Astron. Astrophys. 321, 492-496 (1997)

2. T Tauri stars

The two defining qualities of a TTS are extreme youth and low mass. Appenzeller & Mundt (1989) quoted 10 Myr as the upper age limit, and 3   as the upper mass limit. However, PMS evolution is strongly mass dependent. Stars with masses in the range 2-3   have reached the ZAMS in less than 10 Myr (D'Antona & Mazzitelli 1994, Palla & Stahler 1993), becoming late B and early A-type stars. Most works on low-mass PMS evolution are concerned with stars with masses 1    or less. Therefore, it seems more adequate to use for TTSs an upper mass limit of 2   , instead of 3   . Moreover, since mass is not a direct observable (except in eclipsing binaries, of which none has yet been discovered amont TTSs), it would be better to use a   or a spectral type. According to recent PMS evolutionary tracks (D'Antona & Mazzitelli 1994, Forestini 1994, Martín & Claret 1996, Palla & Stahler 1993) a convenient upper limit to the   of a TTS ( hereafter) would be around log T=3.7 (5011 K). The most widely used   - spectral type scale for TTSs is the one adopted by Cohen and Kuhi (1979), in which 5080 K corresponds to K1-type. Other scales are also found in the literature (see discussion in Martín et al. 1994) which differ from the one of Cohen & Kuhi by up to 150 K for luminosity class V. Furthermore, spectral types are usually determined to an accuracy of half a spectral subclass. It is conservative to adopt a spectral type of K0 for the hot limit of TTSs, and a of 5250 K, which is the hottest value among the different calibrations.

The models of Martín & Claret (1996) with updated opacities and including rotation are cooler than those of Forestini (1994), which in turn are cooler than D'Antona & Mazzitelli's (1994). Despite such systematic differences, the models agree that the entire fully convective (Hayashi) tracks of stars less massive than 2    are cooler than our adopted . For stars less massive than 1.4   , all PMS evolution for ages shorter than 10 Myr takes place at   lower than . The original definition of the T Tauri class includes stars of spectral type late-F and G (Joy 1945, Haro 1983). However, less than 5% of the TTSs listed in the comprenhensive Herbig & Bell (1988) catalogue have spectral types earlier than K0. It is convenient to consider objects intermediate in mass between the TTSs and the Herbig Ae/Be stars as a new class of young stars, which could be called PMS Fe/Ge stars. Thé, de Winter & Pérez (1994) denominated a few of such stars "PMS F-type stars".

The lower limit in the spectral type of a TTS is close to the substellar limit. Recent works have shown that in the Pleiades cluster (age 100 Myr) brown dwarfs have spectral types later than M6.5 (e.g. Martín, Rebolo & Zapatero-Osorio 1996). Very young brown dwarfs are expected to have similar   than their Pleiades-age counterparts as they follow fully-convective evolutionary paths. Therefore, it seems reasonable to set a spectral subclass of M6 as the cool limit for a TTS. This constraint does not exclude any TTS listed in the Herbig & Bell catalogue, which are all earlier than M6. The latest type of a known TTS, namely UX Tau C, is M6 (Magazzù, Martín & Rebolo 1991).

CTTSs can be easily identified because of their characteristic emission-line spectrum (Herbig 1962, Appenzeller & Mundt 1989) and non-photospheric continuum excesses (Kenyon & Hartmann 1987, Bertout 1989). Such properties cannot be explained by conventional stellar activity. The equivalent width of H   should not be used as the only criterion to classify a TTS as classical or weak, because it can vary depending on spectral type, binarity, flare activity, etc. Additional criteria, such as UV and near-IR excesses and presence of forbidden emission lines, should also be used. However, for lack of other data, it is sometimes convenient for statistical purposes to rely on H   as the only criterion for classifying CTTSs. In such case, the CTT H   threshold should be set at such an equivalent width that for higher values chromospheric activity is unlikely to be the source of the emission. For K-type stars a high enough value is 5 Å  because chromospherically active single and binary stars do not show higher H   equivalent widths (Strassmeier et al. 1990, Montes et al. 1995). However, for M-type stars, the value has to be risen, because M-stars in young open clusters and the field do show H   equivalent widths stronger than 5 Å  (Prosser 1994, Zapatero-Osorio et al. 1996). Safe enough values are 10Å  for early-M and 20Å  for late-M types.

