Stellar population astrophysics (SPA) with the TNG -- The Arcturus Lab
C. Fanelli, L. Origlia, E. Oliva, A. Mucciarelli, N. Sanna, E. Dalessandro, D. Romano
AAstronomy & Astrophysics manuscript no. 39397corr © ESO 2020November 26, 2020
Stellar population astrophysics (SPA) with the TNG (cid:63)
The Arcturus Lab
C. Fanelli , , L. Origlia , E. Oliva , A. Mucciarelli , , N. Sanna , E. Dalessandro , and D. Romano Dipartimento di Fisica e Astronomia, Università degli Studi di Bologna, via Piero Gobetti 93 /
2, 40129, Bologna, Italy, e-mail: [email protected] INAF-Osservatorio di Astrofisica e Scienza dello Spazio, via Piero Gobetti 93 /
3, 40129, Bologna, Italy INAF–Osservatorio Astrofisico di Arcetri, Largo Enrico Fermi 5, 50125 Firenze, Italy
ABSTRACT
Context.
High-resolution spectroscopy in the near-infrared (NIR) is a powerful tool for characterising the physical and chemicalproperties of cool-star atmospheres. The current generation of NIR echelle spectrographs enables the sampling of many spectralfeatures over the full 0.9-2.4 µ m range for a detailed chemical tagging. Aims.
Within the Stellar Population Astrophysics Large Program at the TNG, we used a high-resolution (R = Methods.
We inspected several hundred NIR atomic and molecular lines to derive abundances of 26 di ff erent chemical species,including CNO, iron-group, alpha, Z-odd, and neutron-capture elements. We then performed a similar analysis in the optical usingArcturus VLT-UVES spectra. Results.
Through the combined NIR and optical analysis we defined a new thermometer and a new gravitometer for giant stars,based on the comparison of carbon (for the thermometer) and oxygen (for the gravitometer) abundances, as derived from atomic andmolecular lines. We then derived self-consistent stellar parameters and chemical abundances of Arcturus over the full 4800 − ff ected by deviations from thermal equilibrium and / or chromospheric activity, as traced by the observed variability of He I at10830 Å. Key words.
Techniques: spectroscopic - stars: abundances - individual stars: Arcturus - stars: late-type
1. Introduction
The enhanced sensitivity of IR observations to intrinsically red(i.e. cool) and / or reddened (by dust extinction) objects makenear-IR (NIR) spectrographs the ideal instrument for studyingthe physics, chemistry, and kinematics of cool giant and super-giant stars in galaxy fields as well as in star clusters.Cool giant and supergiant stars are among the brightest pop-ulations in any stellar systems and are easily observable at IRwavelengths out to large distances. They are also easy to detectin heavily reddened environments, such as the inner disk andbulge regions, where observations in the visual range are pro-hibitive. These stars are important tracers of the star formationand chemical enrichment history of their hosts.High-resolution spectroscopy of these stars is crucial to ob-tain an exhaustive description of their detailed chemistry andnucleosynthesis. Di ff erent chemical elements are synthesised instars with di ff erent initial masses and thus released into the in- (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 theSpanish Observatorio del Roque de los Muchachos of the Instituto deAstrofisica de Canarias. This study is part of the Large Program titled
SPA - Stellar Population Astrophysics: the detailed, age-resolved chem-istry of the Milky Way disk (PI: L. Origlia), granted observing time withHARPS-N and GIANO-B echelle spectrographs at the TNG. terstellar medium with di ff erent time delays with respect to theonset of star formation. The detailed chemical tagging of key el-ements is therefore crucial to constrain formation and chemicalenrichment scenarios of the Milky Way and other nearby stellarsystem, in which these stars can be individually resolved.In the past two decades high-resolution NIR spectroscopyhas experienced a burst of activity in terms of newly commis-sioned spectrographs and stellar surveys. However, the preciseidentification and characterisation of the optimal atomic andmolecular lines for abundance analysis, as well as their mod-elling over the entire NIR range, is still work in progress.To this purpose, high-resolution spectroscopy in both vi-sual and NIR spectral ranges of suitable chemical calibrators ismandatory. Arcturus is such a calibrator for giant stars, and wepresent a comprehensive study in the YJHK NIR bands using theechelle spectrum at R = α Boo , HR HIP HD ff erentialchemical analysis of giant stars relative to Arcturus can indeedlargely minimise most of systematic errors due to atmospheric Article number, page 1 of 16 a r X i v : . [ a s t r o - ph . S R ] N ov & A proofs: manuscript no. 39397corr
Table 1.
Stellar parameters and metallicity of Arcturus inferred from di ff erent optical and NIR studies. T e f f log ( g ) ξ [ Fe / H ] Range Res Re f erencesK dex kms − dex Å λ/ ∆ λ ±
39 1 .
55 1 . ± . − . ± .
07 5370 − ∗ Fulbright et al. (2006)4290 1 .
55 1 . − .
50 5000 − ∗∗ Ryde et al. (2009)4286 ±
30 1 . ± .
05 1 . − . ± .
02 5000 − ∗∗ Ramírez & Allende Prieto (2011)4275 ±
50 1 . ± .
10 1 . ± . − . ± .
04 15100 − ∗∗ Smith et al. (2013)4286 ±
50 1 . ± .
10 1 . ± . − . ± .
04 15100 − ∗∗ Shetrone et al. (2015)4286 ±
35 1 . ± .
06 1 . ± . − . ± .
06 9300 − ∗∗∗ Kondo et al. (2019) a ±
35 1 . ± .
06 1 . ± . − . ± .
04 9300 − ∗∗∗ Kondo et al. (2019) b Notes. ( a ) Using VALD3 line list: http: // vald.astro.uu.se ( b ) Using Meléndez & Barbuy (1999) line list ( * ) Spectrum observed with the 0 . ( ** ) Spectrum from Hinkle & Wallace (2005) ( *** ) Spectrum observed with WINERED mounted at the 1 . parameters (e.g. McWilliam & Rich 1994; Worley et al. 2009;Alves-Brito et al. 2010; Ramírez & Allende Prieto 2011).However, it is challenging to take a spectrum of Arcturusbecause of its apparent ultra-bright luminosity. Most of thechemical studies of Arcturus are based on high-resolution spec-troscopy. The reference Arcturus spectrum covering the entirespectral range from the UV to the IR is the one made available byHinkle & Wallace (2005, and references therein). This spectrumhas been built using three di ff erent instruments: the Space Tele-scope Imaging Spectrograph (STIS) mounted on board of theHubble Space Telescope in the 1000-3000 Å range, the echelleoptical spectrograph in the 3100-9000 Å range, and the Fouriertransform spectrometer in the 0 . µ m range mounted at the KittPeak National Observatory (KPNO) 4 m telescope.Ryde et al. (2009) used a portion of the Hinkle & Wallace(2005) H-band spectrum, from 15326 to 15705 Å, to study sev-eral clean molecular lines of CO, CN, and OH and derive C, N,and O abundances. Ramírez & Allende Prieto (2011) providedatmospheric parameters and abundances for several metals bymostly using the Hinkle & Wallace (2005) optical spectrum andthe line list by Asplund et al. (2009). Smith et al. (2013) andShetrone et al. (2015) used the H-band portion of the Hinkle &Wallace (2005) spectrum and their detailed line list prepared forthe Sloan Digital Sky Survey III Apache Point Galactic Evolu-tion Experiment (APOGEE) (Majewski et al. 2007) to providestellar parameters, Fe, CNO and other elemental abundances.Arcturus was also studied by Fulbright et al. (2006) usingan optical spectrum taken with the Hamilton spectrograph at the0 .
