A&A 379, L9-L12 (2001)
DOI: 10.1051/0004-6361:20011337

UVES spectra of young brown dwarfs in ChaI: Radial and rotational velocities

V. Joergens¹ - E. Guenther²

1 - Max-Planck-Institut für Extraterrestrische Physik, Giessenbachstrasse 1, 85748 Garching, Germany
2 - Thüringer Landessternwarte Tautenburg, Karl-Schwarzschild-Observatorium, Sternwarte 5, 07778 Tautenburg, Germany

Received 3 September 2001 / Accepted 24 September 2001

Abstract
Based on high-resolution UVES spectra we found that the radial velocity (RV) dispersion of nine of twelve known young bona fide and candidate brown dwarfs in the ChaI dark cloud is 2.0 kms^-1, i.e. significantly smaller than the RV dispersion of T Tauri stars in ChaI (3.6 kms^-1) and only slightly larger than the dispersion of the surrounding molecular gas (1.2 kms^-1) (Mizuno et al. 1999). This result indicates that the majority of these brown dwarfs are not ejected with high velocity out of a dense region as proposed by some formation scenarios for brown dwarfs. The mean RV values are consistent with the objects being kinematic members of ChaI. The RV dispersion of the T Tauri stars confined to the ChaI region is based on a compilation of T Tauri stars with known RVs from the literature plus three T Tauri stars observed with UVES and unpublished RVs for nine T Tauri stars. Time-resolved spectroscopy revealed RV variations for five out of nine of the bona fide and candidate brown dwarfs in ChaI, which could be due to orbiting planets or surface features. Furthermore we derived rotational velocities $v\,\sin i$ and the Lithium 6708Å equivalent width.

Key words: stars: low-mass, brown dwarfs - stars: formation - planetary systems: formation - stars: individual: ChaH$\alpha$1 to 12, B34, CHXR74, Sz23

1 Introduction

It is still an open question how brown dwarfs form. They may form like planets in disks around normal stars. The "brown dwarf desert'' is not supporting this theory unless the brown dwarfs are ejected after their formation and can now be detected as freely floating brown dwarfs. Sterzik & Durison (1999) proposed the ejection of brown dwarfs from cloud cores by three-body encounters.

Brown dwarfs may also form like stars by collapse of a cloud but do not become stars because they formed out of relatively small cores. There are several observational hints to significant circumstellar material around young brown dwarfs (e.g. Comerón et al. 2000 for ChaI) supporting a star-like formation. Recently Reipurth & Clarke (2001) proposed that brown dwarfs form by cloud fragmentation but failed to become stars because they have been ejected in the early accretion phase.

Understanding the formation of brown dwarfs is important because the number of brown dwarfs is at least equal to the number of stars (Reid et al. 1999). The ejection of a (proto-) brown dwarf in the early accretion phase or later might have left observable signs. In this paper we report on high-resolution spectroscopy of young bona fide and candidate brown dwarfs in ChaI, which indicate that there is probably no run-away brown dwarf among them.

The Chamaeleon cloud complex is comprised of three main clouds. ChaI is the most active star forming cloud among them and one of the most promising grounds for observational projects on very low-mass objects. Recently twelve low-mass M6-M8-type objects, ChaH$\alpha$1 to 12, have been detected in the center of ChaI with ages in the range of 1 to 5 Myrs (Comerón et al. 1999, 2000). Their masses are below or near the border line separating brown dwarfs and very low-mass stars. Four of them are confirmed bona fide brown dwarfs ($\sim$30 to 50  $M_{\mbox{\tiny Jup}}$) (Neuhäuser & Comerón 1998, 1999; Comerón et al. 2000).

2 Spectroscopy

Using the high-resolution Echelle spectrograph UVES (Dekkeret al. 2000) at the VLT, we carried out spectroscopic observations of nine bona fide and candidate brown dwarfs (M6-M8) (ChaH$\alpha$1 to 8 and 12) and three mid-M-type T Tauri stars (B34, CHXR74 and Sz23) in ChaI between March and May 2000. For each object at least two spectra separated by a few weeks have been obtained.

