Astronomical Science

The Abundance of Measured for the First Time Beyond Our Galaxy

Alessio Mucciarelli1 In a seminal paper, Spite & Spite (1982) utes after the . In fact, if the Maurizio Salaris2 first noted that dwarf Population II stars ­primordial lithium abundance derived Piercarlo Bonifacio3 in the Solar Neighbourhood (with from the CMB results is incorrect, the Lorenzo Monaco4 [Fe/H] < –1.5 and effective temperatures current SBBN model would have to be Sandro Villanova5 > 5800 K) share the same lithium abun- drastically re-thought, introducing “new dance, regardless of their metallicity physics”, e.g., the inclusion of “new” high- and temperature, a feature known as the energy, non-thermal particles and the 1 Dipartimento di Fisica & Astronomia, Spite Plateau. The derived lithium abun- annihilation/decay of dark matter Universitá degli Studi di Bologna, Italy dance turns out to be in the range ­particles. On the other hand, if the initial 2 Astrophysics Research Institute, A(Li) = 2.1–2.4 dex, depending on the lithium content in Population II stars is ­Liverpool John Moores University, adopted temperature scale. The existence depleted by transport processes, the ­Liverpool, United Kingdom of a narrow lithium plateau has been measurement of the lithium abundance 3 GEPI, Observatoire de Paris, CNRS, ­confirmed by three decades of observa- in these stars could provide robust con- Univ. Paris Diderot, Meudon, France tions, both in halo field stars and in glob- straints on the efficiency of turbulent 4 ESO ular cluster stars. mechanisms occurring at the bottom of 5 Universidad de Concepcion, Casilla, the stellar convective envelope (which Concepcion, Chile The results obtained with the Wilkinson is a free parameter in the stellar models). Microwave Anisotropy Probe (WMAP; Spergel et al., 2007) and Planck (Planck The discrepancy between the primor- collaboration, 2013) satellites have pro- An alternative route: The lower RGB stars dial lithium abundance derived from vided an alternative route to inferring

Population II dwarf stars and from the ΩB. In fact, these high precision measure- Mucciarelli, Salaris & Bonifacio (2012) predictions of standard Big Bang ments of the cosmic microwave back- proposed an alternative/complementary

­nucleosynthesis is one of the most ground (CMB) have allowed ΩB to be esti- route to investigate the initial lithium intriguing and challenging open ques- mated with unprecedented accuracy. abundance in Population II stars (with tions in modern astrophysics. The The derived value of ΩB, coupled with the respect to the observations of dwarf use of lower red giant branch stars, standard stars), by measuring the surface lithium instead of the usual method of observ- (SBBN) model, provided a lithium abun- abundance in lower red giant branch ing dwarf stars, represents a new dance of 2.72 ± 0.06 dex (Coc et al., (RGB) stars. These stars are defined as approach to attacking the problem. 2013). This value is significantly higher, by the giants evolving between the first Lithium in distant, extragalactic stellar at least a factor of three, than the lithium dredge-up and the luminosity level of the systems, for which observations of abundance derived from dwarf stars. RGB Bump. When a star evolves off the dwarf stars are precluded because of Such a discrepancy is referred to as the , the convection propa- their faintness, becomes open for cosmological lithium problem. gates inward (first dredge-up) reaching investigation. From observations with hot regions where lithium has been FLAMES at the VLT, we have been At present, the discrepancy between the burned. Lithium-free material is dredged able to derive for the first time the initial Spite Plateau and the CMB + SBBN up to the surface, with the main effect of lithium abundance in an extragalactic results is still unexplained. Three possible reducing the surface lithium abundance. globular cluster, namely M54 in the explanations appear to be especially When the convective envelope attains Sagittarius galaxy. promising: its maximum penetration, depletion of the 1) the effect of atomic diffusion and surface lithium abundance ends. After- some competing additional mixing, the wards the lithium abundance remains Lithium, together with hydrogen and combined effect of which decreases constant until the star reaches the lumi- helium, is synthesised during the early the lithium abundances in the atmos- nosity level of the RGB Bump; then, an phase of the Universe, in the first few pheres of dwarf stars (see e.g., Korn et additional mixing episode occurs, further minutes after the Big Bang. All the ele- al., 2006); reducing the surface lithium abundance ments heavier than lithium are produced 2) inadequacies in the SBBN models used down to virtually zero, or to effectively later, mainly during the nucleosynthesis to calculate the lithium abundance non-measurable values (see Figure 1). occurring in the stellar interiors. In par- (Iocco et al., 2009); ticular, the primordial lithium abundance 3) lithium depletion driven by Population II The lower RGB stars (both in the Galactic is strictly linked to the baryonic density stars during early Galaxy evolution field and globular clusters) display a

