Spectroscopic observations of blue stars with infrared excesses in NGC 6611
aa r X i v : . [ a s t r o - ph . S R ] J un Astronomy & Astrophysics manuscript no. rbonito c (cid:13)
ESO 2018October 1, 2018
Spectroscopic observations of blue stars withinfrared excesses in NGC 6611
R. Bonito , , L. Prisinzano , M. G. Guarcello , and G. Micela Dipartimento di Fisica e Chimica, Universit`a di Palermo, P.zza del Parlamento 1,90134 Palermo, Italy INAF – Osservatorio Astronomico di Palermo, P.zza del Parlamento 1, 90134Palermo, Italy Smithsonian Astrophysical Observatory - SAOReceived, accepted
ABSTRACT
Context.
The young open cluster NGC 6611 includes among its candidate members aclass of peculiar objects with interesting properties: blue stars with infrared IR excesses.These stars show excesses in IR bands, signature of the presence of a circumstellar disk,but optical colors typical of older field stars. In order to confirm their membership tothe cluster, it is therefore important to use new spectroscopic observations, togetherwith previous photometric data.
Aims.
We aim at confirming the membership of these objects and at investigating theirphysical properties to verify whether the observed colors are intrinsic or altered by thepresence of the disk or by the accretion processes.
Methods.
We analyze the intermediate resolution spectroscopic data obtained for asubsample of blue stars in NGC 6611 with FLAMES. In particular, we focus on thestudy of: 1) the profile of the Hα emission line, to select stars with accretion andoutflow activity; 2) the Li absorption line, used as a youth indicator; 3) the radialvelocity. Results.
Using the spectroscopic analysis, it has been possible to investigate the presenceof the Li absorption line as well as to discriminate between stars with inert or activedisk. In particular, from the analysis of the Hα emission line we were able to infer theactivity due to the accretion and outflow processes and the variability of the emission.We also investigated the binarity of the blue stars and their membership to NGC 6611. Conclusions.
From our spectroscopic analysis, we conclude that half of the sample ofblue stars (10 /
20) are confirmed members of NGC 6611 (with 6 more stars that couldalso be possible members). In conclusion, our results indicate that members of youngclusters can be found also in an anomalous region of the color-magnitude diagram, i.e.outside of the pre main sequence locus were most of the cluster members lie.
Key words. stars: formation - stars: pre-main sequence - accretion, accretion disks -individual: NGC 6611 1. Bonito et al.: BWE stars in NGC 6611
1. Introduction
The evolution of stars in the pre main sequence (PMS) phase can be investigated bystudying the properties of young stellar objects (YSOs) in star forming regions. Spectralanalysis of young clusters can shed a light on the accretion properties of YSOs, in particularin the age range 1 −
10 Myr, i.e. when the YSOs lose their disks and the accretion processstops (Hartmann et al. 1998). Disk evolution and properties are interesting also in thewider context of planetary formation.In this paper we focus on the NGC 6611 cluster. This cluster and its parental cloud M16 (the Eagle Nebula) have been deeply investigated in the past and the main parametershave been derived in detail (Guarcello et al. 2007, 2010). The authors estimated a distanceof 1750 pc, a median age of ∼ ≈ < − Hα emission line, importantsince we can infer from its profile the presence of accretion and outflow. In fact, CTTSsspectra show broad Hα emission lines, due to the accretion and outflow processes, whileWTTSs do not show any feature of accretion neither of outflow. Furthermore, spectracontaining the absorption Li line allow us to discriminate between young and old stars,since the presence of this line is an indicator of youth. In fact, during the PMS, in stars with M < . M ⊙ (where the convective region is more extended) the Li in the stellar surfacereaches the internal region where it is destroyed. The process is very efficient in particular Send offprint requests to : R. Bonitoe-mail: [email protected]
2. Bonito et al.: BWE stars in NGC 6611 in totally convective stars with mass between 0 . − . M ⊙ where the Li is depleted in 5 − V − I ) colors. Therefore the scatteringcan make the emission bluer, while the obscuration makes the optical emission fainter.It is worth noting that a mismatch between the optical and the IR catalogs could mis-interpret MS optical foreground stars as BWE candidate members. This can happen sincethe most probable mismatch occurs between a foreground optical source and a backgroundinfrared source. The combined spectral energy distribution (SED) can resemble that ofa YSO with large infrared excesses, that will be easily detected by the selection criteriaadopted in Guarcello et al. (2007) which combine and compare optical and infrared colors.Note that as all the optical data have been collected during the same night (Guarcello et al.2007), no photometric variability can be considered as a cause for the observed blue V-Icolors.If the BWE stars were old cluster members, this should be evidence of age spread andthis is crucial to constrain the theoretical models of cluster formation (e.g. Shu et al. 1987,Tassis & Mouschovias 2004, Ballesteros-Paredes & Hartmann 2007).The consequences of our study are of general interest: BWE stars are typically excludedfrom cluster memberships based on photometric data. If a significant fraction of the BWEstars analyzed here are spectroscopically confirmed members of NGC 6611, this will implythat not including them among candidate members creates a serious completeness issue,with consequences on studies, for instance, on the IMF and cluster dynamics.We note that stars with properties similar to those of BWE stars have been observednot only in NGC 6611 but also in other clusters (as in the ONC, Hillenbrand 1997; see alsoSect. 4.1 for more examples), therefore the analysis discussed here has a wide application.