UVES is ideal for high precision measurements of RVs of faint objects. The spectra have been taken using the red arm of the two-armed UVES spectrograph equipped with a mosaic of two CCDs (blue side of red arm: EEV; red side of red arm: MIT-LL). The wavelength coverage is 6600Å to 10400 Å and the spectral resolution $\rm\lambda / \Delta \lambda=40\,000$. The spectra have been taken with a slit of 1 $\hbox{$^{\prime\prime}$ }$ to 1.2 $\hbox{$^{\prime\prime}$ }$ during a seeing of 0.4 $\hbox{$^{\prime\prime}$ }$ to 1.0 $\hbox{$^{\prime\prime}$ }$.

The spectra have been optimally extracted, including bias correction, flat fielding, cosmic ray elimination and sky subtraction using the Echelle package of IRAF.

3 Radial velocities

	Figure 1: Small part of an UVES Echelle spectrum of ChaH$\alpha$4.
Open with DEXTER

For the determination of precise RVs we used the telluric O₂ lines as wavelength reference. They are produced in the Earth atmosphere and show up in the red part of the optical spectral range (cp. Fig.1). They are stable up to $\sim$20 ms^-1 (Balthasar et al. 1982; Caccin et al. 1985).

Heliocentric RVs are determined by cross-correlating plenty of stellar lines of the object spectra against a template spectrum and locating the correlation maximum. A mean UVES spectrum of CHXR74 served as template. The zero point of its velocity has been determined by means of the Lithium line at 6708Å.

For the measurements of Doppler shifts of stellar features we carefully selected appropriate wavelength regions, which are not affected by telluric lines, cosmic defects of the CCD or fringes of the CCD in the near IR. We achieved a RV precision of $\sim$200ms^-1 for a S/N of $\sim$20 in agreement with the expectations for this S/N (Hatzes & Cochran 1992). The precision of the RVs is limited by the S/N of the spectra and not by systematic effects.

The mean heliocentric RVs are given in Table1. They are consistent with RVs measured by Neuhäuser & Comerón (1999) within the measurements uncertainties for ChaH$\alpha$1, 3, 4, 5, 7, B34 and CHXR74. However, the values for ChaH$\alpha$8 and Sz23 are discrepant by more than 1$\sigma$ and the RVs for ChaH$\alpha$2 and 6 by 2$\sigma$. This may be a hint to long-term single-lined spectroscopic binaries.

Due to a low S/N for ChaH$\alpha$7 only a mean RV and an upper limit of $v\,\sin i$ and the Lithium equivalent width could be determined.

4 Kinematics of brown dwarfs and T Tauri stars in the ChaI cloud

   

Table 1: Parameters derived from UVES spectra: mean heliocentric radial velocity RV in [kms^-1], projected rotational velocity $v\,\sin i$ and equivalent width EW of lithium (6708Å). Last columns: (sub)stellar radii and upper limits for rotational periods.

object	RV	$v\,\sin i$	EW(Li)	$R_{\star}$	$P_{\rm max}$
	$\pm$0.2	[kms-1]	[Å]	[$R_{\odot}$]	[d]
ChaH$\alpha$1	15.5	$7.6\pm2.2$	$0.14 \pm0.10$	0.46	3.1
ChaH$\alpha$2	15.4	$12.8\pm1.2$	$0.30 \pm0.08$	0.73	2.9
ChaH$\alpha$3	14.2	$21.0\pm1.6$	$0.26 \pm0.08$	0.77	1.9
ChaH$\alpha$4	15.4	$18.0\pm2.3$	$0.46 \pm0.11$	0.89	2.5
ChaH$\alpha$5	15.5	$15.4\pm1.8$	$0.44 \pm0.10$	0.83	2.7
ChaH$\alpha$6	16.1	$13.0\pm2.8$	$0.32 \pm0.08$	0.68	2.6
ChaH$\alpha$7	13.7	$\leq$10	$\leq$0.09	0.37	$\geq$1.9
ChaH$\alpha$8	14.5	$15.5\pm2.6$	$0.33 \pm0.15$	0.59	1.9
ChaH$\alpha$12	13.8	$25.7\pm2.6$	$0.37 \pm0.10$	0.66	1.3
B34	16.5	$15.2\pm1.9$	$0.60 \pm0.13$
CHXR74	15.1	$14.1\pm1.6$	$0.63 \pm0.04$
Sz23	15.8	$17.3\pm3.4$	$0.31 \pm0.08$

Neuhäuser & Comerón (1999) determined a mean RV of $\sim$14.6kms^-1 and a total range of 11kms^-1for ChaH$\alpha$1 to 8 from medium resolution spectra. The measurements of precise RVs for ChaH$\alpha$1 to 8 and ChaH$\alpha$12 with UVES allow us to study the kinematics of these bona fide and candidate brown dwarfs with high accuracy. We find that their RVs lie close together, only spanning a range of 2.4kms^-1. The mean RV is 14.9kms^-1 and the RV dispersion is 2.0 kms^-1 (cp. Table1 and Fig.2).

