Stellar Population Astrophysics (SPA) with TNG. The old open clusters Collinder 350, Gulliver 51, NGC 7044, and Ruprecht 171
G. Casali, L. Magrini, A. Frasca, A. Bragaglia, G. Catanzaro, V. D'Orazi, R. Sordo, E. Carretta, L. Origlia, G. Andreuzzi, X. Fu, A. Vallenari
AAstronomy & Astrophysics manuscript no. 39176corr c (cid:13)
ESO 2020September 16, 2020
Stellar Population Astrophysics (SPA) with TNG (cid:63)
The old open clusters Collinder 350, Gulliver 51, NGC 7044, and Ruprecht 171
G. Casali , , L. Magrini , A. Frasca , A. Bragaglia , G. Catanzaro , V. D’Orazi , R. Sordo , E. Carretta , L. Origlia ,G. Andreuzzi , , X. Fu , A. Vallenari Dipartimento di Fisica e Astronomia, Università degli Studi di Firenze, via G. Sansone 1, 50019 Sesto Fiorentino (Firenze), Italye-mail: [email protected] INAF-Osservatorio Astrofisico di Arcetri, Largo E. Fermi, 5, I-50125 Firenze, Italy INAF-Osservatorio di Astrofisica e Scienza dello Spazio di Bologna, via P. Gobetti 93 /
3, 40129, Bologna, Italy INAF-Osservatorio Astrofisico di Catania, via S. Sofia 78, 95123, Catania, Italy INAF-Osservatorio Astronomico di Padova, vicolo Osservatorio 5, 35122, Padova, Italy Fundación Galileo Galilei - INAF, Rambla José Ana Fernández Pérez 7, 38712 Breña Baja, Tenerife, Spain INAF-Osservatorio Astronomico di Roma, Via Frascati 33, 00078 Monte Porzio Catone, Italy The Kavli Institute for Astronomy and Astrophysics at Peking University, 100871 Beijing, ChinaSeptember 16, 2020
ABSTRACT
Context.
Open clusters are excellent tracers of the chemical evolution of the Galactic disc. The spatial distribution of their elementalabundances, through the analysis of high-quality and high-resolution spectra, provides insight into the chemical evolution and mech-anisms of element nucleosynthesis in regions characterised by di ff erent conditions (e.g. star formation e ffi ciency and metallicity). Aims.
In the framework of the Stellar Population Astrophysics (SPA) project, we present new observations and spectral analysis of foursparsely studied open clusters located in the solar neighbourhood, namely Collinder 350, Gulliver 51, NGC 7044, and Ruprecht 171.
Methods.
We exploit the HARPS-N spectrograph at the TNG telescope to acquire high-resolution optical spectra for 15 member starsof four clusters. We derive stellar parameters ( T e ff , log g , [Fe / H] and ξ ) using both the equivalent width (EW) analysis and the spectralfitting technique. We compute elemental abundances for light, α -, iron-peak, and n-capture elements using the EW measurementapproach. We investigate the origin of the correlation between metallicity and stellar parameters derived with the EW method forthe coolest stars of the sample ( T e ff < Results.
We locate the properties of our clusters in the radial distributions of metallicity and abundance ratios, comparing our resultswith clusters from the Gaia-ESO and APOGEE surveys. We present the [X / Fe]-[Fe / H] and [X / Fe]- R GC trends for elements in commonbetween the two surveys. Finally, we derive the C and Li abundances as a function of the evolutionary phase and compare them withtheoretical models. Conclusions.
The SPA survey, with its high-resolution spectra, allows us to fully characterise the chemistry of nearby clusters.With a single set of spectra, we provide chemical abundances for a variety of chemical elements, which are comparable to thoseobtained in two of the largest surveys combined. The metallicities and abundance ratios of our clusters fit very well in the radialdistributions defined by the recent literature, reinforcing the importance of star clusters to outline the spatial distribution of abundancesin our Galaxy. Moreover, the abundances of C and Li, modified by stellar evolution during the giant phase, agree with evolutionaryprescriptions (rotation-induced mixing) for their masses and metallicities.
Key words.
Open clusters and associations: general - open clusters and associations: individual: Collinder 350, Gulliver 51,NGC 7044, Ruprecht 171 - Galaxy: disc - Galaxy: evolution - stars: abundances
1. Introduction
Star clusters are among the most versatile astronomical objects.Within our Galaxy, they play a key role in the study of both stel-lar and Galactic evolution: they allow us to study the formationand evolution of stars (e.g. Krause et al. 2020; Krumholz & Mc- (cid:63)
Based on observations made with the Italian Telescopio NazionaleGalileo (TNG) operated on the island of La Palma by the FundaciónGalileo Galilei of the INAF (Istituto Nazionale di Astrofisica) at the Ob-servatorio del Roque de los Muchachos. This study is part of the LargeProgram titled SPA – Stellar Population Astrophysics: a detailed, age-resolved chemical study of the Milky Way disc (PI: L. Origlia), grantedobserving time with HARPS-N and GIANO-B echelle spectrographs atthe TNG.
Kee 2020), the dynamics of stellar systems (e.g. Sacco et al.2017b; Kuhn et al. 2019; Piatti et al. 2019), and they providerobust constraints on the formation timescales and the chemicaland dynamical history of the Milky Way (e.g. Friel et al. 2002;Bragaglia & Tosi 2006; Magrini et al. 2009; Reddy et al. 2016;Jacobson et al. 2016; Spina et al. 2017; Casamiquela et al. 2019a;Zhong et al. 2020; Donor et al. 2020; Chen & Zhao 2020).Large spectroscopic surveys, such as for example Gaia-ESO (Gilmore et al. 2012), GALAH (De Silva et al. 2015),and APOGEE (Majewski et al. 2017), and future survey-dedicated spectrographs, such as WEAVE (Dalton et al. 2012)and 4MOST (de Jong 2011), make use of multi-object spec-troscopy at medium-high resolution to characterise the kinemat-
Article number, page 1 of 17 a r X i v : . [ a s t r o - ph . S R ] S e p & A proofs: manuscript no. 39176corr ics and global chemical properties of the di ff erent Galactic stellarcomponents (i.e. disc, bulge, and halo) with high-statistical sig-nificance. Their observations complement the data from the Gaia mission (Perryman et al. 2001; Bailer-Jones et al. 2018a; Linde-gren et al. 2018), which in its second data release (
Gaia
DR2,Gaia Collaboration et al. 2018) has provided positions, paral-laxes, proper motions, and photometry in three bands ( G , BP , RP ) for more than 1.3 billion sources.A large number of star clusters are included among the tar-gets of the large spectroscopic surveys. In particular, the Gaia-ESO survey provides data for 81 open clusters, in most caseswith more than 100 member stars; APOGEE presently has spec-tra for 128 open clusters (Donor et al. 2020), but generally withonly a few stars per cluster; WEAVE, which will begin soon, willtarget about 300 open clusters. The observations of large spectro-scopic surveys, together with Gaia data, have improved our un-derstanding of Galactic chemical evolution (e.g. Magrini et al.2017, 2018; Donor et al. 2020), the Milky Way structure (e.g.Meingast et al. 2019; Cantat-Gaudin et al. 2020; Castro-Ginardet al. 2020; Anders et al. 2020) and the cluster formation anddisruption processes (e.g. Sacco et al. 2017a; Bravi et al. 2018;Piatti et al. 2019). In addition,
Gaia has enabled the discoveryand characterisation of a large number of clusters in the solarneighbourhood and beyond (Cantat-Gaudin et al. 2018a,b, 2020;Sim et al. 2019; Liu & Pang 2019; Castro-Ginard et al. 2018,2020), and at the same time has allowed some candidate clus-ters to be discarded (see e.g. Kos et al. 2018; Cantat-Gaudin &Anders 2020, for discussion), and several extended structures tobe identified, such as strings and filaments, which are often con-nected to known clusters (e.g. Kounkel & Covey 2019), streams(e.g. Meingast et al. 2019), extended halos (e.g in M67 by Car-rera et al. 2019), and tidal tails (e.g. in Praesepe by Röser &Schilbach 2019).
Gaia data, in combination with spectral infor-mation on kinematics (from radial velocities) and abundances,are driving a revolution in the study of open clusters and, in gen-eral, of the whole Milky Way.In this framework, the aims of the Stellar Population Astro-physics (SPA) project, an ongoing Large Programme running onthe 3.6 m Telescopio Nazionale Galileo (TNG) at the Roque delos Muchachos Observatory (La Palma, Spain), are to contributeto our understanding of the star formation and chemical enrich-ment history of our Galaxy by providing high-resolution spectraof a sample of stars in the Solar neighbourhood. The SPA projectis obtaining high-resolution spectra of approximately 500 starsnear to the Sun, covering a wide range of ages and properties (seeOriglia et al. 2019, for a general description), such as Cepheidsand stars in both young and old open clusters of spectral typefrom A to K. These stars are observed in the optical and near-infrared (NIR) bands at high spectral resolution using GIARPS,a combination of HARPS-N and GIANO-B spectrographs. Theaim of SPA is to obtain a large variety of elemental abundancesin order to seek possible global trends, such as for example[X / Fe]–age relations. Chemical characterisation combined withthe kinematic counterpart from
Gaia and other surveys will pro-vide a framework for a comprehensive chemo-dynamical mod-elling of disc formation and evolution in the Solar vicinity. Thepresent study is part of a series of papers dedicated to the re-sults of the SPA project. So far, the series includes a paper onthe red supergiants in Alicante 7 and Alicante 10 (Origlia et al.2019), and studies of the young open clusters ASCC 123 (Frascaet al. 2019) and Praesepe (D’Orazi et al. 2020). Part of our goalis to gain kinematic and chemical information on open clus-ters that have never been studied before or for which very littlespectroscopic data is available. This is the case of the four clus- ters presented here, Collinder 350, Gulliver 51, NGC 7044, andRuprecht 171. One of them, Gulliver 51, is a new cluster discov-ered using
Gaia
DR2 data (Cantat-Gaudin et al. 2018a), whileonly one star was observed spectroscopically in Collinder 350,and NGC 7044 has only been studied with low-resolution spec-tra. The paper is structured as follows: in Sect. 2 we present ourcluster sample, and in Sect 3 we show the spectral data samplecollected with the spectrograph HARPS-N at the TNG telescopeduring the SPA observing campaign. In Sect. 4 we present thephotometric parameters for our star sample, and in Sect. 5, wedescribe our spectral analysis using both the equivalent width(EW) analysis and the spectral fitting. In Sect. 6 we discuss thediscrepancy in metallicity between cool and warm member starsof the same cluster, and in Sects. 7 and 8, we review the chemicalabundance ratios of our clusters in the context of larger data sam-ples which define the abundance gradients [X / Fe] versus [Fe / H]and [ α / Fe] versus [Fe / H] and the radial metallicity gradient. Wepresent the C and Li abundances as a function of log g com-pared with the evolutionary predictions. Finally, in Sect. 9, wesummarise our results and give our conclusions.