A WTTS lacks the exotic properties of a CTTS. It has a spectral energy distribution similar than a MS star, and spectroscopically only the strong Li I 670.8  feature might be a signal of identity. In fact, papers claiming to have found new WTTSs have usually relied on the detection of the Li I   feature. The use of this line is justified because of its accessibility using modern spectrographs with CCD detectors and the well-known property of Li ² as an age indicator (Herbig 1962, Bodenheimer 1965, Magazzù, Rebolo & Pavlenko 1992, Martín et al. 1994). One of the main conclusions of the last two works is that PMS stars are formed with a homogeneous "initial" Li abundance of log N(Li)=3.1 0.1 (in the usual scale of log N(H)=12), which is similar to the abundance found in the interstellar medium. Martín et al. (1994) also concluded that PMS stars do not show significant Li depletion until they are relatively evolved. No Li depletion larger than the uncertainties has been found in any CTTS, whereas some low-luminosity WTTSs clearly showed Li depletion.

Theoretical models agree that Li depletion is less than 50% for ages younger than 5 Myr, independently of mass (D'Antona & Mazzitelli 1994, Forestini 1994, Martín & Claret 1996, Bildsten et al. 1996). For masses 0.5   , Li depletion is not significant until ages larger than 10 Myr. A typical TTS has a mass of 0.5    and has preserved most of its "initial" Li content, implying the presence of a strong Li I 670.8  feature in the spectrum. Quantitatively, the "minimum" Li I   equivalent width () of a TTS is determined by the line formation for "minimum" Li abundance as a function of mainly temperature. Gravity and microturbulence are only second-order effects. In order to derive the "minimum" shown in Fig. 1, I have used the grid of NLTE curves of growth of Pavlenko et al. (1995), and I have adopted the following values: a "minimum" Li abundance for a TTS of log N(Li)=2.8, i.e., 50% lower than the "initial" value; the highest   for each spectral type among those available from different calibrations (Martín et al. 1994 and references therein); a gravity of log g=4.0 for spectral types in the range K0-K5 and log g=3.5 for cooler temperatures; and finally a microturbulent velocity of 2 km s . For   cooler than 4000 K the Pavlenko et al. (1995) computations are not very reliable because of problems with the model atmospheres and the blending effects of molecular features. Hence, a constant "minimum" value has been used for M-type stars in Fig. 1. It is important to note that CTTSs may present apparently small Li I   equivalent widths due to the blurring effect of optical veiling (Basri, Martín & Bertout 1991, Magazzù et al. 1992). Hence, the "minimum" values of Fig. 1 can be used for WTTSs without veiling correction, and for CTTSs with veiling correction. Nevertheless, a CTTS can be identified from its extreme emission lines and continuum excess, without needing to accurately unveil the photospheric spectrum.

Fig. 1. Li I 670.8  equivalent widths of TTSs (empty triangles) and low-mass members of the young open clusters IC 2602, IC2391, IC 4665 and the Pleiades (filled pentagons or, for upper limits, inverted filled triangles) plotted as a function of   . 89% of the TTSs fall on or above the dashed line, which represents the cutoff at 5250 K, and the Li isoabundance line for log N(Li)=2.8 ("minimum" abundance for TTSs). Note the empty region between the TTSs and the cluster stars for   4800 K (PTT-gap).

© European Southern Observatory (ESO) 1997

Online publication: June 30, 1998
helpdesk.link@springer.de