Analysis of the non-LTE lithium abundance for a large sample of F-, G-, and K-giants and supergiants

Abstract

A five-dimensional interpolation method and corresponding computer program are developed for using published calculations to determine the non-LTE correction {Δ\}NLTE to the lithium abundance logε(Li) derived from the Li I 6707.8 Å line. The {Δ\}NLTE values is determined from the following five parameters: {Δ\}NLTE the effective temperature Teff, the acceleration of gravity logg, the metallicity index [Fe/H], the microturbulent velocity Vt, and the LTE Li abundance log ε(Li). The program is used to calculate values of {Δ\}NLTE and the non-LTE Li abundance for 91 single bright giants from the list of Lebre, et al. By combining these results with data for 55 stars, we obtain the non-LTE values of log ε(Li) for 146 FGK-giants and supergiants. We confirm that, in the spectra of most of these stars, it is only possible to estimate for them an upper limit for the Li abundance. A large spread is confirmed in log ε(Li) for stars with masses M<6Msun. A comparison of these results with model calculations of stars confirms the unique sensitivity of the lithium abundance to the initial rotation velocity V0. We discuss the giants and supergiants with lithium abundances log ε(Li) = 1.4 +- 0.3, which could have a rotation velocity V0=0 km/s and have already undergone deep convective mixing. Li-rich giants with lithium abundances log ε(Li)>2 and nearly up to the initial value of log ε(Li) = 3.2 +- 0.1 are examined. It is shown that the fraction of Li-rich giants with V0 = 0 - 50 km/s is consistent with current evolutionary models. The other stars of this type, as well as all of the "super Li-rich" giants, for which the standard theory is untenable, can be explained by invoking the hypothesis of recent lithium synthesis in the star or an alternative hypothesis according to which a giant planet is engulfed by the star.