We present high-resolution and high-quality UVES spectroscopic data of the metal-poor double-lined spectroscopic binary CS 22876--032 ([Fe/H] $\sim -3.7$ dex), with the goal to derive the $^6$Li/$^7$Li isotopic ratio by analysing the \ion{Li}{i} $\lambda$~670.8~nm doublet. We coadd all 28 useful spectra normalised and corrected for radial velocity to the rest frame of the primary star. We fit the Li profile with a grid of the 3D-NLTE synthetic spectra, to take into account the line profile asymmetries induced by stellar convection, and perform Monte Carlo simulations to evaluate the uncertainty of the fit of the Li line profile. We check that the veiling factor does not affect the derived isotopic ratio, $^6$Li/$^7$Li, and only modifies the Li abundance, A(Li), by about 0.15~dex. The best fit of the Li profile of the primary star provides A(Li)~$ = 2.17 \pm 0.01$~dex and $^6$Li/$^7$Li~$=8^{+2}_{-5}$\% at 68\% confidence level. In addition, we improve the Li abundance of the secondary star at A(Li)~$= 1.55 \pm 0.04$~dex, which is about 0.6~dex lower than that of the primary star. The analysis of the Li profile of the primary star is consistent with no detection of $^6$Li and provides an upper-limit to the isotopic ratio of $^6$Li/$^7$Li~$< 10$\% at this very low metallicity, about 0.5~dex lower in metallicity than previous attempts for detection of $^6$Li in extremely metal poor stars. These results do not solve or worsen the cosmological $^7$Li problem, nor support the need for non standard $^6$Li production in the early Universe.

Extremely metal-poor stars are keys to understand the early evolution of our Galaxy. The ESO large programme TOPoS has been tailored to analyse a new set of metal-poor turn-off stars, whereas most of the previously known extremely metal-poor stars are giant stars. Sixty five turn-off stars (preselected from SDSS spectra) have been observed with the X-Shooter spectrograph at the ESO VLT Unit Telescope 2, to derive accurate and detailed abundances of magnesium, silicon, calcium, iron, strontium and barium. We analysed medium-resolution spectra (R ~ 10 000) obtained with the ESO X-Shooter spectrograph and computed the abundances of several alpha and neutron-capture elements using standard one-dimensional local thermodynamic equilibrium (1D LTE) model atmospheres. Our results confirms the super-solar [Mg/Fe] and [Ca/Fe] ratios in metal-poor turn-off stars as observed in metal-poor giant stars. We found a significant spread of the [alpha/Fe] ratios with several stars showing sub-solar [Ca/Fe] ratios. We could measure the abundance of strontium in 12 stars of the sample, leading to abundance ratios [Sr/Fe] around the Solar value. We detected barium in two stars of the sample. One of the stars (SDSS J114424-004658) shows both very high [Ba/Fe] and [Sr/Fe] abundance ratios (>1 dex).

The atmospheres of cool stars are temporally and spatially inhomogeneous due to the effects of convection.The influence of this inhomogeneity, referred to as granulation, on colours has never been investigated over a large range of effective temperatures and gravities. We aim to study, in a quantitative way, the impact of granulation on colours. We use the CIFIST grid of CO5BOLD hydrodynamical models to compute emerging fluxes. These in turn are used to compute theoretical colours in the UBVRI,2MASS,Hipparcos,Gaia and SDSS systems. Every CO5BOLD model has a corresponding one dimensional (1D) plane-parallel LHD model computed for the same atmospheric parameters, which we used to define a "3D correction" that can be applied to colours computed from fluxes computed from any 1D model atmosphere code. The 3D corrections on colours are generally small, of the order of a few hundredths of a magnitude, yet they are far from negligible. We find that ignoring granulation effects can lead to underestimation of Teff by up to 200K and overestimation of gravity by up to 0.5dex, when using colours as diagnostics. We have identified a major shortcoming in how scattering is treated in the current version of the CIFIST grid, which could lead to offsets of the order 0.01mag, especially for colours involving blue and UV bands.We have investigated the Gaia and Hipparcos photometric systems and found that the (G-H_p),(BP-RP) diagram is immune to the effects of granulation. In addition, we point to the potential of the RVS photometry as a metallicity diagnostic. Our investigation shows that the effects of granulation should not be neglected if one wants to use colours as diagnostics of the stellar parameters of F,G,K stars. A limitation is that scattering is treated as true absorption in our current computations, thus our 3D corrections are likely an upper limit to the true effect. (Abridged)

Extremely metal-poor stars provide us with indirect information on the first generations of massive stars. The TOPoS survey has been designed to increase the census of these stars and to provide a chemical inventory that is as detailed as possible. Seven of the most iron-poor stars have been observed with the UVES spectrograph at the ESO VLT Kueyen 8.2m telescope to refine their chemical composition. We analysed the spectra based on 1D LTE model atmospheres, but also used 3D hydrodynamical simulations of stellar atmospheres. We measured carbon in six of the seven stars: all are carbon-enhanced and belong to the low-carbon band, defined in the TOPoS II paper. We measured lithium (A(Li)=1.9) in the most iron-poor star (SDSS J1035+0641, [Fe/H] < -5.2 ). We were also able to measure Li in three stars at [Fe/H]~ -4.0, two of which lie on the Spite plateau. We confirm that SDSS J1349+1407 is extremely rich in Mg, but not in Ca. It is also very rich in Na. Several of our stars are characterised by low alpha-to-iron ratios. The lack of high-carbon band stars at low metallicity can be understood in terms of evolutionary timescales of binary systems. The detection of Li in SDSS J1035+0641 places a strong constraint on theories that aim at solving the cosmological lithium problem. The Li abundance of the two warmer stars at [Fe/H]~ -4.0 places them on the Spite plateau, while the third, cooler star, lies below. We argue that this suggests that the temperature at which Li depletion begins increases with decreasing [Fe/H]. SDSS J1349+1407 may belong to a class of Mg-rich EMP stars. We cannot assess if there is a scatter in alpha-to-iron ratios among the EMP stars or if there are several discrete populations. However, the existence of stars with low alpha-to-iron ratios is supported by our observations.

One of the primary objectives of the TOPoS survey is to search for the most metal-poor stars. Our search has led to the discovery of one of the most iron-poor objects known, SDSS\,J092912.32+023817.0. This object is a multiple system, in which two components are clearly detected in the spectrum. We have analysed 16 high-resolution spectra obtained using the UVES spectrograph at the ESO 8.2m VLT telescope to measure radial velocities and determine the chemical composition of the system. Cross correlation of the spectra with a synthetic template yields a double-peaked cross-correlation function (CCF) for eight spectra, and in one case there is evidence for the presence of a third peak. Chemical analysis of the spectrum obtained by averaging all the spectra for which the CCF showed a single peak found that the iron abundance is [Fe/H]=-4.97. The system is also carbon enhanced with [C/Fe]=+3.91 (A(C)=7.44). From the permitted oxygen triplet we determined an upper limit for oxygen of [O/Fe]<+3.52 such that C/O>1.3. We are also able to provide more stringent upper limits on the Sr and Ba abundances ([Sr/Fe]<+0.70, and [Ba/Fe]<+1.46, respectively).

The solar photospheric oxygen abundance is still widely debated. Adopting the solar chemical composition based on the "low" oxygen abundance, as determined with the use of three-dimensional (3D) hydrodynamical model atmospheres, results in a well-known mismatch between theoretical solar models and helioseismic measurements that is so far unresolved. We carry out an independent redetermination of the solar oxygen abundance by investigating the center-to-limb variation of the OI IR triplet lines at 777 nm in different sets of spectra with the help of detailed synthetic line profiles based on 3D hydrodynamical CO5BOLD model atmospheres and 3D non-LTE line formation calculations with NLTETD. The idea is to simultaneously derive the oxygen abundance,A(O), and the scaling factor SH that describes the cross-sections for inelastic collisions with neutral hydrogen relative the classical Drawin formula. The best fit of the center-to-limb variation of the triplet lines achieved with the CO5BOLD 3D solar model is clearly of superior quality compared to the line profile fits obtained with standard 1D model atmospheres. Our best estimate of the 3D non-LTE solar oxygen abundance is A(O) = 8.76 +/- 0.02, with the scaling factor SH in the range between 1.2 and 1.8. All 1D non-LTE models give much lower oxygen abundances, by up to -0.15 dex. This is mainly a consequence of the assumption of a $\mu$-independent microturbulence.

