Gemini/Phoenix H-band analysis of the globular cluster AL3
B. Barbuy, H. Ernandes, S.O. Souza, R. Razera, T. Moura, J. Meléndez, A. Pérez-Villegas, M. Zoccali, D. Minniti, B. Dias, S. Ortolani, E. Bica
AAstronomy & Astrophysics manuscript no. 39761corr © ESO 2021February 26, 2021
Gemini/Phoenix H-band analysis of the globular cluster AL 3 (cid:63)
B. Barbuy , H. Ernandes , , , S. O. Souza , R. Razera , T. Moura , J. Meléndez , A. Pérez-Villegas , M. Zoccali , , D.Minniti , , B. Dias , S. Ortolani , , and E. Bica Universidade de São Paulo, IAG, Rua do Matão 1226, Cidade Universitária, São Paulo 05508-900, Brazil UK Astronomy Technology Centre, Royal Observatory, Blackford Hill, Edinburgh, EH9 3HJ, UK IfA, University of Edinburgh, Royal Observatory, Blackford Hill, Edinburgh, EH9 3HJ, UK Instituto de Astrofísica, Pontificia Universidad Católica de Chile, Av. Vicuña Mackenna 4860, Santiago , Chile Millennium Institute of Astrophysics, Av. Vicuña Mackenna 4860, 782-0436 Macul, Santiago, Chile Departamento de Ciencias Fisicas, Facultad de Ciencias Exactas, Universidad Andres Bello, Av. Fernandez Concha 700, LasCondes, Santiago, Chile Vatican Observatory, V00120 Vatican City State, Italy Instituto de Alta Investigación, Sede Esmeralda, Universidad de Tarapacá, Av. Luis Emilio Recabarren 2477, Iquique, Chile Università di Padova, Dipartimento di Fisica e Astronomia, Vicolo dell’Osservatorio 2, I-35122 Padova, Italy INAF-Osservatorio Astronomico di Padova, Vicolo dell’Osservatorio 5, I-35122 Padova, Italy Universidade Federal do Rio Grande do Sul, Departamento de Astronomia, CP 15051, Porto Alegre 91501-970, BrazilReceived; accepted
ABSTRACT
Context.
The globular cluster AL 3 is old and located in the inner bulge. Three individual stars were observed with the Phoenixspectrograph at the Gemini South telescope. The wavelength region contains prominent lines of CN, OH, and CO, allowing thederivation of C, N, and O abundances of cool stars.
Aims.
We aim to derive C, N, O abundances of three stars in the bulge globular cluster AL 3, and additionally in stars of NGC 6558and HP 1. The spectra of AL 3 allows us to derive the cluster’s radial velocity.
Methods.
For AL 3, we applied a new code to analyse its colour-magnitude diagram. Synthetic spectra were computed and comparedto observed spectra for the three clusters.
Results.
We present a detailed identification of lines in the spectral region centred at 15555 Å, covering the wavelength range 15525-15590 Å. C, N, and O abundances are tentatively derived for the sample stars.
Key words.
Stars: Abundances, Atmospheres – Galaxy Bulge – Globular clusters: individual: AL 3, NGC 6558, HP 1
1. Introduction
The globular clusters AL 3, NGC 6558, and HP 1 share the char-acteristics of having a metallicity of [Fe / H] ∼ -1.0 and of being lo-cated in the Galactic bulge. They are old and could represent theearliest stellar populations in the Galaxy (Ortolani et al. 2006;Barbuy et al. 2018a; Kerber et al. 2019).The star cluster AL 3 was discovered by Andrews & Lindsay(1967) and was also cataloged as BH 261 by van den Bergh &Hagen (1975), reported as a faint open cluster. It is reported inthe ESO / Uppsala catalogue (Lauberts 1982) as ESO 456-SC78.Ortolani et al. (2006) showed that the star cluster shows B, V,I colour-magnitude diagrams (CMD) typical of a globular clus-ter. It is centred at J2000 α = h m s , δ = − o (cid:48) (cid:48)(cid:48) ,with Galactic coordinates l = o , b = − o , and located at6 ◦ .
25 and 2 kpc from the Galactic centre, hence in the inner bulgevolume. The cluster has a depleted red giant branch (RGB), sim-ilarly to low-mass Palomar clusters, indicating it to have beenstripped along its lifetime. This cluster has not been further ob-served so far.
Send o ff print requests to : B. Barbuy (cid:63) Observations collected at the Gemini Observatory, Proposals GS-2006A-C9 and GS-2008A-Q-23-5, and at the European Southern Ob-servatory (ESO), Proposal 64L-0212(A).
