CFHT Information Bulletin, number 39, Second Semester 1998

Accretion in PMS binaries : high angular resolution spectra with OSIS

J.-L. Monin and G. Duchêne (Grenoble).

During the past five years, many studies have addressed the question of the multiplicity in star-forming regions (SFR). While about 60% of G-K main sequence dwarfs belong to multiple systems in the solar vicinity (Duquennoy & Mayor 1991), several papers (e.g. Leinert et al. 1993) have pointed out that 80%, and maybe up to 100%, of Taurus young stars are not formed singly. More recently, this binarity excess has been found in other SFR (e.g. Ghez et al. 1997). It is a major theoretical challenge to explain several points including why do stars form in multiple system, and why is the degree of multiplicity different in different SFR and in the solar environment.

Various binary formation mechanisms have been proposed. Although fragmentation is the more widely accepted idea nowadays, many details remain to be investigated observationally and numerically to fully understand the details (and initial conditions) of the process. Binary formation has been studied using SPH codes by various authors.

When the formation results in a central binary surrounded by a circumbinary disk, Artymowicz & Lubow (1996) showed that matter could flow through one or two points of the inner ring of the circumbinary disk, and when the stars have very unequal masses, the accretion funnel is preferentially directed towards the star orbiting closer to the circumbinary disk. A different prediction applying to younger objects has been made by Bonnell & Bastien (1992): in high mass ratio systems, accretion of low angular momentum material is directed toward the center of mass which is close to the more massive star.

The study of the accretion activity on one or both components in PMS binary systems will tell us a lot about the way the residual matter flows onto the central star. Such a study can be performed by spectroscopic measurements. However, up to now, spectroscopic studies of PMS binaries have been limited to somewhat wide systems due to the spatial resolution limit of the observations.

Using the unique opportunity of the (O)SIS instrument on the CFHT, we have started in 1995 a spectroscopic survey of young close binaries in Taurus, investigating the classification as classical or weak-line T Tauri stars (C/W TTS: stars with or without disks) of both stars in a pair, along with a more detailed study of spectroscopic signature of accretion activity (Monin et al. 1998, Duchêne et al. 1998).

Observations have been performed using the SIS instrument in 1995, december and OSIS in 1996, november. The tip-tilt correction, allowed us to measure binaries with an angular resolution 0.9", using a 1" slit. We obtained spectra on a 4000 to 7800 Å range with an actual resolution is 9.6 Å.

These spectra first allowed us to estimate the spectral types of each component of the binaries, using the strength of TiO bands for M stars, and relative strengths of Ca I 6122,62, Na I 5893, CaH 6350,80 and CaH 6750-7050 for K stars. We used the standard grids from Allen & Strom (1995) and Kirkpatrick et al. (1991). By observing a grid of spectral type standards, we find that our estimation is most of the time accurate to within one subclass.

To investigate the accretion phenomenon, we selected Balmer H and H, and HeI 6678. [O III] 6300 equivalent width was also measured. Errors were estimated by using the maximum and minimum acceptable value for the continuum near the lines. For some objects, we could also detect [O I] 6363 and the [S II] 6716,31 doublet in emission.

Figure 4 shows as an example the spectrum of both components of HK Tau. The binary separation is 2.4". The HK Tau B spectrum is similar to but fainter and somewhat bluer than the primary's spectrum; this result was later explained by the discovery of a circumstellar disk around the B component (Stappelfeldt et al., 1998). The central star is deeply extincted and the scattering of the stellar light by the disk makes the overall spectral color bluer.

To compare the accretion activity of both components in all systems studied here, we evaluated the H luminosity ratios (see Figure 5), assuming that the H luminosity is propotional to the energy dissipated in the accretion phenomenon, and thus to the accretion rate, and that extinction is of the same order on both stars.

The absence of close mixed systems is a constraint on binary formation mechanisms. Both component must share a common environment, as the stars cannot have been formed independently from different cloud cores and then paired (via capture). Concerning the C/C pairs, we see a trend for closer pairs to have higher activity in the primaries. Spectral types determinations show that the star accreting more matter appears to be the more massive one. This effect cannot be accounted for by a failure of the equal extinction hypothesis. Despite small samples, this provides another hint for close pairs (with separation smaller than 350 AU) to behave in a different way than wider systems. It is unlikely that the difference is due to undiscovered wide, faint companions.