2. Target selection and observations
In the context of the ESO program Obs. ID: 083.C-0837 (P.I.: Guarcello), we ob-served a sample of 194 candidate members of the cluster NGC 6611, selected from theGuarcello et al. (2010) catalog, with the aim of deriving their spectroscopic properties.Disk bearing cluster members have been selected from Guarcello et al. (2007, 2009,2010) adopting two criteria: the IRAC [3 . − [4 .
5] vs. [5 . − [8 .
0] diagram and a set ofreddening-free color indices. In the IRAC color-color diagram the locus of normal-colorstars, reddened photospheres, and disk-bearing objects can be easily separated, resulting
3. Bonito et al.: BWE stars in NGC 6611
Table 1.
List of the 20 BWE stars observed with FLAMES and their B, V, and I magni-tudes.
BWE ID B V I10151 18 .
083 17 .
091 15 . .
171 16 .
949 15 . .
59 17 .
349 15 . .
268 17 .
186 15 . .
598 15 .
733 14 . .
166 17 .
69 16 . .
936 17 .
693 16 . .
93 17 .
626 15 . .
61 17 .
358 15 . .
898 15 .
778 14 . .
409 17 .
971 16 . .
736 16 .
631 15 . .
251 17 .
749 15 . .
45 17 .
874 16 . .
075 16 .
595 14 . .
147 17 .
509 15 . .
462 18 .
609 16 . .
222 16 .
321 14 . .
659 16 .
452 15 . .
18 17 .
674 15 . in a robust selection suffering a small contamination (Allen et al. 2004). The set of red-dening free color indices used for the selection is defined in Damiani et al. (2006) andGuarcello et al. (2007, 2009). Since a detailed explanation of the properties of these indicesis beyond the scope of this paper, we just summarize their main characteristics. The readeris referred to these papers for more details. Each of these indices is defined in order to com-pare an optical color, representative of the photospheric emission, with an infrared color.Normal stars have indices with positive values even at large extinctions, while the indicesbecome more negative when stars have infrared excesses, being possible to build diagramswhere the locus of the reddened stars is well distinguished from that of the disk-bearingobjects. Besides, the indices are particularly effective (Guarcello et al. 2010) in selectingstars whose optical colors are affected by accretion or presence of light scattered into theline of sigh by the disk itself.While most of these sources show ages in agreement with the cluster age ( < V vs. V − I color-magnitude diagram of the cluster members fromthe Guarcello et al. (2010) catalog (crosses), with all the FLAMES targets indicated with
4. Bonito et al.: BWE stars in NGC 6611
Fig. 1. V vs. V − I color-magnitude diagram of NGC 6611 members from theGuarcello et al. (2010) catalog (crosses). The 20 BWE stars analyzed here are markedwith black filled circles, while the empty circles are all the targets observed with FLAMES.The solid and dashed lines are the ZAMS and the isochrone at 0 . − V − I < .
15 mag.
Table 2.
Observations log.
Config. ID Setup t exp ( s ) Run dates OB name354292 HR15N 2760.0167 2009-05-22 OBA354294 HR15N 2760.0153 2009-06-19 OBB354296 HR15N 2760.0152 2009-06-19 OBC354298 HR15N 2760.0146 2009-06-27 OBA-2354300 HR15N 2760.0154 2009-06-27 OBB-2354302 HR15N 2760.0163 2009-06-27 OBC-2 empty circles, and the 20 BWE stars analyzed here marked with black filled circles (theerror associated to V − I requested for the selection of the targets is < .
15 mag).The spectra analyzed in this program have been obtained with theGIRAFFE/FLAMES@VLT instrument operating in MEDUSA mode using the HR15Nsetup. The spectra have a resolution R ≈ S/N ≈ − Li . Hα
5. Bonito et al.: BWE stars in NGC 6611
3. Results
It is important to use independent criteria to confirm the origin of the BWE stars andtheir membership to NGC 6611. Here we focus mainly on the analysis of the Li and Hα lines, the radial velocity, RV , and vsin ( i ) measurements, using also the X-ray detectionpreviously obtained. The presence of a strong absorption Li line is commonly used as a membership criterion asit is indicative of youth of stars. The evidence of a strong Li line is a membership criterionalso for WTTSs (therefore it is a complementary criterion with respect to the study of the Hα profile discussed in Sect. 3.2).In order to subtract the background contribution due to the sky, we computed themedian of several ( ≈ −
20) spectra of the sky for each observation during the same night(same OB) and subtract it to the spectrum of the BWE analyzed (following the method ofJeffries & Oliveira 2005). After the subtraction of the median sky value, we have correctedfor the RV measurements (presented in Sect. 3.4). Then we combined the different OBsby summing the spectra, in order to improve the S/N. To derive the the equivalent width, EW , of the Li line of the BWE stars, EW ( Li ), we first normalized the spectrum of eachBWE star to the continuum using the CONTINUUM task of IRAF in a small (10 ˚A) regioncentered on the Li absorption line at 6707 . EW ( Li ) by using the SPLOTtask of IRAF. In particular, we fit the line with a Voigt profile if its EW is >
100 m˚A andwith a Gaussian if its EW is <
100 m˚A. The contribution due to the nearby Fe I line at6707.441 ˚A blended with the Li line is negligible with respect to the EW ( Li ). In fact,following Soderblom et al. (1993), we derive that EW ( F e ) ranges between 5 and 70 m˚A,while the EW ( Li ) measured for the BWE stars showing significant Li absorption line is >
180 m˚A. Therefore, we do not correct the EW ( Li ) for the contribution of the iron line.See Table 3 for the EW ( Li ) values derived for BWE stars with EW ( Li ) >
100 m˚A. Thestrongest Li lines detected have EW varying between ≈ −
450 m˚A, values consistentwith other very young clusters (e.g. Prisinzano et al. 2007).An example of a strong Li line is shown in Fig. 2 (see also Table 3).We cannot derive the Li abundance since for these stars it is not trivial to derivespectroscopically a reliable estimate of the effective temperature. In fact, classically thetemperatures can be derived from spectra in a wider spectral range: the spectra analyzedhere do not show known spectral features suitable to derive the temperature. Besides, theeffective temperatures derived from the observed colors are not reliable due to the peculiarposition of the BWE stars in the color-magnitude diagram that can be altered by thescattering and/or obscuration effects related to the presence of the disk as well as by theaccretion process at work. Since we cannot constrain the position in the CMD for the BWEstars, we cannot derive the Li abundance.We measured the EW ( Li ) of the BWE stars, focusing on those stars showing a strongabsorption line. For those cases not showing an evident absorption Li line, we investigated
6. Bonito et al.: BWE stars in NGC 6611 (A)
Fig. 2.
