Mono-enriched stars and Galactic chemical evolution -- Possible biases in observations and theory
Camilla Juul Hansen, Andreas Koch, Lyudmila Mashonkina, Mattis Magg, Maria Bergemann, Tatyana Sitnova, Andrew J. Gallagher, Ilya Ilyin, Elisabetta Caffau, Huawei W. Zhang, Klauss G. Strassmeier, Ralf S. Klessen
AAstronomy & Astrophysics manuscript no. 38805corr_arxiv c (cid:13)
ESO 2020September 28, 2020
Mono-enriched stars and Galactic chemical evolution – Possiblebiases in observations and theory (cid:63) , (cid:63)(cid:63) Hansen, C. J. , Koch, A. , Mashonkina, L. , Magg, M. , , Bergemann, M. , Sitnova, T. , Gallagher, A. J. , Ilyin, I. ,Ca ff au, E. , Zhang, H.W. , , Strassmeier, K. G. , Klessen, R. S. , Max Planck Institute for Astronomy, Königstuhl 17, 69117 Heidelberg, Germany Zentrum für Astronomie der Universität Heidelberg, Astronomisches Rechen-Institut, Mönchhofstr. 12, 69120 Heidelberg, Ger-many Institute of Astronomy, Russian Academy of Sciences, Pyatnitskaya 48, 119017, Moscow, Russia Universität Heidelberg, Zentrum für Astronomie, Institut für Theoretische Astrophysik, 69120 Heidelberg, Germany International Max Planck Research School for Astronomy and Cosmic Physics at the University of Heidelberg (IMPRS-HD) Leibniz-Institut für Astrophysik Potsdam (AIP), An der Sternwarte 16, 14482 Potsdam, Germany – e-mail: [email protected], [email protected] [orcid: 0000-0002-6192-6494] GEPI, Observatoire de Paris, Université PSL, CNRS, 5 Place Jules Janssen, 92190 Meudon, France Department of Astronomy, School of Physics, Peking University, Beijing 100871, P.R. China Kavli Institute for Astronomy and Astrophysics, Peking University, Beijing 100871, P.R. China Universität Heidelberg, Interdisziplinäres Zentrum für Wissenschaftliches Rechnen, Im Neuenheimer Feld 205, 69120 Heidelberg,GermanyReceived July 1 2020; accepted September 15, 2020
ABSTRACT
A long sought after goal using chemical abundance patterns derived from metal-poor stars is to understand the chemical evolutionof the Galaxy and to pin down the nature of the first stars (Pop III). Metal-poor, old, unevolved stars are excellent tracers as theypreserve the abundance pattern of the gas from which they were born, and hence they are frequently targeted in chemical taggingstudies. Here, we use a sample of 14 metal-poor stars observed with the high-resolution spectrograph called the Potsdam EchellePolarimetric and Spectroscopic Instrument (PEPSI) at the Large Binocular Telescope (LBT) to derive abundances of 32 elements (34including upper limits). We present well-sampled abundance patterns for all stars obtained using local thermodynamic equilibrium(LTE) radiative transfer codes and one-dimensional (1D) hydrostatic model atmospheres. However, it is currently well-known that theassumptions of 1D and LTE may hide several issues, thereby introducing biases in our interpretation as to the nature of the first starsand the chemical evolution of the Galaxy. Hence, we use non-LTE (NLTE) and correct the abundances using three-dimensional (3D)model atmospheres to present a physically more reliable pattern. In order to infer the nature of the first stars, we compare unevolved,cool stars, which have been enriched by a single event (‘mono-enriched’), with a set of yield predictions to pin down the mass andenergy of the Pop III progenitor. To date, only few bona fide second generation stars that are mono-enriched are known. A simple χ -fit may bias our inferred mass and energy just as much as the simple 1D LTE abundance pattern, and we therefore carried outour study with an improved fitting technique considering dilution and mixing. Our sample presents Carbon Enhanced Metal-Poor(CEMP) stars, some of which are promising bona fide second generation (mono-enriched) stars. The unevolved, dwarf BD + (cid:12) and 0.6 foe (0.6 10 erg) in LTE and 19.2 M (cid:12) and1.5 foe in NLTE, respectively. Finally, we explore the predominant donor and formation site of the rapid and slow neutron-captureelements. In BD-10_3742, we find an almost clean r-process trace, as is represented in the star HD20, which is a ‘metal-poor Sunbenchmark’ for the r-process, while TYC5481-00786-1 is a promising CEMP-r / -s candidate that may be enriched by an asymptoticgiant branch star of an intermediate mass and metallicity. Key words.
Stars: abundances – stars: kinematics & dynamics – galaxy: halo – Nuclear reactions, nucleosynthesis, abundances –early Universe
1. Introduction
Understanding the chemical evolution of the Milky Way (MW)has been a long-standing quest that crucially a ff ects severalbranches of astrophysics. Key aspects in such analyses are stel-lar abundances, yield predictions, and the modelling of how gas Send o ff print requests to : C. J. Hansen, e-mail: [email protected] (cid:63) Based on data acquired with PEPSI using the Large Binocular Tele-scope (LBT). The LBT is an international collaboration among institu-tions in the United States, Italy, and Germany. (cid:63)(cid:63)
The line list will appear online on CDS. flows and condensates into later generations of stars. These top-ics are central to our following analysis, where we target the ear-liest enrichment from the first, so-called Population III (Pop III)stars, up to later enrichment events from asymptotic giant branch(AGB) stars. The vast majority of these studies focus on one-dimensional (1D) local thermodynamic equilibrium (LTE) abun-dances and explore them in the grand scheme of Galactic chemi-cal evolution (GCE), as was observationally done in Edvardssonet al. (1993) and Cayrel et al. (2004), for example. However, re-cent improvements in atomic physics and more computational
Article number, page 1 of 25 a r X i v : . [ a s t r o - ph . S R ] S e p & A proofs: manuscript no. 38805corr_arxiv power allow for higher dimensional model atmosphere calcula-tions and a more sophisticated physical treatment of atomic tran-sitions, including improved radiative and collisional rates. Whencombined, these allow for more precise and accurate abundancedeterminations and they are an improvement from the simpler1D LTE to three-dimensional (3D) and non-LTE (NLTE) results.The chemo-dynamic behaviour of the major Galactic compo-nents, the halo as well as the thick and thin disc, are known to dif-fer in several ways. The MW halo is typically more metal-poor(peaking around [Fe / H] = − .
