The [Y/Mg] clock works for evolved solar metallicity stars
D. Slumstrup, F. Grundahl, K. Brogaard, A. O. Thygesen, P. E. Nissen, J. Jessen-Hansen, V. Van Eylen, M. G. Pedersen
AAstronomy & Astrophysics manuscript no. clusters_le c (cid:13)
ESO 2018July 17, 2018
The [Y/Mg] clock works for evolved solar metallicity stars (cid:63)
D. Slumstrup , F. Grundahl , K. Brogaard , , A. O. Thygesen , P. E. Nissen , J. Jessen-Hansen , V. Van Eylen , andM. G. Pedersen Stellar Astrophysics Centre (SAC). Department of Physics and Astronomy, Aarhus University, Ny Munkegade 120, DK-8000Aarhus, Denmark e-mail: [email protected] School of Physics and Astronomy, University of Birmingham, Edgbaston, Birmingham B15 2TT, UK California Institute of Technology, 1200 E. California Blvd, MC 249-17, Pasadena, CA 91125, USA Leiden Observatory, Leiden University, 2333CA Leiden, The Netherlands Instituut voor Sterrenkunde, KU Leuven, Celestijnenlaan 200D, 3001 Leuven, BelgiumReceived 3 July 2017 / Accepted 20 July2017
ABSTRACT
Aims.
Previously [Y / Mg] has been proven to be an age indicator for solar twins. Here, we investigate if this relation also holds forhelium-core-burning stars of solar metallicity.
Methods.
High resolution and high signal-to-noise ratio (S / N) spectroscopic data of stars in the helium-core-burning phase have beenobtained with the FIES spectrograph on the NOT 2 .
56 m telescope and the HIRES spectrograph on the Keck I 10 m telescope. Theyhave been analyzed to determine the chemical abundances of four open clusters with close to solar metallicity; NGC 6811, NGC 6819,M67 and NGC 188. The abundances are derived from equivalent widths of spectral lines using ATLAS9 model atmospheres withparameters determined from the excitation and ionization balance of Fe lines. Results from asteroseismology and binary studies wereused as priors on the atmospheric parameters, where especially the log g is determined to much higher precision than what is possiblewith spectroscopy. Results.
It is confirmed that the four open clusters are close to solar metallicity and they follow the [Y / Mg] vs. age trend previouslyfound for solar twins.
Conclusions.
The [Y / Mg] vs. age clock also works for giant stars in the helium-core burning phase, which vastly increases thepossibilities to estimate the age of stars not only in the solar neighborhood, but in large parts of the Galaxy, due to the brighter natureof evolved stars compared to dwarfs.
Key words. stars: abundances - stars: fundamental parameters - stars: late-type - Galaxy: evolution - open clusters and associations:individual: NGC 6811, NGC 6819, M67, NGC 188
1. Introduction
Stars in an open cluster can be assumed to originate from thesame molecular cloud and are therefore expected to have thesame chemical composition and age. Age is an especially dif-ficult parameter to determine for single field stars and di ff er-ent methods yield ages that are not necessarily in agreement(Soderblom 2010), whereas stars in clusters have ages deter-mined to much higher precision (e.g., VandenBerg et al. (2013)).da Silva et al. (2012), Nissen (2016) and Tucci Maia et al.(2016) investigated trends of di ff erent chemical abundances withstellar age and found that [Y / Mg] is a sensitive age indicatorfor solar twins. Feltzing et al. (2017) did an independent studyextending the sample to a much larger range in [Fe / H]. Theyfound that the relation depends on [Fe / H]. For solar metallicitystars they clearly see the relation, but for stars with [Fe / H] ∼ -0.5dex the relation is insignificant and cannot be used to determineage. Our targets have solar metallicity but are evolved (in the (cid:63) Based on spectroscopic observations made with two telescopes: theNordic Optical Telescope operated by NOTSA at the Observatorio delRoque de los Muchachos (La Palma, Spain) of the Instituto de As-trofísica de Canarias and the Keck I Telescope at the W.M. Keck Obser-vatory (Mauna Kea, Hawaii, USA) operated by the California Instituteof Technology, the University of California and the National Aeronau-tics and Space Administration. helium-core-burning phase) and therefore do not have the sameatmospheric properties as the solar twins. They are also locatedat larger distances than previously studied. Ages are well deter-mined from cluster studies, which allows us to determine if therelation found for solar twins persists for helium-core-burningstars. Previously, yttrium abundances of open clusters have beenstudied to find possible trends with age, for example, Misheninaet al. (2014) found no clear trend of [Y / Fe] with age most likelydue to large uncertainties, whereas Maiorca et al. (2011) find adeclining trend, but the scatter is large, which could be a ff ectedby the large range in [Fe / H].We have carried out a detailed fundamental parameterand abundance analysis of six targets in four open clusters(NGC 6811, NGC 6819, M67 and NGC 188) based on high-resolution and high signal-to-noise ratio (S / N) spectroscopicdata from the Nordic Optical Telescope (NOT) and the Keck ITelescope. For NGC 6811, NGC 6819 and M67 we have astero-seismic data available from the
Kepler (Borucki et al. 1997) andK2 missions (Howell et al. 2014), which put a strong constrainton the log g value and can thereby constrain the analysis.In Sect. 2 we present the data taken for this project. In Sect. 3,we describe the analysis. In Sect. 4 we present the results. InSect. 5, we test the [Y / Mg] vs. age relation by Nissen (2016)for our sample of solar metallicity helium-core-burning stars. Fi-nally, we conclude on the results in Sect. 6.