6m CAT telescope of the Lick Observatory. Kondo et al. (2019)have analysed a ZYJ spectrum of Arcturus at R (cid:39) . ff erent line lists: the Vi-enna Atomic Line Database (VALD3) (Ryabchikova & Pakho-mov 2015), and the public line list provided by Meléndez & Bar-buy (1999).Finally, we mention the works by Maas et al. (2017); D’Oraziet al. (2011) and Overbeek et al. (2016), who discussed the Arc-turus abundances of P, Y, and Dy, respectively. Table 1 lists thestellar parameters and metallicity [Fe / H] of Arcturus inferredfrom di ff erent optical and NIR studies.This paper is organised as follows. In Section 2 we describethe observation and the data reduction of the Arcturus spectrum.In Section 3 we discuss the method we adopted for spectral anal- ysis, and in Section 4 we describe the procedure we used to de-termine the stellar parameters for Arcturus using new NIR diag-nostics. In Section 5 we report the results of our chemical anal-ysis in the optical and NIR range, and in Section 6 we comparethem with those from previous studies and draw our conclusions.
2. Observations and data reduction
Arcturus was observed on July 2, 2018, with GIANO-B, thehigh-resolution (R = / / ∼ . (cid:48)(cid:48) .The raw spectra were reduced using the data reductionpipeline software GOFIO (Rainer et al. 2018), which processescalibration (darks, flats, and U-Ne lamps taken in daytime) andscientific frames. The main feature of the GOFIO data reduc-tion is the optimal spectral extraction and wavelength calibrationbased on a physical model of the spectrometer that accuratelymatches instrumental e ff ects such as variable slit tilt and ordercurvature over the echellogram (Oliva et al. 2018). The data re-duction package also includes bad pixel and cosmic removal, skyand dark subtraction, flat-field and blaze correction.The spectrum was corrected for telluric absorption using thespectra of an O-type standard star taken at di ff erent air massesduring the same night. The normalised spectra of the telluricstandard taken at low and high air-mass values were combinedwith di ff erent weights to match the depth of the telluric lines inthe Arcturus spectrum. Figs. A.1, A.2, A.3, and A.4 in the ap-pendix show the rest-frame normalised spectra corrected for tel-luric absorption. The average signal-to-noise ratio of the reducedand telluric-corrected spectrum is about 150 per pixel.We also analysed two optical spectra of Arcturus retrievedfrom the ESO archive, in order to cross-check chemical abun-dances over the widest possible spectral range. These opticalspectra were collected with the high-resolution spectrographUVES at the ESO Very Large Telescope (VLT) at a resolution Article number, page 2 of 16. Fanelli et al.: Stellar population astrophysics (SPA) with the TNG of R ∼
3. Spectral analysis
Accurate and precise stellar parameters and chemical abun-dances of Arcturus were determined by means of spectralsynthesis technique applied to the observed spectra. Syn-thetic spectra were computed by using the radiative transfercode
TURBOSPECTRUM (Alvarez & Plez 1998; Plez 2012) withMARCS models atmospheres (Gustafsson et al. 2008), theatomic data from VALD3 and the most updated molecular datafrom the website of B. Plez, . The synthetic spectra were convoluted with aGaussian function in order to reproduce the observed broaderprofile that corresponds to an equivalent resolution of 32 , ≈ kms − ) because the projected rotational velocity ofArcturus is negligible ( ξ rot = . kms − , Gray 1981).For the abundance analysis, we used a selected list of C I,Na I, Mg I, Al I, Si I, P I, S I, K I, Ca I, Sc I, Ti I, V I, Cr I, Mn I,Fe I, Fe II, Co I, Ni I, Cu I, Zn I, Y I; Y II, Ce II, Nd II, andDy II atomic lines and CO, OH, CN, and HF molecular lines.Each line was carefully checked against possible blending withclose contaminants. For this purpose, we developed a code called TurboSLine that identifies as potential contaminants any atomicor molecular j-th line whose centroid λ j is within one full widthat half maximum (FWHM) from the centroid λ i of the analyzedi-th line.For each of these potential contaminants, we computed the the-oretical line equivalent width (EW) and the amount of contami-nation using the following approximation: C j = EW j × (cid:18) − | λ i − λ j | FWHM j (cid:19) . (1)If (cid:80) C j > . × EW i , the i-th line was classified as blended, andit was not normally used for abundance analysis.Some other lines were later on rejected because they are con-taminated by the wings of nearby strong photospheric and / or bydeep telluric lines by visual inspection. In our chemical analy-sis we also rejected strong lines because of the uncertainty in themodelling of their wings, non-local thermal equilibrium (NLTE),and chromospheric e ff ects, etc.Tables A.1-A.4 provide the complete list of optical and NIRatomic and molecular lines used for the abundance analysis. Forthe computation of the chemical abundances we used SALVADOR ,a tool developed by A. Mucciarelli (priv. comm.), which per-forms a χ minimisation between observed and synthetic spectrawhile the normalisation of the observed spectrum around eachline is optimised interactively.As a further check, we also computed the line EWs and de-rived the corresponding abundances. The latter were found to bepractically coincident with those obtained from the spectral syn-thesis, with eventually only a slightly higher dispersion, likelybecause lines with some impurity provide slightly more uncer-tain abundances when the EW method is used. Tables A.1-A.4 are only available in electronic form at the CDS viaanonymous ftp to cdsarc.u-strasbg.fr (130.79.128.5) or via http://cdsweb.u-strasbg.fr/cgi-bin/qcat?J/A+A/
4. Stellar parameters
Arcturus belongs to a kinematic group of several dozen old stars(Eggen 1971). Its metallicity ([Fe / H] (cid:39) − . v t = √ ( U + V + W ) =
106 km s − and its location in the Toomre diagram (see Fig. 1)suggest that it is a thick disk star (Bensby et al. 2014).
200 150 100 50 0 50
V (km/s) ( U + W ) / ( k m / s ) ArcturusThick diskThin diskHalo
Fig. 1.