6; Schörck et al. 2009; Youakimet al. 2020), exhibiting fast moving stars on elliptical orbitsreaching apocentres several kiloparsecs above the plane and be-yond. For comparison, disc stars are on average more metal-rich,and they move on more circular orbits close to and / or withinthe plane (e.g. Bensby et al. 2003). By combining the orbitalproperties with the chemical composition of stars, we can tracethe spatial and chemical origin of the stars with a higher levelof confidence and, in turn, label the evolution and chemistry ofthe Galaxy. Earlier studies, such as Nissen & Schuster (2010),showed how the α − element abundances combined with the 3Dphase space velocities could help determine whether stars wereformed in situ in the MW and likely enriched by more massivesupernovae (SN), or whether they have been accreted, predomi-nantly move on fast, retrograde orbits in the outer halo, and ex-hibit a lower [ α / Fe] (see also Cooper et al. 2013; Pillepich et al.2015; Myeong et al. 2019). The underlying scenario is likely theaccretion of dwarf galaxies onto the MW, which would explainthe counter-rotating stellar orbits. Moreover, the low-mass dwarfgalaxies typically retain a small gas reservoir and may, as a re-sult, predominantly form lower-mass supernovae, which in turnyield smaller amounts of α − elements (Tinsley 1979).The GCE scenarios are best studied in large, statistically sig-nificant samples. However, the precise determination of stellarpatterns o ff ers an intriguing counterpart to help learn a greatlevel of detail about the galacto-chemical origin of stars insmaller samples. As an example, the level of α -content can pro-vide information on the mass of the previous generation of SNprogenitors, while the Fe-peak elements or the odd-even elementabundances can reveal information on the explosion energy (e.g.Kobayashi et al. 2006). The heavy elements ( Z >
30) are formedvia neutron-captures in most cases, and they split into two cat-egories depending on the rates of captures with respect to thefollowing β − decay. Here, the slow neutron-capture ( s -)processforms elements, such as Ba and Pb, while the rapid ( r -)processforms, Eu, Th, and U, for example. The s -process is typicallyassociated with AGB stars or massive rotating stars (Busso et al.1999; Käppeler et al. 2011; Meynet et al. 2006), while the r -process site is still debated (Horowitz et al. 2019; Côté et al.2019). However, recent discoveries have confirmed that neutronstar mergers can produce r -process material in the event (Watsonet al. 2019), while magneto-rotational SN provide a promisingextra site, both at higher and lower metallicities (Winteler et al.2012; Côté et al. 2019).When trying to trace the individual sources of enrichment,be it a rare or common SN, the best way to explore their na-ture in great detail is by studying the chemical abundances in thefollowing generation of old, metal-poor stars and indirectly in-ferring the nature of the SN that provided the enriched material(e.g. Cayrel et al. 2004; Hansen & Primas 2011). The surfacesof old, cool, unevolved stars preserve the gas composition fromwhich they were born and hence o ff er a perfect laboratory forunderstanding how and where the elements formed. This is be-cause we can derive stellar abundances more accurately in thephotospheres of cool stars, especially because calculations of ra- diative transfer for cool stars can now include 3D hydrodynam-ics and detailed NLTE, yielding more robust results compared towhat is currently possible in fast expanding SN ejecta.At the lowest metallicities, we are most likely to find a gen-eration of stars that are mono-enriched, that is to say enrichedby only one supernova. This leaves the cleanest fingerprint ofthe nature of that SN, which is why metal-poor stars are a muchsought after diagnostic. At the lowest metallicities, an increasingnumber of stars with high C-abundances have been discovered inthe past decades (Frebel et al. 2006; Lee et al. 2013; Placco et al.2014a). These stars are commonly referred to as CEMP stars(Carbon Enhanced Metal-Poor stars), and they tend to have a fac-tor of 5–10 times higher [C / Fe] than the Sun (Beers & Christlieb2005; Aoki et al. 2007; Hansen et al. 2016a). These stars comein di ff erent flavours depending on their heavy element composi-tion, where a rich s -process composition likely indicates that thestar is or was in a binary system with an AGB star transferringmass to the companion, while a lack of heavy elements at thelowest metallicities seems to point towards the star being a bonafide second generation star (Ito et al. 2009; Bonifacio et al. 2015;Placco et al. 2014b; Hansen et al. 2019). This indicates that thefirst stars (Pop III stars) likely produced large amounts of C veryearly on (e.g. Meynet et al. 2006).Model predictions have shown that a trace of the Pop IIIstars could be inferred by looking at the [Mg / C] abundancesin the most metal-poor stars where a low value could indicatemono-enriched Population II stars (Hartwig et al. 2018). A sin-gle element ratio is, however, no guarantee that the star is trulymono-enriched, which is why the stellar abundance pattern isoften compared to a set of SN yield predictions with