In the course of the TOPoS (Turn Off Primordial Stars) survey, aimed at discovering the lowest metallicity stars, we have found several carbon-enhanced metal-poor (CEMP) stars. We here present our analysis of six CEMP stars. Calcium and carbon are the only elements that can be measured in all six stars. The range is -5.0<=[Ca/H]< -2.1 and 7.12<=A(C)<=8.65. For star SDSS J1742+2531 we were able to detect three FeI lines from which we deduced [Fe/H]=-4.80, from four CaII lines we derived [Ca/H]=-4.56, and from synthesis of the G-band we derived A(C)=7.26. For SDSS J1035+0641 we were not able to detect any iron lines, yet we could place a robust (3sigma) upper limit of [Fe/H]< -5.0 and measure the Ca abundance, with [Ca/H]=-5.0, and carbon, A(C)=6.90. No lithium is detected in the spectrum of SDSS J1742+2531 or SDSS J1035+0641, which implies a robust upper limit of A(Li)<1.8 for both stars. Our measured carbon abundances confirm the bimodal distribution of carbon in CEMP stars, identifying a high-carbon band and a low-carbon band. We propose an interpretation of this bimodality according to which the stars on the high-carbon band are the result of mass transfer from an AGB companion, while the stars on the low-carbon band are genuine fossil records of a gas cloud that has also been enriched by a faint supernova (SN) providing carbon and the lighter elements. (Abridged)

Moderately r-process-enriched stars (r-I) are at least four times as common as those that are greatly enriched in r-process elements (r-II), and the abundances in their atmospheres are important tools for obtaining a better understanding of the nucleosynthesis processes responsible for the origin of the elements beyond the iron peak. The main aim of this work is to derive abundances for a sample of seven metal-poor stars with classified as r-I stars, to understand the role of these stars for constraining the astrophysical nucleosynthesis event(s) that is(are) responsible for the production of the r-process, and to investigate whether they differ, in any significant way, from the r-II stars. We carried out a detailed abundance analysis based on high-resolution spectra obtained with the VLT/UVES spectrograph. The OSMARCS LTE 1D model atmosphere grid was employed, along with the spectrum synthesis code Turbospectrum. We have derived abundances of light elements Li, C, and N, alpha-elements, odd-Z elements, iron-peak elements, and the trans-iron elements from the first peak, the second peak, the third peak, and the actinides regions. The results are compared with values for these elements for r-II and normal very and extremely metal-poor stars reported in the literature, ages based on radioactive chronometry are explored using different models, and a number of conclusions about the r-process and the r-I stars are presented. Hydrodynamical models were used for some elements, and general behaviors for the 3D corrections were presented.

The metal-weak tail of the metallicity distribution function (MDF) of the Galactic Halo stars contains crucial information on the formation mode of the first generation of stars. To determine this observationally, it is necessary to observe large numbers of extremely metal-poor stars. We present here the Turn-Off Primordial Stars survey (TOPoS) that is conducted as an ESO Large Programme at the VLT. This project has {four} main goals: (i) to understand the formation of low-mass stars in a low-metallicity gas: determine the metal-weak tail of the halo MDF below [M/H]=-3.5. In particular, we aim at determining the critical metallicity, that is the lowest metallicity sufficient for the formation of low-mass stars; (ii) to determine the relative abundance of the elements in extremely metal-poor stars, which are the signature of the massive first stars; (iii) to determine the trend of the lithium abundance at the time when the Galaxy formed; and (iv) to derive the fraction of C-enhanced extremely metal-poor stars with respect to normal extremely metal-poor stars. The large number of stars observed in the SDSS provides a good sample of candidates of stars at extremely low metallicity. Candidates with turn-off colours down to magnitude g=20 were selected from the low-resolution spectra of SDSS by means of an automated procedure. X-Shooter has the potential of performing the necessary follow-up spectroscopy, providing accurate metallicities and abundance ratios for several key elements for these stars. We here present the stellar parameters of the first set of stars. The nineteen stars range in iron abundance between -4.1 and -2.9 dex relative to the Sun. Two stars have a high radial velocity and, according to our estimate of their kinematics, appear to be marginally bound to the Galaxy and are possibly accreted from another galaxy.

The extremely metal-poor stars are the direct descendants of the first generation stars. They carry the chemical signature of the pristine Universe at the time they formed, shortly after the Big Bang. We aim to derive information about extremely metal-poor stars from their observed spectra. Four extremely metal-poor stars were selected from the Sloan Digital Sky Survey (SDSS) and observed during the guaranteed observing time of X-Shooter. The X-Shooter spectra were analysed using an automatic code, MyGIsFOS, which is based on a traditional analysis method. It makes use of a synthetic grid computed from one-dimensional, plane-parallel, hydrostatic model atmospheres. The low metallicity derived from the SDSS spectra is confirmed here. Two kinds of stars are found. Two stars are confirmed to be extremely metal-poor, with no evidence of any enhancement in carbon. The two other stars are strongly enhanced in carbon. We could not derive iron abundance for one of them, while [Ca/H] is below -4.5. Two of the stars are members of the rare population of extremely metal-poor stars low in alpha elements.

In this work we present first results of our current project aimed at combining the 3D hydrodynamical stellar atmosphere approach with non-LTE (NLTE) spectral line synthesis for a number of key chemical species. We carried out a full 3D-NLTE spectrum synthesis of the oxygen IR 777 nm triplet, using a modified and improved version of our NLTE3D package to calculate departure coefficients for the atomic levels of oxygen in a CO5BOLD 3D hydrodynamical solar model atmosphere. Spectral line synthesis was subsequently performed with the Linfor 3D code. In agreement with previous studies, we find that the lines of the oxygen triplet produce deeper cores under NLTE conditions, due to the diminished line source function in the line forming region. This means that the solar oxygen IR 777 nm lines should be stronger in NLTE, leading to negative 3D NLTE-LTE abundance corrections. Qualitatively this result would support previous claims for a relatively low solar oxygen abundance. Finally, we outline several further steps that need to be taken in order to improve the physical realism and numerical accuracy of our current 3D-NLTE calculations.

Detailed abundances of the elements produced by r-process nucleosynthesis in various circumstances are our best observational clues to their origin, since the site(s) of r-element production is(are) still not known with certainty. A small fraction of extremely metal-poor (EMP) stars exhibit excesses of heavy neutron-capture elements produced in the r-process, and CS 31082-001 is among the 4 well-known r-process-enhanced EMP stars. Observations with HST/STIS provide abundances for elements observable only from the UV region. Here we aim to supplement the optical data with abundances from near-UV spectroscopy of the first and second peak of the r-elements, which are crucial to giving insight into the nucleosynthesis of the elements beyond iron. The UVES spectrum provided additional measurements, thereby improving the previous results. The spectra were analyzed with the OSMARCS LTE model atmosphere and with a consistent approach based on the spectrum synthesis code Turbospectrum to derive abundances of heavy elements in CS 31082-001, using updated oscillator strengths from the recent literature. We computed synthetic spectra for all lines of the elements of interest, checking for proper intensities and possible blends. We combined the abundances of heavy elements derived in previous works with the derivation of abundances from all reliable new list of lines, for the first and second peaks of r-elements. We were able to derive new abundances for 23 n-elements, 6 of them - Ge, Mo, Lu, Ta, W, and Re - were not available in previous works, making this star the most complete r-II object studied, with a total of 37 detections of n-capture elements. We also present the first NLTE+3D lead abundance in this star. The results provide improved constraints on the nature of the r-process.

The presence of 6Li in the atmospheres of metal-poor halo stars is usually inferred from the detection of a subtle extra depression in the red wing of the 7Li doublet line at 670.8 nm. However, the intrinsic line asymmetry caused by convective flows in the photospheres of cool stars is almost indistinguishable from the asymmetry produced by a weak 6Li blend on a (presumed) symmetric 7Li profile. Previous determinations of the 6Li/ 7Li isotopic ratio based on 1D model atmospheres, ignoring the convection-induced line asymmetry, must therefore be considered as upper limits. By comparing synthetic 1D LTE and 3D non-LTE line profiles of the Li 670.8 nm feature, we quantify the differential effect of the convective line asymmetry on the derived 6Li abundance as a function of effective temperature, gravity, and metallicity. As expected, we find that the asymmetry effect systematically reduces the resulting 6Li/7Li ratios. Depending on the stellar parameters, the 3D-1D offset in 6Li/7Li ranges between -0.005 and -0.020. When this purely theoretical correction is taken into account for the Asplund 2006 sample of stars, the number of significant 6Li detections decreases from 9 to 5 (2 sigma criterion), or from 5 to 2 (3 sigma criterion). We also present preliminary results of a re-analysis of high-resolution, high S/N spectra of individual metal-poor turn-off stars, to see whether the "second Lithium problem" actually disappears when accounting properly for convection and non-LTE line formation in 3D stellar atmospheres. Out of 8 stars, HD84937 seems to be the only significant (2 sigma) detection of 6Li. In view of our results, the existence of a 6Li plateau appears questionable.

(Abridged) Extremely metal-poor stars contain the fossil records of the chemical composition of the early Galaxy. The NLTE profiles of the calcium lines were computed in a sample of 53 extremely metal-poor stars with a modified version of the program MULTI. With our new model atom we are able to reconcile the abundance of Ca deduced from the Ca I and Ca II lines in Procyon. -We find that [Ca/Fe] = 0.50 $\pm$ 0.09 in the early Galaxy, a value slightly higher than the previous LTE estimations. -The scatter of the ratios [X/Ca] is generally smaller than the scatter of the ratio [X/Mg] where X is a "light metal" (O, Na, Mg, Al, S, and K) with the exception of Al. These scatters cannot be explained by error of measurements, except for oxygen. Surprisingly, the scatter of [X/Fe] is always equal to, or even smaller than, the scatter around the mean value of [X/Ca]. -We note that at low metallicity, the wavelength of the Ca I resonance line is shifted relative to the (weaker) subordinate lines, a signature of the effect of convection. -The Ca abundance deduced from the Ca I resonance line (422.7 nm) is found to be systematically smaller at very low metallicity, than the abundance deduced from the subordinate lines.

