The young, tight and low mass binary TWA22AB: a new calibrator for evolutionary models ? Orbit, spectral types and temperatures determination
M. Bonnefoy, G. Chauvin, C. Dumas, A-M. Lagrange, H. Beust, M. Desort, R. Texeira, C. Ducourant, J-L. Beuzit, I. Song
AAstronomy & Astrophysics manuscript no. Twa22˙Article c (cid:13)
ESO 2018October 28, 2018
The young, tight and low mass binary TWA22AB: a new calibratorfor evolutionary models ? (cid:63)
Orbit, spectral types and temperatures determination
M. Bonnefoy , G. Chauvin , C. Dumas , A-M. Lagrange , H. Beust , M. Desort , R. Texeira , C. Ducourant ,J-L. Beuzit and I. Song Laboratoire d’Astrophysique de Grenoble, BP 53, F-38041 GRENOBLE C´edex 9, France.e-mail: [email protected] ESO, Alonso de Cordova 3107, Vitacura, Casilla 19001, Santiago 19, Chile. Instituto de Astronomia, Geof´ısica e Ciˆencias Atmosf´ericas, Universidade de S˜ao Paulo, Rua do Mat˜ao, 1226 - Cidade Universit´aria,05508-900 S˜ao Paulo - SP, Brazil. Observatoire Aquitain des Sciences de l’Univers, CNRS-UMR 5804, BP 89, 33270 Floirac, France. Department of Physics and Astronomy, University of Georgia, Athens, GA 30602, USA.Received September 08, 2008; accepted May 17, 2009
ABSTRACT
Context.
Tight binaries discovered in young, nearby associations, with known distances, are ideal targets to provide dynamical massmeasurements to test the physics of evolutionary models at young ages and very low masses.
Aims.
We report for the first time the binarity of TWA22. We aim at monitoring the orbit of this young and tight system to determineits total dynamical mass using an accurate distance determination. We also intend to characterize the physical properties (luminosity,e ff ective temperature and surface gravity) of each component based on near-infrared photometric and spectroscopic observations. Methods.
We use the adaptive optics assisted imager NACO to resolve the components, to monitor the complete orbit and to obtainthe relative near infrared photometry of TWA22 AB. The adaptive optics assisted integral field spectrometer SINFONI was also usedto obtain medium resolution ( R λ = − ff ective temperature and the surface gravity for each component of the system. Results.
Based on an accurate trigonometric distance (17.53 ± ±
21 M
Jup for the system. From the complete set of spectra, we find an e ff ective temperature T ef f = + − K for TWA22 A and T ef f = + − K for TWA22 B and surface gravities between 4.0 and 5.5 dex. From our photometry and a M6 ± / L (cid:12) ) = -2.11 ± / L (cid:12) ) = -2.30 ± Key words.
Stars: fundamental parameters, low-mass, brown dwarfs, binary (TWA22 AB): close, formation – Instrumentation: adap-tive optics, spectrographs
1. Introduction
Mass and age are fundamental parameters of stars and browndwarfs that determine their luminosity, e ff ective tempera-ture, atmospheric composition and surface gravity commonlyderived through photometric and spectroscopic observations.Evolutionary models are currently widely used in the commu-nity to infer masses of stars and brown dwarfs, but they relyon equations of states and atmospheric models non calibrated atyoung ages and at very low masses. However, direct mass mea-surements can be obtained by the mean of di ff erent observingtechniques, combining light curve studies to radial velocity oneclipsing binaries, astrometric follow-up with double lined spec-troscopy of tight binaries or measuring the Keplerian motion ofcircumstellar disks. In recent years, direct mass measurementsfor 23 pre-main sequence stars with masses ranging from 0.5 to 2 (cid:63) Based on service-mode observations (072.C-0644, 073.C-0469,075.C-0521, 076.C-0554, 078.C-0510, 080.C-0581) collected at theEuropean Organisation for Astronomical Research in the SouthernHemisphere, Chile. M (cid:12) showed discrepancies with predictions by up to a factor of 2in mass and 10 in ages (Mathieu et al. 2007). Such measurementsare more rare for lower masses ( (cid:54) . M (cid:12) ) systems. Hillenbrand& White (2004) showed that the models tend to understimate themass of the companion UZ Tau Eb (M = ± M (cid:12) , age (cid:118) = ± M (cid:12) , age (cid:118)
75 Myr) is still under debate(Boccaletti et al. 2008). And recently, the surprising discoveryof the unpredicted temperature reversal (Stassun et al. 2007) be-tween 2M035 A (M = ± M (cid:12) , age (cid:118) = ± M (cid:12) , age (cid:118) L , T e f f , g andage) and to explore as much as possible the parameter space cov-ered by evolutionary models. The influence of other parameterssuch as metallicity needs also investigation (Boden et al. 2005;Burgasser 2007). a r X i v : . [ a s t r o - ph . S R ] J un Bonnefoy et al.: The young, tight and low mass binary TWA22AB: a new calibrator for evolutionary models ?
Table 1.
Observing log.
UT Date Name Instrument Mode Filter Camera Airmass Seeing EC a Exp. Note(Grism) Time(arcsec) (%) (s)2004 / /
05 TWA22AB NACO imaging
NB 2.17
S27 1.16 1.23 16.5 202004 / /
05 TWA22AB NACO imaging
NB 1.24
S13 1.15 1.40 27.7 1202004 / /
27 TWA22AB NACO imaging
NB 1.75
S27 1.16 0.84 48.2 62005 / /
06 TWA22AB NACO imaging H S13 1.15 0.72 12.5 52006 / /
08 TWA22AB NACO imaging J S13 1.16 0.84 17.5 502006 / /
08 TWA22AB NACO imaging H S13 1.17 0.65 17.4 502006 / /
26 TWA22AB NACO imaging H S13 1.15 1.10 32.1 1502006 / /
26 TWA22AB NACO imaging K s S27 1.15 1.50 10.2 502007 / /
06 TWA22AB NACO imaging H S13 1.15 0.60 45.7 103.52007 / /
04 TWA22AB NACO imaging H S13 1.28 1.00 34.4 1502007 / /
26 TWA22AB NACO imaging H S13 1.16 0.90 39.9 1502007 / /
12 TWA22AB SINFONI spectroscopy J (2000) 25 1.18 1.56 12.1 10802007 / /
13 TWA22AB SINFONI spectroscopy J (2000) 25 1.16 0.98 18.5 10802007 / /
13 HIP049201 SINFONI spectroscopy J (2000) 25 1.13 0.86 27.0 20 Telluric Standard2007 / /
13 HIP038858 SINFONI spectroscopy J (2000) 25 1.14 0.78 28.5 20 Telluric Standard2007 / /
12 HIP035208 SINFONI spectroscopy J (2000) 25 1.17 1.17 15.2 60 Telluric Standard2007 / /
11 TWA22AB SINFONI spectroscopy H + K (1500) 25 1.16 0.89 29.5 9602007 / /
09 TWA22AB SINFONI spectroscopy H + K (1500) 25 1.52 1.00 17.2 9602007 / /
11 HIP052202 SINFONI spectroscopy H + K (1500) 25 1.19 0.93 24.7 20 Telluric Standard2007 / /
09 HIP052202 SINFONI spectroscopy H + K (1500) 25 1.55 1.22 14.0 40 Telluric Standard a Corresponds to the mean strehl ratio for the spectroscopic observations.