The NGC 6558 cluster is located in a window, identifiedby Blanco (1988), with equatorial coordinates (J2000) α = h m s , δ = -31 ◦
45’ 49" and Galactic coordinates l = ◦ , b = -6.025 ◦ . It was analysed in terms of CMD by Rich et al. (1998).Rossi et al. (2015) obtained a proper-motion-cleaned CMD andpresented a proper motion analysis, from which a study of itsorbits was given in Pérez-Villegas et al. (2018, 2020).The globular cluster Cl Haute-Provence 1 or HP 1, also des-ignated BH 229 and ESO 455-SC11, was discovered by Du-fay et al. (1954). It is located at J2000 α = h m s , δ = − o (cid:48) (cid:48)(cid:48) , with Galactic coordinates l = o , b = o .In the present work, we studied individual stars of these clus-ters in a limited region of the spectrum in the H-band corre-sponding to the wavelength region of the Phoenix spectrographat the Gemini South telescope, centred at 15555 Å, and covering15520-15590 Å, with a high spectral resolution of R ∼ Article number, page 1 of 15 a r X i v : . [ a s t r o - ph . S R ] F e b & A proofs: manuscript no. 39761corr
2. Spectroscopy in the H-band: Atomic andmolecular lines
The H-band will be intensely observed in the near future,given the new instruments placing emphasis on the near-infraredregion, such as the James Webb Space Telescope (JWST),and new spectrographs on ground-based telescopes such asMOONS@VLT (presently CRIRES@VLT is available) andMOSAIC@ELT. The project APOGEE (Apache Point Obser-vatory Galactic Evolution Experiment), with observations at aresolution of R ∼ µ Leo, and created ashortened version of a line list, containing only detectable lines.Meléndez & Barbuy (1999, hereafter MB99) worked on a listof atomic lines in the J and H bands. The list of lines mostly cor-responded to the detectable lines. That previous line list neededto be largely completed. Upon checking the lines detectable inthe wavelength range 15520-16000 Å, this was done by veri-fying the line lists from APOGEE (Shetrone et al. 2015) andVALD (Piskunov et al. 1995, Ryabchikova et al. 2015). We notethat astrophysical oscillator log gf strengths were applied to theAPOGEE line list, and these should be suitable for abundancederivation. Through a line-by-line check of its detectability inthe Arcturus spectrum, we identified lines of Mg i , Si i , Ca i , Ti i ,Mn i , and Ni i , and we were not able to find detectable lines fromthe species C i , O i , Sc i , V i , Cr i , Co i , Cu i , Y i , and Y ii . Thespectra computed including all lines of all these elements areentirely equivalent to the one computed with the shortened linelist, therefore to ensure practicality when identifying which linesreally contribute to a feature, we created a table containing thedetectable lines only. In this table, available upon request, we re-port the oscillator strengths from MB99, APOGEE, VALD, andadopted values, where in order of preference we adopted NISTand MB99.Molecular electronic transition lines of CN A Π -X Σ , andvibration-rotation CO X Σ + , OH X Π lines were included in thesynthetic spectra calculations. The line lists for CN were madeavailable by S. P. Davis, the CO line lists were adopted fromGoorvitch (1994), and the OH line list was made available by S.P. Davis and A. Goldman (Goldman et al. 1998). For more de-tails on CN, CO, and OH molecular lines, we refer the reader toMeléndez & Barbuy (1999), Meléndez et al. (2001), and Melén-dez et al. (2002). TiO φ -system b Π -d Σ lines are also presentin the region. The line list by Jorgensen (1994) is included inthe calculations as described in Schiavon & Barbuy (1999) andBarbuy et al. (2018a). The adopted dissociation potential of OHis 4.392 eV, D = = Table 1.
Log book of observations. Proposals GS-2006A-C9 on 15-16 / / / / / / (2018b), and it is available together with the atomic and molec-ular line lists. We identified the lines in the reference stars Arcturus and µ Leo. For the reference star Arcturus, the spectrum atlas fromHinkle et al. (1995) is used, and for the metal-rich reference giantstar µ Leo, APOGEE spectra are used, and their studies are bepresented elsewhere.The adopted stellar parameters for Arcturus and µ Leo arefrom Meléndez et al. (2003) and Lecureur et al. (2007), and theyare reported in Sect. 5.
3. Observations
The spectra of red giant stars of the bulge globular clustersNGC 6558, AL 3, and HP 1 were observed with the Phoenixspectrograph installed on the 8m Gemini South telescope. Theprogram was tri-national from Brazil (PI: B. Barbuy), Chile (PI:M. Zoccali), and Australia (PI: J. Meléndez).The final suitable spectra include three stars of AL 3, twostars of NGC 6558, and one star of HP 1, centred at 1.555 µ min the H band. Another three stars in NGC 6558, and three inHP 1 were also observed; however, these showed a low Signal-to-Noise (S / N), due to clouds or high airmass. The log of obser-vations is provided in Table 1.This is the first spectroscopic work on AL 3, except for Gaiadata (Gaia Collaboration 2020). In Fig. 1, a 3 min B exposure ofAL 3 is shown for a field extraction of 3.3’ × ×
510 pix-els). The sample stars of AL 3 are identified in the chart. Chartsand identifications of the observed stars in NGC 6558 and HP 1are given in Barbuy et al. (2009, 2018) and Barbuy et al. (2006,2016), respectively.
In order to verify the corresponding membership probability ofobserved stars in AL 3, we performed the cross-match with GaiaEarly Data Release 3 (EDR3; Gaia Collaboration 2020). We se-lected stars within 20’ from the cluster centre and used the renor-malised unit weight error (RUWE) ≤ . http: // trevisanj.github.io / PFANTArticle number, page 2 of 15. Barbuy et al.: Phoenix analysis of the globular cluster AL 3
Fig. 1.