Article number, page 1 of 5 a r X i v : . [ a s t r o - ph . S R ] J u l & A proofs: manuscript no. clusters_le
Table 1.
Cluster information.
Cluster E ( B − V ) ( m − M ) V Age [Gyr]NGC 6811 a ± b ± c ± d ± a Molenda-Zakowicz et al. (2014) b Rosvick & VandenBerg (1998) c Reddening from Taylor (2007), distance modulusand age from Yadav et al. (2008) d Meibom et al. (2009)
2. Targets, observations and data reduction
The metallicity of NGC 6819 is still debated (Bragaglia et al.2001; Lee-Brown et al. 2015) and we aim to establish it securelythrough the analysis in this project. M67 is a very well stud-ied nearby solar-like cluster, which has also been observed withthe K2 mission. NGC 188 is also a solar metallicity cluster thatis older and fairly well studied, however not as well as M67.For these three clusters, we have only chosen one target in eachcluster, but they are all confirmed members, (Tabetha Hole et al.2009; Yadav et al. 2008; Stetson et al. 2004). The data used in ouranalysis for both NGC 188 and NGC 6819 is of higher resolutionand higher S / N than previously used. NGC 6811 is also a solar-metallicity cluster and it is the youngest in the sample. For thiswe have three targets, all confirmed members (Sandquist et al.2016). All targets, except the NGC 188 target, is confirmed byasteroseismology to be in the helium-core-burning phase (Cor-saro et al. 2012; Stello et al. 2016; Arentoft et al. 2017). Wehave adopted literature values for reddening, distance modulusand age for each cluster (see Table 1).The observations of NGC 6819, M67 and NGC 188 were col-lected with the Nordic Optical Telescope (NOT) on La Palma,Spain. Spectra for the three stars were obtained in the summerof 2013 and 2015 with the high resolution FIbre-fed EchelleSpectrograph (FIES) covering the wavelength region from 3700-7300Å, see Telting et al. (2014) for a detailed description of thespectrograph. All observations were carried out in the high res-olution mode, R = , , an automated data reduction software for FIES.Lastly, the spectra for each target were shifted to a commonwavelength scale and merged.The observations of the three targets in NGC 6811 were car-ried out on the night of Aug. 23, 2016 using the HIRES spec-trograph (Vogt et al. 1994) on Keck I. All targets were ob-served with the red cross-disperser, covering the wavelength re-gion from 4000-7900Å, with a few inter-order gaps for the red-dest orders. We used the ”C5” decker, providing a resolution of R =
37 000. The targets were observed with the exposure me-ter in operation (Kibrick et al. 2006) to ensure a uniform S / N inall spectra. The reductions were performed using the MAKEEpipeline . All observations are presented in Table 2.The co-added spectrum for each star was normalized orderby order using RAINBOW , which uses appropriate syntheticspectra to identify continuum points in the observed spectrum,which are then fitted with a spline function. The S / N values givenin Table 2 were estimated from the rms variation of the flux in aregion around 6150 Å. http://sites.google.com/site/vikingpowersoftware
3. Data analysis
Parameters derived from the photometry presented in Fig. 1 wereused as starting estimates, listed in Table 3. The e ff ective tem-peratures for NGC 6819, M67 and NGC 188 were calculatedfrom the color-temperature calibration presented by Ramírez &Meléndez (2005) for di ff erent filter combinations. For each tar-get, an average of the di ff erent filter combinations was used.For NGC 6811 we used the e ff ective temperatures from Arentoftet al. (2017) as first estimates.The log g for the stars in NGC 6811, NGC 6819 and M67 inTable 3 is from asteroseismology, (Arentoft et al. 2017; Corsaroet al. 2012; Stello et al. 2016 respectively) calculated with the ν max scaling relations (Brown1991, Kjeldsen1995):log g = log (cid:32)(cid:18) ν max (cid:19) · (cid:18) T e ff (cid:19) / (cid:33) + . . (1)The asteroseismic log g values are determined to very high preci-sion and have been shown to be in very close agreement with thephysical log g (e.g., Brogaard et al. 2016; Gaulme et al. 2016;Frandsen et al. 2013). This provides a strong constraint on theanalysis. For NGC 188, the log g is calculated with the equa-tion from Nissen et al. (1997) using a mass of 1.1 M (cid:12) from Mei-bom et al. (2009) assuming no significant mass loss on the RGB(Miglio et al. 2012).We have tested di ff erent line lists and di ff erent programsto calculate the equivalent widths, which will be discussed inmore detail in a forthcoming paper. Based on external constraintsfrom especially asteroseismology, the final choice of line listis from Carraro et al. (2014) with astrophysical log g f valuesbased on solar abundances from Grevesse & Sauval (1998). Wehave omitted lines stronger than 100 mÅ for Fe and 120 mÅfor other elements. A few additional magnesium and yttriumlines were added to do a more robust determination of [Y / Mg]for the [Y / Mg] vs. age relation discussed in Sect. 5. The finalline list will be given in a forthcoming paper with the mea-sured equivalent widths for all lines. The equivalent widths weremeasured with DAOSPEC (Stetson & Pancino 2008) and theauxiliary program Abundance with SPECTRUM (Gray & Cor-bally 1994) was used to calculate the atmospheric parametersand abundances based on solar abundances from Grevesse &Sauval (1998) and ATLAS9 stellar atmosphere models (Castelli& Kurucz 2004). Local thermodynamic equilibrium (LTE) isassumed. There may be non-LTE e ff ects on the derived abun-dances, but since the stars have similar parameters, di ff erentialabundances between the stars are reliable.The atmospheric parameters were determined by requiringthat [Fe / H] has no systematic dependence on the excitation po-tential or the strength of the FeI lines and that the mean [Fe / H]values derived from FeI and FeII lines are consistent. The slopeof [Fe / H] as a function of excitation potential is sensitive to thee ff ective temperature and the slope of [Fe / H] as a function ofthe reduced equivalent widths of the lines (log(EW) /λ ) dependson the microturbulence. The surface gravity is determined viaits e ff ect on the electron pressure in the stellar atmosphere withthe FeI and FeII equilibrium, as the FeII lines are more sensitiveto pressure changes than the FeI lines. This is however also af-fected by the temperature and heavier element abundances and itwas necessary to make a number of iterations. For NGC 6811,NGC 6819 and M67, we also calculated a new asteroseismiclog g with the newly found e ff ective temperature, but the vari-ation in e ff ective temperature is low enough, that the asteroseis-mic log g was not significantly a ff ected. Article number, page 2 of 5. Slumstrup et al.: The [Y / Mg] clock works for evolved solar metallicity stars
Table 2.
Observations information.
Target α J δ J V Memb. prop Numberof exp. Total exp.time [h] S / N @6150ÅNGC 6811-KIC9655167 19 37 02.68 +
46 23 13.1 11.32 97 1 0.31 145NGC 6811-KIC9716090 19 36 55.80 +
46 27 37.6 11.53 94 1 0.51 140NGC 6811-KIC9716522 19 37 34.63 +
46 24 10.1 10.65 97 1 0.15 150NGC 6819-KIC5024327 19 41 13.45 +
40 11 56.2 13.11 94 21 17.0 115M67-EPIC211415732 08 51 12.70 +
11 52 42.4 10.39 97 7 3.9 135NGC 188-5085 b
00 46 59.57 +
85 13 15.8 12.31 85 13 11.5 105 b Stetson et al. (2004) V NGC 6811 0.5 1.0B-V121416 V NGC 6819 V M67 V NGC 188
Fig. 1.
From left to right: The CMDs for NGC 6811, NGC 6819, and M67 with photometry from Yontan et al. (2015); Tabetha Hole et al. (2009);Yadav et al. (2008) respectively and lastly a CMD for NGC 188 with photometry from Peter Stetson for one of his standard fields (Stetson et al.2004). The target star(s) for each cluster is marked with a red star.
Table 3.