The TW Hydrae association (TWA) is the first co-movinggroup of young ( ≤
100 Myr), nearby ( ≤
100 pc) stars, thatwas identified near the Sun (Kastner et al. 1997). Ideal obser-vational niche for the study of stellar and planetary formation,TWA was actually the tip of an iceberg composed of hundreds ofyoung stars, spread in di ff erent groups, that were identified dur-ing the last decade (Zuckerman & Song 2004; Torres et al. 2008).TWA counts now 27 members covering a mass regime fromintermediate-mass stars to planetary mass objects (Chauvin et al.2005). Its 8.3 ± + − Myr from thephotometry, the activity and the lithium depletion. Scholz et al.(2007) show that the association is 9 + − Myr old by comparingrotational velocities with published rotation periods for a subsetof stars. Finally, Mentuch et al. (2008) estimated an age of 12 ± λ ± ∼
100 mas) binary. With a projectedphysical separation of 1 . ± .
10 AU (see Fig. 1), this systemo ff ered a unique opportunity to measure its dynamical mass andto possibly test the evolutionary model predictions at young agesusing combined photometric and spectroscopic observations.We report here the discovery of the TWA22 binarity and theresults of a dedicated 4 years observing program, using com- m a s NE Fig. 1.
VLT / NACO image of TWA22 AB obtained in H-bandwith the S13 camera on December 26th, 2007. North is up andEast is left. The field of view is 1 (cid:48)(cid:48) × (cid:48)(cid:48) .bined imaging and 3D-spectroscopy with AO. The purpose wasto measure the dynamical mass of TWA22 AB and to charac-terize the physical properties of the individual components. In §
2, we describe our AO observations with the VLT / NACO im-ager and with the VLT / SINFONI integral field spectrograph. Theassociated data reduction and spectral extraction techniques aredetailed in §
3. In §
4, we present our orbital solutions and ourspectral analysis. In § onnefoy et al.: The young, tight and low mass binary TWA22AB: a new calibrator for evolutionary models ? 3
2. Observations
TWA22 AB was observed at the 8.2 m VLT UT4 Yepun withthe Nasmyth Adaptive Optic (AO) system NAOS (Roussetet al. 2000) coupled to the High-Resolution Near-IR CameraCONICA (Lenzen et al. 1998). NAOS and CONICA (NACO)resolved the system as a tight binary for the first time on March5th, 2004. Follow-up observations were conducted during fouryears from early 2004 to end 2007. To image TWA22 AB,we used the narrow band filters:
NB 1.24 ( λ = µ m,FWHM = µ m), NB 1.75 ( λ = µ m, FWHM = µ m), NB 2.17 ( λ = µ m, FWHM = µ m). The broadband filters J ( λ = µ m, FWHM = µ m), H ( λ = µ m,FWHM = µ m) and K s ( λ = µ m, FWHM = µ m) werealso used coupled to a neutral density (attenuation factor of 80).CONICA was used with the S13 and S27 cameras to Nyquist-sample the PSF depending on the selected filter. The data wererecorded under seeing ranging from 0 . (cid:48)(cid:48) to 1 . (cid:48)(cid:48) (see Table 1).TWA22 AB was bright enough in visible to be used by NAOSfor wave-front analysis. For each observation period, ditheringaround the object in J , H and K s bands combined with noddingwere necessary to run a good sky estimation during the data re-duction process (see part 2.2). PSF references were observed atdi ff erent airmasses with identical setups. The Θ Ori C astromet-ric field (McCaughrean & Stau ff er 1994) was observed at eachepoch to calibrate the detector platescale and orientation when-ever necessary. The results are reported in Table 2. Table 2.
Mean plate scale and detector orientation for our di ff er-ent observing NACO runs. UT Date Camera Plate Scale Orientation of true northeast of the vertical(mas / pixel) (o)2004 / /
05 S27 27 . ± . − . ± . / /
27 S27 27 . ± .
10 0 . ± . / /
06 S13 13 . ± . − . ± . / /
08 S13 13 . ± .
08 0 . ± . / /
26 S13 13 . ± .
06 0 . ± . / /
06 S13 13 . ± .
10 0 . ± . / /
04 S13 13 . ± . − . ± . / /
26 S13 13 . ± . − . ± . The SINFONI instrument (Spectrograph for INtegral FieldObservations in the Near Infrared), located at the Cassegrain fo-cus of the VLT UT4 Yepun, was used to observe TWA22 ABbetween February 9 and 13 2007. SINFONI includes an integralfield spectrometer SPIFFI (SPectrograph for Infrared Faint FieldImaging, see Eisenhauer et al. (2003)), operating in the near-infrared (1.1 - 2.45 µ m). SPIFFI is assisted with the 60 actuatorsMutlti-Applications Curvature Adaptive Optic system MACAO(Bonnet et al. 2003). We used the small SPIFFI field of view of0.8” × × µ m, in-dividual integrations times of 90 s were necessary to image thesystem in the J band (1.1 - 1.4 µ m, R = H + K band (1.45 - 2.45 µ m, R = ∼
3. Data reduction and analysis
For each observation periods, the ESO eclipse reduction soft-ware (Devillard 1997) dedicated to AO image processing wasused on the complete set of raw images. Eclipse computes bad-pixel detection and interpolation, flat field correction and aver-aging pairs of shifted images with sub-pixel accuracy. The soft-ware run sky estimation on object-dithered frames using medianfiltering through the frame sequence.A deconvolution algorithm dedicated to stellar field blurredby the adaptive optics corrected point spread functions (Veran &Rigaut 1998) was applied on TWA22 AB images to accuratelyfind the position and the photometry of the companion relativeto the primary. The algorithm is based on the minimization inthe Fourier domain of a regularized least square objective func-tion using the Levenberg-Marquardt method. We used Nyquist-sampled unsaturated images of standard stars obtained the samenight as TWA22 observations with identical setups under vari-ous atmospheric conditions. These frames captured the variationof AO corrections. They were used as input point spread func-tions (PSF) to estimate the deconvolution process error. The IDL
Starfinder PSF fitting package (Diolaiti et al. 2000) confirmedthese results. IDL procedures can be downloaded athttp: // / ∼ giangi / StarFinder / index.htm Table 3.