The strong Li line in the BWE star 15806. We have subtracted the median skyvalue, we have corrected for the RV measurements, and we combined the different OBs bysumming the spectra.if this line can be affected by veiling effect (which causes a shallowing or a suppressionof absorption lines), monitoring also other expected absorption lines in the region of thespectrum near the Li line. If there is evidence of accretion features in the Hα line profile,and if there are also only a few absorption lines observed in the spectrum, veiling effectscould explain the reduction of the absorption lines. As an example, the BWE star 2062suffers strong veiling effects. In fact, only one absorption line is visible in its spectrum(Fig. 3) at 6613 ˚A, and it is due to diffuse interstellar bands (DIB). In contrast, spectra ofunveiled BWE stars show many absorption lines, e.g. a strong Li line (see Fig. 2). Hα profiles We analyzed the Hα profiles to investigate whether the disks associated with the confirmedBWE members (see last column in Table 3 and discussion on membership in Sect. 4.1) areactive or inert disks. In fact, for stars surrounded by active disks (i.e. with accretion and/oroutflow processes still occurring), this emission line can be: a) symmetric, but showingbroad wings, or b) asymmetric. In extreme cases a P Cygni profile (i.e. a line with a deepblue absorption, under the continuum level, indicating a wind at work) or an inverse PCygni profile (i.e. a deep red absorption, under the continuum level, indicating infall onthe central star) can occur.Following Reipurth et al. (1996), the Hα line profile can be classified as: 1) type I, ifthe line is symmetric; 2) type II if the line is asymmetric, with a second emission peakwhose intensity is greater than half of the main peak; 3) type III if the line is asymmetricwith the second peak intensity lower that half of the main peak; 4) type IV if the line isasymmetric and there is an absorption feature below the continuum level (P Cygni profile).
7. Bonito et al.: BWE stars in NGC 6611
Fig. 3.
The spectrum of the BWE star 2062 (excluding the spectral range near the Hα line shown in Fig. 5) as observed during the OBA-2. This star is strongly influenced byveiling effects, as discussed in the text. In fact, the only one absorption line present in itsspectrum is due to diffuse interstellar bands.These features can be observed in the blue (B) or red (R) part of the main peak of theemission line. Therefore, a IV-R profile corresponds to the inverse P Cygni profile. Multipleprofile (indicated by ’m’) can also occur. A possible physical explanation of these profiletypes is briefly discussed Sect. 4.3.From the analysis of the Hα emission line profile and the comparison with the models,it is possible in general to estimate the accretion rate (White & Basri 2003; Muzerolle et al.2003). Natta et al. (2004) have found that the measurement of the width of the line at 10%of its peak, Hα , can be correlated to the accretion rate. However, this method cannot beused in the case of YSOs located near O stars, as in this case. In fact, since their emissionproduces the formation of HII regions contributing to the Hα emission, the line can containa nebular contribution, concentrated in the region near the peak, which is also spatiallyvariable and that cannot be easily measured independently. Therefore, a good subtractionof the nebular contribution to the emission near the Hα line cannot be achieved. Thisprevent us from measuring the EW ( Hα ) and the Hα of the star excluding the nebularcomponent.However, the contribution to the emission line due to the nebula is narrow, while theCTTSs are expected to show a broad profile, since the Hα line can also show the motionof the surrounding gas. Therefore, we measured the full width zero intensity ( F W ZI ) ofthe Hα line and we investigated its profile to select stars showing accretion or outflowsignatures. The stars with spectra showing narrow Hα line (likely of nebular origin) areclassified as candidate WTTSs, with an inert disk, since all the stars in the sample analyzedhere are BWE, i.e with IR excesses by definition.
