Abundances of neutron-capture elements in CH and Carbon-Enhanced Metal-Poor (CEMP) stars
aa r X i v : . [ a s t r o - ph . S R ] J a n J. Astrophys. Astr. (0000) :
Abundances of neutron-capture elements in CH and Carbon-EnhancedMetal-Poor (CEMP) stars
Meenakshi Purandardas and Aruna Goswami Indian Institute of Astrophysics, Koramangala, Bangalore 560034, India * Corresponding author. E-mail: [email protected], [email protected] received —- March 2020; accepted ———
Abstract.
All the elements heavier than Fe are produced either by slow (-s) or rapid (-r) neutron-capture process.Neutron density prevailing in the stellar sites is one of the major factors that determines the type of neutron-captureprocesses. We present the results based on the estimates of corrected value of absolute carbon abundance, [C / N]ratio, carbon isotopic ratio and [hs / ls] ratio obtained from the high resolution spectral analysis of six stars thatinclude both CH stars and CEMP stars. All the stars show enhancement of neutron-capture elements. Location ofthese objects in the A(C) vs. [Fe / H] diagram shows that they are Group I objects, with external origin of carbon andneutron-capture elements. Low values of carbon isotopic ratios estimated for these objects may also be attributedto some external sources. As the carbon isotopic ratio is a good indicator of mixing, we have used the estimates of C / C ratios to examine the occurance of mixing in the stars. While the object HD 30443 might have experiencedan extra mixing process that usually occurs after red giant branch (RGB) bump for stars with log(L / L ⊙ ) > / N] ratios obtained for these objects also indicate that none of these objects have experienced any stronginternal mixing processes. Based on the estimated abundances of carbon and the neutron-capture elements, and theabundance ratios, we have classified the objects into di ff erent groups. While the objects HE 0110 − −
38 2151 are found to be CEMP-s stars, HE 0308 − −
28 1082 with enhancement of both r- ands-process elements is found to belong to the CEMP-r / s group. Keywords.
Stars—Individual; Stars—Abundances; Stars—Carbon; Stars—Nucleosynthesis.
1. Introduction
Chemical analysis of metal-poor stars such as CH starsand CEMP stars can provide important clues aboutthe nature of nucleosynthesis processes occured in theearly Galaxy. Especially, the abundances of neutron-capture elements can be used to constrain the Galac-tic chemical evolution due to heavy elements. Var-ious sky survey programmes (HK survey and Ham-burg / ESO survey, Beers et al. et al. et al. et al. et al. et al. et al. / H] < - 1) counter parts of CH stars. Both the CH stars andCEMP stars show enhancement of carbon and neutron-capture elements. Hence these objects are ideal can-didates to study the origin and evolution of these ele-ments. Based on the type of enhancement of neutron-capture elements CEMP stars are classified into di ff er-ent groups, such as CEMP-s, CEMP-r, CEMP-r / s andCEMP-no stars (Beers & Christlieb 2005). The evolu-tionary status of CH and CEMP stars do not support theenhancement of carbon and heavy elements observedin these stars. The widely accepted scenario to explainthis enhancement is that these objects are in a binarysystem. The primary companion once passed throughthe Asymptotic Giant branch (AGB) phase and synthe-sized carbon and heavy elements. The synthesized ma-terials are then transferred to the secondary companionthrough some mass transfer mechanisms. The radialvelocity variations exhibited by CH and CEMP stars(McClure 1983, 1984; McClure & Woodsworth 1990;Hansen et al. © Indian Academy of Sciences 1
J. Astrophys. Astr. (0000) :
In this paper, we have presented the results fromthe high-resolution analysis of six stars that includefour CEMP stars and two CH stars. The paper is or-ganized as follows. In section 2, we have presented abrief discussion on the new results obtained for our pro-gramme stars as some of the results from abundanceanalysis of these objects were presented in Purandar-das et al. (2019), and Purandardas, Goswami & Dod-damani (2019). Section 3 presents the sample selection,observations and data reductions. Section 4 describesthe determination of radial velocity and stellar atmo-spheric parameters. Details of the abundance anaysisare presented in section 5. In section 6, interpretationof our results are presented. Conclusions are drawn inSection 7.