Relative positions and contrasts of TWA 22 A and B.Magnitude di ff erences are given in the NACO photometric sys-tem. UT Date Filter Camera ∆ α J ∆ δ J ∆ m (mas) (mas) (mag)2004 / /
05 NB2.17 S27 99 ± ± ± ± / /
27 NB1.75 S13 98 ± ± ± / /
06 H-ND S13 15 ± ± ± / /
08 J S13 0.40 ± ± ± ± / /
26 H S13 -74 ± ± ± s S27 0.46 ± / /
06 H S13 -57 ± ± ± / /
04 H S13 19 ± ± ± / /
26 H S13 26 ± ± ± We used the SINFONI data reduction pipeline (1.7.1 version, seeModigliani et al. (2007)) for raw data processing. The pipelinecarries out cube reconstruction from raw detector images. Theflagging of hot and non-linear pixels is executed in a similar wayas in NACO images. The distortion and wavelength scale arecalibrated on the entire detector using arc-lamp frames. Slitletdistances are accurately measured with North-South scanningof the detector illuminated with an optical fiber. In the caseof standard stars observation, object-sky frame pairs are sub-tracted, flat fielded and corrected from bad pixels and distor-tions. Datacubes are finally reconstructed from clean science im-ages and are merged in a master cube. Spectra of standard starscleaned from stellar lines are finally used to correct the TWA22AB spectra from telluric absorptions.TWA22 A and B are centered and oriented horizontally inthe J and H + K master cubes with a field of view of 1 . (cid:48)(cid:48) × . (cid:48)(cid:48) .Atmospheric refraction induces di ff erent sources positions fordi ff erent wavelengths within the instrument field of view and in-creases with airmass. Combined with the small SINFONI fieldof view, this produces di ff erential flux losses that were noticedin the bright standard stars datacubes. This e ff ect remains lim-ited for TWA22. The cubes of 11 February 2007 appear to havesome spaxels contaminated by flux oscillations of a few ADUs.These oscillations were not negligible and blurred CO bands at2.3 µ m. They are present along the dispersion axis in the raw de-tector images of both HIP052202 and TWA22. Their amplitudesdo not remain constant in time but follow a 15.3 pixels period.We then filtered partially this contribution on each individual im-age in the Fourier space using a pass-band function. The originof the problem is likely to be related to 50 Hz pick-up noise.We used a modified version (Dumas et al. 2001) of theCLEAN algorithm (H¨ogbom 1974; Schwartz 1978) to extractseparately the flux of TWA22 A and B in each monochromaticimages contained in the datacubes. The standard star are usedas initial PSF-references. Once scaled to match the TWA22 Amaximum at the primary position and for all wavelengths, thePSF is subtracted to the TWA22 AB datacube. The sequence isrepeated to model the secondary contribution, cleaned from theprimary wings, and to provide a new PSF-reference. After a fewiterations minimizing the final quadratic residual datacube, thespectra of each individual component are extracted.The algorithm was first adapted to work on cube images.Unfortunately, the di ff erence of sampling between the X andY directions limited the sub-pixel shift accuracy. We thereforecollapsed the cube along the Y-axis in order to obtain the fluxprofile along the X direction. We chose to duplicate the primaryflux profile for the PSF model. The algorithm converged in afew iterations and produced extracted spectra in J and H + K withan extraction error less than 5%. The extracted spectra were di-vided by standard star spectra corrected from intrinsic featuresand multiplied by a black body spectrum at the standard startemperature. The SINFONI pipeline coe ffi cients were used forwavelength calibration.
4. Results
The relative positions of TWA 22 A and B (B with respect toA) at all observation epochs are reported in Table 3. The dataallows a determination of the mutual orbit of the binary. We de-fine a cartesian referential frame ( O , X , Y , Z ) where X points to-wards the north, Y toward the east and Z toward the Earth. The ( OXY ) plane corresponds thus to the plane of the sky. Then in aKeplerian formalism, the ( x ≡ ∆ δ, y ≡ ∆ α ) projected position ofthe binary onto the plane of the sky reads x = ad (cos u − e ) (cos ω cos Ω − sin ω cos i sin Ω ) + ad √ − e sin u ( − sin ω cos Ω − cos ω cos i sin Ω ) , (1) y = ad (cos u − e ) (cos ω sin Ω + sin ω cos i cos Ω ) + ad √ − e sin u ( − sin ω sin Ω + cos ω cos Ω cos i ) , (2)where a is the semi-major axis of the orbit (in AU), d is thedistance of the binary (in pc), e is the eccentricity, i is the incli-nation, Ω is the longitude of ascending node (counted from northtowards east), ω is the argument of periastron, and u is the ec-centric anomaly that describes the current location of the binaryalong its orbit. u is related to the time t by the classical Kepler’sequation2 π T ( t − t p ) = u − e sin u , (3)where T is the orbital period and t p is the time reference for peri-astron passage. Once the distance d is known (17.5 pc, Teixeiraet al. (2009)), the fit of the observational data allows to deter-mine the 7 parameters T , a , e , i , Ω , ω and t p . Then Kepler’s thirdlaw leads to the determination of the total mass M .The fit is performed via a Levenberg-Marquardt χ minimiz-ing algorithm. In practice, instead of ( e , i , Ω , ω ), the equationsare solved for the classical variables k = e cos( Ω + ω ) q = sin i Ω h = e sin( Ω + ω ) p = sin i Ω . (4)which avoids singularities towards small eccentricities and incli-nations. The uncertainties on the fitted parameters are estimatedfrom the resulting covariance matrix at the end of the fit proce-dure.Levenberg-Marquardt is an interative gradient method forconverging towards a mininum of the χ function. Dependingon the starting guess point, many local minima can be found. Inthe present case, all the attemps we made (by letting the start-ing point vary) converge towards the same solution that is listedin Table 4 and viewed in projection onto the plane of the sky inFig. 2. The available astrometric data set appears to cover almostone complete orbital period with a good sampling of the perias-tron passage. We are thus confident in our fitted solution. Theorbit appears then slighly eccentric ( e (cid:39) .