8. Bonito et al.: BWE stars in NGC 6611
Fig. 4. Hα profiles of the 15584 BWE star for all the OBs. For this object, even if wecannot distinguish the central profile, the Hα profile clearly suggests accretion/outflowactivity.We analyzed the 50 ˚A range centered on the Hα emission line of the spectra. Wenormalized the spectra to the continuum level obtained by considering a second orderLegendre function and varying the residual rejection limit in the IRAF task CONTINUUM.We performed the F W ZI measurements for each spectrum in all the different ConfigurationIDs (see Table 2) to investigate also the emission variability, discussed in Sect. 3.3.We considered the original spectra of the BWE stars without performing the sky sub-traction. However, we have taken into account the nebular contribution on the Hα emission.In fact, we evaluated the λ min and λ max wavelength range within which the nebular emis-sion dominates the Hα emission line. These values are quite consistent for all the OBsanalyzed, ranging between 6561 − < Hα line, evenif we cannot distinguish the central profile dominated by the nebular emission (e.g. locatedbetween the two vertical lines superimposed on Fig. 4) and Fig. 5. For example, the BWEstar 2062 has a III-Bm or III-Rm profile .Figure 6 shows the histogram of the profile types observed in the BWE stars analyzedhere. Even if the profile type I (corresponding to no evidence for disk activity) is the mostfrequent (8 stars plus 3 variable, indicated by the crosses), in many cases the emission lineprofile indicates clearly accretion/outflow processes at work. In fact, the IV-B profile isevident in many spectra (3 + 4 variable), followed by the III-B type (3 variable +1 variableand within the nebular range, indicated by the asterisk). The III-R profile occurs twice(observed together with blue-shifted features); IV-R profile occurs once; II-B profile occursonce and it is variable. In 2 cases (11751 and 1320) the spectra show a wide Hα absorptionline (with narrow emission due to the nebular contribution), evidence for these objects tobe early stars, so no classification has been made for these stars. Note that we cannot discriminate between the cases of II and III profiles, both R or B, ingeneral, if this profile lies within the nebular range See also Sect. 3.3, where we account for the variability in the profiles 9. Bonito et al.: BWE stars in NGC 6611
Fig. 5. Hα profiles of the 2062 BWE star for all the OBs. For this object, even if we cannotdistinguish the central profile, the Hα profile clearly suggests accretion/outflow activity. Fig. 6.
Histogram of the Hα profile types. The crosses refer to the variable cases. Theasterisks refer to the profile types different by the I type and not observable because of thenebular emission influence. Hα variability Hα line profiles are expected to be variable since the infall, as well as the outflow, are notstationary processes. To investigate this variability, we compared all the spectra of eachOB (different nights, see Table 2) of the same star.
10. Bonito et al.: BWE stars in NGC 6611
Table 3 shows the profile type for each BWE star analyzed here. Some stars show morethan one profile type (also indicated in the Table), suggesting a time variability of the Hα profile.Among the 20 BWE stars discussed here , 12 do not show variability in the Hα emissionline: 8 show a type I profile, while in 4 cases the line shows a P Cygni (or inverse P Cygni)profile.An example of star showing variable emission line profile is 15584, which varies froma III-Bm and a IV-Bm profile type. In the BWE star 2062 both intensity and position ofthe peaks vary, while the profile type is not variable (type III). The variability of the Hα emission line for the 15584 and the 2062 BWE stars is shown in Fig. 4 and in Fig. 5. In order to measure the radial velocity and projected rotational velocity for each BWEstar, we used the FXCOR task of IRAF and performed a Fourier cross-correlation foreach spectrum with respect to a template spectrum (following Tonry & Davis 1979). Thetemplate star has been chosen among the stars without IR excess (not a BWE star), a“bona fide” photometric member, with X-ray emission, with V − I within the range of theBWE stars ( V − I = 1 . − . V − I ) = 2 .
02 mag), and with a strong Li line. The position of the peak of the cross-correlation function gives a measure of theradial velocity relative to the template (by selecting ranges without emission lines in thespectra).We have chosen the object 1660 (see Fig. 7) as the template to derive the RV and vsin ( i ). The star 1660 has been observed in all the OBs (and shows features of accre-tion/ejection in the Hα profile but an adequate number of absorption lines to be used as atemplate). We compared the RV values derived using the star 1660 as template with thosederived using another possible template, the star 16250. We verified that the two tem-plates provide consistent results. However, the template 1660 ( V = 16 .
103 mag, spectrumOBB-23) is the best choice, since it allows us to obtain smaller associated errors.The cross-correlation function (CCF) can also appear to be broad due to rotation effect.Therefore its width is a measure of vsin ( i ) of the star, if the template shows the minimumbroadening. We derived the projected rotational velocities for our sample by using the CCFwidths and interpolating these values with the relation between the width of the cross-correlation function and the rotational velocities vsin ( i ) derived in Prisinzano et al. (2007)(by using synthetic spectra with spectral resolution very similar to that of our spectra).The error associated with the vsin ( i ) (following Rhode et al. 2001) is ± vsin ( i ) / (1 + r ),where r is the Tonry & Davis (1979) parameter, which is a measure of the S/N of the peakof the CCF. The lower limit for the vsin ( i ) measure is 17 km/s, as this is the instrumentalresolution.For all the BWE stars analyzed here, we measured the RV and vsin ( i ) of each singleobservation (each OB), in order to search also for possible variability in the timescalesinvolved in our observations (almost one month). Table 3 shows our results. For those Excluding the two early-type stars, 11751 and 1320, see Table 3 11. Bonito et al.: BWE stars in NGC 6611
Fig. 7.
Spectrum near the Li line (the strongest absorption line in this enlargement) ofthe template used to compute the RV and vsin ( i ) of all the BWE stars observed withFLAMES.cases with spectra without enough absorption lines we cannot derive reliable values. Forthese objects no value is indicated in Table 3.The vsin ( i ) for most BWE stars are 20 −
40 km/s, a result which is consistent with thenature of T Tauri stars.In addition we have found one case (star 4112) in which the cross-correlation function isnot symmetric and can be fitted with two Gaussians. This is an indication that this objectis a double-lined spectroscopic binary (SB2).In Table 3, the objects showing variability in their RV are highlighted in bold. A variable RV in different OBs (different nights) suggests that the star can be a binary system andthis is the case for 8 objects.