Toomre diagram for thin-disk (blue dots), thick-disk (red dots),and halo (black crosses) stars from Reddy et al. 2003 and Reddy et al.2006 and for the disk stars of Bensby et al. 2014 (grey dots). The blackstar marks the position of Arcturus, and the dashed lines delineate con-stant total space velocities with respect to the LSR of V tot =
85 and 180km s − , respectively (e.g. Nissen 2004, and references therein). Previous determinations of the Arcturus stellar parame-ters (cf. Table 1) suggested temperatures in the 4275-4290 Krange, log(g) in the 1.55-1.70 dex range, and microturbulencein the 1.20-1.85 kms − range. Using the Dartmouth web-toolat http: // stellar.dartmouth.edu (Dotter et al. 2008), we computedisochrones with [Fe / H] = − .
5, [ α / Fe] =+ .
20 and di ff erent ages.At the bolometric luminosity of Arcturus ( L bol = ± L (cid:12) ,Smith et al. (2013), see also Ramírez & Allende Prieto (2011)),we found that old ( ≥
10 Gyr) ages are consistent with an e ff ectivetemperature T e f f in the 4260-4310 K range and gravity log(g) inthe 1.60-1.70 dex range.We used these photometric ranges for T e f f and log(g) to alsoconstrain the microturbulence velocity range with the standardapproach of minimising the slope between the iron abundanceand the reduced EW of the measured lines log ( EW /λ ) (see thediscussion in Mucciarelli 2011). We find microturbulence ve-locities in the 1.50 and 1.70 kms − range. We finally adoptedthe value of ξ = . ± . kms − that best minimizes anytrend between the abundances and the reduced EW of about 400iron lines distributed over the 4800 − e f f and log(g) were fine-tuned using the spectroscopic data. Article number, page 3 of 16 & A proofs: manuscript no. 39397corr A ( F e ) log(EW/ ) Fig. 2.
Iron abundances as a function of the reduced EW for all themeasured optical (blue circles) lines in the UVES spectra and NIR (redcircles) lines in the GIANO-B spectrum. The dotted line marks the best-fit median abundance.
We defined a new powerful diagnostic tool to derive T e f f in oxygen-rich cool stars, based on the balance between thecarbon abundance, derived from atomic lines (see Table 2), andmolecular CO roto-vibration transitions. Hereafter, we refer tothis method as the C-thermometer. Table 2.
Measurable atomic carbon lines in the NIR spectrum of Arc-turus. λ air χ log(gf) ( a ) NoteÅ eV8727 .
140 1 . − .
165 Forbidden9658 .
435 7 . − .
280 NLTE10683 .
080 7 . + .
079 NLTE10685 .
340 7 . − .
272 NLTE10691 .
245 7 . + .
344 NLTE10707 .
320 7 . − .
411 NLTE10729 .
529 7 . − .
420 NLTE11748 .
220 8 . + .
375 NLTE11753 .
320 8 . + .
691 NLTE11754 .
760 8 . + .
542 NLTE17234 .
463 9 . + .
293 LTE17448 .
535 9 . + .
012 LTE17672 .
039 7 . − .
974 LTE17768 .
910 9 . + .
420 LTE17793 .
158 9 . − .
045 LTE
Notes. ( a ) For all the tabulated lines we used the NIST log(gf) values as listedin the VALD3 database.
The basic principle of this thermometer follows from thevery high dissociation potential of the CO molecule (11.1 eV).Thus the CO / C relative abundance of carbon in molecular (CO)and atomic (C ) form has a very strong exponential dependenceon temperature. Like for other diatomic molecules, the abun-dance ratio also depends linearly on the gas pressure (i.e. ongravity); but this e ff ect is far weaker. Figure 3 shows the be-haviour of the Saha equilibrium abundances for gas temperaturesand pressures relevant for this work.When the spectrum is modelled, CO and C I lines must pro-vide the same abundances. If this does not happen, it is becausethe model temperature is incorrect and must be tuned until thetwo abundances match. However, the analysis is complicated Fig. 3.
Selected pressure values are representative of the line-formingregions for surface gravities log(g) = by the fact that some C I lines may show significant departurefrom LTE (see e.g. Fabbian et al. 2006; Takeda & Takada-Hidai2013), while CO lines form under LTE conditions (e.g. Hin-kle & Lambert 1975; Ryde et al. 2009). For example, Takeda& Takada-Hidai (2013) showed that the 10683 / / ∆ A ( C ) = − .
23 dex.NLTE corrections are sensitive to the adopted stellar model anddepend on temperature and gravity, hence they cannot be safelyused to define a reliable thermometer. It is therefore desirable touse only those C I lines that are not a ff ected by NLTE e ff ects. T eff (K) A ( C ) Best-fit CIBest-fit COCI lines in the H band[C I]
Fig. 4.
C abundances as a function of temperature from the C I linesin the H band (green circles), and the forbidden [C I] line at 8727 .
14 Å(empty triangle). The shaded green region is the corresponding best-fit ± σ trend of the C I lines, while the shaded red region is the best-fit ± σ trend of the CO lines. The large black dot marks the intersectionof the two curves, and its x-coordinate provides the best-fit temperature.Article number, page 4 of 16. Fanelli et al.: Stellar population astrophysics (SPA) with the TNG A ( F e ) (eV) Fig. 5.
Iron abundances from neutral lines in the optical UVES spec-tra (blue circles) and in the NIR GIANO-B spectrum (red circles) as afunction of their excitation potential. The dotted line marks the derivedbest-fit median iron abundance.
For this purpose, given that forbidden lines do not su ff er fromNLTE (e.g. Alexeeva & Mashonkina 2015), we used the [CI] at8727 .
14 Å measurable in our UVES red spectrum to derive aproxy of the atomic carbon abundance in LTE. Then, we com-puted LTE abundances for all the C I lines measurable in ourGIANO-B spectrum (Table 2), and we checked their abundancesagainst that from the [C I] line. We found that the C I lines in theH band with excitation potentials above 9 eV provide similar Cabundances, and the C I lines in the Y and J bands with lower ex-citation potentials give LTE abundances that are systematically( ∼ ff ects.We therefore used only the C I lines in the H band for the C-thermometer.Fig. 4 shows the variation in carbon abundance from C I andCO lines in the Arcturus GIANO-B spectrum as a function of T e f f . Both diagnostics are very sensitive to T e f f , but have op-posite trends. The two curves intersect at T e f f = ±
33 K.The quoted error of ±
33 K corresponds to a ± σ variation inthe derived C abundances from CI and CO. The C-thermometeris virtually independent of the other parameters within the un-certainties. Variations of ± . kms − in microturbulence veloc-ity or ± .