the goalof finding the best fitting model and inferring the mass and en-ergy of the SN progenitor. This way, the initial mass function(IMF) in GCE modelling can be retrieved. Several studies haveattempted to place constraints on the nature of the first stars us-ing 1D, LTE abundance patterns of old, Fe-poor ([Fe / H] (cid:46) − Article number, page 2 of 25. J. Hansen et al.: PEPSI studies (1D LTE versus 3D NLTE, Ca ff au et al. 2011; Kelleret al. 2014; Nordlander et al. 2017), where, for example, verydi ff erent masses were inferred for the SN progenitor yieldingthe birth gas in the Fe-poor star when using 1D, LTE versus 3D,NLTE abundance patterns. From a theoretical perspective, it isextremely important to consider whether or not the mixing anddilution of the SN ejecta into the ISM physically make sense.If the dilution of metals in the ISM is left as a free parameter,in e ff ect, only abundance ratios (patterns) are considered. How-ever, low-yield faint SNe may reproduce the abundance ratios(pattern), but they are still unable to produce su ffi cient amountsof metals to explain the observed (absolute) abundances. Hence,it is equally important to carefully treat the SN model predic-tions, their dilution, and mixing, as well as stellar abundances,since simplified views can easily bias the results (e.g. mass andenergy) by tens of solar masses and explosion energies (10 erg- ‘foe’) (see e.g. Ezzeddine et al. 2019).Here, we shed light on how these types of biases easily occurand how they can bias the conclusions when drawn from thesimple χ fitting of SN yields to 1D LTE abundances, as wellas further a ff ecting the first generation of stars and the IMF inGCE modelling. We also attempt to trace the origin of the gascomposition in more metal-rich stars.In Sect. 2 we describe the observations, sample, and datareduction. Section 3 outlines the stellar parameter determina-tion and related uncertainties, followed by Sect. 4, which detailsthe 1D LTE analysis, the NLTE corrections, and finally the 3DNLTE corrected abundances for a subset of elements. The resultsare laid out in Sect. 5 and they are elaborated on in the discus-sion in Sect. 6, where yields, kinematics, and improved fittingtechniques are described. Finally, our conclusions are presentedin Sect. 7.
2. Sample, observations, and data reduction
The spectra were obtained with the Potsdam Echelle Polari-metric and Spectroscopic Instrument (PEPSI, Strassmeier et al.2015) at the 2 × . µ mfibre (1 . (cid:48)(cid:48) on sky) and the five-slice image slicer. PEPSI hasblue and red arms with three cross-dispersers (CD) in every arm.The échelle image recorded, in each arm, on a 10 . × . µ m pixel size and 16 amplifiers.The spectra were observed in December 2017 and January2018 with the pre-selected number of blue and red CDs withthe two LBT mirrors. Depending on the stellar brightness, thesignal-to-noise ratio (S / N) per CCD pixel averaged over each CDwavelength range is about 360 in the red CD6 (8200 Å) and 190in the blue CD2 (4500 Å) attained in one hour integration time.The spectrograph is located in a pressure-controlled chamber ata constant temperature and humidity to ensure that the refractiveindex of the air inside stays constant over a long-term period andproviding the radial velocity stability at about 5 m s − . The data reduction is done via the Spectroscopic Data Systemsfor PEPSI (SDS4PEPSI) with its pipeline adapted to the PEPSIdata calibration flow and image specific content. It is designedbased on Ilyin (2000) and its recent description is given in Strass-meier et al. (2018). The specific steps of image processing include bias subtrac-tion and variance estimation of the source images, super-masterflat field correction for the CCD spatial noise, a definition oféchelle orders from the tracing flats, and scattered light sub-traction. Then what follows is the wavelength solution for theThAr images, the optimal extraction of image slicers and cos-mic spikes elimination of the target image, wavelength calibra-tion, and merging slices in each order. A normalisation to themaster flat field spectrum is carried out to remove CCD fringesand blaze function. A global 2D fit to the continuum of the nor-malised image and a rectification of all spectral orders in theimage to a 1D spectrum for a given cross-disperser is conductedas well.The spectra from two sides of the telescope are averaged withweights into one spectrum and corrected for the barycentric ve-locity of the Solar System. The wavelength scale is preservedfor each pixel, as given by the wavelength solution without re-binning. The wavelength solution uses about 3000 ThAr linesand has an error on the fit at the image centre of 4 m s − . The sample was selected based on a literature compilation cov-ering the Galactic discs and halo. Hence, we selected relativelybright stars (9 . < V < . ∼ − . < [Fe / H] < − . ff for e ffi ciency and exposuretime, we typically chose setting 2 (CD2) in the PEPSI setup toget a blue setting but with higher e ffi ciency than setting 1 (CD1).Details on the sample, observations, and adopted instrument set-ting can be seen in Table 1. Article number, page 3 of 25 & A proofs: manuscript no. 38805corr_arxiv
Fig. 1.