AL 3: 3 min B image with the three sample stars identified.Extraction of 3’ × Having the high-precision EDR3 proper motions ( µ ∗ α = µ α cos δ and µ δ ), we obtain the mean proper motions for the clus-ter of µ ∗ α = . ± .
03 mas yr − and µ δ = − . ± .
04 mas yr − .These values are compatible with those given in Baumgardt et al.(2019). We also computed the Gaussian membership probabilitydistribution of AL 3. We found that the stars AL3-6 and AL3-7have membership probabilities of 100%. Finally, the star AL3-3has a relatively low membership probability of ∼ We were able to derive radial velocities for the sample stars. Weused the low S / N individual observations of each star (S / N ∼ / N ∼ / N ∼ / NiI 15,55.25blend, CN 15,555.25, and FeI 15,591.49. We also used the OHsky lines, as listed in Table 2 by Meléndez et al. (2003). Thesefeatures were used for AL3-3, giving a radial velocity of -67.65 ± − . In the combined spectrum of AL3-7, the same fea-tures result in a radial velocity of -68.93 ± − . The cor-responding heliocentric velocities of -57.29 km.s − and -58.57km.s − lead to a final mean heliocentric velocity of -57.93 km.s − ± ff erent heliocentric radial velocity, as shown in Fig.4, of -29.57 ± − , compatible with the value given byBaumgardt et al. (2019) of -29.38 ± − . The derived radial velocity is of crucial importance for thecomputation of the cluster’s orbits - see Sect. 5. However, weobtained two di ff erent figures: v hel r = -57.93 and -29.57 km.s − .In Fig. 4, we show the spectrum of AL3-6 compared with that ofAL3-7. Therefore, we obtained two di ff erent radial velocities forAL 3, and consequently we considered both of them for the cal-culation of orbits. Since we had already computed the orbits forthe lower value (from Baumgardt et al. 2019), we show the or-bits for the higher value here. Clearly, new observations of thesestars in a more extended wavelength range and with a higher S / Nwould be of great interest.
4. Colour-magnitude diagram of AL 3
Ortolani et al. (2006) presented B , V , and Cousins I images ofAL 3, observed on 2000 March 6 using the 1.54m Danish tele-scope at the European Southern Observatory (ESO) at La Silla.They derived a reddening of E(B-V) = . ± .
03 and a distanceof 6 . ± . / H] = − . ± . ∼ . α / Fe] = − σ . We obtain a reddening of E ( B − V ) = . ± .
04, a distance of d (cid:12) = . ± . / H] = − . ± .
18. Our age determinationindicates an old age of 13 . + . − . Gyr, indicating that AL 3 is an-other relic fossil. It is important to stress that our distance of 6.0kpc is also in agreement, within 1 − σ , with the value of 6.5 kpcfrom Harris (1996, 2010), Rossi et al. (2015), and Baumgardt etal. (2019).Finally, we note that the dispersion of the data could bedue to di ff erential reddening, together with contamination andblends. The reddening of AL 3 is relatively high, and the dif-ferential reddening is certainly an issue, as it is in other bulgeglobular clusters with similar reddening. We expect an amountof about 20% of di ff erential reddening. In principle it can becorrected, but the standard procedures for di ff erential reddeningcorrection in this cluster cannot be applied due to the few bonafide reference stars in the CMDs.Figure 5 shows the solution of isochrone fitting. The solidblue line represents the median solution, while the shaded re-gions indicate the solutions within 1 − σ . The red stars are thethree sample stars. Finally, Figure A.1 exhibits the corner plotsshowing the (anti)correlations between the parameters. Article number, page 3 of 15 & A proofs: manuscript no. 39761corr
Table 2.
Gaia magnitudes, proper motions and membership probability.Star α (J2000) δ (J2000) G G RP µ ∗ α µ δ Membdeg deg mag mag mas yr − mas yr − %AL3-3 273.5288067 -28.6357960 14 . ± .
003 13 . ± .
004 3 . ± . − . ± .
07 59AL3-6 273.5247404 -28.6346067 15 . ± .
003 14 . ± .
011 3 . ± . − . ± .
04 100AL3-7 273.5220767 -28.6380356 13 . ± .
003 13 . ± .
005 3 . ± . − . ± .
03 100 ( Å ) F l u x -67.65 ± AL33
Fig. 2.
AL 3-3: Radial velocity derivation. The solid black line is the observed spectrum, the solid grey line is the noise spectrum, the solid redline is the synthetic spectrum, the dashed red lines are those used to derive the radial velocity, and the dashed black lines are the OH sky lines. ( Å ) F l u x -68.93 ± AL37
Fig. 3.
AL 3-7: Radial velocity derivation. The solid black line is the observed spectrum, the solid grey line is the noise spectrum, the solid redline is the synthetic spectrum, the dashed red lines are those used to derive the radial velocity, and the dashed black lines are the OH sky lines.