1) and viewed closeto pole-on i (cid:39) ◦ from the Earth. Table 3 summarizes the magnitude di ff erences between TWA22A and B, measured with NACO at di ff erent wavelengths.Taking into account the filter tranformations between NACO and2MASS and the photometry of the unresolved system given fromthe 2MASS Survey (Cutri et al. 2003), we derived the apparentJHK magnitudes of each component (see Table 5). Observationsunder bad seeing conditions were excluded. Based on an accu-rate distance (17.53 ± onnefoy et al.: The young, tight and low mass binary TWA22AB: a new calibrator for evolutionary models ? 5 Fig. 2.
Orbital fit of the relative positions of TWA22 AB ob-served from March 2004 to December 2007, as projected ontothe plane of the sky. The crosses represent the observational datawith their error bars, the solid line is the fitted projected orbit,and the dots correspond to the predicted positions of the modelat the times of the observations. The dashed line sketches theprojected direction of the periastron of the orbit. On Jan. 8, 2006,the binary was actually very close to periastron.
Table 4.
TWA22 B orbital parameters as determined from the fitof the astrometric data (see text for the definition of the parame-ters). The reduced χ of the fit is also reported. Reduced χ ± × − ± × − i (degrees) 27.43 ± Ω (degrees) 134.95 ± ω (degrees) 100 ± T p (years) 2006.039 ± ± M Jup ) 220 ± magnitude of the system along time. This variation is reported inthe error bars on our photometry in Table 5.Reported in a color ( J − K ) -magnitude ( M Ks ) diagram, theTWA22 A and B photometry can be compared with the pho-tometry of M dwarfs of the young, nearby associations TWHydrae ( ∼ β Pictoris ( ∼
12 Myr), Tucana-Horologium( ∼
30 Myr) and AB Doradus ( ∼
70 Myr). Predictions of evolu-tionary models of (Bara ff e et al. 1998, also named NEXTGEN)are also given at these young ages (see Fig. 3). Although age-dependent, the near-infrared photometry of TWA22 A and B ap-pears compatible for both components with that expected foryoung mid-M dwarfs but does not allow to give an age estima-tion for the binary. The NEXTGEN tracks also appears bluerthan 10 Myrs old mid-M dwarfs by ∼ Table 5.
TWA22 A and B individual magnitudes converted intothe 2MASS system.
Band m A m B M A M B (mag) (mag) (mag) (mag) J ± ± ± ± H ± ± ± ± K s ± ± ± ± Fig. 3.
Color ( J − K ) - magnitude ( M Ks ) diagram of TWA22 Aand B compared with the photometry of young M dwarfs mem-bers of the TW Hydrae (age = β Pictoris(age =
10 Myr, diamonds), Tuc-Hor (age =
30 Myr, stars) and ABDor (age =
70 Myr, crosses) associations. Distances and spec-tral types were taken from Torres et al. (2008) exepted for themembers of the TW Hydrae moving group (Mamajek 2005).Typical error bars are 0.03 for M K and 0.08 for J-K s . NEXTGEN(Bara ff e et al. 1998) 8 Myr (dashed-dotted black line), 10 Myr(long-dashed-dotted black line), 30 Myr (long-dashed black line)and 70 Myr isochrones (full black line) are overplotted. DUSTYtracks (Chabrier et al. 2000) at identical ages are added (redlines). These tracks were shifted by 0.2 mag to redder J-Ks. Thisartificially compensates the remaining incompleteness, at rela-tively warm temperatures, in the AMES linelists used in DUSTYatmospheric models (Mohanty et al. 2007). The NEXTGENtracks also appears bluer than 10 Myrs old mid-M dwarfs by ∼ To identify the numerous spectral features in the TWA22 A andB spectra between 1.10 to 2.45 µ m, the spectra were comparedwith an homogeneous medium resolution ( R λ ∼ Bonnefoy et al.: The young, tight and low mass binary TWA22AB: a new calibrator for evolutionary models ? of field dwarfs from Cushing et al. (2005) (hereafter C05; seeFig. 4, 5 and 6). The TWA22 A and B spectra appear very simi-lar. In J -band, both TWA22 A and B spectra are dominated bythe strong Na I doublet at 1.138 µ m, the deep K I lines at 1.169,1.177, 1.243 and 1.253 µ m and the presence of a broad H O ab-sorption from 1.32 to 1.35 µ m. Fe I absorptions are also detectedaround 1.170 µ m. One is blended with the 1.177 µ m K I line.We notice additional broad FeH absorptions around 1.20 µ m and1.24 µ m compatible with that expected for mid-M dwarfs as wellas the presence of the weak Q-branch at 1.22 µ m. Finally, the AlI doublet at 1.313 µ m is detected. This doublet is expected todisappear at the M-L transition.In H -band, the spectra are a ff ected by H O absorptions from1.45 to 1.52 µ m and from 1.75 to 1.8 µ m. They exhibit pro-nounced K I atomic lines at 1.517 µ m as well as weak doubletsof Mg I at 1.503 µ m and Al I at 1.675 µ m. Weak FeH absorptionsare also present. They increase from M5 to the M-L transitionCushing et al. (2003) and their depths are here compatible withthose expected for M5 to M7 field dwarfs.In K -band, strong H O absorptions appear from 1.95 to 2.04 µ m and from 2.3 to 2.45 µ m. They are typical from mid-M tomid-L dwarfs. Strong Ca I features are present from 1.9 to 2.0 µ m. They tend to disappear in the spectra of field dwarfs at theM-L transition. We identify firmly the first overtone of CO near2.3 µ m, the rest being a ff ected by the 50 Hz pick-up noise oscil-lations mentionned earlier. Additional weak Mn I, Ti I, Mg I andSi I absorptions are spread over the J, H and K bands. These linesare expected to be rapidly replaced by molecular absorptions fordwarfs later than M5. The 1.106 µ m band seems to be overlap-ping H O and TiO absorptions with increasing depths from earlyto late M dwarfs. Finally, the 1.626 µ m feature corresponds toclose OH lines, as noted in Leggett et al. (1996).To conclude, the features detected over the spectra ofTWA22 A and B between 1.1 and 2.45 µ m suggest that bothcomponents have a cool atmosphere, typical of mid to late-Mdwarfs. The continuum of both TWA22 A and B spectra were comparedto spectra of field M dwarfs obtained by C05 and McLean et al.(2003), hereafter ML03. Least squares were computed on partsof the spectra free from telluric correction residuals. From 1.10 µ m to 1.27 µ m, the TWA22 A and B continuums are well repro-duced by M6 ± H -band spectra are visually poorlyreproduced. Least squares are minimized for M9 dwarfs but with2 subclasses uncertainty. Finally, our K -band spectra are well fit-ted by M5 to M7 dwarfs. From these comparisons, we assign aspectral type M6 ± ff ectour spectral analysis. Intermediate surface gravity reduces thestrength of alkali lines (Lucas et al. 2001; Gorlova et al. 2003;McGovern et al. 2004; Kirkpatrick et al. 2006) and produces tri-angular shape in H -band interpreted as collision induced absorp-tions (CIA) of H . Our spectra were then compared with young(age (cid:46) J and K bands (see Fig. 7 and 8).They are mostly similar to M5, M6 and M7 dwarf spectra, andconsistent with the continuum fit obtained with field dwarfs. Inboth cases, our J-band spectra are slightly redder and our H-bandspectra visually are still poorly reproduced by young and old Mdwarfs. Fig. 4.