2. The cluster sample
The SPA project is observing a large sample of nearby star clus-ters, allowing complete characterisation of the open clusters lo-cated in the Solar neighbourhood. Most of these clusters are lo-cated within 1.5-2 kpc of the Sun in order to match the zonewhere
Gaia data reach their highest precision. We need to ob-tain, in a reasonable amount of time, su ffi ciently high S / N topermit precise abundance determination from spectra at the veryhigh resolution of GIARPS. This means that our magnitude limitis about G = − . / or extincted, only stars on the brighter partof the red giant branch can be observed.In the present work, we discuss the analysis of four openclusters that are deemed important to characterise the nearby re-gions of our Galaxy because of their location and age, but forwhich very little spectroscopic data are available. Their agesrange between 0.3 and 3 Gyr and they are located from ∼
300 pcto about 3300 pc from the Sun at di ff erent Galactocentric dis-tances and altitudes above the Galactic plane.Collinder 350 was listed for the first time in the catalogueof Collinder (1931), but the first study of its properties was pre-sented by Kharchenko et al. (2005). The cluster was also studiedusing Gaia
DR1 / TGAS data by Cantat-Gaudin et al. (2018b) andYen et al. (2018, who derived an age of 1 Gyr). Pakhomov et al.(2009) obtained high-resolution spectroscopy of one star, indi-cated as HD161587 (our Cr350_1), and derived a metallicity of[Fe / H] = + . ± .
06 dex together with the abundances of manyspecies. Blanco-Cuaresma et al. (2015) and Blanco-Cuaresma& Fraix-Burnet (2018), also from one single star (at resolutionR =
80 000, with the NARVAL spectrograph), derived a lowervalue for the metallicity, [Fe / H] = − . ± .
01 (or 0, depend-ing on normalisation) and 0.03 dex, respectively. On the otherhand, there are no high-resolution spectroscopic observations ofthe other three clusters.
Article number, page 2 of 17asali, G. et al.: SPA-OCs
NGC 7044 has been studied using photometry several timesin the past (see Kaluzny 1989; Aparicio et al. 1993; Sagar &Gri ffi ths 1998), with a general consensus on its age of around1.5 Gyr, and a high reddening. An estimate of the metallicity andradial velocity of ten member stars in NGC 7044 is provided byWarren & Cole (2009) using low-resolution spectroscopy andthe IR Ca ii triplet (CaT) technique. These authors derived amean metallicity [Fe / H] = − . ± .
09 dex and a mean radialvelocity RV = − . ± .
18 km s − .Even less information is available for Ruprecht 171: afterthe classification by Ruprecht (1966), the cluster was studied byTadross (2003) who derived an age 3.2 Gyr, a reddening E ( B − V ) = .
12 mag, and a distance d = JHK s photometric data.Finally, Gulliver 51 was only recently discovered by Cantat-Gaudin et al. (2018a), who reported the serendipitous discoveryof 60 candidate clusters based on Gaia
DR2 data, which werenamed ‘Gulliver’. They were identified as groups of stars withcoherent proper motions and parallaxes, and a more concentrateddistribution on the sky than the field population. An age of 0.8Gyr, reddening E ( B − V ) = .
375 mag, and a distance of 1330 pcwere attributed to Gulliver 51 by Monteiro & Dias (2019) basedon automated isochrone fitting to
Gaia
DR2 colour-magnitudediagrams (CMDs).For the first three clusters, distance, reddening, age, andproper motions are reported in the catalogue of Kharchenkoet al. (2013). All have updated values by Cantat-Gaudin et al.(2018a, 2020) using
Gaia
DR2 data. Table 1 gives the coordi-nates, proper motions ( µ α , µ δ ), parallax, distance ( d ), Galacto-centric distance ( R GC ), altitudes above the Galactic plane ( z ), ex-tinction ( A V ), and age of the clusters by Cantat-Gaudin et al.(2020). Figure 1 shows the CMDs of the four clusters, in whichthe stars observed by the SPA project are highlighted.This is the first paper dedicated exclusively to these fourclusters. In this work we provide, for the first time (exceptfor Collinder 350, for which high-resolution spectroscopy of asingle star is available), results based on high-resolution spec-troscopy for a number of candidate members in each cluster. Ouraim is to perform a full characterisation in terms of atmosphericparameters and chemical abundances.
3. Observations and data sample
We select high-probability member stars among the red giantbranch (RGB) and red clump (RC) stars. The membership prob-ability was taken from Cantat-Gaudin et al. (2018a) who used
Gaia
DR2 proper motions and parallaxes. The observations ofthese four clusters were conducted from 18 to 22 August 2018and from 10 to 15 August 2019 with GIARPS using both theoptical echelle spectrograph HARPS-N (R ∼
115 000, spectralrange = . − . µ m , Cosentino et al. 2014) and the NIR spec-trograph GIANO-B (R ∼
50 000, spectral range = . − . µ m ,Oliva et al. 2006). The spectra were acquired with total exposuretimes ranging from 600 to 7200 seconds depending on the starbrightness, in order to reach an S / N per pixel at red wavelengthsof S / N >
30. Exposure times longer than 1800 seconds wereusually split into two or three sub-exposures to reduce the con-tamination of cosmic rays and to avoid saturation. The stars anal-ysed in the present work are shown in Table 2 with their coordi-nates,
Gaia magnitudes, parallax, radial velocity (RV), exposuretime, and S / N. The RV was measured by cross-correlating thetarget spectrum with a template, which was chosen as the syn-thetic BT-Settl spectrum (Allard et al. 2011) with T e ff and log g closer to the target. Very broad lines, such as Na i D and Balmer lines, as well as strong telluric features were excluded from thecross-correlation function CCF analysis. For this task we used adhoc software developed by us in the IDL environment. The CCFpeak was fitted with a Gaussian to evaluate its centroid and fullwidth at half maximum (FWHM). The RV error was estimatedby the fitting procedure accounting for the CCF noise far fromthe peak. Here we only make use of HARPS-N spectra, whichare better suited for our analysis methods (see following section).GIANO spectra will be presented in a forthcoming paper.
4. Photometric parameters
In Table 3, we present the photometric parameters T e ff , Gaia andlog g Gaia obtained from
Gaia
DR2 photometry and parallaxes,and the parameters obtained from the comparison with the best-fit isochrone ( T e ff , iso and log g iso ), projecting the Gaia coloursand magnitudes on a set of PARSEC isochrones (Bressan et al.2012). Our aim is to use these parameters as an input to es-timate the spectroscopic ones. The photometric gravities from
Gaia photometry are obtained using the following equation:log( g Gaia ) = log( M / M (cid:12) ) + . · M bol + · log( T e ff , Gaia ) − . , (1)where M / M (cid:12) is the stellar mass (in solar mass units) obtained asthe mass at the main sequence turn o ff (MSTO) of the isochronesat their literature cluster age; and M bol ( M bol = − . · log( L / L (cid:12) ) + .
75) is the bolometric magnitude computed from the luminositypresent in the
Gaia
DR2 catalogue (Li et al. 2018) and correctedconsidering the average distance for each cluster. In parenthesisin Table 3, we give the log g with M bol computed with the lu-minosity corrected considering the individual distances derivedby Bailer-Jones et al. (2018b) and the mean extinction values ofthe clusters reported in Table 1; T e ff , Gaia is the photometric e ff ec-tive temperature from Gaia obtained with the calibration of gi-ant stars by Mucciarelli & Bellazzini (2020) for the
Gaia colour BP − BR . In first approximation, we consider solar metallicitiesfor our clusters. As a test, we also re-compute the photometrice ff ective temperatures using average metallicities from spectro-scopic analysis, finding a negligible correction. For the cooleststars of the sample (all stars of NGC 7044 and Rup171_1), astheir BP − RP colours fall outside the calibration range of Muc-ciarelli & Bellazzini (2020), we adopt the e ff ective temperatureof Gaia
DR2 as T e ff , Gaia , for which we are aware of the inad-equacy, especially at those very low temperatures and for clus-ter members (see Andrae et al. 2018; Gaia Collaboration et al.2018). For clusters a ff ected by high extinction and in crowdedregions, the photometric parameters are indeed unreliable be-cause of the spatial variations of the reddening, which is usu-ally assumed to be constant, and because of the more di ffi cultextraction of the fluxes of the individual stars. The photomet-ric gravities of the four observed stars of NGC 7044, the mostdistant and extincted cluster of our sample, range from 0.8 to1.34 dex (obtained with the individual distances), which is unex-pected, because they have very similar colours and magnitudes(see Fig. 1). Their parallaxes have percentage errors larger than10%, and they di ff er by more than 3 σ from the mean clusterparallax. In addition, the individual extinctions available in Gaia
DR2 (albeit su ff ering from the same limitation as Gaia
DR2 T e ff )vary from star to star, and are not available for all the clustermembers. Therefore, the photometric parameters of the observedstars in NGC 7044 have to be considered as just a starting pointfor the following spectroscopic analysis. In a similar way, thetwo stars observed in Gul 51 both have high extinction in Gaia
DR2, which varies from star to star ( A G ∼ Article number, page 3 of 17 & A proofs: manuscript no. 39176corr
Fig. 1.
Colour-magnitude diagrams with
Gaia
DR2 photometric data (G mag vs. G BP − G RP ) of the four clusters. The red circles indicate the memberstars observed by SPA project. Table 1.
Parameters of the open clusters sample from Cantat-Gaudin et al. (2020).Cluster R.A. Dec. µ α µ δ parallax log(Age) A V d R GC z (J2000) (mas yr − ) (mas yr − ) (mas) (yr) (mag) (pc) (kpc) (pc)Collinder 350 17:48:14.26 + − . ± . − . ± .