4. Discussion
The criteria requested for a BWE star to be a candidate member of the NGC 6611 clusterare: 1) the EW ( Li ) is greater than 100 m˚A; 2) the Hα line profile indicates activityrelated to the accretion or outflow processes. A prominent absorption Li line demonstratesthe PMS nature of the object and the evidence of asymmetries and absorption feature inthe Hα emission line is a strong signature of the presence of an active disk surroundingthe star. We will also refer to previous detection of the stars in the X-ray band. As forthe RV analysis, a value within 3 σ with respect to the RV of the template chosen is acriterion which is not strong enough in this case to discriminate between members and nonmembers of the cluster because there could be a strong contamination which prevent usfrom using this criterion.
12. Bonito et al.: BWE stars in NGC 6611
Li H α H α NM ? M
NV (10) N (15) Y (5) N (4) e a r l y ( ) V (5) Y (5) Y (4)
Fig. 8.
Membership criteria used to confirm the BWE stars as members or non-members:Y = yes; N = no; V = veiling; NV = no veiling; M = member; NM = non-member; ? =to be further investigated. In brackets the total number of objects for each case.To summarize the final membership of the BWE stars discussed here, we will refer tothe flow chart shown in Fig. 8. First of all, we check for the evidence of the Li absorptionline in the spectra and confirm the BWE star as a cluster member if the EW ( Li ) valueis above the chosen threshold (“Y” in Table 3 and in Fig. 8, 5 stars). This is a strongcriterion (without considering other criteria to support the membership) because: 1) theIR excesses detected in the BWE stars suggest the presence of a circumstellar disk and 2)a symmetric Hα emission line (type I profile) cannot exclude that the star is young. Infact, if the EW ( Li ) values derived are consistent with that object being a member of thecluster, but the Hα emission line profile does not show features related to the accretionor outflow process, we can conclude that the object under inspection is a member of thecluster with an inert disk. Note that if the RV value of the object is not within 3 σ withrespect to the RV of the template, but the EW ( Li ) is above the threshold and there is alsothe detection of this object in X-rays (and this is true in 4 cases over the total of 5 BWEstars with evident Li ), the star could be a binary member with the spectrum of just onecomponent visible. Therefore, we do not reject this object to be a member of the cluster.If the EW ( Li ) measured from the spectra is below the chosen threshold (“N” in Table 3and in Fig. 8, 15 cases), we distinguish two different cases: the Li line is not present possiblydue to veiling (“V” in Fig. 8, 5 objects), or there is no evidence of veiling in the spectra(“NV” in Fig. 8, 10 objects). Then we check the Hα emission line profile of the star. If the Li line is not evident possibly due to veiling effects on the spectra and the Hα emission lineprofile is broad or asymmetric (“Y” in Table 3 and in Fig. 8, 5 stars), then we consider thisstar as a member of the cluster. In fact, the presence of asymmetric features in this line isconsistent with the presence of a circumstellar disk which can cause strong veiling effectsdue to accretion phenomena. If the Li line is not strong and there is no evidence for veiling,then: a) if there is not evidence of features related to accretion/outflow activity (“N” inFig. 8, 4 cases), then this object is considered a non member; b) if the Hα emission lineprofile is broad or asymmetric (“Y” in Fig. 8), the object should be further investigated(“?” in Table 3 and in Fig. 8, 4 cases); c) if the star is early-type (2 stars), in this case the
13. Bonito et al.: BWE stars in NGC 6611 Li line is not expected to be strong, but the object should be further investigated (“?” inTable 3 and in Fig. 8, 2 cases).Following the considerations stated above, we can conclude that half of the sampleanalyzed here (10 /
20 objects) are confirmed members from our spectroscopic analysis ofthe BWE stars. In particular, among these 10 BWE members: in 5 cases there is evidenceof a strong Li absorption line, and four out of these five BWE stars with a strong Li linealso have been detected in X-rays (15805, 15806, 13310, and 15584). The other BWE starsdo not show X-ray detection (11598 was not in the field of view of the X-ray observations);in 5 cases the Hα line criterion is fulfilled, while the Li line could be suppressed due toveiling effects.In 6 cases we cannot discriminate between member or non member (“?” in Table 3 andin Fig. 8), since: in 4 cases there is not evidence of a strong Li line nor of veiling effectseven if there are suggestions of accretion/outflow processes in the Hα line; in 2 cases the Hα line analysis suggests these objects to be early-type stars (so we expect the Li line notto be present in these stars).Four cases are not confirmed as members of the NGC 6611 cluster, since in these objectsboth criteria are not fulfilled. As explained in Sect. 4.1, we do not consider the RV value consistent with the template asa strong membership criterion. However, we use the variability of RV to check the binarityof the BWE stars. From our analysis we derive that 8 BWE stars among the total of 20can be binary systems. In particular, the star 4112 is possibly a SB2 as its cross correlationfunction has a double peaked profile. Among these 8 binary stars: 3 are confirmed membersof NGC 6611; for 3 cases we cannot discriminate between member or non member; 2 objectsare not members of the cluster. Hα profiles In order to study the accretion properties of the BWE stars, we focused on the Hα lineprofiles observed with FLAMES. In general, the Hα line profiles of young stars consistof a symmetrical emission line with one or more absorbing components superimposed,as anticipated in Sect. 3.2. Here we investigate the physical origin of the Hα profile typesdetected in the BWE stars, following Reipurth et al. (1996). The Hα line consists of severalcomponents to which different regions (characterized by different dynamical and physicalconditions) contribute. Magnetically driven winds can explain the central and the blue-shifted absorption. On the other hand, the free fall of material from the disk onto the star(leaded by a magnetic field) causes the infalling gas to reach velocity up to several hundredkm/s, decelerated in strong shocks which can explain the asymmetry of the Hα line due toa red-shifted absorption component. Models of winds accelerated close to the star as blobsand then decelerated can reproduce the double peaks observed. Therefore, type IV-B orP Cygni profiles can be explained by models of spherical winds, while type III-B profilesare reproduced by stochastic decelerating wind models. Infall material can explain type II