The ChaH$\alpha$ objects are located at the periphery of one of the six cloud cores (No.5) in ChaI in a region with a relatively high density of young stellar objects. The mean RV of the molecular gas of the ChaI cloud and also of the cloud core No.5 is 15.3kms^-1 (Mizuno et al. 1999). The mean RV of the studied brown dwarfs is consistent with this velocity of the gas and therefore with the objects being kinematic members of ChaI. Mizuno et al. determined the RV dispersion of the gas of core No.5 to 1.2kms^-1. The brown dwarfs show a slightly larger RV dispersion (2.0kms^-1) than the surrounding molecular gas but basically reflect the motion of the gas.

The relatively small RV dispersion of the studied bona fide and candidate brown dwarfs gives suggestive evidence that there is no run-away brown dwarf among them. We cannot rule out that some of them have a larger space velocity dispersion since RVs are tracing only space motions in one dimension. Nevertheless our finding indicates that the majority of the nine ChaH$\alpha$ objects are not ejected with high velocities out of a dense region as proposed in formation scenarios (Sterzik & Durisen 1999; Reipurth & Clarke 2001). Some or all of the brown dwarfs may still have been "ejected'' with less than escape velocity into an extended orbit around another component of a multiple system. None of the studied brown dwarfs is closer to a known T Tauri star than 4600AU, i.e. it is unlikely that one of them is still bound to a star. There is still the possibility that the parent star itself was later ejected with escape velocity and left an unbound brown dwarf.

	Figure 2: Histogram of mean RVs of nine bona fide and candidate brown dwarfs (hashed) and for 27 T Tauri stars in ChaI.
Open with DEXTER

   

Table 2: List of T Tauri stars in ChaI with known RVs with a precision of 2kms^-1 or better from: ¹ Dubath et al. (1996), ² Covino et al. (1997), ³ Guenther et al., in prep. and from UVES observations (B34, CHXR74, Sz23: see Table1).

object	RV [kms-1]	$v\,\sin i$ [kms-1]
Sz41	$16.5 \pm1.3$	$16.6 \pm3.3$
Sz61	$14.9 \pm0.8$	$38.0 \pm1.5$
Sz91	$14.7 \pm0.3$	$15.2 \pm1.5$
Sz111	$15.1 \pm0.5$	$16.2 \pm2.3$
Sz151	$17.2 \pm0.6$	4.9 $\pm <$7.9
Sz191	$13.5 \pm0.6$	$35.7 \pm1.1$
Sz201	$15.4 \pm1.3$	$23.9 \pm2.7$
Sz361	$12.9 \pm0.9$	7.5 (+2.7-4.8)
Sz411	$13.9 \pm0.4$	$34.4 \pm1.4$
Sz421	$15.1 \pm0.3$	$27.2 \pm1.3$
RXJ1109.4-76272	$13.1 \pm2.0$	$14.5 \pm3$
B332 (CHXR25)	$13.0 \pm2.0$	-
F342	$14.0 \pm2.0$	$55 \pm3$
RXJ1111.7-76202	$19.0 \pm2.0$	$23 \pm3$
RXJ1112.7-76372	$16.0 \pm2.0$	$11 \pm3$
CS Cha3	14.9
CT Cha3	$15.5 \pm1.4$
CV Cha3	$15.6 \pm0.9$
SX Cha3	$13.4 \pm0.9$
SY Cha3	$12.7 \pm0.1$
TW Cha3	$15.7 \pm1.2$
VW Cha3	$15.1 \pm0.1$
VZ Cha3	14.7
WY Cha3	12.1

We compared the RV distribution of the bona fide and candidate brown dwarfs also with those of T Tauri stars. Radio observations by Mizuno et al. (1999) revealed that the three main clouds in the Chamaeleon star forming region differ to a large extent in their star formation properties and also the RVs of the molecular gas vary between the clouds (differences up to 3.6km^-1), whereas they are relatively constant ($\sim$1kms^-1) within each single cloud. Therefore it is reasonable to compare the kinematics of the brown dwarfs in ChaI with those of T Tauri stars also confined to the ChaI star forming cloud.