243 2 . ± .
129 8.77 0.52 371 8.02 94Gulliver 51 02:01:20.40 + − . ± . − . ± .
079 0 . ± .
026 8.56 1.42 1536 9.41 52NGC 7044 21:13:08.16 + − . ± . − . ± .
151 0 . ± .
078 9.22 1.78 3252 8.73 − − + . ± . + . ± .
165 0 . ± .
066 9.44 0.68 1522 9.41 52 and more than 2.5 mag for Gul 51_2). We adopt the averagecluster extinction as in Cantat-Gaudin et al. (2020), howeverdi ff erential extinction can significantly a ff ect photometric stel-lar parameters, with strong e ff ects on the photometric T e ff andlog g . In Ruprecht 171, we observed seven stars, four of themwith very similar colours and magnitudes. This cluster is closerand less a ff ected by extinction. The photometric parameters ofthe four hottest stars (Rup171_4-5-6-7) are indeed very similar.Finally, Collinder 350 is close and not significantly a ff ected byextinction.Regarding the stellar parameters T e ff , iso and log g iso , we se-lect the best isochrone for each cluster, starting from the clusterparameters of Cantat-Gaudin et al. (2020) and fine-tuning themwith our grid of isochrones. The presence of high di ff erentialreddening produces, in some cases, di ff erences between the twomethods for estimating the photometric stellar parameters andconfirms that they can only be used as priors for our spectro-scopic analysis.
5. Spectral analysis
We follow two di ff erent approaches to analyse our sample stars.The first one is a spectral analysis using the EWs with Fast Auto- matic MOOG Analysis (FAMA, Magrini et al. 2013), while thesecond one is a spectral fitting with ROTFIT (Frasca et al. 2006,2019). We perform a spectral analysis based on the EWs to determinethe atmospheric parameters and abundances of our sample stars.We measure the EWs of the spectral absorption lines with theD aospec (Stetson & Pancino 2008) tool (in the form of DOOp- DAOSPEC Output Optimiser pipeline, an automatic wrapper;Cantat-Gaudin et al. 2014). We use the master list of atomic tran-sitions that was prepared for the analysis of the stellar spectrafor the Gaia-ESO survey (Heiter et al. 2015b). This line list in-cludes quality flags, such as ‘Y’ (yes), ‘N’ (no), and ‘U’ (unde-termined). These flags are assigned on the basis of the quality ofthe line profiles (at the spectral resolution of about 47000) andthe accuracy of the log g f derived from the comparison of syn-thetic spectra with a spectrum of the Sun and Arcturus. If theprofile of a given line is unblended in both the Sun and Arcturusand its log g f value is well determined, the flag will be ‘Y / Y’,and ‘N / N’ otherwise. In our analysis, we consider all lines, ex-
Article number, page 4 of 17asali, G. et al.: SPA-OCs
Table 2.
Observed candidate member stars for the four clusters.Gaia DR2 ID ID R.A. Dec.
G BP − RP RV Exposure Time S / N parallax(J2000) (mag) (mag) (km s − ) (s) (mas)4372743213795720704 Cr350_1 266.603650 + − .
57 600 172 2.910 ± + − .
73 1800 158 2.770 ± + − .
36 7200 68 0.725 ± + ± + − .
88 7200 46 0.336 ± + − .
15 7200 36 0.216 ± + − .
45 7200 34 0.280 ± + − .
44 7200 31 0.333 ± − .
55 3600 53 0.746 ± − .
32 7200 66 0.638 ± − .
84 7200 85 0.667 ± − .
46 7200 53 0.626 ± − .
03 7200 45 0.677 ± − .
81 7200 38 0.629 ± − .
28 7200 36 0.639 ± Table 3.
Photometric stellar parameters from
Gaia
DR2 and PARSECisochrones. ID T e ff , Gaia log g Gaia T e ff , iso log g iso (K) (dex) (K) (dex)Cr350_1 4030 1.24 (1.30) 4100 1.35Cr350_2 4620 2.51 (2.54) 4880 2.71Gul51_1 4453 1.90 (2.02) 4630 2.26Gul51_2 6122* 2.93 (2.92) 6610 3.11NGC7044_1 3825** 0.89 (1.04) 3850 0.98NGC7044_2 3898** 0.98 (0.82) 3900 1.06NGC7044_3 3929** 1.13 (1.14) 4000 1.23NGC7044_4 4021** 1.20 (1.34) 4000 1.23Rup171_1 3863** 1.25 (1.39) 3930 1.25Rup171_2 3940 1.52 (1.53) 4090 1.54Rup171_3 4220 1.68 (1.72) 4300 1.76Rup171_4 4600 2.42 (2.41) 4750 2.52Rup171_5 4600 2.43 (2.48) 4750 2.52Rup171_6 4600 2.44 (2.43) 4750 2.52Rup171_8 4600 2.38 (2.39) 4750 2.52 Notes. (*) T e ff from the calibration of dwarf stars in Mucciarelli & Bel-lazzini (2020); (**) T e ff from Gaia
DR2. cept those with ‘N’ for the log g f values. At high-spectral res-olution, the line profile might di ff er from the Gaussian profile,especially for the strongest lines. However, this e ff ect is usuallynegligible at least for EWs < −
120 mÅ, which correspondsto the range we consider for the analysis, as discussed in Spinaet al. (2020), who tested the e ff ect of the use of a Voigt profilesin the measurement of the EWs of strong lines for HARPS spec-tra. The two measurements, that is those made with a Gaussianand those with a Voigt profile, are consistent, and the uncertaintydue to the choice of the profile is negligible with respect to theuncertainty related to continuum placement.We determine the atmospheric parameters with MOOG inthe automatic form (FAMA, Magrini et al. 2013). FAMA usesMOOG in its 2014 version (Sneden et al. 2012) and MARCSmodel atmospheres (Gustafsson et al. 2008) to determine stellarparameters and to calculate abundances. It iteratively searchesfor the three equilibria: the ionisation equilibrium, excitationequilibrium, and minimisation of the trend between the reduced EW log(EW /λ ) and the iron abundance. The e ff ective tempera-ture T e ff is obtained by minimising the slope between iron abun-dance and the excitation potential; the surface gravity log g isobtained by assuming the ionisation equilibrium condition be-tween Fe i and Fe ii ; and the microturbulence ξ is obtained byminimising the trend between iron abundance and the reducedEWs. To avoid saturated and overly weak lines, in our analysiswe only consider lines with 20 mÅ < EW <
120 mÅ for iron,and 5mÅ < EW <
120 mÅ for other elements.The evaluation of the uncertainties on the final stellar pa-rameters is done when the slope of the trend between the Feabundance and reduced EWs, and the di ff erence between Fe i and Fe ii abundances, are not completely minimised. FAMA usesthe dispersion of the abundances to derive the errors on the at-mospheric parameters: T e ff error is the ratio between σ Fe i , thedispersion around the mean of log n (Fe i ), and the range of exci-tation potential EP; log g error is (cid:113) σ i + σ ii ; and ξ error isthe ratio between σ Fe i and the range of reduced EW. There aretwo types of errors in metallicity: (i) the statistical uncertaintydue to the random errors in the EW measurements and to un-certainties on the atomic parameters; and (ii) the errors on theabundances generated by the uncertainties in the determinationof the atmospheric parameters. Both are indicated in Table 4.Finally, FAMA computes the elemental abundances of Li, C,Na, Mg, Al, Si, Ca, Sc, Ti ( i and ii ), V, Cr, Fe ( i and ii ), Co, Ni,Y, Zr ( i and ii ), La, Ce, and Eu.In Table 4 we present the results of our spectral analysis inwhich the four stellar parameters ( T e ff , log g , [Fe / H], and ξ ) arevaried up to convergence. The errors in parenthesis for [Fe / H] arethe errors on the abundances generated by the uncertainties in thedetermination of the atmospheric parameters. One of the stars ofGulliver 51 (Gul51_2) is a fast rotator, and therefore it cannot beanalysed with FAMA; we compute its v sin i using ROTFIT andlist this in Table 5. The results are presented in the left panel ofFig. 2.From the left panel of Fig. 2, we notice a trend betweenmetallicity and T e ff . In particular, the coolest (lowest log g )stars of our sample reach the lowest metallicities. For clus-ters with member stars spanning wide ranges in T e ff and log g (Ruprecht 171 and Collinder 350), the trends are particularlyevident: three stars of Ruprecht 171 (Rup171_1, Rup171_2,Rup171_3) and one of Collinder 350 (Cr350_1) with T e ff < Article number, page 5 of 17 & A proofs: manuscript no. 39176corr
Table 4.
Stellar parameters obtained with FAMA. ID T e ff log g [Fe / H] ξ (K) (dex) (dex) (km / s)Cr350_1 4100 ±
100 1.35 ± − ± ± ± ±
110 2.85 ± − ± ± ± ±
90 2.45 ± − ± ± ± ±
90 0.95 ± − ± ± ± ±
70 1.06 ± − ± ± ± ±
100 1.23 ± − ± ± ± ±
90 1.23 ± − ± ± ± ±
90 1.18 ± − ± ± ± ±
70 1.54 ± − ± ± ± ±
80 1.76 ± − ± ± ± ±
80 2.82 ± ± ± ± ±
110 2.71 ± ± ± ± ±
150 2.91 ± ± ± ± ±
90 2.79 ± ± ± ± Fig. 2. [Fe / H] vs. T e ff for the results obtained with FAMA (left panel) and ROTFIT (right panel). The labels indicate the ID of each star member. K and log g between 1 and 1.8 dex have much lower metallic-ity than the other members of the same clusters. For clusters inwhich only cool giants are observed, such as NGC 7044, their[Fe / H] is lower than the literature value − .
16 dex (Warren &Cole 2009). Moreover, given the location of the clusters closeto the Sun and their age younger than 3 Gyr, we expect a metalcontent near to the solar one (within ± − − for the radial metallicity gradient, see, e.g.Zhong et al. 2020, and references therein). This is indeed true forthe hottest stars of our sample. For instance, the metallicity ofCr350_2 is [Fe / H] = − . ± .
06 in agreement within the errorswith the literature value ( + / H] = + . ± .
