... [C/Fe]

The usual spectroscopic notation ${\rm [A/X]}=\log~({N_{\rm A}/N_{\rm X} })_{*}-\log~({N_{\rm A}/N_{\rm X} })_{\odot}$is used throughout.

.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.

... NIST

http://physics.nist.gov/PhysRefData/

.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.

... approach

Because the fine-structure levels are not directly included in the whole problem.

.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.

... MULTI

Available at http://www.astro.uio.no/~matsc/mul22/

.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.

... function

More explicitly, assuming complete redistribution and in the Wien regime, the ratio of the line source function to the Planck function can be expressed in terms of the ratio of departure coefficients for the upper and lower levels of a transition, as $S^{\rm l}_{\nu}$/ $B_{\nu}\approx \beta_{\rm u}/\beta_{\rm l}$.

.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.

... formed

It is important to remember that, as mentioned, these weak lines form with a significant contribution from a tail of atmospheric layers located much higher up than one would expect in the simple Eddington-Barbier approach, as we discovered by looking at their contribution functions (Magain 1986). For example, at $T_{\rm {eff}}$ = 6000 K, log g = 4.50, [Fe/H] = -3, the average height of formation actually moves out to log $_{10}~\tau_{500}\sim -0.7$, with a significant contribution to the absorption profile still coming from layers as high up as log $_{10}~\tau_{500}\sim -2$. Also note the steep increase of the level population at low metallicity, as one moves outward in the atmospheric model (Fig. 2).

.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.

... routine

Available via FTP from:
ftp://ftp.mso.anu.edu.au/pub/damian/corr_nlte/c/

.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.

... al.

Namely, $\log\epsilon_{C_\odot}=8.39$ and $\log\epsilon_{O_\odot}=8.66$ (Asplund et al. 2004; Asplund et al. 2005b), as compared to $\log\epsilon_{C_\odot}=8.41$ and $\log\epsilon_{O_\odot}=8.74$ in Akerman et al. (2004). We believe the values we adopt are best estimates since they are averaged over different lines and they account for 3D/non-LTE effects.

.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.