Various lines from the PEPSI spectra of some of our sample starsshowing normalised flux versus wavelength [Å].Article number, page 4 of 25. J. Hansen et al.: PEPSI T a b l e . O b s e r v a ti on l og f o r t h e P E PS I s p ec t r a . S t a rI D ( A lt e r n a ti v e / R AV E I D ) R A ( J . ) D ec ( J . ) V m a g D a t e T i m e E xp . ti m e S e tti ng W a v e l e ng t h [ Å ] M e d i a n S / N ∗ B D - . − . . / / : : . : : . : ( R AV E J . - ) / / : : . : : . : B D - . − . . / / : : . : : . : / / : : . : : . : / / : : . : : . : / / : : . : : . : B D - . − . . / / : : . : : . : - / / : : . : : . : B D + . + . . / / : : . : : . : - / / : : . : : . : B D - . − . . / / : : . : : . : - ( R AV E J . - ) / / : : . : : . : B D - . − . . / / : : . : : . : - / / : : . : : . : / / : : . : : . : / / : : . : : . : B D - . − . . / / : : . : : . : - / / : : . : : . : / / : : . : : . : / / : : . : : . : B D - . − . . / / : : . : : . : - ( HD ) / / : : : : . : B D + . + . . / / : : . : : . : - / / : : . : : . : / / : : . : : . : / / : : . : : . : HD . − . . / / : : . : : . : - ( R AV E J . - ) / / : : . : : . : H E + a . + . . / / : : . : : . : - / / : : . : : . : T Y C - - . − . . / / : : . : : . : - / / : : . : : . : T Y C - - . - . . / / : : . : : . : - / / : : . : : . : M A SS J - . − . . / / : : . : : . : - / / : : . : : . : ∗ T h e S / N i s g i v e np e r p i x e l a nd t h e ob s e r v i ng ti m e i s i n t h e f o r m a t hh : mm : ss . Article number, page 5 of 25 & A proofs: manuscript no. 38805corr_arxiv
3. Stellar parameters
As the targets have been previously analysed in the literature, weadopt the photometric temperatures from Hansen et al. (2012)and Ruchti et al. (2013), which have been calculated via the in-frared flux method. We then checked for excitation balance andslightly altered the input literature values to achieve this. In mostcases (12 stars), the adopted photometric temperatures almostdirectly led to a balanced excitation potential (labelled ’Teb’ inTable 2). In two of these stars (BD + interface), bolometric correc-tions (BC), and assuming masses of 1 M (cid:12) , we derived log g .Here, the bolometric correction was calculated using our initialtemperature and metallicity ([Fe / H]). This approach is similarto ’method 2’ in Ruchti et al. (2013), and our values are gener-ally in agreement within 0.1–0.2 dex. Only in a few cases doesthe di ff erence reach 0.4. The stars tagged with ’gib’ in Table 2also fulfil ionisation balance where Fe I LTE and Fe II
LTE agreeto within 0.1dex (typically even within 0.05 dex). If the label’(gib)’ is used, the balance is just above 0.1 dex (in LTE).The metallicity, [Fe / H], was based on an average of Fe I andFe II lines. However, due to the spectral ranges we chose in orderto measure heavy element lines, we missed out on several Fe IIlines. As a result, we measured equivalent widths (EW) of 1–3Fe II and 8–30 Fe I lines. Hence, our [Fe / H] is mainly driven byFe I lines, and it may be biased by deviations from LTE, which,however, we corrected for (see Sect. 4.2). The uncertainties listedin Table 2 are the errors on the mean.The microturbulence ( ξ ) was fixed by requiring that all Felines yield the same abundance, regardless of EW. In three stars,we only measured a few Fe lines that were distributed, such thatoptimising a linear trend was hard to achieve, so we adoptedthe empirical scaling from Mashonkina et al. (2017a) to obtainthe microturbulence in these three cases (labeled with a ’v’ inTable 2). Uncertainties
The uncertainties on the temperatures were estimated based onthe residual slopes when attempting to obtain a perfect excita-tion potential balance with zero slope. For the two stars, wherethe excitation potential balance was not achieved, we adopted anuncertainty of ∼
100 K, which is in agreement with Ruchti et al.(2013).The main source of error in our gravities originates in theparallaxes. We computed the total error by varying the parallax,the initial temperature, and the metallicity by their respective er-rors and we computed new gravities. This change was adoptedas the error on the gravity (see Table 2). For the most distant star(TYC 5329), which is just beyond 3 kpc, we have listed a slightlylarger uncertainty (as indicated by the ’*’ in Table 2) in order toaccommodate uncertainties on parallaxes and distances and tonot simply treat the latter as a ‘1 / parallax’. Using the probabili-ties from Bailer-Jones et al. (2018) and their Bayesian distance This thereby ensures that all Fe lines yield the same Fe abundance,regardless of excitation potential. https: // irsa.ipac.caltech.edu / applications / DUST / computation of this star results in a slightly higher gravity (by0.12 dex), which we have taken into account in the listed uncer-tainty. Most of the stars are within ∼ . ∼ , hence possible distance dis-crepancies are well accounted for by the associated error, whichwe used to estimate the uncertainty on the gravity.For the metallicity, we adopted the line-to-line abundancescatter as the statistical error on [Fe / H]. In cases where an ioni-sation balance is not achieved, the Fe II lines increase this errorslightly.In the case of the microturbulence, ξ , the uncertainty in theslope was adopted as the uncertainty on the value (i.e. the de-viation from a perfect zero slope). Additionally, for three starswhere the empirical scaling relation was used, we varied the in-put parameters ( T e ff , log g , and [Fe / H]) by their respective uncer-tainties and errors.
4. Abundance analysis
Our present analysis consists of both EW measurements andspectrum synthesis. We present the abundances derived underthe assumptions of 1D LTE, as well as NLTE and 3D correctedabundances. The EWs also serve as an extra check when com-puting the NLTE corrections. First we describe the LTE analysis(Sect. 4.1) and in Sect. 4.2 we outline the details of the NLTEcorrections. In addition, we present 3D corrections for C, O, andFe in Sect. 4.3.
