, ApJ 635, 442) + radial velocities

VLTI Summer school 2010 - Inner regions of YSOs revealed by interferometry - F. Malbet 4J.-P.Bergeretal.:FirstastronomicalunitscaleimageofGW Orionis triple system the GW Ori triple system

Berger et al. (2011, A&A 529, L1) Fig. 2. Left:The nearly Reconstructed equal (2:1) H-band image offlux the ratio 2004 of the epoch inner components using the MIRA suggests so ftware.that: Three components are resolved (encircled dots) and identified- either according GW Ori to B theiris undergoing brightness a preferential (A, B and accretion C respecti eventvely that from increases brightest its disk to luminosity dimmest in H). Blue crosses indicate the positions - or that the estimate of the masses has to be revisited in favour of a more equal mass-ratio as obtainedsystem from that model is seen fitting. at lower Centre: inclination. presentation of the data from Table 2. Square point represents the alternative position for epoch 2005. Right: Predicted H-band flux ratio as a function ofthesecondarymasswithaprimarymassof3.6M!.Shadedregion is the flux-ratio constraintVLTI from Summer our school observations2010 - Inner regions of YSOs (epoch revealed by 20 interferometry04). Vertical - F. Malbet dashed lines provide the system inclination. Vertical blue lines are the secondary mass estimate corresponding to distance d ≈ 414 ± 7pc.

6 mary was Teff ≈ 5700K and the age ≈ 10 yrs.Weinterpolate did not find a solution that could simultaneously fit our motions from the 106yrs isochrones that this corresponds to a primary and the spectroscopic orbital elements. If the 20 mas separation mass of ≈ 3.6M!.Therightmostplotinfigure2showstheex- of AC is the true semi-major axis, then we would expect an or- pected flux ratio from the Siess et al. (2000) computation as a bital period of ∼3600 days, which is quite close to the period function of secondary mass. The straight line materializes the seen in the RV residuals. It is far from clear if this hierarchical observed flux ratio (i.e 0.57 ± 0.05, 2004). Additionally, we use triple is stable. Additional interferometric monitoring ofthein- the mass ratio and mass function determined from radial velocity ner pair will allow the inclination to be measured directly and to compute the requested orbital inclination as in Mathieu etal. the matter will be settled quantitatively in the near future. (1991). If we take the most recent and precise determination of the distance by Menten et al. (2007), i.e 414 ± 7pc,wecanextract 5. Conclusion from figure 2 the following information: We have presented the first direct detection of the CTTS GW Orionis as a triple system: both the secondary and tertiary – Our observation can be explained by a binary system with companion are spatially resolved.Ourresultsshowthatthe amassratioclosetounityatlowinclinationandnoexcess inner stars have H-band fluxes within a factor of 2 of each other, from circumstellar emission. Indeed, taking the uppermost which suggests that the exact status of the pair should be revis- distance estimate of 421 pc the observed flux ratio in the H ited. The outer companion is clearly detectable, consistentwitha band can be reproduced by a system seen at low inclination ◦ roughly 3500 day orbital period, in line with hints from radial ve- (i ≈ 9 )withasecondarymassof≈ 3.1M!,i.eamass-ratio locity residuals. GW Orionis offers a unique laboratory to study of the order of 0.86; the dynamical evolution of a close, young and (perhaps) hierar- – With decreasing distance the observed flux ratio requires chical triple system and its interaction with a massive disk and the addition of H-band excess to the secondary and a slight envelope. Rapid progress in determining the full orbital elements increase in inclination. For example, if we take the low- of the three components is possible with current interferometers est distance boundary of 407 pc, the expected secondary- in the coming years. to-primary ratio is ≈ 0.29, which requires a secondary circumstellar excess of roughly the same value ≈ 0.28. The re- Acknowledgements. This work was supported by ASHRA, PNPS/INSU, and ◦ quested inclination is then ≈ 11 Michelson fellowship program. This research has made use of the SIMBAD database, operated at CDS, Strasbourg, France and getCal software from the NASA Exoplanet Science Institute, Caltech. Bibliographic references were pro- Although this analysis is subject to important uncertainties, vided by the SAO/NASA Astrophysics Data System. IONIC-3 has been devel- including the primary spectral type estimate, we tested thatour oped by LAOG and CEA-LETI, and funded by the CNRS and CNES. We thank conclusions hold at a qualitative level, i.e that the GW Ori inner Russel White, Lee Hartmann and S. Meimon for useful discussions. system is probably seen at a much lower inclination (i.e ≈ 10◦) with more circumsecondary H band emission than previously thought (Mathieu et al. 1991). Reconciling our observation with References the observed eclipses by Shevchenko et al. (1998) is a puzzle we Artymowicz, P. & Lubow, S. H. 1994, ApJ, 421, 651 cannot solve. Bate, M. R. & Bonnell, I. A. 1997, MNRAS, 285, 33 Another important discovery of our work is the existence of Berger, J., Haguenauer, P., Kern, P. Y., et al. 2003, in SPIE conf., Vol. 4838, , source C. We detected C in each epoch and found clear evidence 1099–1106 Cohen, M. & Kuhi, L. V. 1979, ApJS, 41, 743 for motion with respect to the inner pair AB. We compared our Hillenbrand, L. A. & White, R. J. 2004, ApJ, 604, 741 motion to the preliminary 3850 day orbit recently inferred from Mathieu, R. D., Adams, F. C., & Latham, D. W. 1991, AJ, 101, 2184 long-term radial velocity monitoring (Latham, priv. comm.)but Menten, K. M., Reid, M. J., Forbrich, J., & Brunthaler, A. 2007, A&A, 474, 515 FUTURE PROSPECTS