RVs of T Tauri stars in ChaI have been measured by Dubath et al. (1996), Covino et al. (1997), Neuhäuser & Comerón (1999) and by us and are listed in Table2. Furthermore unpublished RVs of T Tauri stars based on FEROS spectra have been included (Guenther et al., in prep.). The T Tauri stars have RVs in the range of [12.1$\dots$ 19.0kms^-1] with a mean RV of 14.9kms^-1 and a dispersion of 3.6kms^-1 (cp. Fig.2).

The mean RV of the T Tauri stars matches very well the ones of the bona fide and candidate brown dwarfs in ChaI, whereas the dispersion as well as the total range of RVs of the T Tauri stars is significantly larger than the ones of the brown dwarfs. The stellar activity of T Tauri stars probably account for this discrepancy since it has been shown that T Tauri stars exhibit a "RV noise'' of the order of $\sim$2kms^-1 due to stellar activity (Guenther et al. 2000). Moreover it may also play a role that the brown dwarfs and brown dwarf candidates are all situated in a small area at the periphery of one cloud core whereas the T Tauri stars are distributed over the whole ChaI region. The RVs of the six cloud cores within ChaI differ by $\pm$0.25kms^-1 (Mizuno et al. 1999).

5 Radial velocity variations

The analysis of UVES spectra taken at different times yielded to the detection of significant RV variations for five out of nine bona fide and candidate brown dwarfs in ChaI. They could be caused by motion of the brown dwarf due to an orbiting companion or by magnetically induced surface features. The detected RV variations are of the order of 1kms^-1. If they are caused by an orbiting object this would have a mass of a few Jupiter masses depending on the orbital parameters (Joergens et al. 2001). More details will be given later after additional follow-up observations.

   
6 Projected rotational velocities $v \sin i$

Projected rotational velocities have been measured using the telluric lines for determining the instrumental profile of the spectrograph and assuming a solar-like center-to-limb variation. We derived $v\,\sin i$ values in the range from 8kms^-1 to 26kms^-1 (Table 1). Figure3 shows the $v\,\sin i$ distribution of the brown dwarfs in comparison with that of T Tauri stars listed in Table2. Both distributions peak at a $v\,\sin i$ of 17 to 21kms^-1, indicating that there is no crucial difference between the rotational velocities of the studied brown dwarfs and T Tauri stars.

Based on $v\,\sin i$ and the radius of the object an upper limit of the rotational period can be derived:

$$P_{\rm max}[\mbox{d}]= 50.6145 \frac{R [{R}_{\odot}]}{(v \sin i) [\mbox{km}\,\mbox{s}^{-1}]}\cdot$$ (1)

The radii of the objects are estimated from bolometric luminosities and effective temperatures given by Comerón et al. (2000). The approximate upper limits for the rotational periods are between one and three days (Table1).

	Figure 3: Histogram of $v\,\sin i$ for brown dwarfs (hashed) and T Tauri stars in ChaI. For details see Sect.6.
Open with DEXTER

7 Lithium equivalent width

All brown dwarfs and low-mass stars observed with UVES show lithium absorption at 6708Å. The measured equivalent width is given in Table1 for all objects. For ChaH$\alpha$7 only an upper limit of EW(Li) was determined due to a low S/N. For eight of the twelve bona fide and candidate brown dwarfs in ChaI lithium detection and equivalent width measurements have been reported by Neuhäuser & Comerón (1999). Our UVES spectra allow us to add the Lithium detection of ChaH$\alpha$12.

8 Summary

Based on high-resolution UVES spectra of bona fide and candidate brown dwarfs and of T Tauri stars in ChaI we determined RVs with a precision of $\sim$200 ms^-1.