The 1D LTE abundances were derived using MOOG (Sneden1973, version 2014) and MARCS atmosphere models (Gustafs-son et al. 2008, adopting interpolated plane parallel and spher-ical models for dwarfs and giants, respectively). The line list isprovided online on CDS . Relative abundances were calculatedusing the solar values from Asplund et al. (2009).As a first attempt, we measured all EWs in IRAF by fit-ting Gaussian or Voigt profiles to the absorption features. How-ever, due to the broad range in stellar metallicity and signal-to-noise ratios of the PEPSI spectra (8 < S / N < – . > FWHM
Gaussian or FWHM
Gaussian > . / blends), – | λ measured line − λ line centre | > .
05 Å, – EW >
250 mÅ (saturation), and – log ε line − (cid:104) log ε (cid:105) > ± . λ < i , C i (CH), O i , Na i , Mg i , Si i+ii , S i , K i , Ca i ,Sc ii , Ti i+ii , V i+ii , Cr i , Mn i , Fe i+ii , Co i , Ni i , Cu i , Zn i , Rb i , The parallax error is between 0.5 and 12.8% with the majority below10%. See CDSArticle number, page 6 of 25. J. Hansen et al.: PEPSI
Table 2.
Stellar parameters of the sample. Comments: Here ’Teb’ indicates the excitation equilibrium, ’gib’ is the ionisation equilibrium, and theuncertainty on A(Fe) is the error on the mean absolute Fe abundance (st.dev / √ N lines ). R13 refers to Ruchti et al. (2013) and H12 to Hansen et al.(2012). The ’*’ indicates that gravity is more uncertain due to the larger distance to the star. ID T e ff [K] log g [dex] A(Fe) LTE [Fe / H] LTE ξ [km / s] commentBD-01_2439 5288 ±
100 2 . ± .
09 6 . ± . − .
09 1 . ± . ±
50 2 . ± .
23 6 . ± . − .
44 2 . ± . ±
100 4 . ± .
05 6 . ± . − .
10 1 . ± . + ±
50 4 . ± .
05 4 . ± . − .
75 1 . ± . ±
120 1 . ± .
04 5 . ± . − .
96 1 . ± . ±
50 2 . ± .
05 5 . ± . − .
11 1 . ± . ±
100 1 . ± .
14 6 . ± . − .
48 1 . ± . + ±
50 4 . ± .
14 4 . ± . − .
53 1 . ± . ±
50 2 . ± .
08 6 . ± . − .
89 1 . ± . ±
50 4 . ± .
02 6 . ± . − .
95 1 . ± . + ±
50 2 . ± .
05 5 . ± . − .
45 1 . ± . ±
50 1 . ± .
12 5 . ± . − .
14 1 . ± . ±
90 1 . ± .
11 6 . ± . − .
48 1 . ± . ±
100 3 . ± .
09 5 . ± . − .
40 1 . ± . Fig. 2.
Spectrum synthesis of C, Sm, (Nd), Dy, and Rb in various sample stars. Specifically, molecular and atomic C in TYC5481. We note that[C i / Fe] = ± .
1; [CH / Fe] = . ± .
1; in BD-0.1_2439, [Sm / Fe] = . ± . / Fe] = / Fe] = . ± .
1; and in HD136343, [Rb / Fe] = . ± .
1. In all cases, the greendashed line indicates [X / Fe] = − Sr ii , Y ii , Zr ii , Ba ii , La ii , Ce ii , Pr ii , Nd ii , Sm ii , Eu ii , Gd ii , andDy ii (Pb ii and Th ii ). The abundances of each element (neutralor ionised species) can be found in the online Table and the linesthat were synthesised or if the Gaussian fit EWs were used toderive the final abundances are flagged. A few examples of spec- trum line syntheses are shown in Fig. 2. For consistency with theNLTE analysis, we only used lines which can be NLTE correctedas well in order to ensure a better and more equal foundationfor comparing the LTE versus NLTE abundance behaviour. Weonly deviated from this criterion for three abundances because Article number, page 7 of 25 & A proofs: manuscript no. 38805corr_arxiv
Table 3.
NLTE atomic models used in this study.
Species Reference H i collisionsC i * Amarsi et al. (2019a,c) AKO i * Przybilla et al. (2000),Sitnova & Mashonkina (2018) BVM19Na i Alexeeva et al. (2014) BBD10Mg i - ii Bergemann et al. (2017a) BBS12Si i - ii Mashonkina (2020) BYB14Ca i - ii Mashonkina et al. (2007, 2017b) BVY17Sc ii Zhang et al. (2008) SH84 (0.1)Ti i - ii Sitnova et al. (2016) SYB20Cr i - ii Bergemann & Cescutti (2010) SH84 (0.0)Mn i - ii Bergemann et al. (2019) BV17, BGE19Fe i - ii Mashonkina et al. (2011, 2019) YBK18,YBK19Co i - ii Bergemann et al. (2010) SH84 (0.05)Zr ii Velichko et al. (2010) SH84 (0.1)Nd ii Mashonkina et al. (2005) SH84 (0.1)Ba ii Gallagher et al. (2020) BY17, BY18Eu ii Mashonkina & Gehren (2000) SH84 (0.1)Pb i Mashonkina et al. (2012) SH84 (0.1)Th ii Mashonkina et al. (2012) SH84 (0.1)
Notes.