The early Universe had a chemical composition consisting of hydrogen, helium and traces of lithium1, almost all other elements were created in stars and supernovae. The mass fraction, Z, of elements more massive than helium, is called "metallicity". A number of very metal poor stars have been found some of which, while having a low iron abundance, are rich in carbon, nitrogen and oxygen. For theoretical reasons and because of an observed absence of stars with metallicities lower than Z=1.5E-5, it has been suggested that low mass stars (M<0.8M\odot, the ones that survive to the present day) cannot form until the interstellar medium has been enriched above a critical value, estimated to lie in the range 1.5E-8\leqZ\leq1.5E-6, although competing theories claiming the contrary do exist. Here we report the chemical composition of a star with a very low Z\leq6.9E-7 (4.5E-5 of that of the Sun) and a chemical pattern typical of classical extremely metal poor stars, meaning without the enrichment of carbon, nitrogen and oxygen. This shows that low mass stars can be formed at very low metallicity. Lithium is not detected, suggesting a low metallicity extension of the previously observed trend in lithium depletion. Lithium depletion implies that the stellar material must have experienced temperatures above two million K in its history, which points to rather particular formation condition or internal mixing process, for low Z stars.

Context: The discovery and chemical analysis of extremely metal-poor stars permit a better understanding of the star formation of the first generation of stars and of the Universe emerging from the Big Bang. aims: We report the study of a primordial star situated in the centre of the constellation Leo (SDSS J102915+172027). method: The star, selected from the low resolution-spectrum of the Sloan Digital Sky Survey, was observed at intermediate (with X-Shooter at VLT) and at high spectral resolution (with UVES at VLT). The stellar parameters were derived from the photometry. The standard spectroscopic analysis based on 1D ATLAS models was completed by applying 3D and non-LTE corrections. results: An iron abundance of [Fe/H]=--4.89 makes SDSS J102915+172927 one of the lowest [Fe/H] stars known. However, the absence of measurable C and N enhancements indicates that it has the lowest metallicity, Z<= 7.40x10^{-7} (metal-mass fraction), ever detected. No oxygen measurement was possible. conclusions: The discovery of SDSS J102915+172927 highlights that low-mass star formation occurred at metallicities lower than previously assumed. Even lower metallicity stars may yet be discovered, with a chemical composition closer to the composition of the primordial gas and of the first supernovae.

Context: Extremely metal-poor stars (EMP) are very rare objects that hold in their atmospheres the fossil record of the chemical composition of the early phases of Galactic evolution. Finding these objects and determining their chemical composition provides important constraints on these early phases. Aims: Using a carefully designed selection method, we chose a sample of candidate EMP stars from the low resolution spectra of the Sloan Digital Sky Survey and observed them with X-Shooter at the VLT to confirm their metallicities and determine abundances for as many elements as possible. Method: The X-Shooter spectra are analysed by means of one-dimensional, plane-parallel, hydrostatic model atmospheres. Corrections for the granulation effects are computed using CO5BOLD hydrodynamical simulations. Results: All the candidates are confirmed to be EMP stars, proving the efficiency of our selection method within about 0.5 dex. The chemical composition of this sample is compatible with those of brighter samples, suggesting that the stars in the Galactic halo are well mixed. Conclusions These observations show that it is feasible to observe, in a limited amount of time, a large sample of about one hundred stars among EMP candidates selected from the SDSS. Such a size of sample will allow us, in particular, to confirm or refute the existence of a vertical drop in the Galactic Halo metallicity distribution function around [Fe/H] ~ -3.5.

Heavy element abundances in extremely metal-poor stars provide strong constraints on the processes of forming these elements in the first stars. We attempt to determine precise abundances of strontium in a homogeneous sample of extremely metal-poor stars. The abundances of strontium in 54 very or extremely metal-poor stars, was redetermined by abandoning the local thermodynamic equilibrium (LTE) hypothesis, and fitting non-LTE (NLTE) profiles to the observed spectral lines. The corrected Sr abundances and previously obtained NLTE Ba abundances are compared to the predictions of several hypothetical formation processes for the lighter neutron-capture elements. Our NLTE abundances confirm the previously determined huge scatter of the strontium abundance in low metallicity stars. This scatter is also found (and is even larger) at very low metallicities (i. e. early in the chemical evolution). The Sr abundance in the extremely metal-poor (EMP) stars is compatible with the main r-process involved in other processes (or by variations of the r-process), as we briefly discuss.

Some insight on the first generation of stars can be obtained from the chemical composition of their direct descendants, extremely metal-poor stars (EMP), with metallicity less than or equal to 1/1000 of the solar metalllicity. Such stars are exceedingly rare, the most successful surveys, for this purpose, have so far provided only about 100 stars with 1/1000 the solar metallicity and 4 stars with about 1/10000 of the solar metallicity. The Sloan Digital Sky Survey has the potential to provide a large number of candidates of extremely low metallicity. X-Shooter has the unique capability of performing the necessary follow-up spectroscopy providing accurate metallicities and abundance ratios for several elements (Mg, Al, Ca, Ti, Cr, Sr,...) for EMP candidates. We here report on the results for the first two stars observed in the course of our franco-italian X-Shooter GTO. The two stars were targeted to be of metallicity around -3.0, the analysis of the X-Shooter spectra showed them to be of metallicity around -2.0, but with a low alpha to iron ratio, which explains the underestimate of the metallicity from the SDSS spectra. The efficiency of X-Shooter allows an in situ study of the outer Halo, for the two stars studied here we estimate distances of 3.9 and 9.1 Kpc, these are likely the most distant dwarf stars studied in detail to date.

The chemical composition of extremely metal-poor stars (EMP stars; [Fe/H]<~ -3) is a unique tracer of early nucleosynthesis in the Galaxy. As such stars are rare, we wish to find classes of luminous stars which can be studied at high resolution. We aim to determine the detailed chemical composition of the two EMP stars CS30317-056 and CS22881-039, originally thought to be red horizontal-branch (RHB) stars, and compare it to earlier results for EMP stars as well as to nucleosynthesis yields from various supernova (SN) models. In the analysis, we discovered that our targets are in fact the two most metal-poor RR Lyrae stars known. Our detailed abundance analysis, taking into account the variability of the stars, is based on VLT/UVES spectra (R~ 43000) and 1D LTE OSMARCS model atmospheres and synthetic spectra. For comparison with SN models we also estimate NLTE corrections for a number of elements. We derive LTE abundances for the 16 elements O, Na, Mg, Al, Si, S, Ca, Sc, Ti, Cr, Mn, Fe, Co, Ni, Sr and Ba, in good agreement with earlier values for EMP dwarf, giant and RHB stars. Li and C are not detected in either star. NLTE abundance corrections are newly calculated for O and Mg and taken from the literature for other elements. The resulting abundance pattern is best matched by model yields for supernova explosions with high energy and/or significant asphericity effects. Our results indicate that, except for Li and C, the surface composition of EMP RR Lyr stars is not significantly affected by mass loss, mixing or diffusion processes; hence, EMP RR Lyr stars should also be useful tracers of the chemical evolution of the early Galactic halo. The observed abundance ratios indicate that these stars were born from an ISM polluted by energetic, massive (25 - 40M*) and/or aspherical supernovae, but the NLTE corrections for Sc and certain other elements do play a role in the choice of model.

Sulfur is important: the site of its formation is uncertain, and at very low metallicity the trend of [S/Fe] against [Fe/H] is controversial. Below [Fe/H]=-2.0, [S/Fe] remains constant or it decreases with [Fe/H], depending on the author and the multiplet used in the analysis. Moreover, although sulfur is not significantly bound in dust grains in the ISM, it seems to behave differently in DLAs and in old metal-poor stars. We aim to determine precise S abundance in a sample of extremely metal-poor stars taking into account NLTE and 3D effects. NLTE profiles of the lines of the multiplet 1 of SI have been computed using a new model atom for S. We find sulfur in EMP stars to behave like the other alpha-elements, with [S/Fe] remaining approximately constant for [Fe/H]<-3. However, [S/Mg] seems to decrease slightly as a function of [Mg/H]. The overall abundance patterns of O, Na, Mg, Al, S, and K are best matched by the SN model yields by Heger & Woosley. The [S/Zn] ratio in EMP stars is solar, as found also in DLAs. We obtain an upper limit on the abundance of sulfur, [S/Fe] < +0.5, for the ultra metal-poor star CS 22949-037. This, along with a previous reported measurement of zinc, argues against the conjecture that the light-element abundances pattern in this star, and, by analogy, the hyper metal-poor stars HE 0107-5240 and HE 1327-2326, are due to dust depletion.