PTI IOTA ISI KI VLTI CHARA

VLTI Summer school 2010 - Inner regions of YSOs revealed by interferometry - F. Malbet First steps to imaging

AB Aur

Millan-Gabet et al. (2006, ApJ 645, L77)

Isella & Natta (2005, A&A 438, 899) Closure phase provides information on departure from centro-symmetry

VLTI Summer school 2010 - Inner regions of YSOs revealed by interferometry - F. Malbet 4 S. Renard et al.: Milli-arcsecond images of the Herbig Ae star HD 163296

H band K band 1 1

0.907 0.904 10 10 0.814 0.808

0.72 0.713

0.627 0.617 (mas) (mas) " 0 0.534 " 0 0.521

0.441 0.425

relative 0.348 relative 0.329

0.255 0.234 −10 −10 0.161 0.138

0.0682 0.0419 10 0 −10 10 0 −10 relative ! (mas) First imagesrelative of! (mas) young disks Fig. 1. Reconstructed images of HD 163296 in H (left) and in K band (right), after a convolution by a Gaussian beam at the interferometer resolution. The colors are scaled with the squared root of the intensity with a min cut corresponding to the maximum expected dynamic range (see text for details). The blue ellipse traces the location of the main secondary peaks while the green ellipse corresponds to the location of the rim in the B09 model. North is up and East is left. MWC 275 H band K band Source 10 H 10 K (mas) (mas) " 0 " 0 H K relative relative

−10 −10

10 0 −10 10 0 −10

relative ! (mas) relative ! (mas) Fig. 2. Image contours of the reconstructed images of HD 163296 in H (left) and in K band (right), after a convolution by a Gaussian beam at the Model interferometer resolution. The contour vary linearly between the min cut corresponding toRenard, the maximum Malbet, expected Benisty dynamic et al. range 2010 and the image maximum.VLTI/AMBER North is up and East is left.

In the H band, the dust rim totally disappears in the recon- the Gaussian ring since the emission from the star was computed structedImage image as areconstruction ring of blobs, but still seems tois define tricky: the on the spectral energy distribution. Fig. 4 indicates clearly that outer boundary of the object. The main reasons are that firstly the star alone does not spread more than the Gaussian width of the rim represents‣ # measures only 10% of the ~1500 flux to be compared with VLTI/AMBER: with 4 pixels, which is less3 AT than inconfigs Fig. 3. 30% in the K band, and, secondly the angular resolution is not This analysis performed on existing models allows us to state high enough‣ artefacts in the H band data due while to it is inthe the K(u,v)thanks toplane the which coverage features in the reconstructed images can be trusted. We CHARA data. For the dust rim to be seen in H band, we need believe that the main secondary peaks present around the main data at longer‣ disk baselines + andring a higher model contrast inconsistent the reconstructed withcentral Benisty spot are real. Their2009 spatial distribution along an ellipse image (1000 at least, 2000 to be unambiguously seen). As in K and the intensity present between these peaks and the central band, the‣ bright 1/1000 inner disk contrast is also present in compared the H image under withspot isfirst also real. VLA However images the structure of the ring is probably not the form a big spot in the middle of the image, which would not representative of the reality, but only of the actual (u, v) plane. have existed with the presence of the star only. This inside blob More observations at different spatial frequencies will probably represents# 86%First of the total indices flux. on the morphologychange the actual position ofof these disks peaks along the ellipse and In order to demonstrate that the bright inner disk is well seen may also be smoothed. However, we believe that the inclina- in thearound reconstructed images, youngi.e. that the centralstars spot represents tion and orientation of the observed distribution of peaks along HR 5999 more energy that a star alone, reconstruction of an even simpler an ellipse is actual. Actually, this orientation and inclination is model is performed. This model is the same as the B09 model, very close to the ones fitted by B09. The second feature that we without the bright inner disk, i.e. a star surrounded by a Gaussian think is representative of the reconstructed image is that the cen- ring. The emission of the bright inner disk is then transferred to tral spot is extended and not reduce to an unresolved point as a Benisty et al. (subm. A&A) 49 Conclusion

• A major leap in less than 10 years: – ~100 objects observed so far,

– +70 refereed papers (mainly with one baseline broadband observations, but it is changing).

– new types of observations with spectral resolution, closure phases, imaging • Observations are mature enough to allow detailed modeling.

VLTI Summer school 2010 - Inner regions of YSOs revealed by interferometry - F. Malbet More images?

LkHa 101 Hot accreting planets in disks? Image simulations

Klahr & Kley (2005) Tuthill et al. (2002, ApJ, 577, 826) Actual Imaging Data Imaging Actual

Micro-jets Kloppendorf et al. (2010)

VLTI Summer school 2010 - Inner regions of YSOs revealedThiébaut by interferometry et al. (2005, - F. Ap&SS,Malbet 286, 171) More images?

LkHa 101 Hot accreting planets in disks? Image simulations

Klahr & Kley (2005) Tuthill et al. (2002, ApJ, 577, 826) Actual Imaging Data Imaging Actual

Micro-jets Kloppendorf et al. (2010)

VLTI Summer school 2010 - Inner regions of YSOs revealedThiébaut by interferometry et al. (2005, - F. Ap&SS,Malbet 286, 171) Open issues

• Disks starts to be imaged: shifts from axial symmetry?

• NIR emitting zone larger than corotation / magnetospheric radii . What implications for disk/star connection?

• Which implications do these measurements have for the initial conditions of planetary formation?

• Need to combine NIR+MIR to secure the disk structure.

• Origin of the Br$ emission ?

• Companions, formation of planets

VLTI Summer school 2010 - Inner regions of YSOs revealed by interferometry - F. Malbet