Collisions with H i were treated following AK (Amarsi et al.2019a; Kaulakys 1991), BVM19 (Belyaev et al. 2019), BBD10(Barklem et al. 2010), BYB14 (Belyaev et al. 2014), BVY17 (Belyaevet al. 2017), SH84 (Zhang et al. 2008), SYB20 (Sitnova et al. 2020),YBK18 (Yakovleva et al. 2018), YBK19 (Yakovleva et al. 2019),BGE19 (Bergemann et al. 2019), BV17 (Belyaev & Voronov 2017),BBS12 (Barklem et al. 2012), BY17 (Belyaev & Yakovleva 2017),BY18 (Belyaev & Yakovleva 2018), and SH84 (0.1) (Steenbock & Hol-weger 1984) with a scaling factor of S H = S H = the average remains the same whether we used all or the reducednumber of lines. The present investigation is based on the NLTE methods devel-oped in our earlier studies and documented in a number of papers(see Table 3 for the references) in which the atomic data and thequestions on line formation have been considered in detail. Fora number of chemical species, their model atoms were updatedby employing quantum-mechanical rate coe ffi cients for inelasticprocesses in collisions with neutral hydrogen atoms.We briefly describe the departures from LTE for the investi-gated lines. Figure 3 displays the NLTE abundance corrections, ∆ NLTE = log ε NLTE − log ε LTE , for some representative lines,namely, O i i i i ii i ii ii ii i ii , either with the Detail code (Butler & Giddings 1985) orMULTI2.3 (Carlsson 1986). For a few elements, Mg, Cr, Mn, and Co, we used MAFAGS mod-els instead; however, the di ff erence between the models was tested andshown to be minor, see, e.g. Hansen et al. (2013) and a detailed com-parison of the models in Bergemann et al. (2012, 2019). Fig. 3.
NLTE abundance corrections (dex) for representative lines ofthe NLTE species in the sample stars. O i . Our NLTE calculations show strengthened O i ff erence between thisstudy and Sitnova & Mashonkina (2018) lies in using recent dataon the O i + H i collisions from Belyaev et al. (2019) instead ofthose from Barklem (2018). This update leads to slightly smallerNLTE e ff ects and smaller magnitude of ∆ NLTE , for example, by0.02 dex in the model atmosphere with T e ff = = / H] = − .
1. In the same model atmosphere, we find ∆ NLTE = − . − .
14, and − .
13 dex for O i − . / H] (cid:39) − − . / H] < −
2. The stronger O i ∆ NLTE is (for 3D NLTE see Sect. 4.3.). Na i . The NLTE e ff ects are rather large, with ∆ NLTE rangingbetween − . − . i is subject toover-recombination in the atmospheres of late-type stars, result-ing in strengthened lines and negative NLTE abundance correc-tions. For the four stars BD-07_163, BD-08_619, BD-12_106,and HD136343 with both Na i σ = (cid:112) Σ ( x − x i ) / ( N l − N l isthe number of measured lines. For example, for BD-07_163, σ = .
11 and 0.03 dex in LTE and NLTE, respectively. TheNLTE e ff ects decrease towards lower metallicity due to weaken-ing of the lines. Mg i . The NLTE corrections to the Mg lines (4167.3, 4351.9,4571.1, 4703.0, 5172.7, 5183.6, and 5528.4 Å) originate inBergemann et al. (2017a). The neutral atom is dominated byphotoionisation and hence it is prone to overionisation by a non-local blue radiation field. In turn, this weakens the spectral lineand leads to a positive correction (see also Fe-peak elements be-low for more details). For instance, some lines, such as the high-excitation Mg i lines, are only a ff ected at the level of − .
03 to + .
08 dex (similarly for the Mg b triplet in TYC5329-1927-1).In the specific case of BD-08_619 (with [Fe / H] ≈ − i .These corrections slightly increase with decreasing metallicity to0.06 for Mg i in HE0420 + / H] ∼ − . Article number, page 8 of 25. J. Hansen et al.: PEPSI Si i - ii . In the stellar parameter range, with which we are con-cerned, number densities of neutral and singly ionised siliconhave comparable values in the line-formation layers. In a compe-tition for enhanced photoionisation of the low-excitation ( E exc < i with a photon suction caused by bound-boundtransitions from many levels close to the ionisation limit downto the lower levels, the latter prevails and increases the popula-tions of the ground state and low-lying ( E exc ≤ . i . This results in a strengthened Si i ∆ NLTE of − .
19 to − .
12 dex in di ff erent stars. In BD-15_779,Si ii ∆ NLTE = − .
03 dex. K i . The NLTE e ff ect of potassium is dictated by the sourcefunction and caused by resonance scattering. Similar to thesodium D lines, an overpopulation of the ground states shifts theline formation slightly outwards, which deepens the lines. Thismeans that the e ff ect is governed by the radiation field and rates(Reggiani et al. 2019). Here we interpolate in their grid of NLTEK corrections over all stellar parameters including EW. In do-ing so, we adopt a multi-D linear interpolation, as the grid is notevenly spaced in abundances and EW. We find a slight o ff set inK abundances between our study and Reggiani et al. (2019), andwe furthermore find slightly higher corrections (by ∼ .
12 dex)if we interpolate in LTE abundances rather than EW. This dif-ference is, however, negligible at metallicities below ∼ −
2. Theatomic data (oscillator strength and excitation potential) in thisand the analysis of Reggiani et al. (2019) are identical but theremight be slight di ff erences in the spectrum synthesis code, modelatmospheres, and possible damping treatment. We chose to inter-polate in EW, even if we ended up slightly underestimating theK NLTE corrections. Since we relied on EW, we mainly used the7698 Å line, as the 7664 Å line has a silicon blend. However, fortwo stars (BD-08_619 and BD-12_106), we were forced to usethe 7664 K line since the telluric A-band is interfering with the,otherwise cleaner, 7698 K line. For BD-08_619, we find a cor-rection of − .
