No evidence for intense cold accretion onto YSOs from measurements of Li in T-Tauri stars

This paper searches for highly lithium depleted stars in the young clusters NGC 2264 and the ONC. If found, these stars would offer significant evidence that proto-stellar objects accrete at high rates during their early 'assembly phase', as suggested by some simulations (Vorobyov & Basu 2005) and observations (Kenyon et al. 1990, Enoch et al. 2009). Cold accretion at high-rates, occuring in the embedded phase has been suggested as a possible solution to the luminosity spreads observed in apparently co-eval pre-main-sequence stellar populations at ages < 10 Myr.

Some models predict that depleted lithium and reduced luminosity in T-Tauri stars are key signatures of 'cold' high-rate accretion occurring early in a star's evolution e.g. Baraffe & Chabrier (2010), Hosokawa et al. (2011) and Baraffe et al. (2012). The radius and luminosity of a pre-main-sequence (pre- MS) star are determined by the balance between its Kelvin- Helmholtz (K-H) contraction timescale and accretion activity. If accretion increases the mass of the star more rapidly than the K-H timescale allows energy to be radiated, it will displace the star from thermal equilibrium and cause it to exhibit a radius and luminosity that is inconsistent with non-accreting models (Tout et al. 1999; Baraffe et al. 2009; Hartmann et al. 2011). As the age of a T-Tauri star (of order 5 Myr) can be less than the K-H timescale, its radius and luminosity may still exhibit the influence of its prior accretion history in the embedded phase.

The core conditions of an embedded proto-star are potentially dependent on its mass accretion rate. Accretion of material with low internal energy ('cold' accretion), at rates in excess of M > 5 x 10^4 M_sol/yr may gravitationally compress the star, increasing its core temperature and triggering the early onset of lithium burning (Baraffe & Chabrier 2010). Efficient large-scale convection (seen in pre-MS stars) would then rapidly deplete lithium throughout the star. Observationally, this mechanism may be apparent in the later T-Tauri phase as a surface lithium abundance that is much lower than that expected for a given age.

We have performed a spectral analysis of two young clusters, NGC 2264 and the Orion Nebula Cluster around the 6708 angstrom lithium doublet at a resolution of R~10,000. Correction has been made for accretion veiling, yielding estimates of the equivalent width of the lithium line for 94 stars (76 single and 18 binary) in the ONC field and 74 stars (54 single and 20 binary) in the NGC 2264 field.

We find no evidence in 168 stars (130 unveiled stars and 38 spectroscopic binaries with lower limits) for objects with significantly depleted lithium surface abundances. This finding helps to place limits on cold accretion models, indicating that burst accretion rates of M > 5 x10^4 M_sol/yr occur in less than 0.5 per cent of 0.3< M <1.9 M_sol stars. This mass accretion rate can only be exceeded without depleting lithium if the initial core mass is > 0.03 M_sol, however such models seem incapable of producing the luminosity dispersion seen in CMDs. This initial core mass is also at least an order of magnitude larger than predicted by theory (see Baraffe et al. 2012, and references therein).

We also find that the dispersion in the strength of the 6708 angstrom lithium line might imply an age spread that is similar in magnitude to the apparent age spread implied by the luminosity dispersion seen in colour magnitude diagrams. Evidence for weak lithium depletion (< 10 per cent in equivalent width) that is correlated with luminosity is also apparent, but we are unable to determine whether age spreads or accretion at rates less than 5 x 10^4 M_sol/yr are responsible.