03 dex).The trends with T e ff and log g are thus general consideringour sample clusters as a whole: stars located in the upper RGB,the coolest ones, are more metal-poor than stars located in RC.This topic is not new, since it has already been addressed in theanalysis of cool stars on the upper RGB (see, e.g. Worley et al. 2010; Worley & Cottrell 2010) and is discussed in more detail inSect. 6. We also analyse our sample stars with the code ROTFIT (Frascaet al. 2006, 2019). ROTFIT uses a grid of template spectra andperforms a χ minimisation of the di ff erence between the tem-plate and target spectrum in selected spectral regions. The gridof templates is composed of high-resolution spectra of stars withknown parameters present in the ELODIE archive . In order touse this grid, we need to degrade our HARPS-N spectra to theresolution of ELODIE (R =
42 000) and to resample them on theELODIE spectral points ( ∆ λ = . v sin i to minimise the χ . For each analysed spectral region,the weighted average of the parameters obtained for the ten best http://atlas.obs-hp.fr/elodie/ Article number, page 6 of 17asali, G. et al.: SPA-OCs templates is taken. In particular, we analyse 28 spectral chunks,each of 100 Å in width, in the wavelength range 4000–6800 Å.The final parameters ( T e ff , log g and the iron abundance [Fe / H])are the average of the results of the individual spectral region,weighted according to the χ .The atmospheric parameters, spectral type, and v sin i ob-tained with ROTFIT are presented in Table 5. With ROTFIT,we can also provide the atmospheric parameters of Gul51_2, thefast rotator that cannot be analysed with FAMA. It is a hot starand its T e ff and log g , derived by ROTFIT, are in agreementwith the values of the TESS catalogue (Stassun et al. 2018):log g = . ± .
54 dex and T e ff = / H]of the other member star, Gul51_1. However, a very large er-ror is associated with this value because of the lower S / N andits fast-rotator nature, and therefore we do not consider it in thecomputation of the mean value of metallicity of Gul 51.The right panel of Fig. 2 shows the metallicity [Fe / H] as afunction of T e ff for the ROTFIT results. The analysis with ROT-FIT yields a smaller di ff erence in [Fe / H] between cool and warmstars of the same cluster. However, there are some residual dif-ferences between the cooler stars (ID 1, 2) and the hotter onesin Ruprecht 171. Finally, as shown by Table 5 and Fig. 2, themetallicity obtained for NGC 7044 is in agreement with the CaTvalue.
6. The cool giant stars
In this section, we investigate the causes of the low metallicitiesmeasured with the EWs in the cool giants. In principle, stars inopen clusters should present a homogeneous chemical composi-tion (see, e.g. Bovy 2016), at least within some range, typicallyof few ∼ ff erent evolutionary stages, might display morenotable di ff erences in their chemical patterns. These di ff erencescan be due to physical phenomena, such as atomic di ff usion andmixing (e.g. Lagarde et al. 2019; Casali et al. 2019; BertelliMotta et al. 2017; Semenova et al. 2020), or to analysis e ff ects,such as non-local thermodynamic equilibrium (NLTE) e ff ectsor correlations between atmospheric parameters and abundances(see, e.g. Blanco-Cuaresma et al. 2015, for a review). In addition,the analysis of cool giant stars can be a ff ected by several com-plications, such as the presence of a forest of molecular lines,possible asymmetric shapes of the lines due to mass loss, devi-ations from hydrostatic equilibrium, the presence of giant con-vective cells, and the deviations from LTE (see, e.g. Asplund2005; Bergemann & Nordlander 2014). Here we discuss someaspects of the cool giant analysis, including the choice of thestellar parameters, the correction for NLTE, the selection of theline list, the adoption of the atmosphere models, and the contin-uum placement. We investigate the e ff ect of adopting stellar parameters from Gaia , which are calculated based on their photometry, ratherthan the spectroscopic ones with FAMA which are based onthe EW analysis. If there is indeed a deviation from the hydro-static equilibrium, the gravity derived from the ionisation bal-ance might be incorrect, and thus produce incorrect abundances.The simultaneous determination of the stellar parameters from spectroscopy can produce, for instance, a degeneration amongthose parameters, producing a correlation between log g and[Fe / H] (cf. Blanco-Cuaresma et al. 2014). Therefore, the use oftemperature and log g independent of spectroscopy might solvethis eventual degeneracy. First of all, we fix the gravity to thevalues of Table 3 using both the gravities from Gaia and fromthe comparison with isochrones, and let T e ff , ξ , and [Fe / H] varyup to convergence. We then keep the photometric log g and T e ff constant, while ξ and [Fe / H] are varied up to convergence. Evenwith these choices, the global trends of [Fe / H] versus T e ff , withcool and low-gravity stars having a lower metal content than theother member stars, are still present with the EW method (e.g. adiscrepancy in metallicity of ∼ . ∼ . FAMA uses MOOG to compute abundances in the LTE approx-imation. The photospheres of cool giants, with their low surfacegravities and thus low densities, might depart from LTE, beingtranslucent over large radial extensions. Thus, the radiative ratescan dominate the collisional rates for many atomic transitions(cf. Short & Hauschildt 2003). This e ff ect is usually in place inlow-gravity giant stars, but it is stronger for metal-poor stars. Thedeparture from LTE could be due to the high excitation levels ofFe i , which do not thermally couple to the ground state of Fe ii .Another aspect of the departure could be the treatment of poorlyknown inelastic collisions with hydrogen atoms (Mashonkinaet al. 2011). There are indeed several studies showing that NLTEe ff ects in the ionisation balance of Fe i / Fe ii are larger for giantmetal-poor stars (e.g. Bergemann et al. 2012; Mashonkina et al.2011; Collet et al. 2005, and references therein) than for theircounterparts at higher metallicity. Since iron lines are used toderive stellar parameters, we aim to estimate the e ff ect of NLTEin the spectra of our giant stars, even if their metallicity is not solow that we expect a strong departure from LTE.We estimate the NLTE abundance corrections for each Feline using the calculator by the MPIA NLTE group (Bergemannet al. 2012). The NLTE abundance corrections are computed as ∆ Fe = log A(Fe) NLTE − log A(Fe) LTE , which is the di ff erence be-tween the NLTE and LTE abundances. These corrections on theconsidered Fe lines are of the order 0 . − .
02 dex. They arenegligible as expected for solar metallicity giants (Bergemannet al. 2012), even at the very low surface gravities of the cooleststars of our sample. Thus, these corrections alone cannot justifythe discrepancy in metallicity among members of the same clus-ter, and they do not a ff ect the determination of the spectroscopicstellar parameters from the EW analysis. The spectra of cool stars, with T e ff < =
115 000, whichmakes the problem of the blending less dominant. As pointed outby Tsantaki et al. (2013), the identification of unblended EWsof photospheric lines and the continuum placement are di ffi culttasks in cool stars. This might lead to incorrect measurement ofthe EWs with consequent errors in the derived stellar parame-ters. Following the work of Tsantaki et al. (2013), we adopt theirline list designed for the analysis of cool stars. However, we re-iterate that the line list of Tsantaki et al. (2013) is designed for http://nlte.mpia.de/gui-siuAC_secE.php Article number, page 7 of 17 & A proofs: manuscript no. 39176corr
Table 5.
Stellar parameters derived with ROTFIT. ID T e ff log g [Fe / H] Sp.Type v sin i (K) (dex) (dex) (km s − )Cr350_1 4330 ±
90 1.28 ± . ± .
07 K3II 1.3 ± ±
60 2.99 ± − . ± .
09 G8III 6.7 ± ±
60 2.66 ± − . ± .
08 K0III 2.3 ± ±
298 4.06 ± − . ± .
13 A9IV 251.6 ± ±
50 1.53 ± − . ± .
08 K5III 2.0 ± ±
50 1.51 ± − . ± .
11 K5III 1.8 ± ±
60 1.50 ± − . ± .
09 K5III 1.7 ± ±
60 1.50 ± − . ± .
08 K5III 2.0 ± ±
50 1.56 ± − . ± .
09 K5III 2.0 ± ±
70 1.56 ± − . ± .
10 K3.5IIIb 1.6 ± ±
110 2.13 ± . ± .
10 K2III 1.4 ± ±
70 2.67 ± . ± .
10 K1.5III 0.7 ± ±
60 2.66 ± . ± .
10 K1.5III 0.8 ± ±
60 2.64 ± . ± .
10 K0III 1.0 ± ±
60 2.64 ± . ± .
11 K1.5III 0.7 ± ff ect of crowdingand continuum placement can be even more severe.To understand whether or not the di ff erence in metallicity isdue to the Gaia-ESO line list (used to obtain the results in Ta-ble 4), we perform the EW spectral analysis using the line list byTsantaki et al. (2013). The atmospheric parameters achieved byboth line lists are consistent within the errors. Moreover, thereare still di ff erences between the [Fe / H] in stars in distinct evolu-tionary stages within the same cluster.