2. Novelty of this work
We have presented the abundance analysis results for26 elements in our programme stars in Purandardas etal. (2019), and Purandardas, Goswami & Doddamani(2019). In these works, we have also reported the massand age of these stars as well as the results from thekinematic analysis. The location of these objects inthe H-R diagram shows that they are either sub-giantsor in the ascending stage of the giant branch. Variousmixing processes have been found to operate in giantstars. It is therefore important to understand whetherthe stars have undergone any internal mixing processesbefore interpretting the observed abundances. We hadnot addressed this problem in our previous works. Inthis paper, we have checked whether any internal mix-ing processes have altered the surface chemical compo-sition of these stars based on [C / N] and carbon isotopicratios. While HD 30443 might have experienced an ex-tra mixing process that usually occurs after red giantbranch (RGB) bump for stars with log(L / L ⊙ ) > / H] diagram. Based on the estimated valueof carbon abundances together with the observed en-hancement of neutron-capture elements, we have clas-sified them as Group I objects following Yoon et al. (2016) classification scheme. As the Group I objectsare all binaries, it is likely that our programme starsthat belong to this group are also in binary systems withexternal origin of carbon and neutron-capture elements. Low values of carbon isotopic ratios estimated for theseobjects may also be attributed to external sources. Inthe present work, we have re-calculated the [hs / ls] ratiofor our programme stars without considering the con-tribution from samarium which is an r-process elementwhich we had taken into account for this estimation inour previous works.
3. Observations and Data reduction
Programme stars are selected from the CH star cata-logue of Bartkevicious (1996), Goswami (2005) andGoswami et al. (2010). In the later papers, poten-tial CH star candidates are identified based on thelow resolution (R = λ/δλ ∼ − − ∼ × −
28 1082 and HD 176021, we haveused the high resolution FEROS spectra (Fiber-fed Ex-tended Range Optical Spectrograph (FEROS) of 1.52m telescope of Europian Southern Observatory at LaSilla). The wavelength coverage of the FEROS spec-tra is from 3500 - 9000 Å with a spectral resolutionof ∼
48 000. The detector is a back-illuminated CCDwith 2948 × µ m size. For the objectCD −
38 2151, a high resolution (R ∼
72 000) spectrumwas obtained using the high-resolution fiber fed Echellespectrometer attached to the 2.34 m Vainu Bappu Tele-scope (VBT) at the Vainu Bappu Observatory (VBO),Kavalur. The spectrum covers the wavelegth regionfrom 4100 to 9350 Å with gaps between orders. Thespectrometer operates in two modes. It allows a resolu-tion of 72 000 with a 60 micron slit and a resolution of27 000 without the slit. The spectrum is recorded on aCCD with 4096 × µ m size. The datareduction is carried out using various spectroscopic re-duction packages such as IRAF. Examples of a fewsample spectra of the programme stars are shown inFigure 1. . Astrophys. Astr. (0000) : Figure 1 . Sample spectra of the programme stars in thewavelength region 7980-8030 Å.
4. Determination of radial velocity and stellar at-mospheric parameters
Radial velocity of the programme stars are determinedby measuring the shift in the wavelength for a largenumber of unblended and clean lines in their spectra.The radial velocities range from − . − .The estimated radial velocities of the programme starsare presented in Table 1.Stellar atmospheric parameters are determinedfrom the measured equivalent widths of clean and un-blended Fe I and Fe II lines using the local thermody-namic equilibrium (LTE) analysis. We made use of therecent version of MOOG of Sneden (1973) for our anal-ysis. Model atmospheres are selected from Kurucz gridof model atmospheres with no convective overshooting(http: // cfaku5.cfa.hardvard.edu / ). Solar abundances aretaken from Asplund, Grevesse & Sauval (2009). E ff ec-tive temperature is taken to be that value for which thetrend between the abundance derived from Fe I linesand the corresponding excitation potential gives a zeroslope. At this temperature, microturbulent velocity is Table 1 . Derived atmospheric parameters and radialvelocities of the programme stars.
Star T ef f log g ζ [Fe I / H] [Fe II / H] V r (K) (cgs) (km s − ) (km s − )HE 0110 − − . ± − . ± − . ± − − . ± − . ± ± −
28 1082 5200 1.90 1.42 − . ± − . ± − . ± − . ± − . ± ± −
38 2151 4600 0.90 2.30 − . ± − .
03 139.7 ± − . ± − . ± ± determined for which the abundance derived from FeI lines do not exhibit any dependence on the reducedequivalent width. Corresponding to these values of ef-fective temperature and microturbulent velocity, log gis determined in such a way that the abundances ob-tained from Fe I and Fe II lines are nearly the same.Only those lines with excitation potential from 0.0 - 5.0eV and equivalent widths from 20 - 180 mÅ are con-sidered for the analysis. The derived atmospheric pa-rameters and the radial velocities are listed in Table 1.