5. Stellar parameters
Individual stars of NGC 6558 were analysed with high-resolution spectroscopy by Barbuy et al. (2007, 2018b) and withmoderate-resolution spectroscopy by Dias et al. (2015). Thestars NGC6558-42 and NGC6558-64 are studied here.Similar studies of HP 1 were carried out in Barbuy et al.(2006, 2016) and Dias et al. (2016). In the 2006 article, the brightred giants were labelled with numbers 1 to 6, for the purpose ofidentifying them in the cluster chart. In 2016, we adopted theidentification numbers corresponding to the photometric reduc-tions relative to observations obtained at the New Technology Telescope (NTT) at ESO, in 1994, as described in Ortolani et al.(1997). HP1-4 and HP1-5 are stars 2115 and 2939 in Barbuy etal. (2016). HP1-2 is the same as in Barbuy et al. (2006). In ourstudy, we only analysed HP1-5.
The magnitudes and colours as follows are indicated in Table 3:B, V from Ortolani et al. (2006), V, I from Rossi et al. (2015),
Article number, page 4 of 15. Barbuy et al.: Phoenix analysis of the globular cluster AL 3 ( Å ) F l u x -29.57 ± AL36
Fig. 4.
AL 3-6: Radial velocity derivation. The solid black line is the observed spectrum, the solid grey line is the spectrum of star AL3-7, thedashed red lines are those used to derive the radial velocity, and the dashed black lines are the OH sky lines.
Fig. 5.
AL 3 V vs. V − I CMD. The black dots are the stars within 120pixel from the cluster centre (see Ortolani et al. 2006). The red stars arethe observed stars of the present work. The solid blue line represents themedian solution of the isochrone fitting, while the blue region revealsthe solutions within 1 − σ . JHK from the 2MASS catalogue (Skrutskie et al. 2006), andJHK from the VVV survey (Saito et al. 2012). E ff ective temperatures were initially derived from B − V , V − I , V − K , and J − K using the colour-temperature cali-brations of Alonso et al. (1999). V,I Cousins were transformedto V,I Johnson using ( V − I ) C = V − I ) J (Bessell 1979). http : // ipac . caltech . edu / / releases / allsky / ; https : // irsa . iapc . caltech . edu horus . roe . ac . uk / vsa The J , H , K S magnitudes and colours were transformed from the2MASS system to California Institute of Technology (CIT), andfrom this to Telescopio Carlos Sánchez (TCS), using the rela-tions established by Carpenter (2001) and Alonso et al. (1998).The conversion of JHK VVV colours to the JHK 2MASS systemwas done using relations by Soto et al. (2013).The temperatures resulting from photometry are of the or-der of 5000 K for the three stars. These temperatures, however,are not compatible with another indicator, which is the Hydro-gen Brackett 16 line, centred at 15556.457 Å. A fit of this linefor both AL 3 stars was carried out iteratively, after deriving theirCNO abundances. The resulting temperatures, adopted in the fol-lowing analysis, are 4250 K and 4500 K for AL3-3 and AL3-7,respectively. The fits to the hydrogen line are shown in Figure 6.For AL3-6, the low quality of the spectrum does not allow the fitof the hydrogen line, in particular due to strong telluric absorp-tions in the region. It appears to be cooler and compatible with4150 K. This incompatibly between photometric and hydrogen-wing-derived temperatures is a main source of uncertainty in thepresent study.To derive the gravity, we used the PARSEC isochrones (Bres-san et al. 2012). To inspect the isochrones, we adopted a metal-licity of [Fe / H] = − .
0, or overall metallicity Z = . E ( B − V ) = .
36 (Ortolani et al. 2006, and present results),leading to E ( V − I ) = .
478 and A V = .
12, we transformedthe apparent magnitudes to absolute magnitudes, as well as thecolours ( V − I corr = V − I - E ( V − I )), and we identified the corre-spondence of the observed stars to the theoretical isochrone.The metallicity resulting from the CMD fitting is [Fe / H] = -1.34, which was imposed as a prior. We inspected individuallines of Fe in the AL3-3 spectrum and the fits are more com-patible with [Fe / H] = -1.0. There is also the evidence from othersimilar bulge globular clusters such as NGC 6558, NGC 6522,HP 1, and Terzan 9, which are found to have [Fe / H] ∼ -1.0 fromhigh-resolution spectroscopy. Bica (2016) showed that there is apeak in metallicity at [Fe / H] ∼ -1.0 in the bulge, which we alsoadopted for AL 3. An isochrone fitting with this higher metallic-ity was tried, but appeared di ffi cult to converge. This is a secondsource of uncertainty of the present study. Final adopted stellarparameters for program stars, and of the reference stars Arcturus http: // stev.oapd.inaf.it / cgi-bin / cmd Article number, page 5 of 15 & A proofs: manuscript no. 39761corr (Meléndez et al. 2003) and µ Leo (Lecureur et al. 2007), are re-ported in Table 4.
6. CNO abundances
The atmospheric models were interpolated in the grid of modelsby Gustafsson et al. (2008). The synthetic spectra were com-puted employing the PFANT code described in Barbuy et al.(2018b). In order to derive the C, N, O abundances, we fittedthe CN, OH, and CO lines iteratively.