14. Bonito et al.: BWE stars in NGC 6611 and III-R profiles if different parameters are used in the model (see all the profile type firstpresented in Sect. 3.2).Kurosawa et al. (2006) interpret the profile types defined in Reipurth et al. (1996) alsoin term of inclination and accretion/outflow activity. In particular, the type II-B can bereproduced by models with medium-to-high inclinations and are due to fast wind acceler-ation. The difference between Type II-R profile and type II-B is interpreted as a geomet-rical effect corresponding to different viewing angles. Type III-B profiles are reproducedby models with moderate inclination and due to fast wind acceleration, while type III-Rare associated with high inclination (the disk obscures the central emission). The classicalP Cygni profile (type IV-B) where the component due to the absorption in the blue hasa sufficient velocity to be beyond the emission line wing, can be explained with a bipolarflow observed along its axis, while the inverse P Cygni profile (type IV-R) is related to thelowest mass-accretion rates at high inclinations.In the context of the BWE stars detected in NGC 6611 and analyzed here, absorptionfeatures superimposed on the emission line has been detected in 10 BWE stars among thetotal of 20 discussed, and lead to the profile type II, III, or IV. In some cases also the type Iprofile can be asymmetric itself indicating activity of the surrounding disk, and this occursin the BWE star 1336 (see also Table 3). The absorption can be due to material with alarge range of velocity. In fact, the dips observed can be narrow (as in the case of 11532,OBA) or broad (e.g. 2062). The multiple profiles (’m’) are characterized by several dips orabsorption features (as in the BWE star 15584).
There are several interpretations of the nature of the BWE stars (see Sect. 1). A qualitativeanalysis of the SEDs of this sample of stars suggests that obscuration or scattering due tothe presence of the surrounding disk can explain the observed colors of some BWE stars.On the other hand, these BWE stars could actually be cluster members significantly olderthan the mean cluster age (suggesting an age spread as in Palla et al. 2005), i.e. a firstgeneration of cluster members yet surrounded by a circumstellar disk. In this case, theircolors are purely photospheric, in fact there are other examples of CTTSs older than thetypical lifetime of the disk (e.g. MP Mus, Argiroffi et al. 2007). Our data do not allow usto discard this hypothesis since the EW of the Li is compatible with that of stars where apartial (less than a factor 10) Li depletion occurred. On the other hand, given the absenceof molecular bands, our targets are certainly stars with spectral type earlier or equal toK stars for which a partial and slow Li depletion is expected up to 10 Myr. Nevertheless,the measured EW of the Li cannot be converted accurately in Li abundances since theeffective temperatures of these objects cannot be derived with available data and thus weare not able to constrain their ages by using the Li depletion timescale inferred by stellarmodels.
15. Bonito et al.: BWE stars in NGC 6611
5. Summary and conclusions
In this paper, we have investigated the nature of the BWE stars. This is a class of candidatemembers to young clusters showing IR excesses, signature of the presence of a circumstellardisk, but with a photometric position inconsistent with the locus defined by the youngcluster members. Here we have reported the spectroscopic analysis of twenty BWE starsobserved with FLAMES to confirm their membership to the NGC 6611 cluster and tounderstand the origin of their position in the CMD.We have analyzed the Hα profiles and derived that in 10 /
20 stars the emission lineshows absorption features or asymmetry, indicative of the presence of an active disk. Insummary, half of the analyzed BWE stars presents accretion and outflow with variouscharacteristics.From the study of the Li line, we derived that in 5 cases there is evidence of a strongabsorption Li line, suggesting that these are young objects.The study of the RV allowed us to infer the binarity of 8 BWE stars, 3 of whichconfirmed as members of the cluster.We have taken advantage of the spectroscopic results derived from the analysis of the Li absorption line and the Hα profiles to confirm the membership of the BWE stars toNGC 6611. Ten (50% of the total sample) BWE stars have been confirmed as membersof the cluster. Among these, we interpret the 4 stars showing Hα type I profiles (12168,13310, 15805, and 15806) as WTTSs with an inert disk. The other 6 confirmed membershave IV-B profiles or variable III m or IV-B profile types, therefore these BWE stars canbe classified as CTTSs with active accretion disk. The Hα profiles observed in the 6 caseswhere the membership is uncertain suggest: in 2 cases that the objects are early-type stars;in 4 cases the presence of accretion or outflow activity (with absorption feature in the blueand red part of the wings). Therefore, the latter could be classified as CTTSs with activeaccretion disks, if confirmed as members of the cluster.The origin of the peculiar position of the BWE stars in the color-magnitude diagram canbe explained as due to the scattering and/or obscuration effects related to the presence ofthe disk as well as by the accretion process at work. On the other hand, we cannot excludethe hypothesis of an older population of the cluster (De Marchi et al. in preparation), butwe can conclude unambiguously that about half of the BWE stars are members of thecluster. In conclusion, in fact, a robust result of our work is that the spectroscopic analysis(as in this case, using FLAMES data) of BWE stars unambiguously demonstrates that half(10 stars out of the total 20 stars of the selected sample) of BWE stars has been confirmedas members of the cluster, with 6 more objects that are possible members as well.Several young clusters, apart from NGC 6611 discussed here, host stars with blue op-tical colors consistent with field stars or old cluster members, as Orion Nebula Cluster(Palla & Stahler 2000), NGC 6530 (Prisinzano et al. 2005), NGC 2264 (Flaccomio et al.2000), NGC 1893 (Prisinzano et al. 2011). Our analysis suggests that not including theBWE stars in the membership of young clusters with high disk fraction results in noncompleted membership, and that the investigation of the spectroscopic properties of theBWE stars can be of general interest in the context of young clusters.
