Äîêóìåíò âçÿò èç êýøà ïîèñêîâîé ìàøèíû. Àäðåñ îðèãèíàëüíîãî äîêóìåíòà : http://www.eso.org/~fcomeron/cham_hires.ps Äàòà èçìåíåíèÿ: Thu Aug 19 12:41:54 1999 Äàòà èíäåêñèðîâàíèÿ: Mon Oct 1 22:59:40 2012 Êîäèðîâêà: Ïîèñêîâûå ñëîâà: m 101

A&A manuscript no.
(will be inserted by hand later)
Your thesaurus codes are:
06(08.06.2; 08.12.1; 08.12.2)
ASTRONOMY
AND
ASTROPHYSICS
19.8.1999
Medium­resolution optical spectroscopy of young stellar and
sub­stellar M­dwarfs in the Cha I dark cloud ?
R. Neuh¨auser 1 and F. Comer'on 2
1 MPI f¨ur extraterrestrische Physik, Giessenbachstraúe 1, D­85740 Garching, Germany
2 European Southern Observatory, Karl­Schwarzschild­Straúe 2, D­85748 Garching, Germany
Received 30 June 1999; accepted 21 July 1999
Abstract. We obtained medium­resolution spectra of the
bona­fide brown dwarf Cha Hff 1, the five brown dwarf
candidates Cha Hff 2 to 6, two additional late M­type
brown dwarf candidates, all originally selected by Hff
emission, and four previously known T Tauri stars, all lo­
cated in Chamaeleon I. The spectral types of our targets
range down to M8. We show their spectra and also list
their IJK magnitudes from DENIS. All objects have ra­
dial velocities consistent with kinematic membership to
Cha I and show Li 6708 š A absorption. Our Cha I brown
dwarf candidates with lithium are young or sub­stellar or
both. Cha Hff 1, 3, 6, and 7 are certainly brown dwarfs: Ei­
ther they are as old or older than the Pleiades and should
have burned all their original lithium if they were late
M­type stars, or, if they are younger than the Pleiades,
Ÿ 125 Myrs, they are sub­stellar because of the young age
and the late spectral types (– M7), according to three dif­
ferent sets of evolutionary tracks and isochrones. To clas­
sify Cha Hff 2, 4, 5, and 8 as either stellar or sub­stellar,
evolutionary tracks of higher reliability are needed.
Key words: Stars: formation -- late­type -- low­mass,
brown dwarfs
1. Introduction: Young brown dwarfs
Recently, Comer'on et al. (1999, henceforth CRN99) per­
formed an Hff objective­prism survey in the Chamaeleon
I dark cloud, a site of on­going and/or recent low­ and
intermediate­mass star formation, and presented six new
low­mass late­type (– M6) objects with Hff emission,
called Cha Hff 1 to 6. Their strong Hff emission, if having
the same origin as in T Tauri stars (TTS), is an indicator
of youth, and therefore of membership to the star forming
Send offprint requests to: R. Neuh¨auser, rne@mpe.mpg.de
? Based on observations collected at the European Southern
Observatory 3.5m NTT on La Silla during program 63.L­0023.
region at ¸ 160 pc. Luminosities obtained from VIJHK
photometry made it possible to place the objects in the H­
R diagram where, according to different sets of pre­main
sequence evolutionary tracks, they all lie near or below
the hydrogen burning limit. Follow­up spectroscopy with
high S/N of the lowest­mass object, Cha Hff 1, confirmed
its sub­stellar nature (Neuh¨auser & Comer'on 1998, hence­
forth NC98) and allowed an accurate determination of its
spectral type to be M7.5­M8. Comparision with evolution­
ary tracks and isochrones yielded a mass of ¸ 0:04 M fi and
an age of ¸ 1 Myrs. Cha Hff 1 is the first brown dwarf
(BD) detected in X­rays (NC98).
Young brown dwarfs were also found in other star
forming regions, namely ae Oph (Rieke & Rieke 1990;
Comer'on et al. 1993, 1998; Luhman et al. 1997; Wilking
et al. 1999), Taurus (Luhman et al. 1998; Brice~no et al.