TWA22 A and B spectra compared to spectra of M5to M8 dwarfs. The M6 dwarf spectrum reproduces well the Jbands of TWA22 A and B. However, our spectra seem to havea slightly redder slope. We reported identified atomic features(blue). Molecular absorptions (FeH bands were identified byCushing et al. (2003)) are flagged in green and telluric residu-als in orange. Atomic absorptions are indicated in blue.To complete this spectral type determination, spectral in-dexes developed by ML03 (from H O bands at 1.34 µ m ( H OA ),1.79 µ m ( H OC ), 1.96 µ m ( H OD ) and at 1.2 µ m from the FeHband) were derived for TWA22 A and B (see Fig. 9). The resultswere compared to the values computed from the ML03 and C05spectral libraries of field dwarfs. They were also compared tovalues derived for young dwarfs (Slesnick et al. (2004); Lodieuet al. (2008), hereafter S04 and L08) to test the sensitivity ofthese indexes to surface gravity (age). In fact, The H O D andFeH indexes values tend to increase with age for M5-L2 dwarfs,and could disturb our analysis. We then used a mean weight ofthe individual spectral type estimations from H OA , H OC andthe recently defined Allers H O index at 1.55 µ m (see Allerset al. (2007)) to infer M ± M . ± ± ± H OD and FeH indexes for the 2 objects. Based on the K -band photom-etry and the associated bolometric corrections of Golimowskiet al. (2004), we derive a luminosity of log(L / L (cid:12) ) = -2.11 ± . log(L / L (cid:12) ) = -2.30 ± .
16 dex TWA22 B.Using the T e f f -spectral type conversion scales for intermediate-gravity objects (Luhman et al. 2003), we find an initial estima-tion of T e f f = + − K for both components. onnefoy et al.: The young, tight and low mass binary TWA22AB: a new calibrator for evolutionary models ? 7
Fig. 5.
Same as Fig. 4 but for H band. In this case, di ff erences be-tween M6V field dwarf spectrum and TWA22 spectra are impor-tant. This could arise from low gravity or flux losses introducedeither in the standard star datacube and during the extraction pro-cess. Depths of many narrow lines were studied to provide additionalinformation on the surface gravity of TWA22 AB and particu-larly on its age. Following the Sembach & Savage (1992) methodto measure pseudo-equivalent widths and their associated uncer-tainties, we derived the equivalent widths of strong atomic linesover the J , H and K bands. They were computed for narrowlines at 1.106 µ m (TiO and H O ), 1.220 µ m (FeH - Q branch),1.313 µ m (Al I) and 1.626 µ m (OH), and for the K I doublets at1.169, 1.177, 1.243 and 1.253 µ m. The results were comparedwith pseudo-equivalent widths of old field dwarfs (C05, ML03)and young Upper Sco dwarfs (S04, L08). The use of both li-brairies confirmed the strong surface gravity dependency of theK I lines, more moderate for the Al I, FeH and OH lines. Dueto the degeneracy in terms of e ff ective temperature and surfacegravity, pseudo-equivalent widths alone are not su ffi cient for aprecise spectral type determination of TWA22 AB. They remainhowever compatible with narrow lines depths of young and olddwarfs of spectral types later than M4.If we now assume a spectral type M6 ± ∼ Fig. 6.
Same as Fig. 4 but for the K band. Our spectra are stillcontaminated by 50 Hz pick-up noise residuals. This e ff ects isnoticible around 2.3 µ m. Spectra look like M6V field dwarf.other age indicators, these intermediate surface gravity featuresconfirm that TWA22 AB is likely to be a young binary system.However, their uncertainties remain significantly large to not as-sign a precise age. For a fine determination of the e ff ective temperatures and sur-face gravities of TWA22 A and B, we compared our observedspectra with theoretical templates from the GAIA model v2.6.1(Brott & Hauschildt 2005). This library is updated from Allardet al. (2001). It benefits from improved molecular dissociationconstants, additional dust species with opacities, spherical sym-metry, and a mixing length parameter 2.0 × H p . The temperatureranges in the templates from 2000 to 10 000 K and the gravityfrom -0.5 to 5.5 but we limited our analysis to 2000 K ≤ T e f f ≤ K and 3 . ≤ log ( g ) ≤ .
5. Theoretical spectra were con-volved with a gaussian to match the SINFONI spectral resolu-tion and interpolated to the TWA22 AB wavelength grid. Leastsquares minimization was applied to find templates that fit theTWA22 A and B continuum avoiding zones polluted by remain-ing oscillations.The TWA22 A least-square map in the J band constrainsthe temperature between 2800 to 3100 K and is minimized forlog(g) = T e f f = H + K band, our minimizationfailed to reproduce faithfuly the TWA22 A spectra and makes ussuspect the existence of a constant flux loss in H band duringthe spectral extraction process. To limit this systematic e ff ect, Bonnefoy et al.: The young, tight and low mass binary TWA22AB: a new calibrator for evolutionary models ?
Fig. 7.
Comparison of the TWA22 B J band spectrum (red) tospectra (black) of young Upper Sco and Orion nebulae clusterobjects (Slesnick et al. 2004; Lodieu et al. 2008). We clearlynotice that the TWA22 B spectral slope is redder than that of ref-erence spectra. The TWA22 B spectrum is very similar to thoseof young M6 and M7 dwarfs. We reported identified atomic fea-tures (blue). Molecular absorptions (FeH bands were identifiedby Cushing et al. (2003)) are flagged in green and telluric resid-uals in orange. Atomic absorptions are indicated in blue.