We present VLT-UVES Li abundances for 28 halo dwarf stars between [Fe/H]=-2.5 and -3.5, 10 of which have [Fe/H]<-3. Four different T_eff scales have been used. Direct Infrared Flux Method (IRFM) has been used on the basis of 2MASS infrared photometry. H_alpha wings have been fitted against synthetic grids computed by means of 1D LTE atmosphere models, assuming different self-broadening theories. Finally, a grid of H_alpha profiles has been computed by means of 3D hydrodynamical atmosphere models. The Li I doublet at 670.8 nm has been used to measure A(Li) by means of 3D hydrodynamical NLTE spectral syntheses. An analytical fit of A(Li)(3D, NLTE) as a function of equivalent width, T_eff, log g, and [Fe/H] has been derived and is made available. A(Li) does not exhibit a plateau below [Fe/H]=-3. A strong positive correlation with [Fe/H] appears, not influenced by the choice of the T_eff estimator. From a linear fit, we obtain a strong slope of about 0.30 dex in A(Li) per dex in [Fe/H], significant to 2-3 sigma, and consistent among all the four T_eff estimators. A significant slope is also detected in the A(Li)--T_eff plane, driven mainly by the coolest stars in the sample which appear Li-poor. Removing such stars does not alter the behavior in the A(Li)-[Fe/H] plane. The scatter in A(Li) increases by a factor of 2 towards lower metallicities, while the plateau appears very thin above [Fe/H]=-2.8. The meltdown of the Spite plateau below [Fe/H]\sim-3 is established, but its cause is unclear. If the primordial A(Li) is the one derived from WMAP, it appears difficult to envision a single depletion phenomenon producing a thin, metallicity independent plateau above [Fe/H]=-2.8, and a highly scattered, metallicity dependent distribution below. The fact that no star below [Fe/H]=-3 lies above the plateau suggests that they formed at plateau level and underwent subsequent depletion.

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 (2 sigma criterion), or from 5 to 2 (3 sigma 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 A 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.

LTE abundances of light elements in extremely metal-poor (EMP) stars have been previously derived from high quality spectra. New derivations, free from the NLTE effects, will better constrain the models of the Galactic chemical evolution and the yields of the very first supernovae. The NLTE profiles of the magnesium and potassium lines have been computed in a sample of 53 extremely metal-poor stars with a modified version of the program MULTI and adjusted to the observed lines in order to derive the abundances of these elements. The NLTE corrections for magnesium and potassium are in good agreement with the works found in the literature. The abundances are slightly changed, reaching a better precision: the scatter around the mean of the abundance ratios has decreased. Magnesium may be used with confidence as reference element. Together with previously determined NLTE abundances of sodium and aluminum, the new ratios are displayed, for comparison, along the theoretical trends proposed by some models of the chemical evolution of the Galaxy, using different models of supernovae.

We present FLAMES/GIRAFFE spectroscopy obtained at the Very Large Telescope (VLT). Using these observations we have been able for the first time to observe the Li I doublet in the Main Sequence stars of a Globular Cluster. We also 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 which makes use of both 1D and 3D model atmospheres. We find that SG stars show, on average, a Li abundance higher by 0.1 dex than MS stars. 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 unambigous evidence that the Li abundance changes with evolutionary status. The physical mechanisms that contribute to this are not yet clear, since none of the proposed models seem to describe accurately the observations. Whether such mechanism can explain the cosmological lithium problem, is still an open question.

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 (MS) stars of a Globular Cluster. At the same time we observed Li in a sample of Sub-Giant (SG) 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. We derived the effective temperatures, from H$\alpha$ fitting, and NLTE Li abundances of the stars in our the sample, using 3D and 1D models. We find that SG stars have a mean Li abundance higher by 0.1dex than MS stars, using both 1D and 3D models. We also detect a positive slope of Li abundance with effective temperature. 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.

We present solar photospheric abundances for 12 elements from optical and near-infrared spectroscopy. The abundance analysis was conducted employing 3D hydrodynamical (CO5BOLD) as well as standard 1D hydrostatic model atmospheres. We compare our results to others with emphasis on discrepancies and still lingering problems, in particular exemplified by the pivotal abundance of oxygen. We argue that the thermal structure of the lower solar photosphere is very well represented by our 3D model. We obtain an excellent match of the observed center-to-limb variation of the line-blanketed continuum intensity, also at wavelengths shortward of the Balmer jump.

The presence of convective motions in the atmospheres of metal-poor halo stars leads to systematic asymmetries of the emergent spectral line profiles. Since such line asymmetries are very small, they can be safely ignored for standard spectroscopic abundance analysis. However, when it comes to the determination of the 6Li/7Li isotopic ratio, q(Li)=n(6Li)/n(7Li), the intrinsic asymmetry of the 7Li line must be taken into account, because its signature is essentially indistinguishable from the presence of a weak 6Li blend in the red wing of the 7Li line. In this contribution we quantity the error of the inferred 6Li/7Li isotopic ratio that arises if the convective line asymmetry is ignored in the fitting of the lithium blend at 6707 A. Our conclusion is that 6Li/7Li ratios derived by Asplund et al. (2006), using symmetric line profiles, must be reduced by typically dq(Li) ~ 0.015. This diminishes the number of certain 6Li detections from 9 to 4 stars or less, casting some doubt on the existence of a 6Li plateau.

Most Globular Clusters are believed to host a single stellar populations. They can thus be considered a good place to study the Spite plateau and probe for possible evolutionary modifications of the Li content. We want to determine the Li content of subgiant (SG) and Main Sequence (MS) stars of the old, metal-poor globular cluster NGC 6397. This work was aimed not only at studying possible Li abundance variations but to investigate the cosmological Li discrepancy. Here, we present FLAMES/GIRAFFE observations of a sample of 84 SG and 79 MS stars in NGC 6397 selected in a narrow range of B-V colour and, therefore, effective temperatures. We determine both Teff and A(Li) using 3D hydrodynamical model atmospheres for all the MS and SG stars of the sample. We find a significant difference in the Li abundance between SG stars and MS stars, the SG stars having an A(Li) higher by almost 0.1 dex on average. We also find a decrease in the A(Li) with decreasing Teff, both in MS and SG stars, albeit with a significantly different slope for the two classes of stars. This suggests that the lithium abundance in these stars is, indeed, altered by some process, which is Teff-dependent. The Li abundance pattern observed in NGC 6397 is different from what is found among field stars, casting some doubt on the use of Globular Cluster stars as representative of Population II with respect to the Li abundance. None of the available theories of Li depletion appears to satisfactorily describe our observations.

CONTEXT:The detailed chemical abundances of extremely metal-poor (EMP) stars are key guides to understanding the early chemical evolution of the Galaxy. Most existing data are, however, for giant stars which may have experienced internal mixing later. AIMS: We aim to compare the results for giants with new, accurate abundances for all observable elements in 18 EMP turnoff stars. METHODS:VLT/UVES spectra at R ~45,000 and S/N~ 130 per pixel (330-1000 nm) are analysed with OSMARCS model atmospheres and the TURBOSPECTRUM code to derive abundances for C, Mg, Si, Ca, Sc, Ti, Cr, Mn, Co, Ni, Zn, Sr, and Ba. RESULTS: For Ca, Ni, Sr, and Ba, we find excellent consistency with our earlier sample of EMP giants, at all metallicities. However, our abundances of C, Sc, Ti, Cr, Mn and Co are ~0.2 dex larger than in giants of similar metallicity. Mg and Si abundances are ~0.2 dex lower (the giant [Mg/Fe] values are slightly revised), while Zn is again ~0.4 dex higher than in giants of similar [Fe/H] (6 stars only). CONCLUSIONS:For C, the dwarf/giant discrepancy could possibly have an astrophysical cause, but for the other elements it must arise from shortcomings in the analysis. Approximate computations of granulation (3D) effects yield smaller corrections for giants than for dwarfs, but suggest that this is an unlikely explanation, except perhaps for C, Cr, and Mn. NLTE computations for Na and Al provide consistent abundances between dwarfs and giants, unlike the LTE results, and would be highly desirable for the other discrepant elements as well. Meanwhile, we recommend using the giant abundances as reference data for Galactic chemical evolution models.

Barium is a key element in constraining the evolution of the (not well understood) r-process in the first galactic stars and currently the Ba abundances in these very metal-poor stars were mostly measured under the Local Thermodynamical Equilibrium (LTE) assumption, which may lead in general to an underestimation of Ba. We present here determinations of the barium abundance taking into account the non-LTE (NLTE) effects in a sample of extremely metal-poor stars (EMP stars): 6 turnoff stars and 35 giants. The NLTE profiles of the three unblended Ba II lines (455.4, 585.3, 649.6nm) have been computed. The computations were made with a modified version of the MULTI code, applied to an atomic model of the Ba atom with 31 levels of Ba I, 101 levels of Ba II, and compared to the observations. The ratios of the NLTE abundances of barium relative to Fe are slightly shifted towards the solar ratio. In the plot of [Ba/Fe] versus [Fe/H], the slope of the regression line is slightly reduced as is the scatter. In the interval -3.3 <[Fe/H] < -2.6, [Ba/Fe] decreases with a slope of about 1.4 and a scatter close to 0.44. For [Fe/H] <-3.3 the number of stars is not sufficient to decide whether [Ba/Fe] keeps decreasing (and then CD-38:245 should be considered as a peculiar "barium-rich star") or if a plateau is reached as soon as [Ba/Fe] ~ -1. In both cases the scatter remains quite large, larger than what can be accounted for by the measurement and determination errors, suggesting the influence of a complex process of Ba production, and/or inefficient mixing in the early Galaxy.