5. Abundance analysis
The detailed discussion on the abundance analysis re-sults for the six programme stars are presented in Pu-randardas et al. (2019), and Purandardas, Goswami &Doddamani (2019). Here we present a brief summaryof these results. However here we give more emphasizeto the results based on the absolute carbon abundance,[C / N] ratio, carbon isotopic ratio and the [hs / ls] ratiowhich is recalculated in this work.The abundances of various elements are determinedfrom the measured equivalent widths of absorptionlines due to neutral and ionized elements. We haveused only the symmetric and clean lines for our anal-ysis. Lines are identified by overplotting the arcturusspectra upon the individual spectra of our programmestars. Then a master linelist is prepared using the mea-sured equivalent widths and other line information suchas lower excitation potential and the loggf values takenfrom the Kurucz database. We have also consultedVALD database. We could estimate the abundances of24 elements which include the light elements C, N, O,odd-Z element Na, α - and Fe-peak elements Mg, Si,Ca, Ti, V, Cr, Mn, Co, Ni and Zn and the neutron-capture elements Sr, Y, Zr, Ba, La, Ce, Pr, Nd, Smand Eu. We have also used spectrum synthesis cal-culations for elements such as Sc, V, Mn, Ba, La andEu taking their hyperfine structures into considerations.The hyperfine structures of Sc, V and Mn are takenfrom Prochaska & McWilliam (2000). For Ba, La andEu, the hyperfine structures are taken from McWilliam(1998), Jonsell et al. (2006) and Woorley et al. (2013)respectively.We could estimate oxygen abundance only for J. Astrophys. Astr. (0000) : CD −
38 2151 and HD 30443. For these objects, theabundance of oxygen is determined from the spectrumsynthesis calculations of the [OI] line at 6300.3 and OIline at 6363.8 Å. The carbon abundance could be deter-mined for all the objects from the spectrum synthesiscalculations of the C molecular band at 5165 Å. Wecould estimate the carbon isotopic ratio for all of ourprogramme stars except HD 176021 using the spectrumsynthesis calculation of CN band at 8005 Å (Figure 2).The values lie in the range from 7.4 to 45. Abundanceof nitrogen is estimated using the spectrum synthesiscalculation of CN band at 4215 Å. Nitrogen is found tobe enhanced in CD −
28 1082 and CD −
38 2151. Otherobjects exhibit moderate enhancement in nitrogen. Themolecular lines for C and CN are taken from Brooke etal. (2013), Sneden et al. (2014) and Ram et al. (2014).The estimated values of carbon and nitrogen are pre-sented in Table 2. Table 2 . Abundance results for carbon, nitrogen, C / Oand carbon isotopic ratios.
Star log ǫ (C) log ǫ (C) ∗ [C / Fe] log ǫ (N) [N / Fe] C / O C / CHE 0110 − − −
28 1082 8.16 8.23 2.19 8.10 2.73 - 16.0HD 30443 8.43 8.47 1.68 6.55 0.40 1.02 7.40CD −
38 2151 7.90 8.02 1.50 7.20 1.40 2.95 11.2HD 176021 8.33 8.33 0.52 7.80 0.59 - - ∗ Corrected value of carbon.
Figure 2 . Synthesis of CN band around 8005 Å. Syn-thesized spectra is shown in red colour and the observedspectra is represented in black colour. Synthetic spectracorresponding to C / C ≃
12 (blue) and 1 (magenta) arealso shown.
Sodium is moderately enhanced in all our pro-gramme stars except HD 176021 in which Na is nearsolar. HE 0110 − − − / Fe] ∼ / Fe] ∼ / Fe] ∼ / Fe] ∼ −
38 2151. While Siis found to be enhanced with [Si / Fe] ∼ − / Fe] ∼ − − −
28 1082which is found to be enhanced with [Mn / Fe] ∼ / Fe] ∼ −
38 2151. While chromium is slightlyenhanced and Mn is underabundant with [Mn / Fe] ∼− .
20. In HD 176021, all the Fe peak elements arefound to be near solar.All of our programme stars exhibit enhancementof neutron-capture elements. From our detailed anal-ysis, we found that CD −
28 1082 is a CEMP-r / s starand the objects HE 0110 − −
38 2151 and HD30443 are CEMP-s stars. We could not estimate Eu inHD 30443 and CD −
38 2151. Hence, it is not possi-ble to classify these stars based on the criteria as givenby Beers & Christlieb (2005). In this case, we haveused the criteria for the classification of CEMP-s starsas given by Hansen et al. (2019) based on [Sr / Ba] ra-tio. According to this classification scheme, [Sr / Ba] > − . / s stars. HD 30443 and CD −
38 2151 show[Sr / Ba] ∼ − .