Adopted parameters from photometry and asteroseismology
Target T e ff [K] ν max log g NGC 6811-KIC9655167 a ±
100 99.4 ± ± ±
100 107.8 ± ± ±
100 53.7 ± ± ±
33 43.9 ± b ± ±
43 37.9 ± c ± ±
19 - 2.45 ± a All NGC 6811 values are from Arentoft et al. (2017) b Corsaro et al. (2012) c Stello et al. (2016)
4. Atmospheric parameters and abundances
The final result for the atmospheric parameters are presented inTable 4. The uncertainties are only internal and calculated byvarying a parameter until at least a 3 σ uncertainty is producedon either of the two slopes, [Fe / H] vs. excitation potential and[Fe / H] vs. reduced equivalent width, or on the di ff erence be-tween FeI and FeII. The change in the parameter is then dividedby the highest produced uncertainty to give one standard devi-ation, provided in Table 4. The errorbars on [Fe / H], [ α / Fe] and[Y / Mg] is the standard error of the mean.The spectroscopic log g values are, within the errorbars, inagreement with the results from asteroseismology. The metal-licities for all targets are close to solar, with NGC 188 havinga slightly higher-than-solar abundance and NGC 6811 having aslightly lower-than-solar abundance, which along with the tem-perature di ff erences can be seen in the line depths in Fig. 2.The metallicity of NGC 6819 is lower than that found by Bra-gaglia et al. (2001) ([Fe / H] =+ . ± .
03 dex found by analyz-ing three giants with high resolution spectroscopy), but fits better with the value from Lee-Brown et al. (2015) who performed ananalysis of multiple targets but with low resolution spectroscopy.Their value of [Fe / H] = − . ± .
02 dex is found using main se-quence and turno ff stars.The abundances for [Fe / H], [ α / Fe], and [Y / Mg] are given inTable 4 while several additional elements will be presented ina forthcoming paper. The magnesium lines used are at wave-lengths 5711 .
09 Å, 6318 .
70 Å, 6319 .
23 Å and 6319 .
48 Å. Theyttrium lines are at 4883 .
70 Å, 4900 .
13 Å and 5728 .
89 Å. Notall of the lines were usable for each star. The magnesium line at5711 .
09 Å (see Fig. 2) is, for some of the targets, stronger thanthe limit of 120 mÅ for non-iron lines, but it is well isolated forall stars, and we therefore chose to include it to do a more robustdetermination of Magnesium. The final value for [Y / Mg] usedfor NGC 6811 is a the mean of the three targets.
5. [Y/Mg] vs. age
Magnesium is an alpha-element mostly originating from type IIsupernovae explosions, which gives an increase of [Mg / Fe] withincreasing stellar age because iron is also produced in the latertype Ia supernovae explosions. Yttrium is an s-process elementand [Y / Fe] is observed to decrease with increasing stellar age.This is likely a consequence of intermediate mass asymptoticgiant branch stars not yet being important for the production of Yat early times. The slope of [Y / Fe] with age is steep and oppositeto that of [Mg / Fe]. For solar twin stars, Nissen (2016) found therelation:[Y / Mg] = . ± . − . ± . · Age [Gyr]. (2)This relation is plotted in Fig. 3. Feltzing et al. (2017) confirmedthe relation for dwarfs of solar metallicity but found that it disap-pears for stars with [Fe / H] ≈ − . Article number, page 3 of 5 & A proofs: manuscript no. clusters_le
Table 4.
Atmospheric parameters and abundances from spectroscopy
Target T e ff [K] log g v t [km / s] [Fe / H] n a [ α / Fe] [Y / Mg]NGC 6811-KIC9655167 5085 ±
22 2.96 ± ± ± /
11 0.02 ± ± ±
25 2.93 ± ± ± /
11 0.06 ± ± ±
30 2.60 ± ± ± /
11 0.03 ± ± ±
18 2.52 ± ± ± /
12 -0.02 ± ± ±
18 2.43 ± ± ± /
12 0.03 ± ± ±
23 2.51 ± ± ± /
12 0.00 ± ± a The number of FeI / FeII lines used. N o r m a li z e d F l u x Mg NGC 188NGC 6819M67NGC 6811 4883.5 4884.0Wavelength [ ]0.20.40.60.81.0 N o r m a li z e d F l u x YII NGC 188NGC 6819M67NGC 6811
Fig. 2.
