We present our latest 3D model atmospheres for carbon-enhanced metal-poor (CEMP) stars computed with the CO5BOLD code. The stellar parameters are representative of hot turn-off objects (Teff ~ 6250 K, log g = 4.0, [Fe/H]=−3). The main purpose of these models is to investigate the role of 3D effects on synthetic spectra of the CH G-band (4140-4400 Å), the CN BX-band (3870-3890 Å), and several UV OH transitions (3122-3128 Å). By comparison with the synthetic spectra from standard 1D model atmospheres (assuming local thermodynamic equilibrium, LTE), we derive 3D abundance corrections for carbon and oxygen of up to −0.5 and −0.7 dex, respectively.

We present a critical review of the construction of 3D model atmospheres with emphasis on modeling challenges. We discuss the basic physical processes which give rise to the effects which set 3D models apart from 1D standard models. We consider elemental abundances derived from molecular features, and the determination of the microturbulence with 3D models. The examples serve as illustration of the limitations inherent to 1D, however, also to 3D modeling. We find that 3D models can provide constraints on the microturbulence parameter, and predict substantial corrections for abundances derived from molecular species.

We present a new determination of the solar nitrogen abundance making use of 3D hydrodynamical modelling of the solar photosphere, which is more physically motivated than traditional static 1D models. We selected suitable atomic spectral lines, relying on equivalent width measurements already existing in the literature. For atmospheric modelling we used the co5bold 3D radiation hydrodynamics code. We investigated the influence of both deviations from local thermodynamic equilibrium (non-LTE effects) and photospheric inhomogeneities (granulation effects) on the resulting abundance. We also compared several atlases of solar flux and centre-disc intensity presently available. As a result of our analysis, the photospheric solar nitrogen abundance is A(N) = 7.86 ± 0.12.

In this talk, we review the basic assumptions and physics covered by classical 1D LTE model atmospheres. We will focus on ATLAS and MARCS models of F-G-K stars and describe what resources are available through the web, both in terms of codes and model-atmosphere grids. We describe the advances made in hydrodynamical simulations of convective stellar atmospheres with the CO5BOLD code and what grids and resources are available, with a prospect of what will be available in the near future.

I present ideas of Gaia’s impact on the determination of the properties of stars primarily connected to the study of their atmospheres. This mainly relates to effective temperatures, gravities and high-fidelity chemical abundances obtained by combining envisioned Gaia measurements with ground-based spectroscopy ranging from single objects to well-selected stellar populations. I further discuss the impact of Gaia on the study of the kinematics of atmospheric flows.

During the five years of the mission, the Gaia spectrograph, the Radial Velocity Spectrometer (RVS) will repeatedly survey the celestial sphere down to magnitude V ~ 17–18. This talk presents: (i) the system which is currently developed within the Gaia Data Processing and Analysis Consortium (DPAC) to reduce and calibrate the spectra and to derive the radial and rotational velocities, (ii) the RVS expected performances and (iii) scientific returns.

The calibration of the Radial Velocity Zero-Point (RVZP) of the Radial Velocity Spectrometer (RVS) will be performed with the help of Radial Velocity (RV) standard stars and asteroids. A full-sky list of RV standard stars candidates has been built for this purpose within the Development Unit (DU) 640 of the Gaia DPAC Consortium. A ground-based campaign of RV observations has been initiated to eliminate unsuitable candidates. Simulations have also been performed in order to estimate in advance the number of RV standard stars and asteroids crossing the RVS field of view. These standard stars and asteroids will also be very helpful for controlling any harmful effect (especially the CCD radiation damage) on calibration and RVZP during the lifetime of the Gaia project. Kinematic RV are expected to be published at the end of the project for the brightest stars, taking into account gravitational redshift and convective shifts for the lines present in the RVS spectral wavelength range.

Based on 3D hydrodynamical model atmospheres computed with the CO5BOLD code and 3D non-LTE (NLTE) line formation calculations, we study the effect of the convection-induced line asymmetry on the derived 6Li abundance for a range in effective temperature, gravity, and metallicity covering the stars of the Asplund et al. (2006) sample. When the asymmetry effect is taken into account for this sample of stars, the resulting 6Li/7Li ratios are reduced by about 1.5% on average with respect to the isotopic ratios determined by Asplund et al. (2006). This purely theoretical correction diminishes the number of significant 6Li detections from 9 to 4 (2σ criterion), or from 5 to 2 (3σ criterion). In view of this result the existence of a 6Li plateau appears questionable. A careful reanalysis of individual objects by fitting the observed lithium 6707 Å doublet both with 3D NLTE and 1D LTE synthetic line profiles confirms that the inferred 6Li abundance is systematically lower when using 3D NLTE instead of 1D LTE line fitting. Nevertheless, halo stars with unquestionable 6Li detection do exist even if analyzed in 3D-NLTE, the most prominent example being HD 84937.

Thanks to the high multiplex and efficiency of Giraffe at the VLT we have been able for the first time to observe the Li I doublet in the Main Sequence stars of a globular cluster. At the same time we observed Li in a sample of Sub-Giant stars of the same B-V colour.

Our final sample is composed of 84 SG stars and 79 MS stars. In spite of the fact that SG and MS span the same temperature range we find that the equivalent widths of the Li I doublet in SG stars are systematically larger than those in MS stars, suggesting a higher Li content among SG stars. This is confirmed by our quantitative analysis carried out making use of 1D hydrostatic plane-parallel models and 3D hydrodynamical simulations of the stellar atmospheres.

We derived the effective temperatures of stars in our the sample from Hα fitting. Theoretical profiles were computed using 3D hydrodynamical simulations and 1D ATLAS models. Therefore, we are able to determined 1D and 3D-based effective temperatures. We then infer Li abundances taking into account non-local thermodynamical equilibrium effects when using both 1D and 3D models.

We find that SG stars have a mean Li abundance higher by 0.1 dex than MS stars. This result is obtained using both 1D and 3D models. We also detect a positive slope of Li abundance with effective temperature, the higher the temperature the higher the Li abundance, both for SG and MS stars, although the slope is slightly steeper for MS stars. These results provide an unambiguous evidence that the Li abundance changes with evolutionary status.

The physical mechanisms responsible for this behaviour are not yet clear, and none of the existing models seems to describe accurately these observations. Based on these conclusions, we believe that the cosmological lithium problem still remains an open question.

Much of what we know about the chemical composition of the Universe actually stems from the chemical composition of stars, which is often deciphered from the spectra emerging from their atmospheres. Cool, low-mass and long-living stars allow to study the evolution of the Universe's chemistry from a time shortly after the big bang until today. The observation and interpretation of stellar spectra is a classical field in astronomy but is still undergoing vivid developments. The enormous increase in available computational resources opened-up possibilities which led to a revolution in the degree of realism to which modelers can mimic Nature. High-resolution, high-stability, high-efficiency spectrographs are now routinely providing stellar spectra whose full information content can only be exploited if a very much refined description of a stellar atmosphere is at hand.

We recently presented (Sbordone et al., 2009a) the largest sample to date of lithium abundances in extremely metal-poor (EMP) Halo dwarf and Turn-Off (TO) stars. One of the most crucial aspects in estimating Li abundances is the Teff determination, since the Li I 670.8 nm doublet is highly temperature sensitive. In this short contribution we concentrate on the Teff determination based on Hα wings fitting, and on its sensitivity to the chosen stellar gravity.