The cool red giant, NGC6558-42, shows strong lines and is atypical red giant. For this reason, we show the fits to the spectrumof this star in detail in Figures 8 and 9.The star NGC6558-64 instead, which would have an e ff ec-tive temperature of 4850 K according to the analysis from opti-cal spectra by Barbuy et al. (2018b), could be as hot as 5500 K.This is seen from the profile of its Hydrogen Brackett 16 line;however, this should be taken with caution due to defects in theobserved spectrum. For this reason, we could not converge onCNO abundances for this star. There is a clear contrast between the spectra of HP1-5, AL3-3,and AL3-7, that have shallow lines, and AL3-6 and NGC6558-42, that show strong lines. Whereas NGC6558-42 is a typical redgiant, the stars AL3-3, AL3-7, and HP1-5 show weak molecularlines. From the location of stars AL3-3 and AL3-7 in the CMDof Fig. 5, they should be AGB stars. In Fig. 10, we show theobserved spectrum in the selected wavelength regions contain-ing CN, OH lines, and the CO bandhead at 15578 Å for starsAL3-3, AL3-7, and HP1-5. The molecular lines are very shal-low, due to a combination of warm temperatures and low metal-licities. Clearly, the CNO abundances derived for these stars areless reliable than for the cool star NGC6558-F42. Their CNOabundances are compatible with being close to solar, but giventhe shallowness of the lines, it is clear that the molecular linesare not reliable for abundance measurements.AL3-6 instead shows very strong CNO lines. Figure 7 in-dicates that [C / Fe] = + / Fe] =+ / Fe] =+ ff erent renormalisations to illustrate thedi ffi culty in analysing this spectrum. Additionally, the computa-tions with two di ff erent carbon abundances illustrate the extremesensitivity of the lines. Clearly, however, there is an urgent needto observe this star in the optical and / or in a more extended wave-length region in the H-band to obtain firm conclusions on theCNO abundances of AL 3. The main uncertainty in the derivation of CNO abundances inAL 3 stems from the e ff ective temperatures. Adopting a colourexcess E(B-V) = ∼ = e ff = ff ective temperature errors of ±
250 K for AL3-6, and ±
150 K for AL3-3 and AL3-7, errors of ± ± In order to investigate whether AL 3 belongs to the bulge, thickdisc, or halo, given its solar CNO abundances, we carried outthe calculations of orbits for the cluster. We employed a Galacticmodel that includes an exponential disc made by the superpo-sition of three Miyamoto-Nagai potentials (Smith et al. 2015),a dark matter halo with a Navarro-Frenk-White (NFW) densityprofile (Navarro, Frenk & White 1997), and a triaxial Ferrersbar. The total mass of the bar is 1 . × M (cid:12) , an angle of 25 ◦ with the Sun-major axis of the bar, and a major axis extensionof 3.5 kpc. We assumed three pattern speeds of the bar Ω b = − kpc − . We kept the same bar extension, eventhough we changed the bar pattern speed. Our Galactic modelhas a circular velocity V =
241 km s − at R = . U , V , W ) (cid:12) = (11 . , . , .
25) km s − (Schönrich, Binney &Dehnen 2010).To take into account the e ff ects of the observational uncer-tainties associated with the cluster’s parameters, we generateda set of of 1000 initial conditions using a Monte Carlo methodconsidering the errors of distance, heliocentric radial velocity,and absolute proper motion components. With such initial con-ditions, we integrated the orbits forward for 10 Gyr using theNIGO tool (Rossi 2015).In Fig. 11, we show the probability density map of the or-bits of AL 3 in the x − y and R − z projections co-rotating withthe bar, using the two adopted distances, 6.0 kpc (Ortolani et al.2006) and 6.5 kpc (Baumgardt et al. 2019). The yellow colourdisplays the space region that the orbits of AL 3 cross more fre-quently. The black curves are the orbits using the central val-ues of the observational parameters. In Figure 12, we show his-tograms relating to the number of probable orbits as a function ofpericentric distance (r min ), apocentric distance (r max ), maximumheight above the plane ( | Z | max ), and eccentricity ( e ). In Table 5,we give the median values of the orbital parameters, and the er-rors provided in each column are derived considering the 16thand 84th percentiles of the distribution. The orbital parametersare similar for the three pattern speeds. From the apocentric dis-tance, we can see that AL 3 is mostly confined within ∼ ∼ ∼ ∼ min ), apoc-entric distance (r max ), maximum height above the plane ( | Z | max ),and eccentricity. If a distance of 6.5 kpc is adopted, then it isclear that AL 3 is a very central cluster, with a maximum heightreaching at most Z ∼ . Article number, page 6 of 15. Barbuy et al.: Phoenix analysis of the globular cluster AL 3
Table 3.