06 dex in log(g) have a negligible e ff ect on temperature( ≤
15 K) and C abundance ( ≤ ξ (up to 0.2 kms − ) does thecorresponding variation in temperature lie within the error. In-terestingly, the C-thermometer also works if the molecular car-bon abundance is derived from spectral synthesis of the ∆ v = ±
33 K derived from theC-thermometer also allows us to minimise any trend betweeniron abundances from neutral lines and their excitation poten-tial within the errors, as shown in Fig. 5, which is the standardspectroscopic method for inferring the e ff ective temperature. The relative abundance of OH and atomic oxygen ( O ) dependslinearly on the gas pressure (i.e. on gravity), while it has a weakdependence on temperature (see Figure 3) because of the lowdissociation potential of the OH molecule (4.4 eV). Therefore,the OH / O ratio can be used to estimate the gravity when thetemperature is constrained.After fixing the temperature at T e f f = A ( O ) [O I]OH A ( F e ) FeIFeII
Fig. 6.
Top panel : best-fit trends of Oxygen abundances from the [OI] line (blue) and from the OH lines (red) with gravity. The blackdot marks the intersection of the two curves, and its x-coordinate pro-vides the best-fit log(g) = ± . Bottom panel : best-fit trends of Ironabundances from Fe II (shaded) and Fe I lines (red) with gravity. Theblack dot marks the intersection of the two curves, and its x-coordinateprovides the best-fit log(g) = . ± . at 6300 . = . ± .
06 dex. Thequoted error of ± ± σ variation in thederived O abundances from [OI] and OH.Variations of ±
33 K in temperature a ff ect gravity by ± ± . kms − in microturbulence veloc-ity a ff ect log(g) by ∓ .
02 dex. Using the standard method forinferring spectroscopic log(g), that is, minimising the di ff erencebetween the iron abundances from neutral and ionised opticallines, we obtained a very similar best-fit log(g) = . ± .
10 dex(see Fig. 6, bottom panel).
5. Chemical analysis
The adopted stellar parameters for the chemical analysis of Arc-turus are summarised in Table 3. Abundance errors from the un-
Table 3.
Stellar parameters for Arcturus
Paramater Value ErrorT e f f .
67 dex 0 .
06 dex ξ . kms − . kms − [ Fe / H ] ∗ − .
57 dex 0 .
01 dex
Notes. ( * ) We used the solar A(Fe) (cid:12) = certainties in the stellar parameters were estimated by comput-ing elemental abundances with varying T e f f by ±
33 K, log(g) by ± .
06 dex, and ξ by ± . kms − (see Table 3). Article number, page 5 of 16 & A proofs: manuscript no. 39397corr T a b l e . A r c t u r u s c h e m i ca l a bund a n ce s a nd a ss o c i a t e d m ea s u r e m e n t e rr o r s fr o m N I R a ndop ti ca lli n e s . N I R O P T O P T + N I R X Z l og ( N ) a [ X / H ][ X / H ][ X / F e ][ X / F e ] (cid:15) l og ( N ) a [ X / H ][ X / H ][ X / F e ][ X / F e ] (cid:15) l og ( N ) a [ X / H ][ X / H ][ X / F e ][ X / F e ] (cid:15) li n e s G r e A s p l G r e A s p l li n e s G r e A s p l G r e A s p l G r e A s p l G r e A s p l C b . − . − . + . + . . . − . − . + . + . . . − . − . + . + . . N b . − . − . + . + . . O b . − . − . + . + . . . − . − . + . + . . . − . − . + . + . . F b . − . − . + . + . . F e I . − . − . + . + . . . − . − . + . + . . . − . − . + . + . . F e II . − . − . + . + . . . − . − . + . + . . . − . − . + . + . . V I . − . − . + . + . . . − . − . + . + . . . − . − . + . + . . C rI . − . − . − . − . . . − . − . − . − . . . − . − . − . − . . M n I . − . − . + . − . . . − . − . + . + . . . − . − . + . + . . C o I . − . − . + . + . . . − . − . + . + . . . − . − . + . + . . N i I . − . − . + . + . . . − . − . + . + . . . − . − . + . + . . C u I . − . − . + . + . . . − . − . + . + . . . − . − . + . + . . Z n I . − . − . + . + . . . − . − . + . + . . . − . − . + . + . . M g I . − . − . + . + . . . − . − . + . + . . . − . − . + . + . . S i I . − . − . + . + . . . − . − . + . + . . . − . − . + . + . . S I . − . − . + . + . . . − . − . + . + . . . − . − . + . + . . C a I . − . − . + . + . . . − . − . + . + . . . − . − . + . + . . T i I . − . − . + . + . . . − . − . + . + . . . − . − . + . + . . T i II . − . − . + . + . . N a I . − . − . + . + . . . − . − . + . + . . . − . − . + . + . . A l I . − . − . + . + . . . − . − . + . + . . . − . − . + . + . . P I . − . − . + . + . . K I . − . − . + . + . . . − . − . + . + . . . − . − . + . + . . S c I . − . − . + . + . . . − . − . + . . . . − . − . + . + . . S c II . − . − . + . + . . Y I . − . − . − . − . . Y II . − . − . − . − . . . − . − . − . − . . . − . − . + . + . . C e II . − . − . + . + . . . − . − . + . + . . . − . − . + . + . . N d II . − . − . − . − . . . − . − . − . − . . . − . − . − . − . . D y II . − . − . + . + . . . − . − . + . + . . . − . − . + . + . . N o t e s . ( a ) l og ( N ) = l og ( N X / N H ) + . ( b ) C a bund a n ce h a s b ee nd e r i v e dby c o m b i n i ng t h e m ea s u r e m e n t o f C I a t o m i c li n e s a nd14 C O m o l ec u l a r li n e s . N , O , a nd F a bund a n ce s h a v e b ee nd e r i v e d fr o m t h e m ea s u r e m e n t o f m o l ec u l a r li n e s on l y . Article number, page 6 of 16. Fanelli et al.: Stellar population astrophysics (SPA) with the TNG
On average, these errors amount to a few hundredths of adex at most. The only notable exception is the higher sensitivityof OH and HF lines to the e ff ective temperature: a variation of ±
33 K indeed implies an error in the derivation of oxygen andfluorine abundances of ± ± ff ect of the abundances of the main electron donors on thosederived from ionised species. We estimate that these e ff ects nor-mally yield errors below 0.1 dex (see also e.g. Ryde et al. 2009).Measurement errors include uncertainty in the continuumpositioning and photon noise. For elements with more thantwo measurable lines, we computed the dispersion around themean abundance, while for those with one or two measurablelines, we computed the dispersion from a Monte Carlo sim-ulation, taking into account an error in the measured EW of ≈ ± FWHM / signal-to-noise ratio for a line FWHM sampledwith 2-3 pixels.The measurement errors (cid:15) quoted in Table 4 are the σ dispersiondivided by the square root of the number of lines.In the course of the chemical analysis we checked a few prob-lematic lines against possible NLTE e ff ects using the online webtool http://nlte.mpia.de/gui-siuAC_secE.php , M. Ko-valev et al. 2018.Abundances and corresponding measurement errors for all thesampled chemical elements in the log(N X / N H ) +
12 and in the[X / H] solar scales, adopting as solar reference both Grevesse &Sauval (1998) (Gre98) and Asplund et al. (2009) (Aspl09), arelisted in Table 4.In Fig. 5.2 we compare the derived abundances from NIR andoptical lines for all the measured elements.