15; while for the more metal-poor HE0420 + − .
31 dex. Ca i . As shown in previous studies (see, for example,Mashonkina et al. 2007), Ca i is subject to the ultraviolet (UV)overionisation in the atmospheres of late-type stars. The overion-isation tends to weaken the lines. However, in mildly metal-poorstars, another NLTE mechanism is working in the opposite direc-tion. This e ff ect is the lowering of the line source function ( S ν )below the Planck function ( B ν ) in the uppermost atmosphericlayers, where the cores of strong Ca i lines form, and it tendsto make the lines stronger. In a given star, ∆ NLTE is positive forweak lines, which form in the layers a ff ected by the overioni-sation, and it can be negative for strong lines. The net e ff ect isthat the di ff erence between average NLTE and LTE abundancesis slightly negative for the [Fe / H] > − . Ti i - ii . Compared with Sitnova et al. (2016), the model atomwas updated by implementing quantum-mechanical rate coe ffi -cients for the Ti i + H i and Ti ii + H i collisions. Calculations ofcollisional data and their impact on the NLTE results are pre-sented by Sitnova et al. (2020). The NLTE computations leadto weakened lines of Ti i and positive NLTE abundance correc-tions, which vary from 0 to 0.23 dex, depending on the spectralline and stellar parameters. For Ti ii , ∆ NLTE is negative (downto − .
13 dex) for strong lines with an equivalent width of EW >
80 mÅ, but it is positive (up to 0.07 dex) for weak lines. To cal-culate average titanium abundances, we employed lines of Ti ii , as recommended by Bergemann (2011), Sitnova (2016), and Sit-nova et al. (2020). Fe-peak (Cr i , Mn i , and Co i ): The NLTE corrections forthe lines of Cr, Mn, and Co were computed based on modelatoms of Bergemann & Cescutti (2010), Bergemann et al.(2019), and Bergemann et al. (2010), respectively. All of theseelements can be observed in neutral and singly-ionised stage;however, in our spectra, only lines of neutral species could beused as a diagnostic, owing to the limited wavelength rangeof the spectra. The neutral atoms of all three elements arephotoionisation-dominated ions (see Bergemann & Nordlander2014 for a detailed discussion on the physics behind NLTE),which means that they are sensitive to overionisation that isdriven by the non-local high-energy (near-ultraviolet to blue)radiation field. This generally leads to weakening of the low-excitation spectral lines of these species, and, therefore, to posi-tive NLTE abundance corrections. In other words, the LTE anal-ysis underestimates the abundances of these elements. However,the amplitude of NLTE corrections di ff ers amongst the spectrallines of the same element. Some spectral lines, such as the MnI resonance triplet at 4030 Å, show NLTE corrections of up to0 . e ff . For an F-type main-sequence star with [Fe / H] ≈ − , such as BD-08_619, we obtain, on average, only modest NLTEcorrections of 0.07 for Cr i and 0.19 for Mn i . The NLTE valuesfor Mn I are comparatively high because only resonance lines areavailable to us and these are very sensitive to the e ff ects of NLTEand 3D inhomogeneities (Bergemann et al. 2019). On the otherhand, the average estimates of NLTE corrections for an RGBstar with a metallicity of [Fe / H] ≈ − . + i and 0.43 dex for Mn i . Fe i - ii . NLTE mechanisms for Fe i are very similar to those ofCa i , resulting, in most cases, in weakened lines of Fe i and pos-itive NLTE abundance corrections, which grow towards lowermetallicity (see, e.g. Bergemann et al. 2012, for a discussion).In one of our most metal-poor stars, BD + ∆ NLTE variesfrom 0.11 to 0.35 dex for di ff erent Fe i lines. In mildly metal-poor stars, ∆ NLTE can be slightly negative (of − .
05 dex) forsome strong Fe i lines. We note that Fe ii is a majority speciesin the atmospheres of our sample stars, and the NLTE e ff ectsare, in general, minor for lines of Fe ii : ∆ NLTE does not exceed0.02 dex, in absolute value, for 4233, 6238, and 6247 Å, and itreaches a maximal value of − .
10 dex for Fe ii Zr ii , Nd ii , Eu ii , and Th ii . The ionised species are the ma-jority ones for these corresponding chemical elements and theyare subject to similar NLTE mechanisms. The investigated spec-tral lines arise in the transitions from the low-excitation ( E exc < ff ectsis small. In a given star and for a given species, the di ff erence in ∆ NLTE between di ff erent lines amounts to 0.02 to 0.10 dex. Thestrongest line exhibits the maximal value for ∆ NLTE , however, itdoes not exceed 0.16, 0.16, 0.08, and 0.08 dex for Zr ii ii ii ii Ba ii . We computed new 1D NLTE corrections for the stars inour sample using the Ba model atom described in Gallagher et al.(2020). Deviations from LTE are caused by strong-line scatteringand radiative pumping, causing a level population of Ba that isfar from LTE. On average, we obtain small NLTE corrections
Article number, page 9 of 25 & A proofs: manuscript no. 38805corr_arxiv ( − .
07 dex) for the 4554 Å line in our sample and slightly highercorrections for the 6496 Å line. Pb i . Lead is strongly ionised in the atmospheres of the sam-ple stars, and the UV overionisation is the main NLTE mecha-nism for Pb i , resulting in depleted absorption in the Pb i −3.0 −2.5 −2.0 −1.5 −1.0 −0.5−0.40−0.200.000.200.40 −3.0 −2.5 −2.0 −1.5 −1.0 −0.5[Fe/H]
1D LTE −0.40−0.200.000.200.40 A bundan c e − D L T E
1D non−LTE3D non−LTE C IO IFe II (3D LTE)
Fig. 4.