1998; White et al. 1999; Reid & Hawley 1999), and Orion
(B'ejar et al. 1999).
The sub­stellar nature of suspected brown dwarfs
can usually be confirmed by detection of lithium
6708 š A absorption, because lithium is burned rapidly and
fully by proton capture in low­mass, i.e. fully convective
stars (Magazz`u et al. 1993) with masses above ¸ 0:06 M fi .
This lithium test cannot be applied to very young objects,
which are not old enough, yet, to have burned all their
initial lithium: for example, objects younger than ¸ 10
Myr with a mass below ¸ 0:3 M fi have burned less than
1% of their original lithium, and the percentage is even
lower for less massive objects of the same age (D'Antona
& Mazzitelli 1997, Baraffe et al. 1998, and A. Burrows,
private communication). For this reason, a lithium detec­
tion in the spectrum of a suspected brown dwarf mem­
ber of a star forming region does not directly confirm
its substellar character, but it does support strongly its
membership in the star forming region (the alternative
possibility, namely the object is evolved and foreground,
automatically implies that it is a brown dwarf because of
the lithium detection). Once such membership and, hence,
distance is established, the mass can be inferred from the

2 Neuh¨auser & Comer'on: Young M­dwarfs in Cha I
Fig. 1. Parts of our medium­resolution spectra around the lithium 6708 š A line, which is detected in all objects
spectral type and luminosity by comparison to theoretical
evolutionary tracks for very low mass objects.
In this paper, we present medium­resolution spectra of
Cha Hff 1 to 6, two additional late M­type objects (Cha
Hff 7 and 8), also first detected in an Hff objective prism
survey (Comer'on et al., in preparation), as well as four
M­type TTS in Cha I. Signal­to­noise ratio (S/N) and
resolution of these new spectra are better than those in
CRN99, so that we can investigate not only Hff emission,
but also lithium absorption and radial velocities.
2. Observations and data reduction
We obtained medium­resolution spectra of the Cha I low­
mass objects Cha Hff 1 to 8 as well as the TTS CHXR
74, B 34, Sz 23, and CHXR 78C. The spectra were ob­
tained in the nights of 17 and 18 April 1999 at the 3.5m­
New Technology Telescope (NTT) of the European South­
ern Observatory (ESO) on La Silla with the ESO Multi
Mode Instrument (EMMI) in the red medium dispersion
mode. Grating #6 was used covering the range 6485 to
7105 š A, providing a dispersion of 0.32 š A per each 0.27''
pixel (CCD #36), and a resolution of ¸ 5800 (with the
1.0'' slit used). One or several spectra were obtained for
each object, with exposure time and number of spectra
according to the measured instrumental throughput and
the available R­band photometry. To keep the cosmic ray
density in the resulting images reasonably low, the longest
exposure times per frame were never longer than 45 min,
always placing two objects into the 330'' long slit.