In our chemical analysis of Arcturus we first computed the abun-dances of CNO and then those of the other elements. CNO arethe most abundant metals, and in red giant and supergiant starspectra the many molecular CO, CN, and OH lines are the mostimportant potential contaminants. Following Ryde et al. (2009)and Smith et al. (2013), we adopted an iterative method to de-rive CNO abundances in order to consider the interplay amongthese three elements in setting the molecular equilibrium. Theresulting abundances are listed in first three lines of Table 4.Most interestingly, the carbon abundance derived from the ∆ v = ∆ v = C using two ∆ v = ∆ v = CO. Thisyielded a C / C isotopic abundance ratio of 7 ±
1, consistentwith the e ff ects of the second dredge-up coupled to some addi-tional mixing in low-mass giants (see e.g. Charbonnel & Lagarde2010).The fluorine abundance was derived from one HF molecularline at 23358 .
33 Å in the K band using the transition parametersof Jönsson et al. 2014.
The iron abundance was derived for more than 400 lines of Fe Iand a dozen Fe II lines. Fully consistent values were obtainedfrom all these line sets, with a σ dispersion of about 0.1 dex. InFig. 7 we show the inferred iron abundances from the measurediron lines over the full 4800-24500 spectral range as a functionof the line central wavelength. We adopted the average value of6 . ± .
01 (see Table 4) as the reference iron abundance of Arc-turus. However, it is interesting to note that many of the lines inthe YJ bands provide abundances that are systematically lower(on average by 0.04 dex) than the reference abundance. We in-spected some of these lines for possible NLTE e ff ects and foundindeed that some positive corrections of ≤ (Å) A ( F e ) Fig. 7.
Iron abundances for all the measured optical and NIR lines asa function of the line central wavelength. The dashed line marks theaverage abundance.
For manganese we used the few measurable lines both in theoptical and NIR. In particular, in the NIR we used the J-bandlines at 13218 .
49 Å and 13415 .
64 Å that need an NLTE correc-tion of ∆ A ( Mn ) = − .
16 and ∆ A ( Mn ) = − .
04, respectively,and two lines in the H band with HFS. For cobalt we used afew dozen optical lines with HFS and a few NIR lines. The lat-ter show a larger scatter that can be explained with small NLTEe ff ects. For example, we applied an NLTE correction of + . .
64 Å line. For copper wewere only able use one optical and one NIR line. The NIR line at16005 .
75 Å has HFS and needs to be used with caution becauseit might be blended. For zinc we used one line in the optical andtwo in the NIR. The NIR line at 11054 .
28 Å is partially blendedwith CN, but the line at 13053 .
63 Å is free of contamination.Altogether, the iron-peak elements show fully consistent op-tical and NIR abundances. They homogeneously scale as iron,with the possible exception of copper and zinc, which areslightly enhanced. α -capture elements Dozens of unblended lines of Si I, Ca I, Ti I, Mg I, and S Iare available in the NIR spectrum of Arcturus for an abundanceanalysis. The NIR Mg I and S I lines are known to experienceNLTE e ff ects (Zhang et al. 2017; Takeda et al. 2016). However,at the metallicity, temperature, and gravity of Arcturus, the cor-rections are negligible. Sulphur also shows a forbidden line [S I] Article number, page 7 of 16 & A proofs: manuscript no. 39397corr C N O F N a M g A l S i P S K C a S c T i V C r M n F e C o N i C u Z n Y C e N d D y [ X / F e ] NIR NeutralOPT NeutralNIR ionizedOPT ionized
Atomic Number
Fig. 8.
Derived [X / Fe] chemical abundances for Arcturus from the GIANO-B NIR spectrum (red symbols) and UVES optical spectra (bluesymbols). The circles indicate the neutral species, and the triangles indicate the ionised species. Error bars are from Table 4. at 10821 Å that provides a fully consistent abundance with theone derived from the selected S I lines. This further proves thatNLTE e ff ects are negligible.The inferred NIR abundances for these alpha elements arefully consistent with the optical ones, as detailed in Table 4 andshown in Fig. 5.2.The derived abundances of Mg and Si, and to a lesser extent, ofTi, S, and Ca, suggest some [ α / Fe] enhancement, as for oxygen.This is typical of thick-disk stars.
A few optical and NIR lines of Na, K, and Sc can be safely usedto derive reliable abundances. We found consistent optical andNIR solar-scaled abundances of Na and Sc and some enhancedK abundance. However, the K lines can show significant NLTEe ff ects with negative corrections to the LTE abundances (see e.g.Zhang et al. 2006; Osorio et al. 2020), thus implying a lower[K / Fe] relative abundance, about solar or even subsolar scaled.Two NIR lines of phosphorus at 10529 .
52 Å and 10581 .
58 Å arealso measurable, giving an abundance of 5 . ± .
08. A third lineat 10596 .
90 Å gives a unexpected higher abundance (see alsoMaas et al. 2017), probably blended because its profile is clearlyasymmetric. We therefore rejected it.For aluminium, ten optical lines and four NIR lines at10782 .
05 Å, 10768 .
37 Å, 10872 .
97 Å , and 10891 .
77 Å in the Yband with small (if any) NLTE corrections provide homoge-neous abundances that are higher by almost a factor of twothan the solar-scaled value. For the NIR lines we used log(gf)from NIST, which is slightly di ff erent from the lines adopted inVALD3.The strong lines at 13123 .
41 Å and 13150 .
75 Å and the K-band line at at 21163 .
76 Å have HFS, show significant NLTEe ff ects (Nordlander & Lind 2017), and require a negative abun-dance correction of 0 . − .
30 dex. Although when these lines are corrected for NLTE, they provideAl abundances that are reasonably consistent with those of the Yband and optical lines, we did not use them.The three strong lines at 16718 .
96 Å, 16750 .
56 Å , and16763 .
36 Å also show NLTE e ff ects and have HFS. The line at16750 .
56 Å also has strong and blended wings.In Arcturus-like stars, the abundances derived from these linescan be quite uncertain, therefore we did not use them (seeSect. 5.6).
We measured NIR lines for five neutron-capture elements: yt-trium (mostly an s-process element), cerium (an s-process ele-ment), neodimium (mostly an s-process element), and dyspro-sium (an r-process element). The NIR Ce II and Nd II lineswere identified for the first time by Cunha et al. (2017) and Has-selquist et al. (2016), respectively.One neutral and five ionised lines of yttrium were measuredin the optical, but only one ionised line is measured in the NIR Yband (see also Matsunaga et al. 2020). We find that the Y abun-dance is slightly depleted with respect to the solar scale value,in agreement with the disk chemistry at the Arcturus metallicity(see e.g. Reddy et al. 2006; Bensby et al. 2014).We finally used a few optical and NIR-ionised lines ofcerium, neodymium, and dysprosium, and we derived aboutsolar-scaled Ce and Nd abundances and slightly enhanced Dywith respect to the solar-scaled value.