Fig. 8.
Comparison of the TWA22 B K band spectrum (red)to spectra of young Upper Sco standards (black) at R = Fig. 9. H O spectral indexes computed on libraries of field(McLean et al. 2003; Cushing et al. 2005) and young dwarfsspectra (Slesnick et al. 2004; Lodieu et al. 2008). Redundanciesbetween libraries have been checked. Young dwarfs M4-L2dwarfs follow trends of field dwarfs exept for H O D. The dis-persion of values for young upper Sco brown dwarfs does notseem to arise from redenning e ff ects. When overplotted, TWA22A and B values point an M6 ± σ errors derived from theresiduals to the linear fit.the minimization was applied separately in H and K bands. In H -band, the e ff ective temperature is minimized between 2600 Kto 3000 K in the full space of surface gravities explored. Theminimum is located at 2800 K and log(g) = K band iswell reproduced by 2900 and 3000 K templates irrespective ofgravity. Summing the three bands, we estimate an e ff ective tem-perature T e f f = + − K for TWA22 A. Conducting a similaranalysis for the component TWA22 B, we derive an e ff ectivetemperature T e f f = + − K. Using the Luhman et al. (2003)scale, these temperatures estimations correspond respectively toM7 + − and M7 + . − spectral types for TWA22 A and B. This is alsoconsistent with spectral types estimated in part. 4.3.2.For a fine determination of the surface gravity from syntheticspectra, we computed the equivalent widths of K I lines in the Jband on each spectra of TWA22 A and B. We then comparedthe values to TWA22 A and B to restrain the acceptable gravitydomain (see Fig. 11). We then estimate that the surface gravityis located between log(g) = onnefoy et al.: The young, tight and low mass binary TWA22AB: a new calibrator for evolutionary models ? 9 Fig. 10.
Equivalent widths in the K I lines computed on librariesof young and field dwarfs spectra. Young dwarfs spectra haveweak K I lines and therefore low equivalent widths comparedto field dwarfs as a consequence of their low surface gravity.TWA22 A and B values are reported as red and green lines re-spectively with their associated uncertainties (dashed lines). KI equivalent widths of field dwarfs point a M5V spectral typedi ff erent from our M6 ± ff ect is probably the result of weakening of KI lines associated to the estimated very young age of the system.
5. Discussion
The membership of TWA22 AB to TW Hydrae constrains theage of the system to 3-20 Myr (Barrado Y Navascu´es 2006;Scholz et al. 2007; de la Reza et al. 2006). Based on our astro-metric observations combined with an accurate distance deter-mination, we were able to derive the dynamical mass of this tightbinary. From photometry and spectroscopy, we derived near-IR fluxes, luminosity, spectral type, e ff ective temperatures andthe surface gravity of each component. Finally, spectroscopytends to indicate that both components have intermediate sur-face gravity features in their spectra, supporting a young age forTWA22 AB. Assuming the TWA age for this system, we can nowcompare the measured total dynamical mass of the binary with Fig. 11.
Iso-contours plots of K I lines equivalent widths com-puted on each spectral templates of the GAIA library v2.6.1 Redcolor indicates high values. The contours for TWA22 A (red)and B (blue) pseudo-equivalents widths values are overplotted.The long-dashed lines represent limits on TWA22 A (red) andTWA22 B (blue) temperatures. Gravity is estimated inside thesetemperatures boxes.the total mass predicted by evolutionary models of Bara ff e et al.(1998; hereafter BCAH98). Model predictions are based on theJHK photometry, the luminosity and the e ff ective temperatureof both components (see Fig 12). At 8 Myr, BCAH98 modelssystematically under-estimate the total mass by a factor of ∼ Fig. 12. Top-Left : Confrontation of the binary direct mass mea-surement to predictions of the BCAH98 masses for di ff erentages from M K . Di ff erent estimations of the age of TW Hydrae(de la Reza et al. 2006; Barrado Y Navascu´es 2006; Scholz et al.2007) are reported on the graph. Errors on the photometry arepropagated on predictions (dotted lines). Bottom-Left : Sameas top-left but for predictions from M H . Top-right : Same asTop-Left but for predictions from M J . Bottom-right : Same asTop-Left but for predictions from our estimated TWA22 A / B lu-minosities. In this plan, predictions from BCAH98, DM94 andDM97 models are nearly the same within our uncertainties.3. Evolutionary model predictions are correct and the age es-timate of TWA22 AB is currently incorrect. TWA22 ABwould be then slightly older and aged of 30 Myr,4. Finally, the TWA22 AB age is 8 Myr and evolutionary mod-els themselves do not predict correctly the physical proper-ties of very low mass stars at young ages.Before drawing important conclusions on the validity of evo-lutionary models at young ages and very low masses, we con-sider below the three first explanations.