27 and [Sr / Ba] ∼ − −
38 2151 andHD 176021. The abundance results for neutron-captureelements are listed in Table 3. In this table, ls standsfor light s-process elements (Sr, Y and Zr) and hs rep-resents the heavy s-process elements (Ba, La, Ce, andNd).
6. Interpretation of results
The interpretation of the results based on the correctedvalue of absolute carbon abundance, [C / N], carbon iso-topic ratios and [hs / ls] ratios obtained for our pro-gramme stars are presented here in detail. Understand-ing the possible source of origin of the enhancement . Astrophys. Astr. (0000) : Table 3. Ratios of light and heavy s-process elements
Star [Fe / H] [ls / Fe] [hs / Fe] [hs / ls]HE 0110 − − .
30 1.03 1.36 0.33HE 0308 − − .
73 1.11 1.62 0.51CD −
28 1082 − .
45 1.52 1.90 0.38HD 30443 − .
69 1.24 1.93 0.69CD −
38 2151 − .
03 1.24 1.10 − . − .
64 1.50 1.37 − . of neutron-capture elements is very important to under-stand the type of nucleosynthesis process which pro-duced it. One of the best ways to get any clues re-garding the source is to locate the stars in the A(C)vs. [Fe / H] diagram. It is found that CEMP stars ex-hibit a bimodal distribution in the A(C) vs. [Fe / H] dia-gram. Spite et al. (2013) claims the occurance of twoplateau at A(C) = = et al. (2016) confirmed two peaks at A(C) = = et al. (2016) into three groups basedon the morphology in the A(C) vs. [Fe / H] diagram.Group I objects are mainly composed of CEMP-s andCEMP-r / s stars. These objects show a weak depen-dence of A(C) on [Fe / H]. The absolute carbon abun-dance, A(C) of Group II objects show a clear depen-dence on [Fe / H]. While for Group III objects, A(C) isfound to be independent of [Fe / H], Group II and GroupIII objects are mainly composed of CEMP-no stars. InFigure 3, the location of our programme stars are shownin the A(C) vs. [Fe / H] diagram. We have applied cor-rections to the estimated carbon abundances using thepublic online tool by Placco et al. (2014) availableat http / : // vplacco.pythonanywhere.com / . The correctedcarbon values are listed in Table 2. Figure 3 shows thatall of our programme stars are Group I objects. Hencewe assume that the observed enhancement of carbonmay be attributed to the binary companion.The low values of C / C ratio observed in our pro-gramme stars also support the extrinsic origin of car-bon. Various mixing processes such as first dredgeup, thermohaline mixing and rotation induced mixing(Charbonnel 2005; Dearborn et al. etal. C abun-dance. We have estimated the [C / N] ratio of our pro-
Figure 3 . Corrected A(C) vs. [Fe / H] diagram for thecompilation of CEMP stars taken from Yoon et al. (2016).CEMP-s stars are represented by Open circles. CEMP-r / sand CEMP-no stars are shown by open squares and opentriangles respectively. Programme stars are representedby black coloured symbols, HE 0110 − − −
28 1082 (filled triangle), HD30443 (open pentagon), CD −
38 2151 (filled square) and HD17602 (filled pentagon) gramme stars to derive any clues if they have under-gone any mixing processes. Since nitrogen is producedat the expense of carbon, [C / N] ratio acts as a sensitiveindicator of mixing. Spite et al. (2005) analysed CNOand Li abundances for a sample of extremely metal-poor stars. Their analysis shows that the stars that ex-hibit clear evidence of mixing have a low value of [C / N]ratio ( < − . / N] > − .
60 and lie on the lowerRGB. Figure 4 shows that all of our programme starshave [C / N] > − .