We realised that some strong lines in the GIANO-B spectrum aredeeper than the corresponding lines in the FTS winter and sum-mer spectra of Arcturus by Hinkle et al. (2000). A few examplesare shown in Fig. 9.
Article number, page 8 of 16. Fanelli et al.: Stellar population astrophysics (SPA) with the TNG
Fig. 9.
Chromospheric He I line in the Y band and a few strong pho-tospheric Al, Si, and Mg lines in the in the GIANO-B spectrum (toppanel) and in the winter, January (middle panel) and summer, June (bot-tom panel) FTS spectra by Hinkle & Wallace (2005).
Chromospheric activity can fill the core of strong lines andmimic shallower absorptions (e.g. Shcherbakov et al. 1996). Wetherefore wondered whether a variation for the chromosphericactivity in Arcturus might cause the di ff erent line depth in theGIANO-B and FTS spectra.For this purpose, we used the He I line at 10830 Å , whichis a good indicator of chromospheric activity (Danks & Lambert1985). As shown in Fig. 9, when the winter and especially thesummer FTS spectra were acquired, chromospheric activity washigher, as suggested by some He I emission and shallower photo-spheric lines, while when the GIANO-B spectrum was acquired,the activity was low, without He I emission and with deeper pho-tospheric lines. Because strong lines can be problematic also be-cause they might be weakened by this chromospheric activity,they should be used with great caution for an abundance analy-sis. We excluded these lines from our abundance analysis.
6. Discussion and conclusions
Detailed high-resolution optical and NIR spectroscopy of stel-lar calibrators is fundamental for defining optimal diagnosticsfor atmospheric parameters and chemical analysis of stars andstellar populations with di ff erent ages, metallicities, and evolu-tionary properties.While diagnostic tools from high-resolution optical spec-troscopy are well established and have been calibrated for a longtime, those from NIR spectroscopy have been begun to be ex-plored only recently with the new generation of NIR echellespectrographs, whose performances are suitable for such quanti-tative studies.We used Arcturus as a laboratory to explore optical and NIRspectroscopic diagnostics for chemical analyses over the fullspectral range from 4800 to 24500. We then provided a compre-hensive and self-consistent determination of the stellar parame-ters and chemical abundances of Arcturus.The value of this combined optical and NIR study is multi-fold and is summarised below. i ) The study maximises the set of diagnostic lines so thatalmost all the chemical elements of interest can be sampled froma statistically significant number of lines for most of them. ii ) The study enables sampling lines of a given species atdi ff erent wavelengths, which extends the range of excitation po-tentials and transition probabilities for a better understanding ofthe physics of line formation and the modelling of the observedspectrum. iii ) The study drives the analysis towards a physical self-consistent solution over the entire spectrum of the degeneracyproblem among stellar parameters and chemical abundances.Taking advantage of our optical and NIR analysis, i ) we wereable to set an optimal value for the microturbulence velocitythat works over the full spectral range from 4800 to 24500 Å , and ii ) we were able to define a new spectroscopic thermome-ter and new spectroscopic gravitometer for cool giants, basedon atomic and NIR molecular diagnostics of carbon and oxy-gen abundances, as detailed in Sect. 3 and Figs. 4 and 6. Us-ing these diagnostic tools, we infer a temperature and gravity forArcturus that are fully consistent with photometric estimates andwith the values obtained from the standard tools used in opticalspectroscopy (see Figs. 5 and 6).As discussed in Sect. 5 and shown in Fig. 5.2, we find fullyconsistent optical and NIR abundances for all elements we anal-ysed. This demonstrates that i ) the current generation of NIRechelle spectrographs is fully adequate to deliver high-qualitydata for quantitative spectroscopy as in the optical, and ii ) theavailable atomic and molecular data for the NIR lines are gener-ally accurate enough for a reliable chemical abundance analysis.Carbon, sodium, potassium, and iron-peak elements (withthe exception of copper and zinc, which are slightly higher) areconsistent with solar-scaled values, with abundances betweenone-fourth and one-third solar. Nitrogen, oxygen, fluorine, andalpha elements (with the only exception of Ca, which is aboutsolar-scaled), are higher by a few tenths of a dex than the corre-sponding solar-scaled values. The values inferred for [F / O] and[F / Fe] agree with literature values for thick-disk giant stars ofsimilar metallicity (see e.g. Grisoni et al. 2020, their figure 2).Among the neutron-capture elements, cerium and neodymiumare about solar-scaled, dysprosium is slightly higher, and Yt-trium is slightly lower than the corresponding solar-scaled val-ues.Our inferred stellar parameters and chemical abundances arenormally fully consistent with those obtained in previous opticalor NIR studies, as shown in Fig. 10. We stress here that abun-dance di ff erences of a few hundredths dex in di ff erent studies isintrinsic to the analysis process because di ff erent studies may usedi ff erent codes, model atmospheres, and / or line lists and transi-tion probabilities as well as some di ff erent assumptions for thestellar parameters.In particular, when we compare our optical abundances withthose obtained by Ramírez & Allende Prieto (2011), we foundsome notable discrepancies only for C and Mn abundances. Theydetermined the C abundance from four C I lines. In our analysiswe rejected these lines because the two at 9078 Å and 9111 Åare a ff ected by NLTE and the other two at 8335 Å and 5380 Åare blended. We thus used only the forbidden line at 8727 Å. Asdiscussed in Sect. 4.1, our [C I] abundance is fully consistentwith the C abundance derived from CO and C I lines in the Hband.Regarding Mn, the authors mostly measured blue lines that areabsent from our UVES spectrum, while we measured lines in thered part of the optical spectrum. As discussed in Sect. 5.2, ouroptical Mn abundance is fully consistent with our NIR estimatesand the values obtained in other NIR studies.It is also interesting to compare our results on stellar parame-ters and iron abundances with those obtained by the NIR studiesof Kondo et al. (2019) in the in the 9300 − − Article number, page 9 of 16 & A proofs: manuscript no. 39397corr [ X / F e ] O P T Fulbright+2007Ramirez+2011D'orazi+2011Overbeek+2016Cunha+2017 C N O F N a M g A l S i P S K C a S c T i V C r M n F e C o N i C u Z n Y C e N d D y [ X / F e ] N I R Ryde+2009Smith+2013Shetrone+2015 (DR12)Maas+2017
Atomic Number
Fig. 10.