Each spectrum was extracted individually from
the corresponding bias­subtracted, flat­fielded two­
dimensional image using tasks in the APEXTRACT pack­
age layered on IRAF. Wavelength calibration was per­
formed by obtaining spectra of a He­Ar lamp immediately
before and after each individual exposure, to ensure that
wavelength shifts due to the rotation of the instrument
during the exposure in the Nasmyth focus were kept to a
minimum. A full dispersion solution was obtained for one
of the He­Ar spectra, which was then used as the master
wavelength calibration frame. Zero­point shifts with re­
spect to this spectrum were then obtained for each of the
other He­Ar lamp exposures and applied to the extracted
spectra of the targets. These zero­point shifts were based
on spectra of the calibration lamp extracted over a narrow
range of rows of the detector centered on the position of
the object. In this way, the determined shifts included the

Neuh¨auser & Comer'on: Young M­dwarfs in Cha I 3
Table 1. M­dwarfs targets and results (*)
Object ff; ffi (J2000.0) spec (1) T (2)
eff
I J K W– (Hff) [ š A] (4) W– (Li) RV (5)
lsr
Classification
designation 11h \Gamma77 ffi type [K] DENIS (3) [mag] CRN99 this work [ š A] [km s \Gamma1 ]
Cha Hff 1 07:17.0 35:54 M7.5 2805 16.4 13.3 12.3 \Gamma59 \Gamma34:5 0.63: +0:9 \Sigma 5:3 bona­fide BD
Cha Hff 2 07:43.0 33:59 M6.5 2940 15.3 12.1 10.6 \Gamma39 \Gamma71:0 0.43 +8:4 \Sigma 1:9 candidate BD
Cha Hff 3 07:52.9 36:56 M7 2890 15.0 12.3 11.1 \Gamma4:5 \Gamma14:4 0.43 +6:9 \Sigma 4:9 bona­fide BD
Cha Hff 4 08:19.6 39:17 M6 2990 14.4 12.0 11.1 \Gamma4:7 \Gamma9:7 0.48 +6:1 \Sigma 3:2 candidate BD
Cha Hff 5 (6) 08:25.6 41:46 M6 2990 14.7 12.0 10.7 \Gamma7:6 \Gamma8:0 0.42 +4:4 \Sigma 3:0 candidate BD
Cha Hff 6 08:40.2 34:17 M7 2890 15.1 12.0 10.9 \Gamma59 \Gamma61:7 0.43 +0:2 \Sigma 2:3 bona­fide BD
Cha Hff 7 (7) 07:38.4 35:30 M8 2720 16.7 13.5 12.4 (8) \Gamma35:0 0.80: \Gamma1:8 \Sigma 6:7 bona­fide BD
Cha Hff 8 (7) 07:47.8 40:08 M6.5 2940 15.6 12.7 11.5 (8) \Gamma8:4 0.49 \Gamma2:9 \Sigma 5:5 candidate BD
CHXR 74 (9) 06:57.4 42:10 M4.5 3200 13.7 11.6 10.2 \Gamma13 \Gamma8:3 0.69 +5:9 \Sigma 1:0 T Tauri star
B 34 (10) 07:35.4 34:51 M5 3125 14.4 12.1 10.9 \Gamma5:5 \Gamma7:2 0.80 +7:0 \Sigma 1:7 T Tauri star
Sz 23 (11) 07:59.4 42:40 M2.5 3500 14.6 12.0 9.9 \Gamma45 \Gamma56:5 0.34 \Gamma2:8 \Sigma 6:0 T Tauri star
CHXR 78C 08:54.4 32:12 M5.5 3060 14.8 12.2 10.9 \Gamma3:2 \Gamma13:4 0.48 +9:5 \Sigma 2:9 T Tauri star
(*) Approximate errors are \Sigma0:25 sub­classes in spectral type, \Sigma50 K in T eff , \Sigma0:1 mag in IJK, \Sigma5 š A in W– (Hff), and \Sigma0:05 š A in
W–(Li) (but \Sigma0:1 š A in Cha Hff 1 and 7 because of lower S/N).
Notes: (1) Spectral types (ST) for Cha Hff 1 to 8 assigned from high­S/N, low­resolution spectra obtained for a much broader
wavelength range (Comer'on et al., in preparation) than in Fig. 1; ST for the four TTS are taken from CRN99. (2) Effective
temperatures T eff converted from ST using Fig. 7 in Luhman (1999) for objects intermediate between dwarfs and giants, as
appropriate for young M­dwarfs. (3) DENIS data for Cha Hff 1 to 8 are from L. Cambr'esy (private communication) and for
TTS from Cambr'esy et al. (1998), they all compare well with data given in CRN99 and Lawson et al. (1996). (4) Negative for
emission. (5) Local standard of rest (lsr) radial velocity (RV). (6) Coordinates of Cha Hff 5 were slightly incorrect in NC98 and
CRN99. (7) For finding charts, see Comer'on et al., in preparation. (8) Not listed in CRN99. (9) Lawson et al. (1996) detected Li
at low S/N and gave W– (Hff) = \Gamma2:7 š A. Alcal'a (1994) gave W– (Hff) = \Gamma4:1 š A. (10) Alcal'a (1994) gave W– (Hff) = \Gamma7:7 š A and
detected Li. (11) Double­peaked Hff emission, possibly a companion to VW Cha (Reipurth & Zinnecker 1993).