Finally, we test the influence of the choice of model atmosphereson the determination of stellar parameters and abundances. Toperform our test, we recompute the stellar parameters using theKurucz models (Castelli & Kurucz 2003) instead of the MARCSspherical models. The di ff erences are negligible for all parame-ters. However, as discussed by Short & Hauschildt (2003), forgiants of spectral types G to M, the failure in reproducing therelation between T e ff and the colours in the blue and violet re-gions of the spectrum, such as B − V , is typical of most modelatmospheres. Bessell et al. (1998) indicated the incomplete orerroneous opacity in the blue-violet region as the origin of thatdiscrepancy. Since this problem a ff ects di ff erent types of modelatmospheres, the change of model does not help to solve the dis-crepancy. The problem of obtaining overly low metallicity in cool stars hasbeen known for a long time, but a clear solution is still missing.An example of this e ff ect in open clusters is the star at the tip ofthe RGB of the cluster Collinder 261. This star is cooler than theother studied members, and its measured metal content is lowerthan the other stars (Carretta et al. 2005; Friel et al. 2003).There is not a single explanation to clarify the di ff erences inmetallicities in stars of the same cluster belonging to di ff erentevolutionary phases, in particular the upper RGB. For the clus-ter with a larger number of observed member stars, Rup 171,we severely underestimate [Fe / H] with the EW method (maxi-mum di ff erences > . ∼ g < . ff ect is smaller with parameters and metallicities derived through ROTFIT. In this case, the maximumdi ff erences in [Fe / H] between the coolest and hottest stars are ofthe order of ∼ ffi cult to evaluate whether or not their [Fe / H]values are underestimated. However, the [Fe / H] measured withROTFIT is higher by about 0.2 dex with respect to [Fe / H] de-termined with the EWs. In addition, the former metallicity is inbetter agreement with the literature value (albeit based on theCaT method, which is not free from large uncertainties and bi-ases). We analyse two member stars of Cr 350. For the hottestone, spectral fitting and EW methods are in agreement, while forthe coolest one the EW method tends to underestimate its [Fe / H],as in the other cool giants.The combination of several aspects makes the determina-tion of metallicity from the EW analysis in the cool giants(T e ff < g < . ff erence in [Fe / H] between FAMA andROTFIT as a function of T e ff , colour-coded by log g . From ouranalysis described in the previous sections, we consider the prin-cipal causes to be as follows: the erroneous opacity in modelatmospheres of cool G to M giant stars (see, e.g Bessell et al.1998; Short & Hauschildt 2003), and the large number of linesin the spectrum, both atomic ones and molecular bands for cooland metal rich stars, which makes it di ffi cult to define the con-tinuum near to the lines of interest and thus to measure reli-able EWs. Regarding the continuum placement, we recall thatD aospec adopts a global and not local continuum and this as-pect can only increase the di ffi culty in the continuum setting forcool giants. A clear example of this problem is seen in Fig. 4,where two normalised spectra are compared: the spectrum of thecool star Cr350_1 (top panel) and the spectrum of the warm starCr350_2 (bottom panel). The continuum computed by D aospec for Cr350_1 produces a normalised spectrum slightly above 1, inwhich the EWs are underestimated, in contrast to Cr350_2. Asexplained in Cantat-Gaudin et al. (2014), the continuum definedby D aospec , on which the EW fit is based, is not the true contin-uum of the spectrum (i.e. the continuous star emission after allthe lines are excluded), but an e ff ective continuum, which is thetrue continuum depressed by a statistical estimate of the contam-inating lines (the unresolved or undetected ones, producing a sort Article number, page 8 of 17asali, G. et al.: SPA-OCs of line blanketing). The use of the e ff ective continuum improvesthe measurement of unblended lines, as demonstrated in Cantat-Gaudin et al. (2014), but it is sometimes perceived as being toolow, especially in spectra dominated by line crowding (i.e. inparticular high-metallicity giant stars) or with decreasing S / N of the spectra. The case of Cr350_1 is a typical example of ourlimits in measuring the EWs of metal-rich giant stars, for whichwe likely underestimate the EWs and, consequently, we derivea lower [Fe / H] than in warmer stars. Moreover, it is known alsofrom other works that the analysis based on EW measurementstends to underestimate the [Fe / H] of cool giant stars, as shownby the EW analyses of benchmark stars performed by di ff erentgroups (Jofré et al. 2014; Heiter et al. 2015a). Consequently, forthe spectra of stars cooler than 4300 K, ROTFIT, which is lessprone to continuum setting and blending e ff ects, produces moresolid determinations of the stellar parameters than the EW anal-ysis. Fig. 3. Di ff erence in [Fe / H] between FAMA and ROTFIT as a functionof T e ff , colour-coded by log g . In what follows, we consider the results from the EW analy-sis only for stars hotter than 4300 K, namely one star in Cr 350and in Gul 51, and four stars in Rup 171. In Table 6, we comparethe mean metallicities of our cluster sample from ROTFIT (allsample) and from FAMA (only stars with T e ff > / H] is reported inparentheses. The two determinations of the mean cluster metal-licity are in good agreement within the uncertainties (1- σ stan-dard deviation). Table 6.
Mean metallicities of our sample open clusters.Cluster [Fe / H] ROTFIT (all) [Fe / H] FAMA (warm)Collinder 350 0.00 ± − ± − ± − ± − ± ± ±
7. Chemical abundances
Elemental abundances are computed with FAMA using the rou-tines abfind and blends of MOOG. The latter is used for ele-ments that present hyperfine splitting in their lines. To computethe Solar-scaled abundances and abundance ratios [X / H] and [X / Fe], we define our Solar scale measuring the element abun-dances on a solar spectrum. For this task we use a spectrum ofCeres collected by the twin HARPS spectrograph at the 3.6 mESO telescope. The Solar abundances are listed in Table 7, inwhich we show our Solar abundances and the photospheric So-lar abundances from Grevesse et al. (2007). For the elements thatcannot be measured in the Ceres spectrum, we use the valuesfrom Grevesse et al. (2007).
Table 7.
Solar chemical abundances.
Element Sun (Ceres) Sun (G07)Li i – 1 . ± . i – 8 . ± . i . ± .
04 6 . ± . i . ± .
02 7 . ± . i . ± .
02 6 . ± . i . ± .
04 7 . ± . i . ± .
05 6 . ± . ii . ± .
03 3 . ± . i . ± .
05 4 . ± . ii . ± .
07 –V i . ± .
07 4 . ± . i . ± .
03 5 . ± . i . ± .
09 7 . ± . ii . ± .
09 –Co i . ± .
08 4 . ± . i . ± .
08 6 . ± . ii . ± .
09 2 . ± . i – 2 . ± . ii . ± .
06 –La ii . ± .
36 1 . ± . ii – 1 . ± . ii – 0 . ± . Notes.
G07: Grevesse et al. (2007).
We cannot produce abundances for NGC 7044 and for thecool stars of Rup 171 and Cr 350 with the stellar parametersderived via EW analysis. In Table 8 we list the elemental abun-dances for each warm member star ( T e ff > / H]. Theyare reported in Table 4 and range from 0.01 to 0.06 dex.In Table 9, we report the mean abundance ratios for eachcluster. The errors for Rup 171 are the standard deviations of themean, as we have four warm stars for these clusters; whereas, forGul 51 and Cr 350 we can use only one star, and therefore theuncertainties are the errors on each individual measurement.
Collinder 350 is the only cluster of our sample that has previousdeterminations of chemical abundances from high-resolutionspectral analysis. The previous works on this cluster present-ing high-resolution spectroscopy are Pakhomov et al. (2009),Blanco-Cuaresma et al. (2015), and Blanco-Cuaresma & Fraix-Burnet (2018). The first work analysed a spectrum of the star
Article number, page 9 of 17 & A proofs: manuscript no. 39176corr
Fig. 4.
Portion of the spectra, normalised with D aospec , of the cool star Cr350_1 (top panel) and the warm star Cr305_2 (bottom panel) with thecontinuum (red line).
Table 8.
Elemental abundances of each star with T e ff > + log(X / H).Star A(Li i ) n(Li i ) A(C i ) n(C i ) A(Na i ) n(Na i ) A(Mg i ) n(Mg i ) A(Al i ) n(Al i ) ...Cr350_2 1 . ± .
03 1 7 . ± .
07 2 6 . ± .
02 2 7 . ± .
12 2 6 . ± .
01 2Gul51_1 1 . ± .
07 2 7 . ± .
06 1 6 . ± .
07 4 7 . ± .
01 2 6 . ± .
03 2Rup171_4 – – 7 . ± .
03 2 6 . ± .
07 4 7 . ± .
02 3 6 . ± .
03 2Rup171_5 – – 7 . ± .
08 1 6 . ± .
06 4 7 . ± .
01 2 6 . ± .
01 2Rup171_6 – – 7 . ± .
06 2 6 . ± .
03 4 7 . ± .
02 2 6 . ± .
07 2Rup171_8 0 . ± .
06 1 7 . ± .
02 2 6 . ± .
06 4 7 . ± .
01 2 6 . ± .
04 1
Notes. n(X) represents the number of lines for each element. The full version of this table is available online at the CDS.
Cr350_1 collected with the red branch spectrograph ( R ∼
50 000) at the 2.16 m Telescope at the Xinglong Observa-tory in China. Pakhomov et al. (2009) found a metallicity of[Fe / H] =+ . ± .
06 dex. Blanco-Cuaresma et al. (2015) andBlanco-Cuaresma & Fraix-Burnet (2018) analysed an archivalNARVAL spectrum of a giant star ( R ∼
80 000), obtainingslightly di ff erent values of [Fe / H] = − . ± .
01 (or [Fe / H] = . ff erent normalisation) and 0.03 dex, respectively. Table 10shows the abundance ratios for Cr350_1 by Pakhomov et al.(2009) and for a giant star in Collinder 350, for which the co-ordinates are not available in the papers (Blanco-Cuaresma et al.2015; Blanco-Cuaresma & Fraix-Burnet 2018), compared withour results of Cr350_2. We exclude from the comparison the ele-ments that are modified by stellar mixing, such as C and Na, be-cause we do not compare the same star. Also, without taking intoaccount the possible di ff erences in solar reference abundancesadopted in the studies mentioned above, Fig. 5 shows that mostof our abundance ratios are in agreement within the errors be-tween our work and those papers. The only exception is [La / Fe],for which we find a lower value.
8. Results
Open clusters are among the best tracers of the radial metallicitydistribution in the Galactic disc. Recent works within the Gaia-ESO and APOGEE surveys (e.g. Magrini et al. 2018; Donoret al. 2020) have investigated the shape of the radial metallic-ity gradient using relatively large numbers of homogeneouslyanalysed clusters. We therefore have the opportunity to compareour sample clusters with the combined APOGEE DR16 (listedby Donor et al. 2020, 128 clusters), and old and young Gaia-ESO iDR5 open clusters samples listed by Magrini et al. (2018,22 clusters) and Baratella et al. (2020, 4 clusters), respectively.Regarding the APOGEE data, we consider only open clusterswith at least three member stars and with a high-quality flag,and therefore we reduce their number to 38 clusters. The resultis shown in Fig. 6. The metallicity of our cluster sample is ob-tained as an average between the two values shown in Table 6.The metallicities of our four clusters, as well as ASCC 123 andPraesepe located around the Solar position, agree very well withthe results of Gaia-ESO and APOGEE, and confirm an intrinsicdispersion of [Fe / H] at each Galactocentric radius. The inner-most cluster, Rup 171, shows instead a slightly lower metallic-
Article number, page 10 of 17asali, G. et al.: SPA-OCs
Table 9.