Derived [X / Fe] chemical abundances for Arcturus from some optical (top panel) and NIR (bottom panel) studies quoted in Table 1 andin Sect. 1. Blue symbols are our abundances from the optical UVES spectra, and red symbols our abundances from the NIR GIANO-B spectrum.Blue and red dots refer to neutral species, and blue and red triangles to ionised species. a microturbulence velocity of 1.2 kms − , which is lower thanany previous study and also lower than our adopted value of 1.6 kms − . They also used two di ff erent line lists, VALD3 and thelist by Meléndez & Barbuy (1999) (hereafter MB99). Our listhas 48 lines in common with the Kondo et al. (2019) VALD3line list. Our and the Kondo et al. (2019) VALD3 abundancesare similar, although we adopted a 0.4 kms − higher microturbu-lence velocity.The other 20 lines in the Kondo et al. (2019) VALD3 line listhave not been used in our analysis because they are blendedor contaminated by nearby strong photospheric or telluric lines.These rejected lines give an average abundances that di ff ers bymore than 0.1 dex and has a significantly larger ( > ff ectscan significantly mitigate if not solve the problem.When we use the MB99 astrophysical log ( g f ), which are givenfor lines in the 10 , − ,
000 Å range and are on averagelower by 0.2-0.3 dex than those in VALD3, we find correspond-ing larger abundances.When compared to the iron abundances obtained in previous op-tical studies (see Table 1) and also in the present one, the abun-dances from the MB99 log ( g f ) are in excess by more than 1 σ in the YJ band and more than 2 σ in the H band. When we alsouse the lower microturbulence velocity of Kondo et al. (2019),we obtain even larger and unlikely iron abundances.Smith et al. (2013) used a microturbulence velocity of 1.85 kms − and nine H-band lines with astrophysical log ( g f ) cali- brated on the Sun and Arcturus IR FTS spectra by Livingston& Wallace (1991) and Hinkle & Wallace (2005), respectively.Their log ( g f ) values are somewhat in between those of VALD3and MB99. In our analysis we used 84 H band lines. Five out ofthe nine lines listed by Smith et al. (2013) are in common withour sample. The other four lines in the Smith et al. (2013) listhave been rejected because they are problematic (i.e. partiallyblended and / or with strong wings). Our and their average abun-dances from the five lines in common are very similar. The lower log ( g f ) values used by Smith et al. (2013) are somewhat com-pensated for by their slightly higher microturbulence velocity.The slightly revised iron abundance in Shetrone et al. (2015) ispractically coincident with our estimate.The problem of the imperfect modelling of a line is highlydegenerate. A given variation in the abundance from a givenline can be obtained by modifying its log(gf), but also byslightly varying the stellar parameters and / or by using di ff erentprescriptions for the damping, HFS, NLTE corrections, etc.This degeneracy (see also e.g. Takeda 1992) cannot be easilyremoved.The astrophysical calibration of the log(gf) is becoming verypopular. However, this calibration is model dependent (i.e. itdepends on the adopted model atmospheres, spectral code, lines,etc.), and it also depends on the calibrator itself, that is, on theselected star and observed spectrum as well as on the adoptedstellar parameters. None of the proposed astrophysically cali-brated log(gf) can therefore be safely adopted in studies that userecipes and tools for chemical analyses that are di ff erent fromthose used for the astrophysical calibration.Our combined optical and NIR analysis shows that it is notnecessary to systematically tune the log(gf) of the NIR linesin the VALD3 database to obtain reliable abundances if theappropriate set of lines is chosen and self-consistent stellarparameters are derived. Article number, page 10 of 16. Fanelli et al.: Stellar population astrophysics (SPA) with the TNG
Acknowledgements.
We thank the anonymous referee forhis / her detailed report and useful suggestions. C. Fanelli wouldlike to thank A. Minelli for useful discussions. We acknowledgethe support by INAF / Frontiera through the "Progetti Premiali"funding scheme of the Italian Ministry of Education, University,and Research. We acknowledge support from the project Light-on-Dark granted by MIUR through PRIN2017-000000 contractand support from the mainstream project SC3K - Star clusters inthe inner 3 kpc funded by INAF.
References
Alexeeva, S. A. & Mashonkina, L. I. 2015, MNRAS, 453, 1619Alvarez, R. & Plez, B. 1998, A&A, 330, 1109Alves-Brito, A., Meléndez, J., Asplund, M., Ramírez, I., & Yong, D. 2010, A&A,513, A35Asplund, M., Grevesse, N., Sauval, A. J., & Scott, P. 2009, ARA&A, 47, 481Bensby, T., Feltzing, S., & Oey, M. S. 2014, A&A, 562, A71Charbonnel, C. & Lagarde, N. 2010, A&A, 522, A10Cunha, K., Smith, V. V., Hasselquist, S., et al. 2017, ApJ, 844, 145Danks, A. C. & Lambert, D. L. 1985, A&A, 148, 293D’Orazi, V., Gratton, R. G., Pancino, E., et al. 2011, A&A, 534, A29Dotter, A., Chaboyer, B., Jevremovi´c, D., et al. 2008, ApJS, 178, 89Eggen, O. J. 1971, PASP, 83, 271Fabbian, D., Asplund, M., Carlsson, M., & Kiselman, D. 2006, A&A, 458, 899Fulbright, J. P., McWilliam, A., & Rich, R. M. 2006, ApJ, 636, 821Gray, D. F. 1981, ApJ, 245, 992Grevesse, N. & Sauval, A. J. 1998, Space Sci. Rev., 85, 161Grisoni, V., Romano, D., Spitoni, E., et al. 2020, MNRAS[ arXiv:2008.00812 ]Gustafsson, B., Edvardsson, B., Eriksson, K., et al. 2008, A&A, 486, 951Hasselquist, S., Shetrone, M., Cunha, K., et al. 2016, ApJ, 833, 81Hinkle, K. & Wallace, L. 2005, in Astronomical Society of the Pacific Confer-ence Series, Vol. 336, Cosmic Abundances as Records of Stellar Evolutionand Nucleosynthesis, ed. I. Barnes, Thomas G. & F. N. Bash, 321Hinkle, K., Wallace, L., Harmer, D., Ayres, T., & Valenti, J. 2000, in IAU JointDiscussion, Vol. 24, 26Hinkle, K. H. & Lambert, D. L. 1975, MNRAS, 170, 447Jönsson, H., Ryde, N., Harper, G. M., et al. 2014, A&A, 564, A122Kondo, S., Fukue, K., Matsunaga, N., et al. 2019, ApJ, 875, 129Livingston, W. & Wallace, L. 1991, An atlas of the solar spectrum in the infraredfrom 1850 to 9000 cm-1 (1.1 to 5.4 micrometer)M. Kovalev, S. Brinkmann, M. Bergemann, & MPIA IT-department. 