(ΩB), a cosmological parameter that (Piau et al., 2006). ­constant lithium abundance, defining a quantifies the amount of ordinary matter. plateau that mirrors the Spite Plateau, but Hence, the study of the lithium abun- Whatever the solution of the lithium at a lower abundance (A(Li) ~ 0.9–1.0 dance in the oldest stars is crucial for ­problem is, the investigation of this dis- dex). Since the amount of lithium deple- several different astrophysical topics, crepancy (and its solution) is a formidable tion after the first dredge-up can be pre- i.e., cosmology, stellar evolution and chance to understand SBBN in greater dicted easily from stellar models, the globular cluster formation. depth and refine our description of the ­lithium abundance measured among the processes occurring in the first few min- lower RGB stars can be used to infer or

The Messenger 158 – December 2014 45 Astronomical Science Mucciarelli A. et al., Lithium Measured for the First Time Beyond Our Galaxy

",! 2!!-

2OHSD/K@SD@T   TW +NVDQ1&! Ck 3$ -+ +H

   -NQL@KHRD

 %HQRS #QDCFDlTO 1&!!TLO  

        

KNF++ी 6@UDKDMFSGÄ

Figure 1. Behaviour of the surface lithium abundance Figure 2. The observed lithium doublet of the aver- as a function of luminosity for the stars in the globular age spectrum obtained by combining 51 member cluster M4 (from Mucciarelli et al., 2011). The value stars of M54 observed with FLAMES GIRAFFE. of the initial lithium abundance as derived by the CMB Superimposed are the best-fit synthetic spectrum + SBBN is marked as a blue dashed line. The value and two synthetic spectra calculated with lithium of the lithium abundance as derived from dwarf stars abundance variations of ± 0.2 dex. (the Spite Plateau) is marked by the blue solid line. constrain the initial lithium abundance of where the observation of dwarf stars is Figure 2 shows the lithium doublet at those stars. precluded because of their distance. 6707 Å observed in an average spectrum, obtained by combining together the spec- This new diagnostic has extraordinary tra of the 51 stars, in order to enhance potential and advantages. In particular: Lithium abundance in M54 the spectral quality. Superimposed are 1) The derived abundance of lithium is the best-fit synthetic spectrum and two totally independent of the abundances We have recently approached the lithium synthetic spectra calculated with lithium obtained from dwarf stars and there- problem by adopting this new diagnostic. abundance variations of ± 0.2 dex. fore does not suffer from the large We used the lower RGB stars to investi- uncertainties on the turbulent mixing gate for the first time the lithium abun- We measured A(Li) = 0.93 ± 0.11 dex, affecting that evolutionary stage. Note dance outside the Milky Way (Mucciarelli in agreement with measurements in that the uncertainties related to the effi- et al., 2014). We obtained high-resolution lower RGB stars of the Galactic Halo. By ciency of diffusion processes amount spectra of 51 member stars of the globu- ­considering the dilution due to the first to 0.4 dex or more for the lithium abun- lar cluster M54, with the FLAMES facility dredge-up, we established that the initial dance in dwarf stars, while they are mounted at the Kueyen Unit Telescope of lithium abundance of this stellar system smaller than 0.07 dex for lower RGB the Very Large Telescope (VLT) and the (in the range A(Li) = 2.29–2.35 dex) was stars; GIRAFFE spectrometer. M54 is a massive compatible with those derived in dwarf 2) The use of lower RGB stars will allow globular cluster located at ~ 25 kpc from stars. This is the most distant measure- estimates of the lithium content in the Sun and immersed in the nucleus ment of the initial lithium abundance in ­stellar populations more distant than of the Sagittarius dwarf galaxy. The dwarf old, metal-poor stars obtained so far. In those usually observed to investigate stars in M54 and the Sagittarius galaxy fact, all the previous studies of the lithium the Spite Plateau. The obvious benefits are too faint (at V ~ 22 mag) to measure abundance in dwarf stars are restricted are not only to enlarge the sample of their lithium abundance, even with the to distances within ~ 8 kpc from the Sun. Galactic field and cluster stars to study VLT, hence the study of lower RGB stars the primordial lithium abundance, but currently represents the only possible Figure 3 summarises the state of the also to offer the formidable opportuni- route to infer the lithium abundance in art of our current knowledge of the ties to investigate the initial lithium this galaxy. ­lithium abundance. The initial lithium abundance in extragalactic systems abundance obtained in M54 matches