corrections due both to flexures of the instrument, and to
the slight curvature of the spectral lines across the surface
of the detector, which amounts to a significant four­pixel
difference between the center and the edges. Correction
for telluric features was performed by observing the bright
star fi Cha at intervals of about 2.5 hours per night. The
high rotation velocity of this star (255 km/s, Hoffleit et
al. 1982) allows an easy separation between the narrow
telluric and the broad photospheric lines. The spectra of
fi Cha were interpolated linearly in the wavelength inter­
vals containing photospheric lines. Then, each individual
spectrum of our targets was ratioed by the one of fi Cha
obtained at the most similar airmass. The next step was
the modification of the wavelength axis of each object to
correct for the observer's motion with respect to the lo­
cal standard of rest (lsr). The lsr velocities for each object
were computed with the RVCORRECT task under IRAF,
and the spectra were then accordingly Doppler­shifted.
Finally, all the wavelength­calibrated, telluric­corrected
spectra in the lsr frame for each object were added to­
gether to produce the combined spectra (Fig. 1). Radial
velocities are derived by Fourier cross­correlation using the
6600 to 7100 š A part of the spectrum of CHXR 74 as a refer­
ence template. The zero­point is derived from the lithium
doublet, using 6707.815 š A as central wavelength. The qual­
ity and S/N of the spectra were insufficient, however, to
obtain reliable rotational velocities.
Results are listed in Table 1. In addition, we obtained
ground­based, high­S/N, low­resolution optical and in­
frared spectra over a much broader wavelength range than
shown in Fig. 1 or in CRN99, in order to better constrain
the spectral types of the Cha Hff objects, as well as ad­
ditional photometry and ISO data, all of which are to
be presented elsewhere (Comer'on et al., in preparation).
Spectral types -- as listed in Table 1 -- are determined
by comparison with late M­type standards (Kirkpatrick
et al. 1991, 1995) and spectra indices as in CRN99. We
converted them to effective temperatures T eff using the
relation recently presented by Luhman (1999) for young
M­type objects in IC 348.
3. Results and discussion
To investigate kinematic membership to Cha I, we com­
pare the radial velocities of our objects with those of
known TTS members of Cha I and the local molecular gas,
both having lsr radial velocities peaking at ¸ 4 km/s with
a 1 oe scatter of ¸ 6 km/s (Dubath et al. 1996, Covino et
al. 1997, see also our Fig. 2). The TTS observed here also
have such velocities. Hence, within the errors, Cha Hff 1 to
8 all have radial velocities in this range, i.e. are kinematic
members of the Cha I association (see Fig. 2).
Cha I membership can also be investigated by looking
for youth signatures, such as Hff emission or lithium ab­
sorption. Gravity sensitive features and thermal infrared
excess emission observed by ISOCAM will be discussed in
a companion paper (Comer'on et al., in preparation).
For most of our objects, W – (Hff) is variable by a factor
of ¸ 2, comparing the results from CRN99 with our data.

4 Neuh¨auser & Comer'on: Young M­dwarfs in Cha I
Fig. 2. Local standard of rest (lsr) radial velocities of Cha
TTS (Dubath et al. 1996, Covino et al. 1997, and Table 1),
candidate BDs (hatched) and bona­fide BDs (filled)
This can be taken as indication for strong chromospheric
activity, i.e. young age. Furthermore, Cha Hff 1, 3, and 6,
classified here as bona­fide BDs, are X­ray sources (NC98),
an indication for magnetic activity, i.e. youth.