2018, NLTEMPIA web server, [Online]. Available: http: // nlte.mpia.de Max Planck Insti-tute for Astronomy, Heidelberg.Maas, Z. G., Pilachowski, C. A., & Cescutti, G. 2017, ApJ, 841, 108Majewski, S. R., Skrutskie, M. F., Schiavon, R. P., et al. 2007, in American As-tronomical Society Meeting Abstracts, Vol. 211, 132.08Matsunaga, N., Taniguchi, D., Jian, M., et al. 2020, ApJS, 246, 10McWilliam, A. & Rich, R. M. 1994, ApJS, 91, 749Meléndez, J. & Barbuy, B. 1999, ApJS, 124, 527Mucciarelli, A. 2011, A&A, 528, A44Navarro, J. F., Helmi, A., & Freeman, K. C. 2004, ApJ, 601, L43Nissen, P. E. 2004, in Origin and Evolution of the Elements, ed. A. McWilliam& M. Rauch, 154Nordlander, T. & Lind, K. 2017, A&A, 607, A75Oliva, E., Biliotti, V., Ba ff a, C., et al. 2012a, in Society of Photo-Optical In-strumentation Engineers (SPIE) Conference Series, Vol. 8453, Proc. SPIE,84532TOliva, E., Origlia, L., Maiolino, R., et al. 2012b, in Society of Photo-OpticalInstrumentation Engineers (SPIE) Conference Series, Vol. 8446, Proc. SPIE,84463TOliva, E., Sanna, N., Rainer, M., et al. 2018, in Society of Photo-Optical In-strumentation Engineers (SPIE) Conference Series, Vol. 10702, Proc. SPIE,1070274Origlia, L., Oliva, E., Ba ff a, C., et al. 2014a, in Society of Photo-Optical Instru-mentation Engineers (SPIE) Conference Series, Vol. 9147, Ground-based andAirborne Instrumentation for Astronomy V, 91471EOriglia, L., Oliva, E., Ba ff a, C., et al. 2014b, in Society of Photo-Optical In-strumentation Engineers (SPIE) Conference Series, Vol. 9147, Proc. SPIE,91471EOsorio, Y., Allende Prieto, C., Hubeny, I., Mészáros, S., & Shetrone, M. 2020,A&A, 637, A80Overbeek, J. C., Friel, E. D., & Jacobson, H. R. 2016, ApJ, 824, 75Plez, B. 2012, Turbospectrum: Code for spectral synthesisRainer, M., Harutyunyan, A., Carleo, I., et al. 2018, in Society of Photo-OpticalInstrumentation Engineers (SPIE) Conference Series, Vol. 10702, Proc. SPIE,1070266Ramírez, I. & Allende Prieto, C. 2011, ApJ, 743, 135Reddy, B. E., Lambert, D. L., & Allende Prieto, C. 2006, MNRAS, 367, 1329Reddy, B. E., Tomkin, J., Lambert, D. L., & Allende Prieto, C. 2003, MNRAS,340, 304Ryabchikova, T. & Pakhomov, Y. 2015, Baltic Astronomy, 24, 453Ryde, N., Edvardsson, B., Gustafsson, B., et al. 2009, A&A, 496, 701Shcherbakov, A. G., Shcherbakova, Z. A., Tuominen, I., & Jetsu, L. 1996, A&A,309, 655Shetrone, M., Bizyaev, D., Lawler, J. E., et al. 2015, ApJS, 221, 24Smith, V. V., Cunha, K., Shetrone, M. D., et al. 2013, ApJ, 765, 16Takeda, Y. 1992, A&A, 253, 487Takeda, Y., Omiya, M., Harakawa, H., & Sato, B. 2016, PASJ, 68, 81Takeda, Y. & Takada-Hidai, M. 2013, PASJ, 65, 65Tozzi, A., Oliva, E., Iuzzolino, M., et al. 2016, in Society of Photo-Optical In-strumentation Engineers (SPIE) Conference Series, Vol. 9908, Proc. SPIE,99086CWorley, C. C., Cottrell, P. L., Freeman, K. C., & Wylie-de Boer, E. C. 2009,MNRAS, 400, 1039Zhang, H. W., Gehren, T., Butler, K., Shi, J. R., & Zhao, G. 2006, A&A, 457,645Zhang, J., Shi, J., Pan, K., Allende Prieto, C., & Liu, C. 2017, ApJ, 835, 90 Article number, page 11 of 16 & A proofs: manuscript no. 39397corr
Appendix A: Arcturus atlas
Article number, page 12 of 16. Fanelli et al.: Stellar population astrophysics (SPA) with the TNG . . . . . . . . . . . . . . . . . . Normalized Flux . . . . . . λ ( Å ) . . . . . . F i g . A . . O b s e r v e d Y - b a nd A r c t u r u ss p ec t r u m ( b l ac k li n e ) w it h t h e t e ll u r i cc o rr ec ti on (r e d li n e ) Article number, page 13 of 16 & A proofs: manuscript no. 39397corr . . . . . . . . . . . . . . . . . . Normalized Flux . . . . . . λ ( Å ) . . . . . . F i g . A . . O b s e r v e d J - b a nd A r c t u r u ss p ec t r u m ( b l ac k li n e ) w it h t h e t e ll u r i cc o rr ec ti on (r e d li n e ) Article number, page 14 of 16. Fanelli et al.: Stellar population astrophysics (SPA) with the TNG . . . . . . . . . . . . . . . . . . Normalized Flux . . . . . . λ ( Å ) . . . . . . F i g . A . . O b s e r v e d H - b a nd A r c t u r u ss p ec t r u m ( b l ac k li n e ) w it h t h e t e ll u r i cc o rr ec ti on (r e d li n e ))
Article number, page 12 of 16. Fanelli et al.: Stellar population astrophysics (SPA) with the TNG . . . . . . . . . . . . . . . . . . Normalized Flux . . . . . . λ ( Å ) . . . . . . F i g . A . . O b s e r v e d Y - b a nd A r c t u r u ss p ec t r u m ( b l ac k li n e ) w it h t h e t e ll u r i cc o rr ec ti on (r e d li n e ) Article number, page 13 of 16 & A proofs: manuscript no. 39397corr . . . . . . . . . . . . . . . . . . Normalized Flux . . . . . . λ ( Å ) . . . . . . F i g . A . . O b s e r v e d J - b a nd A r c t u r u ss p ec t r u m ( b l ac k li n e ) w it h t h e t e ll u r i cc o rr ec ti on (r e d li n e ) Article number, page 14 of 16. Fanelli et al.: Stellar population astrophysics (SPA) with the TNG . . . . . . . . . . . . . . . . . . Normalized Flux . . . . . . λ ( Å ) . . . . . . F i g . A . . O b s e r v e d H - b a nd A r c t u r u ss p ec t r u m ( b l ac k li n e ) w it h t h e t e ll u r i cc o rr ec ti on (r e d li n e )) Article number, page 15 of 16 & A proofs: manuscript no. 39397corr . . . . . . . . . . . . . . . . . . Normalized Flux . . . . . . λ ( Å ) . . . . . . F i g . A . . O b s e r v e d K - b a nd A r c t u r u ss p ec t r u m ( b l ac k li n e ) w it h t h e t e ll u r i cc o rr ec ti on (r e d li n e ))