Systematics on the estimation of the relative position and near-IR fluxes of TWA22 A and B seems very unlikely. Our analysisrelies on the use of several imaging analysis techniques (aper-ture photometry, PSF fitting, deconvolution), already used andtested in various contexts. The tight binary TWA22 AB does notrepresent itself a di ffi cult case. In addition, at each epoch, consis-tent results were found on several observing sequences obtainedduring the night.Systematics in the spectroscopic observation and extractionseem more probable for the determination of e ff ective temper-ature and surface gravity. Di ff erential flux losses over the J or H + K spectral range may have occurred due to the limited sizeof the SINFONI field of view. The impact of this e ff ect can besimulated by adding a linear slope in our spectral minimizationover the di ff erent spectral bands. The results do not change sig- nificantly our analysis based on continuum fitting or spectralindexes. It does not a ff ect at all the study of narrow lines andour surface gravity estimation. A non-linear di ff erential flux losscould be responsible for our failure to faithfully reproduce theTWA22 A and B spectra in H-band using either empirical or syn-thetic libraries. Finally, the atmosphere models were also used invarious conditions (metallicity, mixing length, di ff erent opacitytables) without drastically changing our results. Considering our data reduction and analysis as robust, we maywonder whether our basic assumptions concerning the systemitself are correct. Actually, we cannot exclude from our observa-tions that TWA22 AB is a multiple system of higher order. Oneor even both components could be in fact unresolved binaries.In such case, the derived e ff ective temperatures as well as theestimated spectral types would not be strongly modified.Dynamically speaking, the stability of the system would re-quire the separation of the individual sub-components to be sig-nificantly less than the size of the main orbit, typically by a factor3-4 (Artymowicz & Lubow 1994). To refine this estimate in thepresent case, we performed 3-body simulations using the sym-plectic code HJS (Beust 2003) dedicated to hierarchical systems.We assume the fitted orbit and split one of the two componentsinto 2 equal mass bodies, with a coplanar orbit with repect tothe wide orbit and a given semi-major axis, and assuming initialzero eccentricity. We find that the system remains stable up to aseparation of ∼ . ∼
1, leading to a physical colli-sion between the two individual components. Hence 0.4 AU canbe considered as the widest possible separation for hypotheti-cal sub-components. This is in agreement with Artymowicz &Lubow (1994).0.4 AU ( ∼
22 mas) remains below the PSF of the VLT / NACOimages ( ∼ ± − ifwe take into account the 27 ◦ inclination with respect to the planeof the sky. Even unlikely, this modulation could not to have beendetected during the monitoring (split up into two periods of 6and 1 months). But if we assume a separation of ∼ . ±
14 km s − over a 0.1 yr period. Sucha variation was not detected in the radial velocity dataset (seeTeixeira et al. (2009)). Finally, no photocenter scatter is presentaround our two-body orbital solution. A motion of ∼ ∼ onnefoy et al.: The young, tight and low mass binary TWA22AB: a new calibrator for evolutionary models ? 11 possible separations is fairly narrow, typically 0.1–0.2 AU. Alsothe system needs to be at least roughly coplanar. Given the good agreement between observations and model pre-dictions at 30 Myr, we can consider that the current age esti-mate of TWA22 AB is possibly incorrect. This age is currentlyinfered from the membership to TWA. Since the age of TWA iswell established at 8 Myr from various age diagnostics, a reliableexplanation concerns the membership to TWA itself.S03 identified TWA22 as a new member of TWA mainlyfrom the observed Li absorption line at 6708 Å and H α emis-sion line. The Li line is stronger (EW =
510 mÅ) than those ofearly-M dwarfs members of β Pic and leads S03 to suggest anage ≤
10 Myr (see Fig. 8 of S03). They derived in addition aphotometric distance of 22 pc for TWA22, confirming the prox-imity of this young system. More recently, Mamajek (2005) dis-cussed the membership of TWA22 AB to TWA based on its kine-matics properties. Mamajek (2005) estimates a probability of2% for TWA22 to be a member of TWA from an implementedconvergence point technique (de Bruijne 1999). However, Songet al. (2006) mentioned that the strong Li line of TWA22 ABis observed only for young active M dwarfs in the direction ofTWA with the exception of a very few M-type members of the β Pictoris moving group (BPMG). Finally, Mentuch et al. (2008)obtained a new visible high resolution spectrum of TWA22 AB.They confirmed the strong equivalent line of the 6708 Å lithiumabsorption (EW = ±
21 mÅ; the strongest measured in theirsample composed of young association members). They esti-mated Te ff of 2990 ±
13 K (compatible with the individual Te ff derived in Part 4.4) and log(g) = ± ≤
30 Myr; see BCAH98 predictions).To conciliate past and present results, we can consider thepossibility that TWA22 AB is a member of the BPMG. With anM6 ± β Pic, which could possi-bly explain a significantly stronger EW(Li) than those observedfor early-M dwarfs of these two associations. We can also no-tice that the λ α ) ofTWA22, used as a second indicator of youth, is compatible withthose of GJ799 A and B, M4.5 members of β Pic (Jayawardhanaet al. 2006).Finally, the projected position of TWA22 AB reveals that thesystem is isolated from other members of TWA. Its distance ismore compatible with the mean distance of the BPMG mem-bers. Teixeira et al. (2009) have recently measured the propermotion, the trigonometric parallax and the mean radial veloc-ity of TWA22 AB. They determined for the first time the helio-centric space motion of TWA22 AB. From a detailed kinematicanalysis they did not ruled out TWA22 from TW Hydrae but theydemonstrated that it was a more probable member of the BPMG.
6. Conclusions
NACO resolved for the first time the young object TWA 22 asa tight binary with a projected separation of 1.76 AU. 80 % ofthe binary orbit was covered during a 4 years observation pro-gram conducted with this instrument. We inferred a 220 ± M Jup total mass for the system and we obtained the individualmagnitudes of each component in the near infrared. This places TWA22 A and B at the substellar boundary. We complete thecharacterization of the system components with medium res-olution individual SINFONI spectra in the J , H and K bands.Our spectra were compared with empirical library of young andfield M dwarfs. We derived a M6 ± T e f f = ±
200 K forTWA22 A and T e f f = + − K for TWA22 B while the surfacegravity was constrained to 4.0 < log(g) < Acknowledgements.
