Astron. Astrophys. 351, 954-962 (1999)

6. Discussion and summary

We have shown that there exists only few mixed systems in Taurus PMS binaries in the separation range 130-1800 AU. This result extends those of PS97 who did not find any mixed system in a sample including binaries with separations of 40-360 AU. This indicates that the accretion history of the two stars are not independent, even for binaries with separations up to AU (from our new spectroscopic observations) and even AU if we take into account the results from H94.

What can explain such a correlation in binaries with separations as large as AU? This "twinning" trend, together with the fact that circumstellar disk dissipation times from optically thick to optically thin are short (Simon & Prato 1995), led PS97 to propose that both components of a close binary system accrete over the same time span because their circumstellar disks are replenished by material from a common (circumbinary) environment. As soon as this environment is cleared, both disks dissapear over a short viscous timescale. However, the circumbinary environment hypothesis appears difficult to apply to wide binaries, and if such envelopes have been detected around a few close binaries, they generally remain elusive. Similarly, it appears unprobable that the binary as a whole can sweep enough material during its wander through the parent cloud: at 1 km.s^-1, a 100 AU radius wide binary sweeps only in a cm^-3 density cloud.

On the other hand, we find that the primaries have larger H fluxes than their secondaries. We call `primary' the brightest component in the V band, which always has an earlier spectral type than the secondary so that it is likely the most massive star. The H luminosity is assumed to be proportional to the accretion luminosity:

Baraffe et al.'s (1998) evolutionary models show that two 2 Myr-old TTS with masses of 1 and 0.1 have ratios only differing by a factor of 4 (the most massive star also has a larger radius). Our measured H luminosity ratios vary by over 2 orders of magnitude and therefore cannot be accounted for by extreme mass ratios. The difference in the accretion luminosities is thus likely to reveal that, in most cases, the primary accretes more than its companion: . It is also noticeable that the mixed systems in our sample all have a CTTS primary, so that in the case of CW pairs, the more massive star again seems to be more active than the other one.

If both components have similar circumstellar disk lifetimes (), these results suggest that the circumprimary disk is preferentially feeded in the early binary formation process by a common circumbinary reservoir of mass. This is in agreement with the prediction of Bonnell et al.'s (1992) model.

Another possibility is that the accretion rate on the star, , is proportional to the disk mass, itself related to the mass of the central star. In the canonical accretion disk theory, the accretion rate is related to the surface density , itself evidently linked to the disk mass. This mechanism would explain simultaneously why and why the disk lifetime does not depend on the mass of the central star. If true, such a relation should hold for single TTS but current mass determination lack the precision needed to study this point further.

Observations of closer binaries down to separations of the order of the peak value in the PMS separation distribution ( AU, Mathieu 1994) should shed more light on this question. Such observations are within reach of current adaptive optics systems equipped with spectroscopic capabilities. Such a peak separation is of the order of the size of a canonical accretion disk and these observations would allow to study systems where the star-disk and disk-disk interactions are strong, and also where the eventual leftover circumbinary environment has a major influence.

© European Southern Observatory (ESO) 1999

Online publication: November 16, 1999
helpdesk.link@springer.de