Comparing W – (Li) of our candidate and bona­fide BDs
with TTS (Fig. 3) shows that the lithium strength seems
to drop after M5 (B 34, log T eff = 3:5) and to increase
again from ¸ M7 to M8. However, objects with Ÿ 0:3 M fi ,
i.e. log T eff Ÿ 3:5, burn less than 1% of their initial
lithium in the first ¸ 10 Myrs, so that the apparent gap
may be due either to missing objects or to an effect of
molecular lines (eg. TiO in late M­type objects) on the
lithium curves­of­growth, i.e. on how the lithium abun­
dance depends on T eff and W – (Li), so that the lithium­
iso­abundance lines do not increase monotonically any­
more in the W – (Li) versus T eff plot, as they do for G­
and K­type stars (see Fig. 3). Moreover, the similar po­
sitions of our objects and of the Pleiades brown dwarfs
(Basri et al. 1996, Rebolo et al. 1996, Stauffer et al. 1998)
in the T eff versus W – (Li) diagram, despite a difference of
more than one order of magnitude in age between both
populations, seem to rule out evolutionary factors as the
cause of the apparent gap.
Any object with detected lithium is young or sub­
stellar or both. However, for any specific age (or age range
or upper age limit), any object later than some specific
spectral type is a BD. For the Pleiades (125 Myrs), Mart'in
et al. (1996) and Stauffer et al. (1998) found that any
member later than M6.5 is a BD. One can also see from
evolutionary tracks and isochrones that any object with an
age of Ÿ 125 Myrs which is cooler than the temperature
corresponding to M6.5 lies below the hydrogen burning
limit (D'Antona & Mazzitelli 1997, Burrows et al. 1997,
Baraffe et al. 1998).
Cha I members are younger than the Pleiades, i.e.
lie higher above the main sequence, so that they are
slightly hotter at any given spectral type (Luhman et al.
1997), corresponding to roughly \Sigma0:25 sub­classes. Con­
sidering also \Sigma0:25 error in assigning the spectral classes,
we should be sure that any Cha I member with spectral
type M7 (i.e. T eff = 2890 K, Luhman 1999) or later is
sub­stellar. This is the case for Cha Hff 1, 3, 6, and 7,
classified as bona­fide BDs in Table 1. This classification
is correct, regardless of whether the objects are young: ei­
ther they are younger than ¸ 125 Myrs and, hence, lie
below the sub­stellar limit in the H­R diagram because
of the late spectral type and the young age (see tracks
and isochrones by D'Antona & Mazzitelli 1997, Burrows
et al. 1997, Baraffe et al. 1998); or, alternatively, if they
are not young, then they are sub­stellar because of the
lithium. Actualy, because most Cha I TTS (see Lawson
et al. 1996) as well as Cha Hff 1 to 6 (see CRN99) are
¸ 1 Myrs old, all our bona­fide and candidate BDs are
probably of that young age, too, i.e. much younger than
¸ 125 Myrs. Given the spectral types of Cha Hff 2, 4, 5,
and 8, they are located too close to the limit for definitive
discrimination.
Acknowledgements. We would like to thank Olivier Hainaut,
Stephane Brillant and the NTT staff for the perfect support,
Eduardo Mart'in and Isabelle Baraffe for very useful discus­
sion, Adam Burrows for providing us with unpublished lithium
depletion calculation, and Laurent Cambr'esy with the whole
DENIS team for contributing unpublished DENIS data.
References
Alcal'a J., 1994, PhD dissertation, Ruprecht­Karls­Universit¨at,
Heidelberg
Baraffe I., Chabrier G., Allard F., Hauschildt P., 1998, A&A
337, 403
Basri G., Marcy G., Graham J., 1996, ApJ 458, 600
B'ejar V., Zapatero Osorio M., Rebolo R., 1999, ApJ, in press
(astro­ph/9903217)
Bessell M.S., 1979, PASP 543, 589
Bessell M.S., 1991, AJ 101, 662
Brice~no C., Hartmann L., Stauffer J., Mart'in E., 1998, AJ 115,
2074
Burrows A., Marley M., Hubbard W., et al., 1997, ApJ 491,
856
Cambr'esy L., Copet E., Epchtein N., et al., 1998, A&A 338,
977
Comer'on F. Rieke G.H., Burrows A., Rieke M.J., 1993, ApJ
416, 185
Comer'on F., Rieke G.H., Claes P., Torra J., Laureijs R.J., 1998,
A&A 335, 522
Comer'on F., Rieke G., Neuh¨auser R., 1999, A&A 343, 477
(CRN99)
Covino E., Alcal'a J., Allain S., et al., 1997, A&A 328, 187
D'Antona F., Mazzitelli I., 1997, Evolution of Low Mass Stars.