We thank the referee for an excellent and thorough review,which helped to improve our manuscript greatly. We thank the ESO Paranal sta ff for performing the service mode observations. We also acknowledge partial fi-nancial support from the Agence National de la Recherche and the
ProgrammesNationaux de Plan´etologie et de Physique Stellaire (PNP & PNPS), in France.We are grateful to Andreas Seifahrt, Laird Close and Eric Nielsen, Catherine L.Slesnick, Nadya Gorlova, Katelyne N. Allers and Nicolas Lodieu for providingtheir spectra. This work would have not been possible without the NIRSPECand UKIRT libraries provided by Ian S. McLean, Michael C. Cushing and JohnT. Rayner. We also would like to thank Peter H. Hauschildt, France Allard andIsabelle Bara ff e for their inputs on evolutionary models and synthetic spectrallibraries. Finally, we thank Carlos Torres, Michael Sterzik and Ben Zuckerman,who gave use precious insights for the discussion. References
Allard, F., Hauschildt, P. H., Alexander, D. R., Tamanai, A., & Schweitzer, A.2001, ApJ , 556, 357Allers, K. N., Ja ff e, D. T., Luhman, K. L., et al. 2007, ApJ , 657, 511Artymowicz, P. & Lubow, S. H. 1994, ApJ , 421, 651Bara ff e, I., Chabrier, G., Allard, F., & Hauschildt, P. H. 1998, A&A , 337, 403Barrado Y Navascu´es, D. 2006, A&A , 459, 511Beust, H. 2003, A&A , 400, 1129Beust, H., Corporon, P., Siess, L., Forestini, M., & Lagrange, A.-M. 1997, A&A, 320, 478Boccaletti, A., Chauvin, G., Baudoz, P., & Beuzit, J.-L. 2008, A&A , 482, 939Boden, A. F., Sargent, A. I., Akeson, R. L., et al. 2005, ApJ , 635, 442Bonnet, H., Str¨obele, S., Biancat-Marchet, F., et al. 2003, in Presented at theSociety of Photo-Optical Instrumentation Engineers (SPIE) Conference, Vol.4839, Adaptive Optical System Technologies II. Edited by Wizinowich, PeterL.; Bonaccini, Domenico. Proceedings of the SPIE, Volume 4839, pp. 329-343 (2003)., ed. P. L. Wizinowich & D. Bonaccini, 329–343 Brott, I. & Hauschildt, P. H. 2005, in ESA Special Publication, Vol. 576, TheThree-Dimensional Universe with Gaia, ed. C. Turon, K. S. O’Flaherty, &M. A. C. Perryman, 565– + Burgasser, A. J. 2007, AJ , 134, 1330Chabrier, G., Bara ff e, I., Allard, F., & Hauschildt, P. 2000, ApJ , 542, 464Chauvin, G., Lagrange, A.-M., Dumas, C., et al. 2005, A&A , 438, L25Close, L. M., Lenzen, R., Guirado, J. C., et al. 2005, Nature , 433, 286Close, L. M., Thatte, N., Nielsen, E. L., et al. 2007, ApJ , 665, 736Cushing, M. C., Rayner, J. T., Davis, S. P., & Vacca, W. D. 2003, ApJ , 582, 1066Cushing, M. C., Rayner, J. T., & Vacca, W. D. 2005, ApJ , 623, 1115Cutri, R. M., Skrutskie, M. F., van Dyk, S., et al. 2003, 2MASSAll Sky Catalog of point sources. (The IRSA 2MASS All-Sky Point Source Catalog, NASA / IPAC Infrared ScienceArchive. http: // irsa.ipac.caltech.edu / applications / Gator / )D’Antona, F. & Mazzitelli, I. 1994, ApJS , 90, 467D’Antona, F. & Mazzitelli, I. 1997, Memorie della Societa Astronomica Italiana,68, 807de Bruijne, J. H. J. 1999, MNRAS , 306, 381de la Reza, R., Jilinski, E., & Ortega, V. G. 2006, AJ , 131, 2609Devillard, N. 1997, The Messenger, 87, 19Diolaiti, E., Bendinelli, O., Bonaccini, D., et al. 2000, in Presented at the Societyof Photo-Optical Instrumentation Engineers (SPIE) Conference, Vol. 4007,Proc. SPIE Vol. 4007, p. 879-888, Adaptive Optical Systems Technology,Peter L. Wizinowich; Ed., ed. P. L. Wizinowich, 879–888Dumas, C., Terrile, R. J., Brown, R. H., Schneider, G., & Smith, B. A. 2001, AJ, 121, 1163Eisenhauer, F., Abuter, R., Bickert, K., et al. 2003, in Presented at theSociety of Photo-Optical Instrumentation Engineers (SPIE) Conference, Vol.4841, Instrument Design and Performance for Optical / Infrared Ground-basedTelescopes. Edited by Iye, Masanori; Moorwood, Alan F. M. Proceedingsof the SPIE, Volume 4841, pp. 1548-1561 (2003)., ed. M. Iye & A. F. M.Moorwood, 1548–1561Golimowski, D. A., Leggett, S. K., Marley, M. S., et al. 2004, AJ , 127, 3516Gorlova, N. I., Meyer, M. R., Rieke, G. H., & Liebert, J. 2003, ApJ , 593, 1074Hillenbrand, L. A. & White, R. J. 2004, ApJ , 604, 741H¨ogbom, J. A. 1974, A&AS, 15, 417Jayawardhana, R., Co ff ey, J., Scholz, A., Brandeker, A., & van Kerkwijk, M. H.2006, ApJ , 648, 1206Kastner, J. H., Zuckerman, B., Weintraub, D. A., & Forveille, T. 1997, Science,277, 67Kirkpatrick, J. D., Barman, T. S., Burgasser, A. J., et al. 2006, ApJ , 639, 1120Kozai, Y. 1962, AJ , 67, 591Leggett, S. K., Allard, F., Berriman, G., Dahn, C. C., & Hauschildt, P. H. 1996,ApJS , 104, 117Lenzen, R., Hofmann, R., Bizenberger, P., & Tusche, A. 1998, in Presented atthe Society of Photo-Optical Instrumentation Engineers (SPIE) Conference,Vol. 3354, Proc. SPIE Vol. 3354, p. 606-614, Infrared AstronomicalInstrumentation, Albert M. Fowler; Ed., ed. A. M. Fowler, 606–614Lodieu, N., Hambly, N. C., Jameson, R. F., & Hodgkin, S. T. 2008, MNRAS ,383, 1385Lucas, P. W., Roche, P. F., Allard, F., & Hauschildt, P. H. 2001, MNRAS , 326,695Luhman, K. L., Stau ff er, J. R., Muench, A. A., et al. 2003, ApJ , 593, 1093Mamajek, E. E. 2005, ApJ , 634, 1385Mathieu, R. D., Bara ff e, I., Simon, M., Stassun, K. G., & White, R. 2007, inProtostars and Planets V, ed. B. Reipurth, D. Jewitt, & K. Keil, 411–425McCaughrean, M. J. & Stau ff er, J. R. 1994, AJ , 108, 1382McGovern, M. R., Kirkpatrick, J. D., McLean, I. S., et al. 2004, ApJ , 600, 1020McLean, I. S., McGovern, M. R., Burgasser, A. J., et al. 2003, ApJ , 596, 561Mentuch, E., Brandeker, A., van Kerkwijk, M. H., Jayawardhana, R., &Hauschildt, P. H. 2008, ApJ , 689, 1127Modigliani, A., Hummel, W., Abuter, R., et al. 2007, ArXiv Astrophysics e-printsMohanty, S., Jayawardhana, R., Hu´elamo, N., & Mamajek, E. 2007, ApJ , 657,1064Prato, L., Simon, M., Mazeh, T., Zucker, S., & McLean, I. S. 2002, ApJL , 579,L99Rousset, G., Lacombe, F., Puget, P., et al. 2000, in Presented at the Societyof Photo-Optical Instrumentation Engineers (SPIE) Conference, Vol. 4007,Proc. SPIE Vol. 4007, p. 72-81, Adaptive Optical Systems Technology, PeterL. Wizinowich; Ed., ed. P. L. Wizinowich, 72–81Scholz, A., Co ffff