Two parts of the normalized spectra for one target in each cluster to illustrate the high quality of the spectra. NGC 6819, M67 and NGC 188are very similar, which is also illustrated in the atmospheric parameters in Table 4. NGC 6811 spectrum is for KIC9655167. [ Y / M g ] Fig. 3. [Y / Mg] vs. age for the four clusters. The line is the relation foundby Nissen (2016) in Eq. 2 for solar twins. The red point is an averageof the results from Önehag et al. (2014) and the errorbar is the standarderror of the mean for their 14 stars. the parameter range to helium-core-burning giants at close tosolar metallicity, and they also follow the relation from Nissen(2016) as seen in Fig. 3. This result is of particular interest forgalactic archaeology studies as giants are much brighter thandwarfs, which allows us to study farther regions of the galaxyand not only the solar neighborhood.Önehag et al. (2014) carried out an abundance analysis of 14stars in M67 at di ff erent evolutionary stages with high resolutionspectra. They find an average metallicity of [Fe / H] =+ / Mg] = − . ± .
05, which is a little lower than ourresult, but still close to being in agreement with the relation fromNissen (2015), marked in Fig. 3
6. Conclusion
Atmospheric parameters and abundances have been determinedfor the four open clusters, NGC 6811, NGC 6819, M67 andNGC 188 with an equivalent width analysis of individual spec-tral lines from high-resolution, high S / N observations from theNOT and the Keck I Telescope. The parameters obtained fit verywell with the literature, and especially, the log g values fits withpredictions from asteroseismology. The metallicities of all fourclusters are nearly solar, with NGC 6811 being slightly sub-solar. The empirical relation between [Y / Mg] and age as presentedby Nissen (2016) was found to hold also for helium-core-burninggiants of close to solar metallicity. This is of great importanceto galactic chemical evolution studies, as the brighter nature ofgiants allows us to probe the Galaxy to greater distances and notonly the solar neighborhood.
Acknowledgements.
The authors wish to recognize and acknowledge the verysignificant cultural role and reverence that the summit of Mauna Kea has al-ways had within the indigenous Hawaiian community. We are most fortunate tohave the opportunity to conduct observations from this mountain. Funding for theStellar Astrophysics Centre is provided by The Danish National Research Foun-dation (Grant DNRF106). This research has made use of the SIMBAD database,operated at CDS, Strasbourg, France. MGP is funded from the European Re-search Council (ERC) under the European Union’s Horizon2020 research andinnovation program (grant agreement N o References
Arentoft, T., Brogaard, K., Silva Aguirre, V., et al. 2017, ApJ, 838, 115Borucki, W. J., Koch, D. G., Dunham, E. W., & Jenkins, J. M. 1997, in PlanetsBeyond the Solar System and the Next Generation of Space Missions, ed.D. R. Soderblom, 21Bragaglia, A., Carretta, E., Gratton, R., et al. 2001, AJ, 121, 327Brogaard, K., Jessen-Hansen, J., Handberg, R., et al. 2016, AstronomischeNachrichten, 337, 793Carraro, G., Villanova, S., Monaco, L., et al. 2014, A&A, 562, A39Castelli, F. & Kurucz, R. L. 2004, ArXiv Astrophysics e-prints, 1Corsaro, E., Stello, D., Huber, D., et al. 2012, ApJ, 757, 190da Silva, R., Porto de Mello, G. F., Milone, A. C., et al. 2012, A&A, 542, A84Feltzing, S., Howes, L. M., McMillan, P. J., & Stonkute, E. 2017, MNRAS, 465,L109Frandsen, S., Lehmann, H., Hekker, S., et al. 2013, A&A, 556, A138Gaulme, P., McKeever, J., Jackiewicz, J., et al. 2016, ApJ, 832, 121Gray, R. O. & Corbally, C. J. 1994, AJ, 107, 742Grevesse, N. & Sauval, A. J. 1998, Space Sci. Rev., 85, 161Howell, S. B., Sobeck, C., Haas, M., et al. 2014, PASP, 126, 398Kibrick, R. I., Clarke, D. A., Deich, W. T. S., & Tucker, D. 2006, in Proc. ofSPIE vol 6274, ed. H. Lewis & A. Bridger, 62741ULee-Brown, D. B., Anthony-Twarog, B. J., Deliyannis, C. P., Rich, E., & Twarog,B. A. 2015, AJ, 149, 121Maiorca, E., Randich, S., Busso, M., Magrini, L., & Palmerini, S. 2011, ApJ,736, 120Meibom, S., Grundahl, F., Clausen, J. V. ., et al. 2009, AJ, 137, 5086Miglio, A., Brogaard, K., Stello, D., et al. 2012, MNRAS, 419, 2077Mishenina, T., Pignatari, M., Carraro, G., et al. 2014, MNRAS, 446, 3651
Article number, page 4 of 5. Slumstrup et al.: The [Y / Mg] clock works for evolved solar metallicity stars