Mean abundance ratios for the clusters.Ratio Cr 350 Gul 51 Rup 171odd-elements[Na / Fe] 0 . ± .
10 0 . ± .
12 0 . ± . / Fe] − . ± .
10 0 . ± .
09 0 . ± . α -elements[Mg / Fe] 0 . ± .
15 0 . ± .
09 0 . ± . / Fe] 0 . ± .
12 0 . ± .
10 0 . ± . / Fe] − . ± . − . ± . − . ± . / Fe] 0 . ± .
13 0 . ± .
11 0 . ± . / Fe] ∗∗ . ± . − . ± .
13 0 . ± . / Fe] 0 . ± .
15 0 . ± .
14 0 . ± . / Fe] 0 . ± . − . ± . − . ± . / Fe] − . ± . − . ± .
14 0 . ± . / Fe] − . ± . − . ± .
12 0 . ± . / Fe] 0 . ± .
16 0 . ± .
12 0 . ± . / Fe] ∗ , ∗∗ . ± .
14 0 . ± .
12 0 . ± . / Fe] − . ± . − . ± . − . ± . / Fe] ∗ − . ± .
14 0 . ± . − . ± . / Fe] ∗ − . ± .
12 0 . ± . − . ± . Notes.
Abundance ratios on our Solar scale, with the exception of ( ∗ ), which are calculated using the solar value from Grevesse et al. (2007). ( ∗∗ )indicates the average value between [TiI / Fe] and [TiII / Fe], and [ZrI / Fe] and [ZrII / Fe].
Table 10.
Abundances for Collinder 350Ratio BC18 BC15 P09 This work[Fe / H] 0 . − . ± .
01 0 . ± .
06 0 . ± . / Fe] − .
07 0 . ± .
03 0 . ± .
07 0 . ± . / Fe] − .
08 – 0 . ± . − . ± . / Fe] − .
07 0 . ± .
06 0 . ± .
12 0 . ± . / Fe] 0 .
04 0 . ± .
11 0 . ± . − . ± . / Fe] 0 .
00 – − . ± . ∗ . ± . / Fe] − .
02 0 . ± . − . ± . − . ± . / Fe] − .
09 0 . ± . − . ± .
08 0 . ± . / Fe] − .
02 0 . ± .
14 0 . ± .
08 0 . ± . / Fe] − . − . ± . − . ± . − . ± . / Fe] − . − . ± . − . ± . − . ± . / Fe] 0 .
18 0 . ± .
17 0 . ± .
07 0 . ± . / Fe] 0 .
13 – 0 . ± . ∗ . ± . / Fe] 0 .
21 – 0 . ± . − . ± . / Fe] 0 .
15 – 0 . ± . − . ± . / Fe] 0 .
02 – 0 . ± . − . ± . Notes.
BC18: Blanco-Cuaresma & Fraix-Burnet (2018); BC15: Blanco-Cuaresma et al. (2015); P09: Pakhomov et al. (2009). Errors with ∗ arelower limits. ity than the other clusters located at similar Galactocentric dis-tances. However, considering only the hottest member stars ofRup 171 analysed through ROTFIT (ID from 3 to 8), we obtain[Fe / H] ∼ / Fe] versus [Fe / H] for the elementin common among our analysis, the APOGEE results in Donoret al. (2020), and the Gaia-ESO results in Magrini et al. (2017,for Ca, Sc, V, Cr), Magrini et al. (2018, for Zr, La, Ce, Eu),and Casali et al. (2020, for Mg, Al, Si, Ti, Y). We also addthe abundance ratios for two young SPA clusters: Praesepe fromD’Orazi et al. (2020), and ASCC 123 from Frasca et al. (2019)– ASCC 123 abundances scaled using the solar reference of
Fig. 5. Di ff erence between our abundance ratios and the values fromthe literature: in green the comparison with Pakhomov et al. (2009), inblue the comparison with Blanco-Cuaresma et al. (2015), and in red thecomparison with Blanco-Cuaresma & Fraix-Burnet (2018). In the lastcase, the error bars take into account only the errors on the abundanceratios computed in this work, and therefore they are lower limits. Grevesse et al. (2007). For the latter, we exclude the abundanceratios displayed as overabundant or under-abundant with respectto the field stars of Gaia-ESO DR4 in Frasca et al. (2019) .Our clusters follow the main trends of the APOGEE andGaia-ESO clusters: open clusters are a thin disc population, andthey do not reach very low metallicities: their [Fe / H] is in therange [ − + α elements (Mg, Ca, Si, Ti) all show anenhancement in [X / Fe] towards lower [Fe / H]. Their productionis essentially due to core collapse supernovae (see, e.g. Woosley& Weaver 1995). ASCC 123 is overabundant in [Ca / Fe] andunder-abundant in [Si / Fe] with respect to other clusters. Alu-minium and sodium are quite scattered, which is an e ff ect of Article number, page 11 of 17 & A proofs: manuscript no. 39176corr
Fig. 6.
Radial metallicity gradient: in red the results of Gaia-ESO iDR5 from Magrini et al. (2018), in green the results of APOGEE DR16 (clusterswith at least three member stars and high-quality flag), and in blue our SPA clusters. The blue square is the mean [Fe / H] of Rup 171 obtained withonly the hottest stars analysed by ROTFIT (ID from 3 to 8). The clusters in common between APOGEE and Gaia-ESO are linked by a black line.The magenta and black circles are the SPA clusters ASCC 123 and Praesepe studied by Frasca et al. (2019) and D’Orazi et al. (2020), respectively.The light green diamonds are the young Gaia-ESO open clusters by Baratella et al. (2020). the internal mixing that causes some sodium overabundance atthe surface of red giants more massive than ∼ (cid:12) . A sim-ilar e ff ect, even if not predicted by stellar evolution models, isobserved for Al (cf. Smiljanic et al. 2016). The iron-peak ele-ments (Sc, V, Cr, Co, Ni) follow an almost flat trend, with di ff er-ent degrees of scattering due to the di ffi culty in measuring someelements, such as for example V and Co. Finally, we comparefive neutron capture elements, four of them predominantly pro-duced by the slow (s) process (Y, Zr, La, Ce) and one by therapid (r) process, Eu. As in Casali et al. (2020), [Y / Fe] shows apeak at solar metallicity, and decreases at sub- and super-solarmetallicities. [Zr / Fe] tends to increase towards lower metallici-ties, a signature of an important production in massive stars atearly epochs. For [La / Fe] and [Ce / Fe], the general trend is sim-ilar to that of Zr, with an increasing trend at low [Fe / H]. Theabundance ratios [La / Fe] of our clusters has an o ff set with re-spect to the Gaia-ESO sample. It might be due to the di ff erentsolar scale for La used in Magrini et al. (2018): 1 .
00 dex (basedon the giant stars of M67) against 1 .
13 dex, determined in oursolar spectrum. Finally, Eu is an almost pure r-process elementproduced on shorter timescales and with a behaviour similar tothe α -elements. This behaviour is confirmed by the combinationof our sample with the literature one. In general, the abundanceratios of the SPA clusters follow the main trends.In Fig. 8, we show the same abundance ratios as in Fig. 7,plotted as a function of Galactocentric distance, R GC . As inFig. 7, our open clusters follow the main trends. For the α -elements, we have a slight enhancement increasing towards the Galactic outskirts. This enhancement is an indication of theinside-out formation of the disc, in which inner regions formedat high star formation rates, thus being quickly enriched by prod-ucts of SNe Ia with a consequent lower [ α / Fe]. The gradients ofthe iron-peak elements (V, Cr, Co, Ni) are flat, with di ff erent lev-els of scatter, indicating the expected similarity with the [Fe / H]gradient. The [Sc / Fe] gradient resembles the [ α / Fe] gradient, in-dicating a contribution from core-collapse supernovae in its nu-cleosynthesis. The [Y / Fe] gradient shows a peak at solar Galac-tocentric distance, and a decreasing trend towards the inner disc,as already discussed in Casali et al. (2020). Even the innermostSPA open clusters do not manifest any high [Y / Fe], but followthe trend of the literature, confirming a less e ffi cient productionof Y in the inner disc and at high metallicity. This behaviour isin common with the other neutron-capture elements with a pre-dominance from s-process. The elements La and Ce share thesame behaviour in the inner disc, while they are more enhancedin the outskirts as a consequence of their partial production alsoin massive stars. Finally [Eu / Fe] versus R GC shows an increasein the outer disc, which is characteristic of a double mechanismof production of this element (Van der Swaelmen et al. in prep.).In Fig. 9, we present [ α / Fe] versus [Fe / H] (where [ α / Fe]is the sum of [Mg / Fe], [Ca / Fe], [Si / Fe] and [Ti / Fe] divided byfour) for our cluster sample as well as for ASCC 123 and Prae-sepe, together with the Gaia-ESO and APOGEE clusters, and thefield stars from Gaia-ESO DR4. The cluster population is colourcoded by age using 790 Myr for Praesepe (Brandt & Huang2015; D’Orazi et al. 2020), the mean value of the range of 100–
Article number, page 12 of 17asali, G. et al.: SPA-OCs
Fig. 7.
Abundance ratios vs. [Fe / H]: blue indicates our SPA clusters with [Fe / H] as an average between the two values shown in Table 6; lightgreen indicates Praesepe; and orange indicates ASCC 123. The transparent colours indicate the following: green circles are APOGEE DR16 openclusters listed in Donor et al. (2020); red circles are the open clusters of Gaia-ESO DR4 with the [X / Fe] ratios calculated in Magrini et al. (2017);magenta circles are the open clusters of Gaia-ESO iDR5 with the [X / Fe] ratios calculated in Casali et al. (2020), not present in Magrini et al.(2017); and finally, black circles are the open clusters of Gaia-ESO iDR5 with the [X / Fe] ratios calculated in Magrini et al. (2018), not present inMagrini et al. (2017) and Casali et al. (2020). The open clusters in common between APOGEE and Gaia-ESO are linked by a black line.