Two problems are discussed here. The first one is the 0.4 dex discrepancy between the 7Li abundance derived from the spectra of metal-poor halo stars on the one hand, and from Big Bang nucleosynthesis, based on the cosmological parameters constrained by the WMAP measurements, on the other hand. Lithium, indeed, can be depleted in the convection zone of unevolved stars. The understanding of the hydrodynamics of the crucial zone near the bottom of the convective envelope in dwarfs or turn-off stars of solar metallicity has recently made enormous progress with the inclusion of internal gravity waves. However, similar work for metal-poor stars is still lacking. Therefore it is not yet clear whether the depletion occurring in the metal-poor stars themselves is adequate to produce a 7Li plateau. The second problem pertains to the large amount of 6Li recently found in metal-poor halo stars. The convection-related asymmetry of the 7Li line could mimic the signal attributed so far to the weak blend of 6Li in the red wing of the 7Li line. Theoretical computations show that the signal generated by the asymmetry of 7Li is 2.0, 2.1, and 3.7 per cent for [Fe/H]= -3.0, -2.0, -1.0, respectively (Teff =6250 K and log g=4.0 [cgs]). In addition we re-investigate the statistical properties of the 6Li plateau and show that previous analyses were biased. Our conclusion is that the 6Li plateau can be reinterpreted in terms of intrinsic line asymmetry, without the need to invoke a contribution of 6Li. (abridged)

The solar oxygen abundance has undergone a major downward revision in the last decade, the most noticeable one being the update including 3D hydrodynamical simulations to model the solar photosphere. Up to now, such an analysis has been carried out only by one group using one radiation-hydrodynamics code. We investigate the photospheric oxygen abundance considering lines from atomic transitions. We also consider the relationship between the solar model used and the resulting solar oxygen abundance, to understand whether the downward abundance revision is specifically related to 3D hydrodynamical effects. We perform a new determination of the solar photospheric oxygen abundance by analysing different high-resolution high signal-to-noise ratio atlases of the solar flux and disc-centre intensity making use of the latest generation of CO5BOLD 3D solar model atmospheres. We find 8.73 < logNoxygen/Nhydrogen +12 < 8.79. The lower and upper value represent extreme assumptions on the role of collisional excitation and ionisation by neutral hydrogen for the NLTE level populations of neutral oxygen. The error of our analysis is +-0.04 +- 0.03 dex, the last being related to NLTE corrections, the first error to any other effect. 3D ``granulation effects'' do not play a decisive role in lowering the oxygen abundance. Our recommended value, considering our present ignorance of the role of collisions with hydrogen atoms on the NLTE level populations of oxygen, is log(Noxygen/Nhydrogen) = 8.76 +- 0.07. The reasons which have led to lower O abundances in the past are identified as (1) the lower equivalent widths adopted, and (2) the choice of neglecting collisions with hydrogen atoms in the statistical equilibrium calculations for oxygen.

Aims: Aluminium is a key element to constrain the models of the chemical enrichment and the yields of the first supernovae. But obtaining precise Al abundances in extremely metal-poor (EMP) stars requires that the non-LTE effects be carefully taken into account. Methods: The NLTE profiles of the blue resonance aluminium lines have been computed in a sample of 53 extremely metal-poor stars with a modified version of the program MULTI applied to an atomic model of the Al atom with 78 levels of Al I and 13 levels of Al II, and compared to the observations. Results: With these new determinations, all the stars of the sample show a ratio Al/Fe close to the solar value: [Al/Fe] =-0.06 +- 0.10 with a very small scatter. These results are compared to the models of the chemical evolution of the halo using different models of SN II and are compatible with recent computations. The sodium-rich giants are not found to be also aluminium-rich and thus, as expected, the convection in these giants only brings to the surface the products of the Ne-Na cycle.

In ESO period 65 (April-September 2000) the large programme 165.N-0276, led by Roger Cayrel, began making use of UVES at the Kueyen VLT telescope. Known within the Team and outside as ``First Stars'', it was aimed at obtaining high resolution, high signal-to-noise ratio spectra in the range 320 nm -- 1000 nm for a large sample of extremely metal-poor (EMP) stars identified from the HK objective prism survey.In this contribution we highlight the main results of the large programme.

We have used high-resolution, high-S/N ratio spectra from the UVES spectrograph at the ESO VLT telescope. Long-term radial-velocity measurements and broad-band photometry allow us to determine improved orbital elements and stellar parameters for both components. We use OSMARCS 1D models and the {{\scshape turbospectrum}} spectral synthesis code to determine the abundances of Li, O, Na, Mg, Al, Si, Ca, Sc, Ti, Cr, Mn, Fe, Co and Ni. We also use the CO$^5$BOLD model atmosphere code to compute the 3D abundance corrections, especially for Li and O. We find a metallicity of [Fe/H]$\sim -3.6$ for both stars, using 1D models with 3D corrections of $\sim -0.1$ dex from averaged 3D models. We determine the oxygen abundance from the near-UV OH bands; the 3D corrections are large, -1 and -1.5 dex for the secondary and primary respectively, and yield [O/Fe] $\sim 0.8$, close to the high-quality results obtained from the [OI] 630 nm line in metal-poor giants. Other [$\alpha$/Fe] ratios are consistent with those measured in other dwarfs and giants with similar [Fe/H], although Ca and Si are somewhat low ([X/Fe]$\la 0$). Other element ratios follow those of other halo stars. The Li abundance of the primary star is consistent with the Spite plateau, but the secondary shows a lower abundance; 3D corrections are small. The Li abundance in the primary star supports the extension of the {{\em Spite Plateau}} value at the lowest metallicities, without any decrease. The low abundance in the secondary star could be explained by endogenic Li depletion, due to its cooler temperature. If this is not the case, another, yet unknown mechanism may be causing increased scatter in A(Li) at the lowest metallicities.

Our aim is to measure accurate, homogeneous neutron-capture element abundances for the sample of 32 EMP giant stars studied earlier in this series, including 22 stars with [Fe/H] $< -$3.0. Based on high-resolution, high S/N spectra from the ESO VLT/UVES, 1D, LTE model atmospheres, and synthetic spectrum fits, we determine abundances or upper limits for the 16 elements Sr, Y, Zr, Ba, La, Ce, Pr, Nd, Sm, Eu, Gd, Dy, Ho, Er, Tm, and Yb in all stars. As found earlier, [Sr/Fe], [Y/Fe], [Zr/Fe] and [Ba/Fe] are below Solar in the EMP stars, with very large scatter. However, we find a tight anti-correlation of [Sr/Ba], [Y/Ba], and [Zr/Ba] with [Ba/H] for $-4.5 <$ [Ba/H] $< -2.5$, also when subtracting the contribution of the main $r$-process as measured by [Ba/H]. The huge, well-characterised scatter of the [n-capture/Fe] ratios in our EMP stars is in stark contrast to the negligible dispersion in the [$\alpha$/Fe] and [Fe-peak/Fe] ratios for the same stars found in Paper V. These results demonstrate that a second (``weak'' or LEPP) $r$-process dominates the production of the lighter neutron-capture elements for [Ba/H] $< -2.5$. The combination of very consistent [$\alpha$/Fe] and erratic [n-capture/Fe] ratios indicates that inhomogeneous models for the early evolution of the halo are needed. Our accurate data provide strong constraints on future models of the production and mixing of the heavy elements in the early Galaxy.

Context: Line asymmetries are generated by convective Doppler shifts in stellar atmospheres, especially in metal-poor stars, where convective motions penetrate to higher atmospheric levels. Such asymmetries are usually neglected in abundance analyses. The determination of the 6Li/7Li isotopic ratio is prone to suffering from such asymmetries, as the contribution of 6Li is a slight blending reinforcement of the red wing of each component of the corresponding 7Li line, with respect to its blue wing. Aims: The present paper studies the halo star HD 74000 and estimates the impact of convection-related asymmetries on the Li isotopic ratio determination. Method: Two methods are used to meet this aim. The first, which is purely empirical, consists in deriving a template profile from another element that can be assumed to originate in the same stellar atmospheric layers as Li I, producing absorption lines of approximately the same equivalent width as individual components of the 7Li I resonance line. The second method consists in conducting the abundance analysis based on NLTE line formation in a 3D hydrodynamical model atmosphere, taking into account the effects of photospheric convection. Results: The results of the first method show that the convective asymmetry generates an excess absorption in the red wing of the 7Li absorption feature that mimics the presence of 6Li at a level comparable to the hitherto published values. This opens the possibility that only an upper limit on 6Li/7Li has thus far been derived. The second method confirms these findings. Conclusions: From this work, it appears that a systematic reappraisal of former determinations of 6Li abundances in halo stars is warranted.

Abundance ratios in extremely metal-poor (EMP) stars are a good indication of the chemical composition of the gas in the earliest phases of the Galaxy evolution. It had been found from an LTE analysis that at low metallicity, and in contrast with most of the other elements, the scatter of [Na/Fe] versus [Fe/H] was surprisingly large and that, in giants, [Na/Fe] decreased with metallicity. Since it is well known that the formation of sodium lines is very sensitive to non-LTE effects, to firmly establish the behaviour of the sodium abundance in the early Galaxy, we have used high quality observations of a sample of EMP stars obtained with UVES at the VLT, and we have taken into account the non-LTE line formation of sodium. The profiles of the two resonant sodium D lines (only these sodium lines are detectable in the spectra of EMP stars) have been computed in a sample of 54 EMP giants and turn-off stars (33 of them with [Fe/H]< -3.0) with a modified version of the code MULTI, and compared to the observed spectra. With these new determinations in the range -4 <[Fe/H]< -2.5, both [Na/Fe] and [Na/Mg] are almost constant with a low scatter. In the turn-off stars and "unmixed" giants (located in the low RGB): [Na/Fe] = -0.21 +/- 0.13 or [Na/Mg] = -0.45 +/- 0.16. These values are in good agreement with the recent determinations of [Na/Fe] and [Na/Mg] in nearby metal-poor stars. Moreover we confirm that all the sodium-rich stars are "mixed" stars (i.e., giants located after the bump, which have undergone an extra mixing). None of the turn-off stars is sodium-rich. As a consequence it is probable that the sodium enhancement observed in some mixed giants is the result of a deep mixing.