16. Bonito et al.: BWE stars in NGC 6611
Acknowledgements.
We would like to thank the referee for her/his useful comments. This work was sup-ported in part by Agenzia Spaziale Italiana under contract No. ASI-INAF (I/009/10/0). L. P. and G. M.acknowledge the PRIN-INAF (PI. Desidera) for financial support. M.G.G. is supported by the Chandragrant GO0-11040X. We would like to thank Dr. D. Randazzo for her suggestions during the preparationof the manuscript.
References
Allen, L. E., Calvet, N., D’Alessio, P., et al. 2004, ApJS, 154, 363Argiroffi, C., Maggio, A., & Peres, G. 2007, A&A, 465, L5Ballesteros-Paredes, J. & Hartmann, L. 2007, Rev. Mexicana Astron. Astrofis., 43, 123Bodenheimer, P. 1965, ApJ, 142, 451Damiani, F., Prisinzano, L., Micela, G., & Sciortino, S. 2006, A&A, 459, 477Flaccomio, E., Micela, G., Sciortino, S., et al. 2000, A&A, 355, 651Guarcello, M. G., Damiani, F., Micela, G., et al. 2010, A&A, 521, A18Guarcello, M. G., Micela, G., Damiani, F., et al. 2009, A&A, 496, 453Guarcello, M. G., Prisinzano, L., Micela, G., et al. 2007, A&A, 462, 245Hartmann, L., Calvet, N., Gullbring, E., & D’Alessio, P. 1998, ApJ, 495, 385Hartmann, L. W. & Kenyon, S. J. 1990, ApJ, 349, 190Hillenbrand, L. A. 1997, AJ, 113, 1733Indebetouw, R., Robitaille, T. P., Whitney, B. A., et al. 2007, ApJ, 666, 321Jeffries, R. D. & Oliveira, J. M. 2005, MNRAS, 358, 13Kurosawa, R., Harries, T. J., & Symington, N. H. 2006, MNRAS, 370, 580Munari, U. & Carraro, G. 1996, A&A, 314, 108Muzerolle, J., Hillenbrand, L., Calvet, N., Brice˜no, C., & Hartmann, L. 2003, ApJ, 592, 266Natta, A., Testi, L., Muzerolle, J., et al. 2004, A&A, 424, 603Oliveira, J. M., Jeffries, R. D., van Loon, J. T., Littlefair, S. P., & Naylor, T. 2005, MNRAS, 358, L21Palla, F., Randich, S., Flaccomio, E., & Pallavicini, R. 2005, ApJ, 626, L49Palla, F. & Stahler, S. W. 2000, ApJ, 540, 255Prisinzano, L., Damiani, F., Micela, G., & Pillitteri, I. 2007, A&A, 462, 123Prisinzano, L., Damiani, F., Micela, G., & Sciortino, S. 2005, A&A, 430, 941Prisinzano, L., Sanz-Forcada, J., Micela, G., et al. 2011, A&A, 527, A77Reipurth, B., Pedrosa, A., & Lago, M. T. V. T. 1996, A&AS, 120, 229Rhode, K. L., Herbst, W., & Mathieu, R. D. 2001, AJ, 122, 3258Sacco, G. G., Franciosini, E., Randich, S., & Pallavicini, R. 2008, A&A, 488, 167Shu, F. H., Adams, F. C., & Lizano, S. 1987, ARA&A, 25, 23Siess, L., Dufour, E., & Forestini, M. 2000, A&A, 358, 593Soderblom, D. R., Jones, B. F., Balachandran, S., et al. 1993, AJ, 106, 1059Tassis, K. & Mouschovias, T. C. 2004, ApJ, 616, 283Throop, H. B., Bally, J., Esposito, L. W., & McCaughrean, M. J. 2001, Science, 292, 1686Tonry, J. & Davis, M. 1979, AJ, 84, 1511White, R. J. & Basri, G. 2003, ApJ, 582, 1109 . B o n i t o e t a l.: B W E s t a r s i n N G C Table 3.