250 Myr suggested in Frasca et al. (2019) for ASCC 123, andage determinations from Table 1 for our SPA clusters from Ma-grini et al. (2017) for Gaia-ESO and from Donor et al. (2020)for APOGEE. While the field population is well separated in thetwo components of thin and thick discs, open clusters are es-sentially a thin-disc population: most clusters are located in thelow- α thin disc with ages lower than 5 Gyr and the SPA clustersare in agreement with the thin-disc field stars within their errors.There are a few exceptions: the three clusters with high [Fe / H]are also slightly enhanced in [ α / Fe]. As discussed for NGC 6705in Magrini et al. (2014, 2015) and Casamiquela et al. (2018),they might be part of a young metal-rich and α -enhanced popu-lation similar to the ones found in the disc (see, e.g. Chiappiniet al. 2015), and very recently also in the bulge (Thorsbro et al.2020). Also in this case, the SPA open clusters follow the maintrend of the thin-disc population. We measure C and Li in some stars of our sample. Carbon ismeasured from the atomic lines (the adopted C i at 5052.144Å and 5380.325 Å are not a ff ected by strong NLTE; see e.g.Franchini et al. 2020)), while lithium is measured from the EWsof the resonance doublet at 6708 Å, which is unblended at thehigh-spectral resolution of HARPS-N. The photospheric abun-dances of these elements are a ff ected by stellar evolution duringthe RGB phase because of the first dredge up (FDU). During theFDU, the stellar convective envelope penetrates into the inner re-gions and brings previously processed materials to the surface,enriching the external layers in N and He, and diluting the Liand C abundances (see Lagarde et al. 2012; Masseron & Gilmore2015; Salaris et al. 2015; Casali et al. 2019, for more details). Af-ter the FDU, the star evolves along the RGB where extra mixingsuch as that caused by the thermohaline mechanism likely dom- Article number, page 13 of 17 & A proofs: manuscript no. 39176corr
Fig. 8.
Abundance ratios vs. Galactocentric distance R GC . Symbols as in Fig. 7. inates the abundance change of these elements (Lagarde et al.2012). These e ff ects are mainly expected along the upper RGB,after the RGB bump phase. The incidence of thermohaline mix-ing dominates at low metallicity and for low-mass stars, whileit is weaker in the Solar metallicity regime. Observational evi-dence of such extra-mixing processes are the very low lithiumabundance after the RGB bump and the variation of the isotopicratio C / C (e.g. Lind et al. 2009; Mucciarelli et al. 2011; La-garde et al. 2019).In Fig. 10, we show C / H and Li / H for the stars of Cr 350,Gul 51, and Rup 171 as a function of their log g . The curvesare the models of Lagarde et al. (2012) for three di ff erent masses(1.5, 2, 3 M (cid:12) , which encompass the range of turn-o ff masses forour clusters) for the standard (solid lines, s-models) and rotation-induced mixing models (dashed lines, r-models). The value C / Hdecreases with decreasing log g for the stars in Rup 171, fol-lowing the theoretical models of Lagarde et al. (2012) (see leftpanel of Fig. 10). Rup171_5 is located out of the locus of thetheoretical models, however its C / H determination exhibits thelargest uncertainties. The right panel shows the evolution of Liversus log g . The stars of Rup 171, the oldest cluster of our sam- ple, with a MSTO mass of ∼ (cid:12) , are in better agreement withmodels with rotation-induced mixing, while the stars in Cr 350and Gul 51, the youngest clusters of our sample, with the highestMSTO masses ( ∼ (cid:12) ), are in agreement with the predictions ofthe standard models, as expected for more massive stars at Solarmetallicity.
9. Summary
The SPA survey is extremely useful to characterise the Solarneighbourhood, in particular the nearby clusters, providing awide variety of elements thanks to the high spectral resolutionand large spectral coverage. In the framework of SPA, we presentthe analysis of four poorly studied clusters: Collinder 350, Gul-liver 51, NGC 7044, and Ruprecht 171. We analyse the high-resolution HARPS-N spectra taken at the TNG for, respectively,two, one, four, and seven candidate member stars. These stars be-long to the RGB or RC of the cluster evolutionary sequence. Weperform, for the first time for these clusters (except for Cr 350),a spectral analysis based on the EW measurements and spectralfitting to chemically characterise these clusters. With the EW
Article number, page 14 of 17asali, G. et al.: SPA-OCs
Fig. 9. [ α / Fe] vs. [Fe / H] for the open clusters, colour-coded by stellar age. The SPA clusters of this work are marked with a star symbol, whilethe cross and the triangle are the SPA clusters ASCC 123 and Praesepe, respectively; the transparent circles are the Gaia-ESO clusters with theabundances from Magrini et al. (2017) and the transparent diamonds are the APOGEE clusters with the abundances from Donor et al. (2020).The magenta dots (not colour-coded by age) in transparency are the field stars of the Gaia-ESO survey (public release, available at ESO archive).Finally, clusters in common between APOGEE and Gaia-ESO are linked by a black line. analysis, we find a correlation between stellar parameters andmetallicity for stars belonging to the same cluster, but in di ff erentevolutionary phases. In particular, the coolest stars (T e ff < g < ff ects, and the use of di ff erent linelists and model atmospheres. We conclude that the continuumplacement is extremely challenging for these stars, and mightlead to the derivation of incorrect metallicities. This is combinedwith the known inaccuracy of model atmospheres to reproducesome features of cool giants (Bessell et al. 1998). On the otherhand, ROTFIT provides results which do not strongly dependon the evolutionary phase. For this reason we adopt them forthe coolest stars of our sample. We derive chemical abundancesfor several elements for the stars with T e ff > / Fe] versus [Fe / H]for elements in common with the SPA, APOGEE, and Gaia-ESO samples. The SPA open clusters follow the trends shownby the other clusters. For instance, the α -elements (Mg, Ca, Si,Ti) show an enhancement in [X / Fe] towards lower metallicity; the iron-peak elements (Sc, V, Cr, Co, Ni) follow an almost flattrend; and the s-process elements and Eu follow the general be-haviour relatively well. We show the [X / Fe] ratios as a functionof the Galactocentric distance and the SPA clusters follow themain trends. α -elements increase slightly towards the Galacticoutskirts, implying an inside-out formation of the disc.We also derive the abundances of C and Li, which are com-pared with the models of Lagarde et al. (2012). Models withrotation-induced mixing are necessary to explain the Li abun-dances for the stars in the older cluster, Rup 171. No Li-richstars are detected in our sample.Further developments are expected from the homogenousanalysis of the whole sample of SPA clusters observed in theforthcoming runs. In addition, we plan to analyse the GIANOspectra, both in combination with HARPS-N (as done in Prae-sepe by D’Orazi et al. 2020) and alone, concentrating on ele-ments that are inaccessible in the optical spectra, such as fluo-rine. Acknowledgements.
We thank Giuseppe Bono for his useful com-ments. This work exploits the Simbad, Vizier, and NASA-ADSdatabases. We thank the TNG personnel for help during the obser-vations. This work has made use of data from the European SpaceAgency (ESA) mission Gaia (https: // / gaia), pro-cessed by the Gaia Data Processing and Analysis Consortium (DPAC,https: // / web / gaia / dpac / consortium). Funding for the DPAChas been provided by national institutions, in particular the institutions Article number, page 15 of 17 & A proofs: manuscript no. 39176corr
Fig. 10. C / H and Li / H versus log g . The circles are the stars of our sample and the curves are the theoretical models by Lagarde et al. (2012).Continuous lines are the standard models with masses of 1.5-2-3 M (cid:12) , and dashed lines are the models with rotation-induced extra-mixing for thesame masses. participating in the Gaia Multilateral Agreement. We made use of data fromthe Gaia-ESO Survey Data Archive (version 2), prepared and hosted by ESOat eso.org / qi / catalogQuery / index / References