Aims. This study aims to determine the level and constancy of the Spite plateau as definitively as possible from homogeneous high-quality VLT-UVES spectra of 19 of the most metal-poor dwarf stars known. Methods. Our high-resolution (R ~ 43000), high S/N spectra are analysed with OSMARCS 1D LTE model atmospheres and turbospectrum synthetic spectra to determine effective temperatures, surface gravities, and metallicities, as well as Li abundances for our stars. Results. Eliminating a cool subgiant and a spectroscopic binary, we find 8 stars to have -3.5 < [Fe/H] < -3.0 and 9 stars with -3.0 < [Fe/H] < -2.5. Our best value for the mean level of the plateau is A(Li) =2.10 +- 0.09. The scatter around the mean is entirely explained by our estimate of the observational error and does not allow for any intrinsic scatter in the Li abundances. In addition, we conclude that a systematic error of the order of 200 K in any of the current temperature scales remains possible. The iron excitation equilibria in our stars support our adopted temperature scale, which is based on a fit to wings of the Halpha line, and disfavour hotter scales, which would lead to a higher Li abundance, but fail to achieve excitation equilibrium for iron. Conclusions. We confirm the previously noted discrepancy between the Li abundance measured in extremely metal-poor turnoff stars and the primordial Li abundance predicted by standard Big-Bang nucleosynthesis models adopting the baryonic density inferred from WMAP. We discuss recent work explaining the discrepancy in terms of diffusion and find that uncertain temperature scales remain a major question. (abridged)

From high-resolution VLT/UVES spectra (R~43,000), we determine abundances or upper limits for Li, C, N, O, and other important elements, as well as 12C/13C isotopic ratio for three Carbon enhanced metal poor stars. All three stars have -3.30 <= [Fe/H] <= -2.85 and moderate to high CNO abundances. CS 22958-042 is one of the most carbon-rich CEMP stars known ([C/Fe] = +3.2), while CS 29528-041 (one of the few N-enhanced metal-poor stars known) is one of the most nitrogen rich ([N/Fe] = +3.0). Oxygen is very high in CS 31080-095 ([O/Fe] = +2.35) and in CS 22958-042 ([O/Fe] = +1.35). All three stars exhibit [Sr/Fe] < 0; Ba is not detected in CS 22958-042 ([Ba/Fe] < -0.53), but it is moderately enhanced ([Ba/Fe] ~ 1) in the other two stars. CS 22958-042 displays one of the largest sodium overabundances yet found in CEMP stars ([Na/Fe] = +2.8). CS 22958-042 has 12C/13C = 9, similar to most other CEMP stars without enhanced neutron-capture elements, while 12C/13C <= 40 in CS 31080-095. CS 31080-095 and CS 29528-041 have A(Li) ~ 1.7, below the Spite Plateau, while Li is not detected in CS 22958-042. CS 22958-042 is a CEMP-no star, but the other two stars are in no known class of CEMP star and thus either constitute a new class or are a link between the CEMP-no and CEMP-s classes, adding complexity to the abundance patterns for CEMP stars. We interpret the abundance patterns in our stars to imply that current models for the presumed AGB binary progenitors lack an extra-mixing process, similar to those apparently operating in RGB stars.

Extremely metal-poor (EMP) stars preserve a fossil record of the composition of the ISM when the Galaxy formed. It is crucial, however, to verify whether internal mixing has modified their surface. We aim to understand the CNO abundance variations found in some, but not all EMP field giants analysed earlier. Mixing beyond the first dredge-up of standard models is required, and its origin needs clarification.The 12C/13C ratio is the most robust diagnostic of deep mixing, because it is insensitive to the adopted stellar parameters and should be uniformly high in near-primordial gas. We have measured 12C and 13C abundances in 35 EMP giants from high-quality VLT/UVES spectra. Correlations with other abundance data are used to study the depth of mixing.The 12C/13C ratio is found to correlate with [C/Fe] (and Li/H), and clearly anti-correlate with [N/Fe]. Evidence for such deep mixing is observed in giants above log L/Lsolar = 2.6, brighter than in less metal-poor stars, but matching the bump in the luminosity function in both cases. Three of the mixed stars are also Na-and Al-rich, another signature of deep mixing, but signatures of the ON cycle are not clearly seen in these stars. Extra mixing processes clearly occur in luminous RGB stars. The Na-and Al-rich giants could be AGB stars themselves, but an inhomogeneous early ISM or pollution from a binary companion remain possible alternatives (abridged).

We compute the evolution of the abundances of barium and europium in the Milky Way and we compare our results with the observed abundances from the recent UVES Large Program "First Stars". We use a chemical evolution model which already reproduces the majority of observational constraints. We confirm that barium is a neutron capture element mainly produced in the low mass AGB stars during the thermal-pulsing phase by the 13C neutron source, in a slow neutron capture process. However, in order to reproduce the [Ba/Fe] vs. [Fe/H] as well as the Ba solar abundance, we suggest that Ba should be also produced as an r-process element by massive stars in the range 10-30 solar masses. On the other hand, europium should be only an r-process element produced in the same range of masses (10-30 solar masses), at variance with previous suggestions indicating a smaller mass range for the Eu producers. As it is well known, there is a large spread in the [Ba/Fe] and [Eu/Fe] ratios at low metallicities, although smaller in the newest data. With our model we estimate for both elements (Ba and Eu) the ranges for the r-process yields from massive stars which better reproduce the trend of the data. We find that with the same yields which are able to explain the observed trends, the large spread in the [Ba/Fe] and [Eu/Fe] ratios cannot be explained even in the context of an inhomogeneous models for the chemical evolution of our Galaxy. We therefore derive the amount by which the yields should be modified to fully account for the observed spread. We then discuss several possibilities to explain the size of the spread. We finally suggest that the production ratio of [Ba/Eu] could be almost constant in the massive stars.

We have investigated the poorly-understood origin of nitrogen in the early Galaxy by determining N abundances in 35 extremely metal-poor halo giants (22 stars have [Fe/H]<-3.0) using the C and O abundances determined in Paper V. Because any dredge-up of CNO processed material to the surface may complicate the interpretation of CNO abundances in giants, we have also measured the surface abundance of lithium. Our sample shows a clear dichotomy between two groups of stars. The first group shows evidence of C to N conversion through CN cycling and strong Li dilution, a signature of mixing. The second group shows no evidence for C to N conversion, and Li is only moderately diluted, and we conclude that their C and N abundances are very close to those of the gas from which they formed in the early Galaxy. These "unmixed" stars reflect the abundances in the early Galaxy: the [C/Fe] ratio is constant (about +0.2 dex) and the [C/Mg] ratio is close to solar at low metallicity, favouring a high C production by massive zero-metal supernovae. The [N/Fe] and [N/Mg] ratios scatter widely. The larger values of these ratios define a flat upper plateau ([N/Mg]= 0.0, [N/Fe]= +0.1), which could reflect higher values within a wide range of yields of zero-metal Sne II. Alternatively, by analogy with the DLA's, the lower abundances ([N/Mg]= -1.1, [N/Fe]= -0.7) could reflect generally low yields from the first Sne II, the other stars being N enhanced by winds of massive Asymptotic Giant Branch (AGB) stars. At present it cannot be decided whether primary N is produced primarily in SNe II or in massive AGB stars, or in both. The stellar N abundances and [N/O] ratios are compatible with those found in Damped Lyman-alpha (DLA) systems.

In a previous paper we were able to measure the abundance of uranium and thorium in the very-metal poor halo giant BPS CS 31082-001, but only obtained an upper limit for the abundance of lead (Pb). We have got from ESO 17 hours of additional exposure on this star in order to secure a detection of the minimum amount of lead expected to be present in CS 31082-001, the amount arising from the decay of the original content of Th and U in the star. We report here this successful detection. We find an LTE abundance log(Pb/H)+12=-0.55 \pm 0.15 dex, one dex below the upper limits given by other authors for the similar stars CS 22892-052 and BD +17d3248, also enhanced in r-process elements. From the observed present abundances of Th and U in the star, the expected amount of Pb produced by the decay of 232Th, and 238U alone, over 12-15 Gyr is -0.73\pm 0.17 dex. The decay of 235U is more difficult to estimate, but is probably slightly below the contribution of 238U, making the contribution of the 3 actinides only slightly below, or even equal to, the measured abundance. The contribution from the decay of 234U has was not included, for lack of published data. In this sense our determination is a lower limit to the contribution of actinides to lead production. We comment this result, and we note that if a NLTE analysis, not yet possible, doubles our observed abundance, the decay of the 3 actinides will still represent 50 per cent of the total lead, a proportion higher than the values considered so far in the literature.