AL 3: coordinates, magnitudes, and colours of sample stars.Star 2MASS α (J2000) δ (J2000) V I J H K J H K2MASS VVVB11 18101902-3144506 18 10 19.01 -31 44 50.64 15.902 14.275 13.136 12.440 12.280 13.017 12.397 12.248B64 18101803-3145435 18 10 18.03 -31 45 43.55 15.703 14.180 13.064 12.456 12.277 13.055 12.529 12.384B73 18102150-3145268 18 10 21.50 -31 45 26.77 15.709 14.187 13.128 12.449 12.316 13.047 13.047 12.313F42 — 18 10 17.65 -31 45 38.93 16.054 14.442 — — — — — —F97 18101520-3146014 18 10 15.21 -31 46 00.67 16.037 14.467 13.183 12.481 12.338 — 12.503 12.378HP1-2 17310585-2958354 17 31 05.60 -29 58 34.00 16.982 14.332 12.210 11.268 10.969 14.588 13.675 13.368HP1-4 17310538-2959199 17 31 05.30 -29 59 20.00 17.070 14.281 — 11.67 — 11.258 11.392 10.688HP1-5 17310729-2959021 17 31 07.20 -29 59 02.00 17.131 14.395 11.901 10.869 10.595 12.021 11.285 10.898AL3-3 18140691-2839087 18 14 06.90 -28 38 09.0 14.524 13.204 12.214 11.631 11.469 12.211 11.714 11.544AL3-6 18140592-2838049 18 14 05.80 -28 38 06.0 15.563 14.203 12.763 12.272 12.256 — — —AL3-7 18140529-2838168 18 14 05.30 -28 38 19.0 14.313 12.963 11.878 11.296 11.170 11.920 11.472 11.195 Table 4.
Adopted stellar parameters for individual stars in NGC 6558,AL 3, and HP 1, and resulting C, N, O abundances. For NGC 6558, thestellar parameters are from Barbuy et al. (2007), and for HP 1 they arefrom Barbuy et al. (2016). Stellar parameters for the Sun, Arcturus, and µ Leo are also included.
Name T e ff log g [Fe / H] v t [C / Fe][N / Fe][O / Fe](K) km / sProgram starsF42 3800 0.5 − − + + − − − − + + + − − + + + + − + + + µ Leo 4540 2.3 + − + + ∼
60 % probability of belongingto the bulge and ∼
40% to the disc. With this distance, the proba-bility of being part of the disc increases significantly, and maybethis result could be more consistent with the solar CNO abun-dances.
7. Conclusions
We analysed spectra of individual stars of the globular clustersAL 3, NGC 6558, and HP 1, obtained with the PHOENIX spec-trograph at the Gemini South telescope. With a high spectralresolution of R ∼ ffi -cult to use atomic lines to deduce the stellar parameters e ff ec-tive temperature, gravity, and metallicity. For this reason, we obtained the e ff ective temperature from the Hydrogen Brack-ett 16 line, gravity from photometric data, and isochrones. Themetallicity [Fe / H] ∼ -1.3 ± / H] = -1.0 for the analysed stars, due to spectro-scopic evidence. For NGC 6558 and HP 1, the stellar parame-ters were adopted from previous analyses from optical spectra(Barbuy et al. 2007, 2018), and Barbuy et al. (2006, 2016) re-spectively. Adopting these stellar parameters, we computed thesynthetic spectra in order to derive the abundances of C, N, andO. Since they vary interdependently, the fit was done iteratively,where particular attention was given to the CO bandhead. Thestars analysed in NGC 6558 and HP 1 show typical CNO abun-dances of red giants, and confirm previous oxygen abundancederivation.AL 3 is a more complex case: two stars analysed in AL 3show solar CNO abundance ratios, but based on very shallowlines, and the location of these two stars in the CMD point tothem being AGB stars. The star AL3-6 shows instead very strongCNO abundances of the order of [C / Fe] =+ / Fe] =+ / Fe] =+ Acknowledgements.
BB, HE, RR, JM, and EB acknowledge grants fromFAPESP, CNPq and CAPES - Financial code 001. TM acknowledges FAPESPpostdoctoral fellowship no. 2018 / / Article number, page 7 of 15 & A proofs: manuscript no. 39761corr
AL3-3Teff=4250,4500,4750,5000 KAL3-3Teff=4250 K AL3-7Teff=4250,4500,4750,5000 KAL3-7Teff=4500 K
Fig. 6.
AL 3-3 and AL 3-7: Hydrogen Brackett 16 line computed for T ef f = Project ICN12_009, awarded to the Millenium Institute of Astrophysics (MAS).S.O. acknowledges the partial support of the research program DOR1901029,2019, and the project BIRD191235, 2019 of the University of Padova.
References
Alonso, A., Arribas, S., Martínez-Roger, C. 1998, A&AS, 131, 209Alonso, A., Arribas, S., Martínez-Roger, C. 1999, A&AS, 140, 261 (AAM99)Andrews, A.D., Lindsay, E.M. 1967, IrAJ, 8, 126Barbuy, B., Zoccali, M., Ortolani, S. et al. 2006, A&A, 449, 349 Barbuy, B., Zoccali, M., Ortolani, S. et al. 2007, AJ, 134, 1613Barbuy, B., Chiappini, C., Cantelli, E. 2014, A&A, 570, A76Barbuy, B., Friaça, A.C.S., da Silveira, C.R. et al. 2015, A&A, 580, A40Barbuy, B., Cantelli, E., Vemado, A. et al. 2016, A&A, 591, A53Barbuy, B., Chiappini, C., Gerhard, O. 2018a, ARA&A, 56, 223Barbuy, B., Trevisan, J., de Almeida, A. 2018b, PASA, 35, 46Barbuy, B., Muniz, L., Ortolani, S. et al. 2018c, A&A, 619, A178Bessell, M.S. 1979, PASP, 91, 589Bica, E., Ortolani, S., Barbuy, B. 2016, PASA, 33, 28Blanco, V.M. 1988, AJ, 95, 1400Bland-Hawthorn, J., & Gerhard, O. 2016, ARA&A, 54, 529Bressan, A., Marigo, P., Girardi, L. 2012, MNRAS, 427, 127
Article number, page 8 of 15. Barbuy et al.: Phoenix analysis of the globular cluster AL 3
Table 5.