6. This implies that none of our pro-gramme stars have undergone any significant internalmixing processes.Stellar models predict that when a low mass star as-cends the red giant branch, the outer convective enve-lope expands inwards and penetrates into the region ofCN processed materials (First dredge up (FDU)). Theluminosity at which the FDU occurs in low mass fieldstars is log(L / L ⊙ ) ∼ et al. / L ⊙ ) > / metallicity stars, C / C ra-tio decreases by a factor of 20-30 from the original
J. Astrophys. Astr. (0000) :
Figure 4 . Position of the programme stars in the [C / N] vs.T e f f diagram. Programme stars are represented using redcoloured symbols. The symbols used for the programmestars are same as in Figure 3. Open squares represent thestars from literature (Spite et al. et al. et al. et al. et al. value and surface abundance of nitrogen increases afterFDU (Iben & Renzim 1984), it is found to be less ef-ficient in metal-poor stars (VandenBerg & Smith 1988,Charbonnel 1994). This implies that the changes in thesurface compositions of C and N after the occurance ofFDU are very small in metal-poor stars. The carbonabundance is found to be decreased only by about 0.05dex and the decrease in the C / C ratio is not largeenough for these stars (Gratton et al. / L ⊙ ) ∼ C abundancedecreases by a factor of ∼ C / C reaches avalue of ∼ et al. −
38 2151 havelog(L / L ⊙ ) ∼ / N] ra-tio. Spite et al. (2006) suggest that [C / N] raio is nota clean indicator of mixing as the abundance of C andN in the interstellar medium from which these stars areformed show large variations. In that case, one can usethe C / C as a good indicator of mixing, since it ishigh in primordial matter ( >
70) and the determinationof carbon isotopic ratio is found to be insenstive to the choice of atmospheric parameters for the stars (Spite etal. C / C ratio and T e f f ofour programme stars as shown in Figure 5. From thefigure, it is clear that the object HD 30443 share the re-gion occupied by the stars that have undergone internalmixing processes. This means that HD 30443 had ex-perienced internal mixing processes that might have al-tered its initial surface chemical compositions. CD − et al. Figure 5 . Position of the programme stars in C / C vs.T e f f diagram. Programme stars are represented using redcoloured symbols. The symbols used for the programmestars are same as in Figure 3. Open squares represent thestars from Spite et al. (2006) and Aoki et al. (2007).
From the estimated carbon isotopic ratio and the lo-cation of our programme stars in the A(C) vs. [Fe / H]diagram, we assume that the observed enhancement ofneutron-capture elements may be attributed to an ex-trinsic source. As Group I objects are mostly foundto be associated with binary systems, the enhancementmay be justified by the mass transfer from the binarycompanion.According to Busso et al. (2001), the light s-processelements such as Y, Zr and Sr are predominantly pro-duced in AGB stars with near solar metallicity. Hence,mass transfer from such a companion can produce moreenhancement of light s-process elements than the heavys-process elements. The AGB stars with [Fe / H] ∼ − . . Astrophys. Astr. (0000) : La, Ce and Nd than the light s-process elements, andthis leads to lower [ls / Fe] than [hs / Fe] in the stars thathave accreted materials from the metal-poor AGB stars.The reason is that, in the metal-poor AGB stars, thenumber of Fe seed nuclei available for neutron-captureprocess is less and hence the neutron exposure for eachFe seed nuclei will be more. This leads to the forma-tion of more heavier elements. The binary mass transferscenario can similarly be applied to justify the observedenhancement of neutron-capture elements in CD − / s star. Several authors (Ham-pel et al. et al. / s stars are consistent with the modelcalculations of i-process in low-mettalicity AGB stars.
7. Conclusion
The possible source of the origin of enhancement ofneutron-capture elements in six carbon stars are exam-ined in the light of absolute carbon abundances. Lo-cation of our programme stars in the A(C) vs. [Fe / H]diagram shows that these objects belong to Group Icategory. Various studies on radial vecities of theseobjects show that most of them exhibit radial veloc-ity variations. That is, these objects are mostly foundto be associated with binary systems. Hence the ob-served enhancement of carbon and neutron-capture el-ements in our programme stars may be attributed to thebinary companion. But none of the programme starsare known to be confirmed binaries. The low values ofcarbon isotopic ratio also supports the extrinsic origin.Elemental abundance ratios also bear important sig-natures about the source of enrichment of neutron-capture elements. We have re-estimated the [hs / ls] ratiofor our programme stars. Abate et al. (2015) predict an[ls / hs] ratio less than zero for AGB models. All theprogramme stars except CD −
38 2151 and HD 176021show an [hs / ls] ratio characteristics of AGB progeni-tors.We have also examined whether the programmestars have undergone any internal mixing based on[C / N] and carbon isotopic ratios. Because it is impor-tant to understand whether the mixing have modifiedthe surface composition of the star before interprettingthe observed abundance patterns. We found that noneof the proramme stars have experienced internal mix-ing, except HD 30443. The estimated values of [C / N]ratio also show that the objects have not gone throughany significant mixing processes. Thus, the observedvalues of low carbon isotopic ratio in the unmixed starsmay be due to the strong mixing processes occured inthe AGB progenitors. In other words, the observed sur- face chemical compositions of our programme stars ex-cept HD 30443, preserve the fossil records of the ma-terials synthesised in the AGB stars from which theyhave accreted the materials.