We have carried out a new determination of abundances in the very metal-poor CH/CN strong stars CS 29497-34 and CS 22948-27, using high-resolution spectra obtained with the HARPS spectrograph at the 3.6m telescope of ESO, La Silla, that covers the range 400 - 690 nm at a resolution of R = 100,000. Both stars are found to be long period binaries. It is confirmed that the abundance patterns show an enhancement of all the alpha-elements (like Mg, Ca), of the proton capture elements (like Na and Al) and a strong enrichment in "r" and "s" process elements, where the s-enrichment is probably due to a mass transfer episode from a companion in its AGB phase. The possible origins of the abundance pattern and especially of the strong enhancement of both "s" and "r" elements are discussed.

We computed the evolution of the abundances of O, Mg, Si, Ca, K, Ti, Sc, Ni, Mn, Co, Fe and Zn in the Milky Way. We made use of the most widely adopted nucleosynthesis calculations and compared the model results with observational data with the aim of imposing constraints upon stellar yields. In order to fit at best the data in the solar neighborhood, when adopting the Woosley and Weaver (1995) yields for massive stars and the Iwamoto et al. (1999) ones for type Ia SNe, it is required that: i) the Mg yields should be increased in stars with masses from 11 to 20 M_sun and decreased in masses larger than 20 M_sun. The Mg yield should be also increased in SNe Ia. ii) The Si yields should be slightly increased in stars above 40 M_sun, whereas those of Ti should be increased between 11 and 20 M_sun and above 30 M_sun. iii) The Cr and Mn yields should be increased in stars with masses in the range 11-20 M_sun, iv) the Co yields in SNe Ia should be larger and smaller in stars in the range 11-20 M_sun, v) the Ni yield from type Ia SNe should be decreased, vi) the Zn yield from type Ia SNe should be increased. vii) The yields of O (metallicity dependent SN models), Ca, Fe, Ni, and Zn (the solar abundance case) in massive stars from Woosley and Weaver (1995) are the best to fit the abundance patterns of these elements since they do not need any change. We adopted also the yields by Nomoto et al. (1997) and Limongi and Chieffi (2003) for massive stars and discussed the corrections required in these yields in order to fit the observations. Finally, the small spread in the [el/Fe] ratios in the metallicity range from [Fe/H]=-4.0 up to -3.0 dex (Cayrel et al. 2003) is a clear sign that the halo of the Milky Way was well mixed even in the earliest phases of its evolution.

In the framework of an ESO Large Programme high quality spectra of 70 very metal-poor dwarfs and giants were obtained with the VLT-UVES combination. This paper reports the detailed analyses of the 35 giants of the sample. At the exception of a few template stars, all the targets are from the H&K Beers et al. survey. No less than 22 stars have [Fe/H] below -3.0. The imprint of yields of the first galactic or protogalactic supernovae is sampled in the photospheres of these stars. If [C/Fe], [Na/Fe] and [Al/Fe] show a spread significantly above the observational scatter, several abundance ratios present an amazing low scatter when plotted against [Fe/H], as [Cr/Fe] but also odd elements, as [K/Fe] or [Sc/Fe]. The abundance of oxygen has been derived from the [O I] line for 16 of these giants, in spite of the extreme weakness of the line. No significant slope with [Fe/H] is found. Similar plots have been made using Mg instead of Fe as reference element. With Mg, no clear trend with metallicity is obvious below [Mg/H]=-3., except perhaps for [Zn/Mg]. It is contemplated that the mild slope found in the very low scatter diagram [Cr/Fe] versus [Fe/H] might be spurious due for example to the LTE approximation plus use of 1-D models. This would relax the absence of scatter in the dilution of the SN ejectas within the ambient medium, as needed if [Fe/H] is a well defined function of [Cr/Fe].

We present an abundance analysis of the very metal-poor, carbon-enhanced star CS 29497-030. Our results indicate that this unusually hot turnoff star (Teff = 6650K, log g = 3.5) has a metallicity [Fe/H] = -2.8, and exhibits large overabundances of carbon ([C/Fe] = +2.38), nitrogen ([N/Fe]= +1.88), and oxygen ([O/Fe]= +1.67). This star also exhibits a large enhancement in its neutron-capture elements; the pattern follows that expected to arise from the s-process. In particular, the Pb abundance is found to be very high with respect to iron ([Pb/Fe]= +3.5), and also with respect to the second peak s-process elements (e.g., Ba, La, Ce, Nd), which fits into the newly introduced classification of lead (Pb) stars. The known spectroscopic binary status of this star, along with the observed s-process abundance pattern, suggest that it has accreted matter from a companion, which formerly was an Asymptotic Giant-Branch(AGB) star. In a preliminary analysis, we have also identified broad absorption lines of metallic species that suggest a large axial rotational velocity for this star, which may be the result of spin-up associated with the accretion of material from its previous AGB companion. In addition, this star is clearly depleted in the light element Li. When considered along with its rather high inferred temperature, these observations are consistent with the expected properties of a very low metallicity halo blue straggler.

We present here the abundance analysis of a metal poor carbon rich lead star, CS29497-030. High resolution and high signal to noise spectra were obtained using the UVES spectrograph on the 8.2m VLT-Kueyen telescope. The observations were made as a part of the Large Programme 165.N-0276, P.I. R. Cayrel. Abundance analysis was done using the latest version of the MARCS model atmospheres (Plez et. al. 1992) and the turbospectrum spectrum synthesis code. We have derived Teff = 6650K from the FeI lines. Visible and infrared broad band colours using the Alonso et al. (1996) calibration, gives similar temperatures. A log g value of 3.5 was obtained from the ionisation equilibrium of FeI and FeII, we remark that this gravity also satisfies the MgI/MgII, TiI/TiII and MnI/MnII equilibria, within errors. The abundance analysis indicates a metallicity, [Fe/H] = --2.7. A large overabundance of carbon ([C/Fe]=2.7) was found. We have also found large enhancement in the s-process elements and in particular lead shows an extremely high abundance of [Pb/Fe]=3.5, which makes this the star with the highest Pb/Fe ratio, up to date. The Pb/Ba ratio is found to be high ([Pb/Ba]=1.2) and the same s true for other second-peak s-process elements(e.g La, Ce, Nd). The star is a known spectroscopic binary with a period of 346 days (Preston & Sneden 2000). The abundance pattern suggests that CS 29497-30 has accreted matter from its companion, when it was in the AGB phase.

CS 22949--037 is one of the most metal-poor giants known ([Fe/H]$ \approx-4.0$), and it exhibits large overabundances of carbon and nitrogen (Norris et al.). Using VLT-UVES spectra of unprecedented quality, regarding resolution and S/N ratio, covering a wide wavelength range (from $\lambda = 350$ to 900 nm), we have determined abundances for 21 elements in this star over a wide range of atomic mass. The major new discovery is an exceptionally large oxygen enhancement, [O/Fe] $= 1.97\pm0.1$, as measured from the [OI] line at 630.0 nm. We find an enhancement of [N/Fe] of $2.56\pm 0.2$, and a milder one of [C/Fe] $= 1.17\pm$0.1, similar to those already reported in the literature. This implies $Z_{\star}=0.01 Z_{\odot}$. We also find carbon isotopic ratios $^{12}$C/$^{13}$C$ =4\pm2.0$ and $^{13}$C/$^{14}$N$ =0.03 ^{+0.035}_{-0.015}$, close to the equilibrium value of the CN cycle. Lithium is not detected. Na is strongly enhanced ([Na/Fe] $= +2.1 \pm 0.2$), while S and K are not detected. The silicon-burning elements Cr and Mn are underabundant, while Co and Zn are overabundant ([Zn/Fe]$ = +0.7$). Zn is measured for the first time in such an extremely metal-poor star. The abundances of the neutron-capture elements Sr, Y, and Ba are strongly decreasing with the atomic number of the element: [Sr/Fe] $\approx +0.3$, [Y/Fe] $\approx -0.1$, and [Ba/Fe] $\approx -0.6$. Among possible progenitors of CS 22949--037, we discuss the pair-instability supernovae. Such very massive objects indeed produce large amounts of oxygen, and have been found to be possible sources of primary nitrogen. Other scenarios are also discussed. A 30-40$M_{\odot}$ supernova, with fallback, seems the most likely progenitor for CS 22949--037.

We present a high-resolution spectroscopic analysis of the bright (V=11.7), extreme halo giant CS31082-001 ([Fe/H] = -2.9), obtained in an ESO-VLT Large Programme dedicated to very metal-poor stars. We find CS31082-001 to be extremely rich in r-process elements, comparable in this respect only to the similarly metal-poor, but carbon-enriched, giant CS22892-052. As a result of the extreme overabundance of the heaviest r-process elements, and negligible blending from CH and CN molecular lines, a reliable measurement is obtained of the U II line at 386 nm, for the first time in a halo star, along with numerous lines of Th II, as well as lines of 25 other r-process elements. Abundance estimates for a total of 43 elements are reported in CS31082-001, almost half of the entire periodic table. All elements with 56 \leq Z \leq 72 follow the Solar r-element pattern, reduced by about 1.25 dex ([r/Fe]=+1.7 dex, a factor 50). Pb, in contrast, seems to be below the shifted Solar r-process distribution, possibly indicating an error in the latter, while thorium is more enhanced than the lighter nuclides. Thus, while a universal production ratio for the r-process elements seems to hold in the interval 56 \leq Z \leq 72, it breaks down in the actinide region. When available, the U/Th is thus preferable to Th/Eu for radioactive dating: (i) because of its faster decay rate and smaller sensitivity to observational errors, and (ii) because the inital production ratio of the neighboring nuclides 238U and 232Th is more robustly predicted than the 151Eu/232Th ratio. Our current best estimate for the age of CS31082-001 is 14.0+/-2.4 Gyr.