Allard, F., Homeier, D., & Freytag, B. 2011, in Astronomical Society of the Pa-cific Conference Series, Vol. 448, 16th Cambridge Workshop on Cool Stars,Stellar Systems, and the Sun, ed. C. Johns-Krull, M. K. Browning, & A. A.West, 91Anders, F., Cantat-Gaudin, T., Quadrino-Lodoso, I., et al. 2020, arXiv e-prints,arXiv:2006.01690Andrae, R., Fouesneau, M., Creevey, O., et al. 2018, A&A, 616, A8Aparicio, A., Alfaro, E. J., Delgado, A. J., Rodriguez-Ulloa, J. A., & Cabrera-Cano, J. 1993, AJ, 106, 1547Asplund, M. 2005, ARA&A, 43, 481Bailer-Jones, C. A. L., Rybizki, J., Fouesneau, M., Mantelet, G., & Andrae, R.2018a, AJ, 156, 58Bailer-Jones, C. A. L., Rybizki, J., Fouesneau, M., Mantelet, G., & Andrae, R.2018b, AJ, 156, 58Baratella, M., D’Orazi, V., Carraro, G., et al. 2020, A&A, 634, A34Bergemann, M., Lind, K., Collet, R., Magic, Z., & Asplund, M. 2012, MNRAS,427, 27Bergemann, M. & Nordlander, T. 2014, arXiv e-prints, arXiv:1403.3088Bertelli Motta, C., Salaris, M., Pasquali, A., & Grebel, E. K. 2017, MNRAS,466, 2161Bessell, M. S., Castelli, F., & Plez, B. 1998, A&A, 333, 231Blanco-Cuaresma, S. & Fraix-Burnet, D. 2018, A&A, 618, A65Blanco-Cuaresma, S., Soubiran, C., Heiter, U., et al. 2015, A&A, 577, A47Blanco-Cuaresma, S., Soubiran, C., Heiter, U., & Jofré, P. 2014, A&A, 569,A111Bovy, J. 2016, ApJ, 817, 49Bragaglia, A. 2008, Mem. Soc. Astron. Italiana, 79, 365Bragaglia, A. & Tosi, M. 2006, AJ, 131, 1544Brandt, T. D. & Huang, C. X. 2015, ApJ, 807, 24Bravi, L., Zari, E., Sacco, G. G., et al. 2018, A&A, 615, A37Bressan, A., Marigo, P., Girardi, L., et al. 2012, MNRAS, 427, 127 Cantat-Gaudin, T. & Anders, F. 2020, A&A, 633, A99Cantat-Gaudin, T., Anders, F., Castro-Ginard, A., et al. 2020, A&A, 640, A1Cantat-Gaudin, T., Jordi, C., Vallenari, A., et al. 2018a, A&A, 618, A93Cantat-Gaudin, T., Vallenari, A., Sordo, R., et al. 2018b, A&A, 615, A49Cantat-Gaudin, T., Vallenari, A., Zaggia, S., et al. 2014, A&A, 569, A17Carrera, R., Pasquato, M., Vallenari, A., et al. 2019, A&A, 627, A119Carretta, E., Bragaglia, A., Gratton, R. G., & Tosi, M. 2005, A&A, 441, 131Casali, G., Magrini, L., Tognelli, E., et al. 2019, A&A, 629, A62Casali, G., Spina, L., Magrini, L., et al. 2020, A&A, 639, A127Casamiquela, L., Blanco-Cuaresma, S., Carrera, R., et al. 2019a, MNRAS, 490,1821Casamiquela, L., Blanco-Cuaresma, S., Carrera, R., et al. 2019b, MNRAS, 490,1821Casamiquela, L., Carrera, R., Balaguer-Núñez, L., et al. 2018, A&A, 610, A66Castelli, F. & Kurucz, R. L. 2003, in IAU Symposium, Vol. 210, Modelling ofStellar Atmospheres, ed. N. Piskunov, W. W. Weiss, & D. F. Gray, A20Castro-Ginard, A., Jordi, C., Luri, X., et al. 2020, A&A, 635, A45Castro-Ginard, A., Jordi, C., Luri, X., et al. 2018, A&A, 618, A59Chen, Y. Q. & Zhao, G. 2020, MNRAS, 495, 2673Chiappini, C., Anders, F., Rodrigues, T. S., et al. 2015, A&A, 576, L12Collet, R., Asplund, M., & Thévenin, F. 2005, A&A, 442, 643Collinder, P. 1931, Annals of the Observatory of Lund, 2, B1Cosentino, R., Lovis, C., Pepe, F., et al. 2014, Society of Photo-Optical In-strumentation Engineers (SPIE) Conference Series, Vol. 9147, HARPS-N @TNG, two year harvesting data: performances and results, 91478CDalton, G., Trager, S. C., Abrams, D. C., et al. 2012, Society of Photo-OpticalInstrumentation Engineers (SPIE) Conference Series, Vol. 8446, WEAVE:the next generation wide-field spectroscopy facility for the William HerschelTelescope, 84460Pde Jong, R. 2011, The Messenger, 145, 14De Silva, G. M., Freeman, K. C., Bland-Hawthorn, J., et al. 2015, MNRAS, 449,2604Donor, J., Frinchaboy, P. M., Cunha, K., et al. 2020, AJ, 159, 199D’Orazi, V., Oliva, E., Bragaglia, A., et al. 2020, A&A, 633, A38Franchini, M., Morossi, C., Di Marcantonio, P., et al. 2020, ApJ, 888, 55Frasca, A., Alonso-Santiago, J., Catanzaro, G., et al. 2019, A&A, 632, A16Frasca, A., Guillout, P., Marilli, E., et al. 2006, A&A, 454, 301Friel, E. D., Jacobson, H. R., Barrett, E., et al. 2003, AJ, 126, 2372Friel, E. D., Janes, K. A., Tavarez, M., et al. 2002, AJ, 124, 2693Gaia Collaboration, Brown, A. G. A., Vallenari, A., et al. 2018, A&A, 616, A1Gilmore, G., Randich, S., Asplund, M., et al. 2012, The Messenger, 147, 25Grevesse, N., Asplund, M., & Sauval, A. J. 2007, Space Sci. Rev., 130, 105Gustafsson, B., Edvardsson, B., Eriksson, K., et al. 2008, A&A, 486, 951Heiter, U., Jofré, P., Gustafsson, B., et al. 2015a, A&A, 582, A49
Article number, page 16 of 17asali, G. et al.: SPA-OCs
Heiter, U., Lind, K., Asplund, M., et al. 2015b, Phys. Scr, 90, 054010Jacobson, H. R., Friel, E. D., Jílková, L., et al. 2016, A&A, 591, A37Jofré, P., Heiter, U., Soubiran, C., et al. 2014, A&A, 564, A133Kaluzny, J. 1989, Acta Astron., 39, 13Kharchenko, N. V., Piskunov, A. E., Röser, S., Schilbach, E., & Scholz, R. D.2005, A&A, 438, 1163Kharchenko, N. V., Piskunov, A. E., Schilbach, E., Röser, S., & Scholz, R.-D.2013, A&A, 558, A53Kos, J., de Silva, G., Buder, S., et al. 2018, MNRAS, 480, 5242Kounkel, M. & Covey, K. 2019, AJ, 158, 122Krause, M. G. H., O ff ner, S. S. R., Charbonnel, C., et al. 2020, Space Sci. Rev.,216, 64Krumholz, M. R. & McKee, C. F. 2020, MNRAS, 494, 624Kuhn, M. A., Hillenbrand, L. A., Sills, A., Feigelson, E. D., & Getman, K. V.2019, ApJ, 870, 32Lagarde, N., Decressin, T., Charbonnel, C., et al. 2012, A&A, 543, A108Lagarde, N., Reylé, C., Robin, A. C., et al. 2019, A&A, 621, A24Li, C., Zhao, G., Zhai, M., & Jia, Y. 2018, ApJ, 860, 53Lind, K., Primas, F., Charbonnel, C., Grundahl, F., & Asplund, M. 2009, A&A,503, 545Lindegren, L., Hernández, J., Bombrun, A., et al. 2018, A&A, 616, A2Liu, F., Yong, D., Asplund, M., Ramírez, I., & Meléndez, J. 2016, MNRAS, 457,3934Liu, L. & Pang, X. 2019, ApJS, 245, 32Magrini, L., Randich, S., Donati, P., et al. 2015, A&A, 580, A85Magrini, L., Randich, S., Friel, E., et al. 2013, A&A, 558, A38Magrini, L., Randich, S., Kordopatis, G., et al. 2017, A&A, 603, A2Magrini, L., Randich, S., Romano, D., et al. 2014, A&A, 563, A44Magrini, L., Sestito, P., Randich, S., & Galli, D. 2009, A&A, 494, 95Magrini, L., Spina, L., Randich, S., et al. 2018, A&A, 617, A106Majewski, S. R., Schiavon, R. P., Frinchaboy, P. M., et al. 2017, AJ, 154, 94Mashonkina, L., Gehren, T., Shi, J. R., Korn, A. J., & Grupp, F. 2011, A&A, 528,A87Masseron, T. & Gilmore, G. 2015, MNRAS, 453, 1855Meingast, S., Alves, J., & Fürnkranz, V. 2019, A&A, 622, L13Monteiro, H. & Dias, W. S. 2019, MNRAS, 487, 2385Mucciarelli, A. & Bellazzini, M. 2020, Research Notes of the American Astro-nomical Society, 4, 52Mucciarelli, A., Salaris, M., Lovisi, L., et al. 2011, MNRAS, 412, 81Oliva, E., Origlia, L., Ba ff a, C., et al. 2006, Society of Photo-Optical Instru-mentation Engineers (SPIE) Conference Series, Vol. 6269, The GIANO-TNGspectrometer, 626919Origlia, L., Dalessandro, E., Sanna, N., et al. 2019, A&A, 629, A117Pakhomov, Y. V., Antipova, L. I., Boyarchuk, A. A., et al. 2009, AstronomyReports, 53, 660Perryman, M. A. C., de Boer, K. S., Gilmore, G., et al. 2001, A&A, 369, 339Piatti, A. E., Angelo, M. S., & Dias, W. S. 2019, MNRAS, 488, 4648Reddy, A. B. S., Lambert, D. L., & Giridhar, S. 2016, MNRAS, 463, 4366Röser, S. & Schilbach, E. 2019, A&A, 627, A4Ruprecht, J. 1966, Bulletin of the Astronomical Institutes of Czechoslovakia, 17,33Sacco, G. G., Randich, S., Bravi, L., Je ff ries, R., & Rigliaco, E. 2017a,Mem. Soc. Astron. Italiana, 88, 660Sacco, G. G., Spina, L., Randich, S., et al. 2017b, A&A, 601, A97Sagar, R. & Gri ffi ths, W. K. 1998, MNRAS, 299, 777Salaris, M., Pietrinferni, A., Piersimoni, A. M., & Cassisi, S. 2015, A&A, 583,A87Semenova, E., Bergemann, M., Deal, M., et al. 2020, arXiv e-prints,arXiv:2007.09153Short, C. I. & Hauschildt, P. H. 2003, ApJ, 596, 501Sim, G., Lee, S. H., Ann, H. B., & Kim, S. 2019, Journal of Korean AstronomicalSociety, 52, 145Smiljanic, R., Romano, D., Bragaglia, A., et al. 2016, A&A, 589, A115Sneden, C., Bean, J., Ivans, I., Lucatello, S., & Sobeck, J. 2012, MOOG: LTEline analysis and spectrum synthesisSpina, L., Meléndez, J., Casey, A. R., Karakas, A. I., & Tucci-Maia, M. 2018,ApJ, 863, 179Spina, L., Nordlander, T., Casey, A. R., et al. 2020, ApJ, 895, 52Spina, L., Randich, S., Magrini, L., et al. 2017, A&A, 601, A70Stassun, K. G., Oelkers, R. J., Pepper, J., et al. 2018, AJ, 156, 102Stetson, P. B. & Pancino, E. 2008, PASP, 120, 1332Tadross, A. L. 2003, New A, 8, 737Thorsbro, B., Ryde, N., Rich, R. M., et al. 2020, ApJ, 894, 26Tsantaki, M., Sousa, S. G., Adibekyan, V. Z., et al. 2013, A&A, 555, A150Warren, S. R. & Cole, A. A. 2009, MNRAS, 393, 272Woosley, S. E. & Weaver, T. A. 1995, ApJS, 101, 181Worley, C. C. & Cottrell, P. L. 2010, MNRAS, 406, 2504Worley, C. C., Cottrell, P. L., McDonald, I., & van Loon, J. T. 2010, MNRAS,402, 2060Yen, S. X., Re ff ert, S., Schilbach, E., et al. 2018, A&A, 615, A12Zhong, J., Chen, L., Wu, D., et al. 2020, arXiv e-prints, arXiv:2006.06929ert, S., Schilbach, E., et al. 2018, A&A, 615, A12Zhong, J., Chen, L., Wu, D., et al. 2020, arXiv e-prints, arXiv:2006.06929