In: Proceedings of the meeting on Cool stars in Clusters and
Associations, Vol. 68 of the Mem. Soc. Astron. Italia con­

Neuh¨auser & Comer'on: Young M­dwarfs in Cha I 5
Fig. 3. Equivalent width W–(Li) versus effective temperature T eff for Pleiades stars (Soderblom et al. 1993), Pleiades BDs
(Basri et al. 1996, Rebolo et al. 1996, Stauffer et al. 1998), TTS in Cha (from Covino et al. 1997 and our Table 1), and Cha I
bona­fide and candidate BDs. The lines indicate the primordial lithium log N(Li)= 3:5 for sub­giants (log g = 3:0) and dwarfs
(log g = 4:5) from Pavlenko & Magazz`u (1996) fitting well as upper envelope to the G­ and K­type TTS; lithium curves­of­growth
for M­type objects are not yet available. For the F­ to K­type TTS and Pleiades, we have converted spectral types to T eff using
Bessell (1979, 1991), for the M­type objects using Fig. 7 in Luhman (1999) for objects intermediate between dwarfs and giants,
taking into account that Pleiades BDs are slightly colder than Cha I BDs at any given spectral type (Luhman et al. 1997)
ference series, Micela G., Pallavicini R., Sciortino S. (Eds.).
Mem. Soc. Astron. Italia, Firence, 1997. pp. 807­822
Dubath P., Reipurth B., Mayor M., 1996, A&A 308, 107
Hoffleit D., Jaschek C., Warren W., Charles E.H., Hauck B.,
1982, The bright star catalogue, 4th revised edition, Yale
Univ. Obs. (5th revised edition available from CDS)
Kirkpatrick J., Henry T., McCarthy D., 1991, ApJS 77, 417
Kirkpatrick J., Henry T., Simons D., 1995, AJ 109, 797
Lawson W., Feigelson E., Huenemoerder D., 1996, MNRAS
280, 1071
Luhman K., 1999, ApJ, in press (astro­ph/9905287)
Luhman K., Liebert J., Rieke G., 1997, ApJ 489, L165
Luhman K., Brice~no C., Rieke G., Hartmann L., 1998, ApJ
493, 909
Magazz`u A., Mart'in E., Rebolo R., 1993, ApJ 404, L17
Mart'in E.L., Rebolo R., Zapatero­Osorio M.R., 1996, ApJ 469
Neuh¨auser R., Comer'on F., 1998, Science 282, 83 (NC98)
Pavlenko Y., Magazz`u A., 1996, A&A 311, 961
Rebolo R., Mart'in E., Basri G., Marcy G., Zapatero­Osorio
M., 1996, ApJ 469, L53
Reid I.N., Hawley S.L., 1999, AJ 117, 343
Reipurth B., Zinnecker H., 1993, A&A 278, 81
Rieke G.H., Rieke M.J., 1990, ApJ 362, L21
Soderblom D., Jones B., Balachandran S., et al., 1993, AJ 106,
1059
Stauffer J., Schultz G., Kirkpatrick J., 1998, ApJ 499, L199
White R.J., Ghez A.M., Reid I.N., Schulz G., 1999, ApJ 520,
811
Wilking B.A., Greene T.P., Meyer M.R., 1999, AJ 117, 469
This article was processed by the author using Springer­Verlag
L A T E X A&A style file L­AA version 3.