46 The Messenger 158 – December 2014 the Spite Plateau well but is lower than Figure 3. Abundance of the CMB + SBBN value by ~ 0.3 dex.  lithium as a function of ",! 2!!- [Fe/H] for the Spite This demonstrates that old stars, regard- ­Plateau (grey circles) less of their birthplace, were born with and lower RGB stars the same initial lithium abundance. (grey squares) of the Galactic field. The empty red circle denotes the Also, an important question can be surface lithium abun- addressed by our study: is the lithium dance measured in the problem a local problem, limited to our  lower RGB stars of M54, Galaxy, or is it independent of the envi- while the filled red circle shows the initial lithium ronment? The analysis of M54 confirms +H abundance of M54. The the findings inω Centauri (Monaco et al., blue dashed line is the 2010) considered as the remnant of an initial lithium abundance accreted dwarf galaxy: the lithium prob- from CMB + SBBN. lem seems to be a universal problem, regardless of the parent galaxy. 

A new tool for new instruments

The result obtained for M54 demon- strates the potential of lower RGB stars in l  l l  l l  l the investigation of the initial lithium :%D'< abundance in stellar systems for which the observation of dwarf stars is pre- cluded. This study has allowed a giant as the European Extremely Large Tele- powerful science case for the future giant leap in this kind of study, pushing our scope or the Giant Magellan Telescope). telescopes. view to ~ 25 kpc from the Sun. Besides the natural impact on the theoretical The current generation of high-resolution SBBN model and stellar evolutionary spectrographs mounted on 8–10-metre- References models, the accurate investigation of the class telescopes (like the VLT, Keck and Coc, A., Uzan, J.-P. & Vangioni, E. 2013, lithium abundance in lower RGB stars Subaru) allows us to reach lower RGB arXiv1307.6955 also has an important benefit in terms of stars in stellar systems out to ~ 25 kpc Korn, A. J. et al. 2006, Nature, 442, 657 the definition of science cases for the from the Sun, with a typical magnitude of Iocco, F. et al. 2009, Phys. Rev., 472, 1 next generation of spectroscopic instru- V ~ 18.5. The advent of the 30–40-metre- Monaco, L. et al. 2010, A&A, 519L, 3 Mucciarelli, A. et al. 2011, MNRAS, 412, 81 mentation. In fact, the proper calibration class telescopes will allow observation Mucciarelli, A., Salaris, M. & Bonifacio, P. 2012, of this diagnostic using a wide sample of of lower RGB stars out to the closest MNRAS, 419, 2195 Galactic stars is a fundamental step in dwarf spheroidal galaxies (~ 80 kpc), Mucciarelli, A. et al. 2014, MNRAS, 444, 1812 the process of observing very distant while dwarf stars will be observed out to Piau, L. et al. 2006, ApJ, 653, 300 Planck Collaboration, 2013, arXiv1203.5076 lower RGB stars with the next generation the Magellanic Clouds (~ 50–60 kpc). Spergel, D. N. et al. 2007, ApJS, 170, 377 of 30–40-metre-class telescopes (such Thus, this new diagnostic represents a Spite, M. & Spite, F. 1982, Nature, 297, 483

VLT Survey Telescope u-, g- and r-band colour image of the globular cluster M54 situated at the core of the Sagittarius Dwarf Galaxy. See eso1428 for details.

The Messenger 158 – December 2014 47