Results for the 20 BWE stars discussed here, for all the OBs.ID of each BWE star; OB in which each BWE star has been observed(number of the spectrum in brackets); Hα profile type (in brackets theprofiles not visible because of the nebular influence, as explained in thetext); F W ZI of this emission line ( − if the profile is within the nebularrange); EW ( Li ); RV and vsin ( i ) ( − for cases in which there are notenough lines that can be used in FXCOR). In bold are highlighted theobjects showing a variable RV . Detection in X-rays ( − if outside of thefield of view); total membership using the criteria: RV , Li >
100 m˚A,and Hα emission line (N (V?): veiling?; Y: YES; N: NO; SKY: emissionline within the λ range effected by the nebular emission, B?: binary?;SB2: binary system with the two components resolved); the values thatare the same for each of the OBs are shown in the first OB in which thestar is detected; otherwise is indicated). ID OB (Sp)
Hα F W ZI EW ( Li ) RV vsin ( i ) X RV MEMBERSHIPprofile type (˚A) (m˚A) (km/s) (km/s)
Li Hα
TOT10151 OBA-2 (62) (I) − − − . ± .
27 17 . ± . − − − . ± . < − − − . ± . < . − . − − . ± .
43 21 . ± . − − . ± .
50 19 . ± . − − − OBA (75) (III-B) − − . ± .
46 20 . ± . . B o n i t o e t a l.: B W E s t a r s i n N G C Table 3. continued.
ID OB (Sp)
Hα F W ZI EW ( Li ) RV vsin ( i ) X RV MEMBERSHIPprofile type (˚A) (m˚A) (km/s) (km/s)
Li Hα
TOT
OBB (76) (I) − − − −
OBC (76) (I) − − . ± . < . − . − − . ± . <
17 - N N (V?) Y Y11598 OBB (71) IV-B − − . ± . <
17 -11598 OBC (80) IV-B − − . ± . <
17 -11751 OBA (88) early − − −
N - N early ?11751 OBA-2 (82) early − − − − − −
OBA (51) (I) − − − . ± . <
17 N B? N N N
OBA-2 (42) (I) − − − . ± . < OBB (51) (I) − − − . ± . < OBB-2 (42) (I) − − − . ± .
35 33 . ± . OBC (49) (I) − − − . ± . < OBC-2 (42) (I) − − − . ± .
72 34 . ± . OBA (52) (I) − − . ± .
29 17 . ± . OBA-2 (72) (I) − − . ± .
11 21 . ± . OBC (62) (I) − − . ± . < − − − . ± .
30 19 . ± . − − − . ± .
35 22 . ± . . B o n i t o e t a l.: B W E s t a r s i n N G C Table 3. continued.
ID OB (Sp)
Hα F W ZI EW ( Li ) RV vsin ( i ) X RV MEMBERSHIPprofile type (˚A) (m˚A) (km/s) (km/s)
Li Hα
TOT1233 OBC-2 (40) (I) − − − . ± .
57 21 . ± . OBA (63) (I) − − − . ± .
02 17 . ± . OBA-2 (76) (I) − − − . ± .
33 21 . ± . OBB (63) (I) − − − . ± . < OBB-2 (73) (I) − − − . ± . < OBC (70) (I) − − − . ± . < OBC-2 (73) (I) − − − . ± . < − − . ± .
87 20 . ± . − − . ± .
85 17 . ± . − − . ± .
47 18 . ± . − − . ± .
49 37 . ± . − − . ± .
75 33 . ± . − − − OBA (41) IV-B 2 . − . − − . ± .
28 17 . ± . OBA-2 (34) IV-B − − . ± .
38 17 . ± . OBB (42) IV-B − − . ± .
56 20 . ± . OBB-2 (38) IV-B − − . ± .
37 19 . ± . OBC (33) (I asymm.) − − . ± .
56 17 . ± . OBC-2 (38) IV-B − − . ± . < . − − . ± .
53 20 . ± . . B o n i t o e t a l.: B W E s t a r s i n N G C Table 3. continued.
ID OB (Sp)
Hα F W ZI EW ( Li ) RV vsin ( i ) X RV MEMBERSHIPprofile type (˚A) (m˚A) (km/s) (km/s)
Li Hα
TOT13407 OBC (65) IV-B − − . ± .
39 22 . ± . − − . ± .
80 22 . ± . . − . − − − N N N (V?) Y Y1455 OBB-2 (36) IV-Bm − − . ± .
52 23 . ± . − − . ± .
47 24 . ± . OBA (86) III-Bm 11 . − . − . ± .
06 40 . ± . OBB (86) IV-Bm + III-R 0 . ± .
33 38 . ± . OBC (92) IV-Bm 0 . ± .
12 41 . ± . −
392 (OBA-2,OBC-2) 5 . ± .
35 40 . ± . − . ± .
04 43 . ± . − . ± .
79 42 . ± . OBA (96) (I) − − . ± .
14 34 . ± . OBB (85) (I) − − . ± . . ± . OBC (86) (I) − − . ± .
16 33 . ± . . − . − − − N - N (V) Y Y2062 OBA-2 (29) III-B or Rm − − − − − − − − − − − − . B o n i t o e t a l.: B W E s t a r s i n N G C Table 3. continued.
ID OB (Sp)
Hα F W ZI EW ( Li ) RV vsin ( i ) X RV MEMBERSHIPprofile type (˚A) (m˚A) (km/s) (km/s)
Li Hα
TOT2062 OBC-2 (24) III-B or Rm − − − . − . − − . ± . <
17 N N N Y ?2594 OBB-2 (4) III-Bm − − . ± .
61 25 . ± . − − . ± .
77 23 . ± . OBA (7) IV-Rm 3 . − . − . ± .
68 26 . ± . OBB (7) IV-Rm − . ± .
90 23 . ± . OBC (5) IV-Rm − . ± .
50 20 . ± .1