Median orbital parameters and membership probability of AL 3. Ω b (cid:104) r min (cid:105) (cid:104) r max (cid:105) (cid:104)| z | max (cid:105) (cid:104) e (cid:105) P bulge P disc (km s − kpc − ) (kpc) (kpc) (kpc) (%) (%) d (cid:12) = . . + . − . . + . − . . + . − . . + . − . . + . − . . + . − . . + . − . . + . − . . + . − . . + . − . . + . − . . + . − . d (cid:12) = . . + . − . . + . − . . + . − . . + . − . . + . − . . + . − . . + . − . . + . − . . + . − . . + . − . . + . − . . + . − . Fig. 7.
AL 3-6: Tentative fit to this low S / N spectrum, showing that it iscompatible with a high CNO abundance. Blue and black lines: Sameobserved spectrum normalized in two di ff erent ways. Red and cyanlines: Synthetic spectra computed with [C / Fe] = / Fe] =+ / Fe] =+ Carpenter, J.M. 2001, AJ, 121, 2851Dias, B., Barbuy, B., Saviane, I. et al. 2015, A&A, 573, A13Dias, B., Barbuy, B., Saviane, I. et al. 2016, A&A, 590, A9Dotter, A., Chaboyer, B., Jevremovi´c, D. et al. 2008, ApJS, 178, 89Dufay, J., Berthier, P., Morignat, B. 1954, Publ. of the OHP, 3, No. 17Gratton, R.G., Sneden, C. 1990, A&A, 234, 366Gravity Collaboration, Abuter, R., Amorim, A., et al. 2019, A&A, 625, L10Grevesse, N. & Sauval, J.N. 1998, SSRev, 35, 161Grevesse, N., Noels, A., Sauval, J. 1996, ASP, 99, p.117Grevesse, N., Scott, P., Asplund, M., Sauval, A.J. 2015, A&A, 573, 25Goldman, A., Schoenfeld, W.G., Goorvitch, D. et al. 1998, JQSRT, 59, 453Goorvitch, D. 1994, ApJS, 95, 535Gustafsson, B., Edvardsson, B., Eriksson, K., Jørgensen, U. G., Nordlund, Å.;Plez, B. 2008, A&A, 486, 951Harris, W.E. 1996, AJ, 112, 1487Hinkle, K., Wallace, L., Livingston, W. 1995, PASP, 107, 1042Huber, K.P., Herzberg, G. 1979, Constants of diatomic molecules, Springer,Boston, MA.Jorgensen, U.G. 1994, A&A, 284, 179Kurucz, R. 1993, CD-ROM 23Lauberts, A. 1982, ESO / Uppsala Survey of ESO(B) Atlas, ESO, GarchingLecureur, A., Hill, V., Zoccali, M., et al. 2007, A&A, 465, 799 Livingston, W., Wallace, L. 1991, An atlas of the solar spectrum in the infraredfrom 1850 to 9000 cm-1 (1.1 to 6.5 micrometer), NSO Technical Report,NOAO, TucsonLodders, K., Palme, H., Gail, H.-P. 2009, Landolt-Börnstein - Group VI As-tronomy and Astrophysics Numerical Data and Functional Relationships inScience and Technology Volume 4B: Solar System. Edited by J.E. Trümper,2009, 4.4., 44Majewski, S.R., Schiavon, R.P., Frinchaboy, P.M. 2017, AJ, 154, 94Martin, W.C., Fuhr, J.R., Kelleher, D.E., et al. 2002, NIST Atomic Database(version 2.0), http: // physics.nist.gov / asd. National Institute of Standards andTechnology, Gaithersburg, MD.McWilliam, A., Wallerstein, G., Mottini, M. 2013, 778, 149Meléndez, J., Barbuy, B. 1999, ApJS, 124, 527Meléndez, J., Barbuy, B., Spite, F. 2001, ApJ, 556, 858Meléndez, J., Barbuy, B. 2002, ApJ, 575, 474Meléndez, J., Barbuy, B., Bica, E. et al. 2003, A&A, 411, 417Navarro, J. F., Frenk, C. S., & White, S. D. M. 1997, ApJ, 490, 493Ortolani, S., Bica, E., Barbuy, B. 1997, MNRAS, 284, 692Ortolani, S., Bica, E., Barbuy, B. 2006, ApJ, 646, L115Pérez-Villegas, A., Rossi, L., Ortolani, S., et al. 2018, PASA, 35, 21Pérez-Villegas, A., Barbuy, B., Kerber, L. et al. 2020, MNRAS, 491, 3251Pietrinferni, A., Cassisi, S., Salaris, M., Castelli, F. 2006, ApJ, 642, 797Piskunov, N., Kupka, F., Ryabchikova, T., Weiss, W., Je ff ery, C., 1995, A&AS,112, 525Ramírez, I., Allende Prieto, C. 2011, ApJ, 743, 136Ryabchikova, T., Piskunov, N., Kurucz, R.L. et al. 2015, PhyS, 90, 054005Rich, R.M., Ortolani, S., Bica, E., Barbuy, B. 1998, AJ, 116, 1295Rossi, L., Ortolani, S., Bica, E., Barbuy, B., Bonfanti, A. 2015, MNRAS, 450,3270Rossi, L. J. 2015, Astronomy and Computing, 12, 11Saito, R. et al. (VVV collaboration) 2012, A&A, 537, 107Schiavon, R.P., Barbuy, B. 