During an ESO-VLT large program devoted to high-resolution spectroscopy of extremely metal-poor stars selected from the H&K survey of Beers and colleagues, a [Fe/H] giant star was found to be as enriched in neutron-capture r-process elements as CS 22892-052, but with a much reduced masking by molecular lines. This allowed the detection and the measurement of the uranium line at 3859 A, for the first time in an extremely old star. Making use of the short 238U decay time (4.47 Gyr) we obtained a radioactive dating of the formation of U and Th in of this star, born in the early days of the Galaxy.

The ages of the oldest stars in the Galaxy indicate when star formation began, and provide a minimum age for the Universe. Radioactive dating of meteoritic material and stars relies on comparing the present abundance ratios of radioactive and stable nuclear species to the theoretically predicted ratios of their production. The radioisotope $^{232}$Th (half-life 14 Gyr) has been used to date Galactic stars, but it decays by only a factor of two over the lifetime of the Universe. $^{238}$U (half-life 4.5 Gyr) is in principle a more precise age indicator, but even its strongest spectral line, from singly ionized uranium at a wavelength of 385.957 nm, has previously not been detected in stars. Here we report a measurement of this line in the very metal-poor star CS31082-001, a star which is strongly overabundant in its heavy elements. The derived uranium abundance, log(U/H) = -13.7+/-0.14+/-0.12 yields an age of 12.5+/-3 Gyr, though this is still model dependent. The observation of this cosmochronometer gives the most direct age determination of the Galaxy. Also, with improved theoretical and laboratory data, it will provide a highly precise lower limit to the age of the Universe.

The very-metal-poor halo star CS31082-001 was discovered to be very strongly r-process-enhanced during the course of a VLT+UVES high-resolution follow-up of metal-poor stars identified in the HK survey of Beers and colleagues. Both the strong n-capture element enhancement and the low carbon and nitrogen content of the star (reducing the CN molecular band contamination) led to the first $^{238}$U abundance measurement in a stellar spectrum (Cayrel et al. 2001), and the opportunity to use both radioactive species $^{238}$U and $^{232}$Th for dating the progenitor to this star. However, age computations all rely on the hypothesis that the r-process pattern is solar, as this was indeed observed in the other famous r-process-enhanced very metal poor stars CS22892-052 (Sneden et al. 1996, 2000) and in HD115444 (Westin et al. 2000). Here, we investigate whether this hypothesis is verified also for CS31082-001, using a preliminary analysis of over 20 abundances of n-capture elements in the range Z=38 to Z=92.

Based on detailed 2D radiation hydrodynamics (RHD) simulations, we have investigated the effects of photospheric temperature inhomogeneities induced by convection on spectroscopic determinations of the lithium abundance. Computations have been performed both for the solar case and for a metal-poor dwarf. NLTE effects are taken into account, using a five-level atomic model for LiI. Comparisons are presented with traditional 1D models having the same effective temperature and gravity. The net result is that, while LTE results differ dramatically between 1D and 2D models, especially in the metal-poor case, this does not remain true when NLTE effects are included: 1D/2D differences in the inferred NLTE Li abundance are always well below 0.1 dex. The present computations still assume LTE in the continuum. New computations removing this assumption are planned for the near future.

From an initial sample of a few hundred stars selected for an Hipparcos program, a subsample of 33 objects with direct determination of their effective temperature, bolometric magnitude and [Fe/H] has been derived. An accurate observational (logTeff, Mbol) HR diagram has been obtained, and compared to theoretical isochrones. Attention has been concentrated on the cool part of the HR diagram, for which evolutionary effects are negligible. Whereas solar metallicity stars show a reasonable agreement with theoretical expectation, a clear discrepancy occurs for stars more metal-poor than [Fe/H] = -0.5. This discrepancy is reduced to practically zero, if corrections to LTE abundances are taken into account for Fe, and if sedimentation effects are included in stellar evolution computations.

High S/N ratio spectra have been obtained at CFHT at a spectral resolution of 100 000 for HD 84397 and two other metal poor stars. For HD 84 937, the S/N of 650 per pixel was obtained, yielding Li-6/Li-7 = 0.052 with a sigma of 0.019. The Li-6 abundance is compared with recent models of formation of light elements LiBeB. This comparison shows that Li-6 is either undepleted or only moderatly depleted in the star of interest. The depletion of Li-7 in this star is estimated to be less than 0.1 dex. The abundance of lithium on the Spite plateau accordingly can be takien as representative of the cosmological primordial lithium. See the joint paper astro-ph/9811327.

The relationship between the respective depletion factors for 6Li and 7Li in metal-poor stars is studied for two simple models of depletion by nuclear burning by mixing with deeper layers, below the bottom of the convective zone. The depletion of 6Li is usually much larger than the depletion of 7Li because it occurs during the pre-mainsequence already.

The primordial abundances of Deuterium, he4, and li7 are crucial to determination of the baryon density of the Universe in the framework of standard Big Bang nucleosynthesis (BBN). li6 which is only produced in tiny quantities and it is generally not considered to be a cosmological probe. However, recent major observational advances have produced an estimate of the li6/li7 ratio in a few very old stars in the galactic halo which impacts the question whether or not the lithium isotopes are depleted in the outer layers of halo stars, through proton induced reactions at the base of (or below) the convective zone. li6 is a pure product of spallation through the major production reactions, fast oxygen and alphas interacting on interstellar H, He (especially in the early Galaxy). The rapid nuclei are both synthesized and accelerated by SN II. In this context, the \li6 evolution should go in step with that of beryllium and boron, recently observed by the Keck and HST telescopes. Li6 adds a new constraint on the early spallation in the Galaxy. In particular, if confirmed, the Li6/Be9 ratio observed in two halo stars (HD 84937, BD +263578) gives strong boundary conditions on the composition and the spectrum of the rapid particles involved. We show that Li6 is essentially intact in halo stars, and a fortiori \li7. We can define a range of the Li6 abundance in the very early Galaxy consistent with Big Bang nucleosynthesis (5.6 10(-14) to 3. 10(-13) . Following the evolution at increasing metallicity, we explain the abundance in the solar system within a factor of about 2.

A library of 211 echelle spectra taken with ELODIE at the Observatoire de Haute-Provence is presented. It provides a set of spectroscopic standards covering the full range of gravities and metallicities in the effective temperature interval [4000 K, 6300 K]. The spectra are straightened, wavelength calibrated, cleaned of cosmic ray hits, bad pixels and telluric lines. They cover the spectral range [440 nm, 680 nm] with an instrumental resolution of 42000. For each star, basic data were compiled from the Hipparcos catalogue and the Hipparcos Input Catalogue. Radial velocities with a precision better than 100 m/s are given. Atmospheric parameters (Teff, log g, [Fe/H]) from the literature are discussed. Because of scattered determinations in the bibliography, even for the most well-known stars, these parameters were adjusted by an iterative process which takes account of common or different spectral features between the standards, using our homogeneous set of spectra. Revised values of (Teff, log g, [Fe/H]) are proposed. They are still consistent with the literature, and also lead to the self-consistency of the library, in the sense that similar spectra have similar atmospheric parameters. This adjustment was performed by using step by step a method based on the least square comparison of carefully prepared spectra, which was originally developed for the on-line estimation of the atmospheric parameters of faint field stars (companion paper in the main journal). The spectra and corresponding data will only be available in electronic form at the CDS (ftp cdsarc.u-strasbg.fr or http://cdsweb.u-strasbg.fr/Abstract.html).

We present a method estimating the atmospheric parameters Teff, log g, [Fe/H] for stars observed, even at low signal to noise ratio, with the echelle spectrograph ELODIE on the 193cm telescope at Observatoire de Haute-Provence. The method relies on the least-square comparison of the spectrum of a target star to a library of 211 spectra of reference stars for which the atmospheric parameters are well known and which were observed with the same instrument. In order to obtain a meaningful comparison between target and reference spectra, all features which are not intrinsic to the objects must be removed. This treatment involves the correction of the blaze efficiency for each order, cosmic rays hits and telluric line removal, convolution of the spectra to a common spectral resolution, wavelength scale and flux level adjustment. The library available at the present time covers the effective temperature range [4000K, 6300K], the metallicity range [-2.9, +0.35] and the gravities of both unevolved and evolved stars existing at these temperatures and metallicities. Tests performed with the actual library allow us to estimate the internal accuracy to be 86 K, 0.28 dex and 0.16 dex for Teff, log g, [Fe/H] for a target star with S/N = 100 and 102 K, 0.29 dex and 0.17 dex at S/N = 10. This accuracy will improve in the future as the number of reference stars in the library will increase. The software (named TGMET) has been installed at Observatoire de Haute-Provence for the on-line analysis of the high-resolution spectra of ELODIE, which was originally conceived for accurate radial velocity measurements.