We found that the RV dispersion of nine of the twelve bona fide and candidate brown dwarfs in ChaI is 2.0 kms^-1, i.e. significantly smaller than the RV dispersion of T Tauri stars (3.6 kms^-1) in this cloud and slightly larger than the dispersion of the surrounding molecular gas (1.2 kms^-1) (Mizuno et al. 1999). This result indicates that the majority of the bona fide and candidate brown dwarfs in ChaI are not ejected with high velocities out of a dense region as proposed in some formation scenarios (Sterzik & Durisen 1999; Reipurth & Clarke 2001). Some or all of the brown dwarfs may still have been thrown with less than escape velocity into an extended orbit around another component of a multiple system.

The kinematic study of the T Tauri stars in ChaI was based on a compilation of all T Tauri stars located in ChaI where RVs were known to better than 2 kms^-1(including our UVES data for three T Tauri stars as well as unpublished data taken with FEROS). The mean RV of the T Tauri stars is 14.9 kms^-1, i.e. the same as for the brown dwarfs. The larger RV dispersion of the T Tauri stars can at least partly be attributed to "RV noise'' caused by stellar activity (Guenther et al. 2000).

Time-resolved spectroscopy revealed significant RV variations for five of the bona fide and candidate brown dwarfs in ChaI, which may be caused by orbiting planets or spots on the surface (Joergens et al. 2001).

Determination of $v\,\sin\,i$ showed that the brown dwarfs do not rotate crucially faster than the T Tauri stars in the same cloud. The $v\,\sin\,i$ values together with radii derived from the literature constrained the maximum rotational periods for the individual brown dwarfs to one to three days. Last not least we detected lithium in absorption for all studied objects and measured the equivalent width - for ChaH$\alpha$12 this is the first lithium detection.

The data presented in this paper showed once more that the brown dwarfs in ChaI form a very homogenous sample and are highly interesting astrophysical objects.

Acknowledgements

We acknowledge helpful discussions on the topic of this paper with R. Neuhäuser, R. Durisen, K. Tachihara and F. Comerón. Furthermore we like to acknowledge the brilliant work of the ESO staff at Paranal. V.J. acknowledges grant from the Deutsche Forschungsgemeinschaft (Schwerpunktprogramm "Physics of star formation'').

References

 

• Balthasar, H., Thiele, U., & Wöhl, H. 1982, A&A, 114, 357 In the text NASA ADS  
• Caccin, B., Cavallini, F., Ceppatelli, G., Righini, A., & Sambuco, A. M. 1985, A&A, 149, 357 In the text NASA ADS  
• Comerón, F., Rieke, G. H., & Neuhäuser, R. 1999, A&A, 343, 477 In the text NASA ADS  
• Comerón, F., Neuhäuser, R., & Kaas, A. A. 2000, A&A, 359, 269 In the text NASA ADS  
• Covino, E., Alcalá, J. M., Allain, S., et al. 1997, A&A, 328, 187 In the text NASA ADS  
• Dekker, H., D'Odorico, S., Kaufer, A., Delabre, B., & Kotzlowski, H. 2000, in SPIE 4008, ed. M. Iye, A. Moorwood, 534 In the text  
• Dubath, P., Reipurth, B., & Mayor, M. 1996, A&A, 308, 107 In the text NASA ADS  
• Guenther, E. W., Joergens, V., Neuhäuser, R., et al. 2000, in Birth and Evolution of Binary Stars, ed. B. Reipurth, & H. Zinnecker, IAU Symp., 200, in press In the text  
• Hatzes, A. P., & Cochran, W. D. 1992, High Resolution Spectroscopy with the VLT, ESOworkshop, ed. M. H. Ulrich, 275 In the text  
• Joergens, V., Guenther, E., Neuhäuser, R., et al. 2001, in Origins of stars and planets: The VLT view, ESO workshop, in press [astro-ph/0106185] In the text  
• Mizuno, A., Hayakawa, T., Tachihara, K., et al. 1999, PASJ, 51, 859 In the text NASA ADS  
• Neuhäuser, R., & Comerón, F. 1998, Science, 282, 83 In the text NASA ADS  
• Neuhäuser, R., & Comerón, F. 1999, A&A, 350, 612 In the text NASA ADS  
• Reid, I. N., Kirkpatrick, J. D., Liebert, J., et al. 1999, ApJ, 521, 613 In the text NASA ADS  
• Reipurth, B., & Clarke, C. 2001, ApJ, 122, 432 In the text  
• Sterzik, M. F., & Durisen, R. H. 1999, in Proc. Star Formation, ed. T. Nakamoto (Nobeyama Radio Obs.), 387