1999, ApJ, 510, 934Schönrich, R., Binney, J., & Dehnen, W. 2010, MNRAS, 403, 1829Souza, S.O., Kerber, L.O., Barbuy, B. et al. 2020, ApJ, 890, 38Scott, P., Grevesse, N., Asplund, M. et al. 2015a, A&A, 573, 26Scott, P., Asplund, M., Grevesse, N. et al. 2015b, A&A, 573, 27Shetrone, M., Bizyaev, D., Lawler, J.E. et al. 2015, ApJS, 221, 24Skrutskie, M.F., Cutri, R.M., Stiening, R. et al. 2006, AJ, 131, 1163Smith & Ruck 2000, A&A, 356, 570SSmith, V. Cunha, K., Shetrone, M.D., et al. 2013, ApJ, 765, 16Smith, R., Flynn, C., Candlish, G. N., Fellhauer, M., & Gibson, B. K. 2015,MNRAS, 448, 2934Soto, M., Barbá, R., Gunthardt, G., Minniti, D., Lucas, P. et al. 2013, A&A, 552,A101Ste ff en, M., Prakapavicius, D., Ca ff au, E. et al. 2015, A&A, 583, A57van den Bergh, S., Hagen, G.L. 1975, AJ, 80, 11 Article number, page 9 of 15 & A proofs: manuscript no. 39761corr
Fig. 8.
NGC 6558-42: Line identification in the range 15527-15555 Å. Dashed line: Observed spectrum. Solid red line: Synthetic spectrum.Synthetic spectrum computed with [C / Fe] = -0.5, [N / Fe] = / Fe] =+ Fig. 9.
NGC 6558-42: Same as Fig. 8, in the range 15555-15587 Å. Article number, page 11 of 15 & A proofs: manuscript no. 39761corr
Fig. 10.
HP 1-5, AL 3-3, and AL 3-7: Spectrum in selected wavelength regions containing CN, OH lines and the CO bandhead. Synthetic spectraare computed for the CNO abundances given in Table 4.Article number, page 12 of 15. Barbuy et al.: Phoenix analysis of the globular cluster AL 3
Fig. 11.
Probability density map for x − y and R − z projections of the set of orbits for AL 3 for distances of 6.0 kpc (left panels) and 6.5 kpc (rightpanels), using three di ff erent values of Ω b = , , and 50 km s − kpc − . The orbits are co-rotating with the bar frame. Yellow corresponds to thelarger probabilities. The black lines show the orbits using the central values. Article number, page 13 of 15 & A proofs: manuscript no. 39761corr r min [kpc] n o r b b = 40 b = 45 b = 50 r max [kpc] |z| max [kpc] e d = 6.0 kpc r min [kpc] n o r b b = 40 b = 45 b = 50 r max [kpc] |z| max [kpc] e d = 6.5 kpc Fig. 12.
Distribution of orbital parameters for AL 3, for distances of 6.0 kpc (top panels) and 6.5 kpc (bottom panels), with pericentric distance r min , apocentric distance r max , maximum vertical excursion from the Galactic plane | z | max , and eccentricity e . The colours show the di ff erent patternspeed of the bar, Ω b =
40 (blue), 45 (orange), and 50 (green) km s − kpc − .Article number, page 14 of 15. Barbuy et al.: Phoenix analysis of the globular cluster AL 3 Appendix A: Isochrone fitting - corner plot
The
SIRIUS code performs the isochrone fitting through the con-struction of the Markov chain Monte Carlo (MCMC) algorithms.The purpose of the MCMC method is to obtain the posterior dis-tribution of each parameter via the generation of random sam-ples. Figure A.1 shows the corner plots, that is, a graphical rep-resentation of the 4D parameter space of the isochrone fitting(age, reddening, distance, and metallicity). The histograms rep-resent the cumulative best solutions (posterior distribution) foreach parameter, while the 2D density maps show the correla-tions between the parameters as well as the region with the high-est density. To represent the posterior distributions of each pa- rameter, we adopt the region of highest density as the centralvalue and the uncertainties calculated from the 16th and 84thpercentiles.
Fig. A.1.
Corner plots representing the 4D parameter space of theMarkov chain constructed through the Monte Carlo algorithm for theisochrone fitting. To represent the solutions, we adopt the peak of thedistribution as the best value, and the uncertainties are computed usingthe 16 th and 84 thth