Acknowledgements
We thank the sta ff members at IAO, CREST andVBO for their assistance and cooperation during theobservations. Funding from DST SERB project No.EMR / / // / gaia),processes by the Gaia Data Process-ing and Analysis Consortium (DPAC,https: // / web / gaia / dpac / consortium). References
Abate C., Pols O.R., Izzard R.G., Karakas A.I., 2015, A&A,581, A22Aoki W., Beers T.C., Christlieb N., Norris J.E., Ryan S.G etal., 2007, ApJ, 655, 492Asplund M., Grevesse N., Sauval A.J., 2009, Ann. Rev. As-tron. Astrophy., 47:481Bartkevicious A., 1996, Baltic Astron, 5, 217Beers T.C., Preston G.W., Shectman S.A., 1985, AJ, 90, 2089Beers T.C., Christlieb N., 2005, ARA&A, 43, 531Brooke J.S., Bernath P.F., Schmidt T.W., Bacskay G.B.,2013, J.Quant. Spectrosc. Radiat. Transfer, 124, 11Busso M., Gallino R., Lambert D.L., Travaglio C., SmithV.V., 2001, ApJ, 557, 802Charbonnel C., 2005, ASP Conference Series, Vol. 336Christlieb N., Green P. J., Wisotzki L., Reimers D., 2001,A&A, 375, 366Dearborn D. S. P., Lattanzio J. C., Eggleton P. P., 2006, ApJ,639, 405Eggleton P. P., Dearborn D. S. P., Lattanzio J. C., 2006, Sci-ence, 314, 1580Frebel A., Aoki W., Christlieb N., Ando H., Asplund M. etal., 2005, Nature, 434, 871Goswami A., 2005, MNRAS, 359, 531Goswami A., Karinkuzhi D., Shantikumar N.S., 2010, MN-RAS, 402, 1111Goswami A., Aoki W., Karinkuzhi D., 2016, MNRAS, 455,402
J. Astrophys. Astr. (0000) : ff e R.J., Lugaro M., Meyer B.S., 2016,ApJ, 831, 171Hansen C.J., Nordstr¨om B., Hansen T.T., Kennedy C.R.,Placco V M. et al., 2016b, A&A, 588, A37Hansen T. T., Andersen J., Nordstr¨om B., Beers T.C., PlaccoV.M. et al. 2016a, A&A, 586, A160Hansen C.J., Hansen T.T., Koch A., Beers T.C., Nordstr¨omB. et al., 2019, A&A, 623, 128Iben I.Jr., Renzini A., 1984, Phys. Letters 105, 329Jonsell K., Barklem P.S., Gustafsson B., Christlieb N., HillV. et al., 2006, A&A, 451, 651McClure R.D., 1983, ApJ, 208, 264McClure R.D., 1984, ApJ, 280, 31McClure R.D., Woodsworth W., 1990, ApJ, 352, 709McWilliiam A., 1998, AJ, 115, 1640Norris J.E., Christlieb N., Korn A.J., Eriksson K., BessellM.S. et al., 2007, ApJ, 670, 774Placco V.M., Frebel A., Beers T. C., Stancli ff e R. J., 2014,ApJ, 797, 21Prochaska J.X., McWilliam A. 2000, ApJ, 537, 57Purandardas M., Goswami A., Goswami P.P., ShejeelammalJ., Masseron T., MNRAS, 2019, 486,3266Purandardas M., Goswami A., Doddamani V.H., 2019,BSRSL, 88, 207Ram R.S., Brooke James S.A., Bernath P.F., Sneden C., Lu-catello S., 2014, ApJS, 211, 5Sneden C., 1973, PhD thesis, Univ. TexasSneden C., Lucatello S., Ram R.S., Brook J.S.A., Bernath P.,2014, Ap. J. Supp., 214, 26Spite M., Cayrel R., Plez B., Hill V., Spite F. et al., 2005,A&A 430, 655Spite M., Cayrel R., Hill V, Spite F., Francois P. et al., 2006,A&A 455, 291Spite, M., Ca ffff