Line-averaged C i , O i , and Fe ii abundance corrections (1D non-LTE −
1D LTE, and 3D non-LTE −
1D LTE, abundance di ff erences)for the sample of stars where available. The Fe ii results presented herestrictly assume zero departures from LTE. Vertical lines between twodata points indicate that they correspond to the same star. The atmospheres, and thus the emergent spectra, of late-type stars are susceptible to the e ff ects of convection occurringjust below the visible surface. These e ff ects can be accountedfor through the use of ab initio 3D hydrodynamic stellar atmo-sphere simulations (e.g. Collet et al. 2007; Freytag et al. 2012).So-called 3D NLTE methods arguably allow for the most realis-tic spectral synthesis and thus most reliable abundance analysis.However, these methods are extremely computationally expen-sive; additionally, for many chemical species, the lack of accu-rate atomic data makes pursuing this approach even more chal-lenging. Hence, to date, 3D NLTE abundance analyses remainsparse.Nevertheless, while the majority of the analysis presented inthis paper is based on 1D NLTE methods (Sect. 4.2), 3D NLTEabundance results are available for a handful of important chemi-cal species. Amarsi et al. (2019b,c) recently presented 3D NLTEabundance corrections across the stagger -grid of 3D model stel-lar atmospheres (Magic et al. 2013) for the chemical species C i and O i as well as 3D LTE abundance corrections for Fe ii .In Fig. 4 we show the line-averaged 3D NLTE abundancecorrections (3D NLTE −
1D LTE abundance di ff erences) for C i and O i , based on the grids from Amarsi et al. (2019c), interpo-lated onto the sample of stars. Unfortunately, these abundancecorrections are only available for the dwarfs and subgiants in thesample (only four out of the 14 stars). For comparison, for C i and O i , we also show 1D NLTE abundance corrections that werecalculated using the same atomic models for the entire sample. Table 4.
Absolute, 3D LTE or 3D NLTE abundances for Fe, C, and O(and their standard deviation).
Star ID A(Fe) , LTE
A(C) , NLTE
A(O) , NLTE
BD-08_619 6.74 (0.10) 8.21 (0.10) 8.50 (0.03)BD + + ii for whichabundance corrections are available for the entire sample.The 3D NLTE abundance corrections for C i are slightly, butsignificantly dissimilar from the 1D NLTE abundance correc-tions (from the spectrum modelling perspective), at least accord-ing to these particular models and for the particular four dwarfsand subgiant shown in Fig. 4. The mean absolute di ff erence be-tween the 3D NLTE and 1D NLTE abundances is 0 .
06 dex, andthe largest discrepancy is 0 .
10 dex for BD-08_619. These dif-ferences are much smaller for O i : The mean absolute di ff erencebetween the 3D NLTE and 1D NLTE abundances is 0 .
01 dex,and the largest discrepancy is 0 .
02 dex, again for BD-08_619. Itis unknown how large the 3D NLTE e ff ects are for these species,for the giants in the sample.The 1D NLTE abundance corrections for C i and O i are neg-ative for these stars: The 1D NLTE abundances are smaller thanthe 1D LTE abundances. For O i , the 3D NLTE abundance cor-rections (with respect to 1D LTE) are also all negative: The 3De ff ect and the NLTE e ff ect go in the same direction (Amarsi et al.2016a). For C i , however, the 3D NLTE abundance correctionscan be positive: The 3D e ff ect and the NLTE e ff ect can go inopposite directions (Amarsi et al. 2019a).The 3D LTE abundance corrections for Fe ii shown in Fig. 4can be larger than the ’typical’ upper limit of + .
15 dex reportedin Amarsi et al. (2019c). For most stars in our sample, applyingthese corrections to our 1D LTE abundances from Fe ii wouldtypically worsen the agreement with Fe i (i.e. the ionisation bal-ance). There are several possible reasons for this. For one, thismay reflect neglected 3D NLTE e ff ects in Fe i lines (Amarsi et al.2016b), keeping in mind that LTE is expected to be a fairly goodapproximation for Fe ii (Lind et al. 2012). Another possibility isthat it reflects issues with the adopted microturbulence relation.We note that the Fe ii lines used in this study are rather strong,with reduced EWs (log(EW / λ )) of greater than − i andO i in this work. We use 1D NLTE abundances for the remainingspecies (including Fe) discussed in Sect. 4.2, and 1D LTE abun-dances otherwise (see Figs.3–4, and A.1). Hence, the 3D NLTEpattern shows two elements, the 1D NLTE up to 17 elements,and finally the 1D LTE pattern includes 32 elements (34 if lim-its are counted as well). The various methods are never mixedwithin a pattern.
5. Results
The sample spans a broad range of metallicities, and we there-fore briefly discuss the trends in the grand scheme of GCE. Foreach element in Fig. 5 and 6, we compare the element to liter-ature samples (Cayrel et al. 2004; François et al. 2007; Roed-erer et al. 2014); furthermore, to illustrate the di ff erence in LTE Article number, page 10 of 25. J. Hansen et al.: PEPSI and NLTE trends, we overplotted a locally weighted scatter-plotsmoothed (LOWESS) line.Starting with our lightest element Li, we find a value ofA(Li) = + ∼ + / Fe] (cid:46) . / Fe] almost classifies as a CEMP star, but it would morelikely be a CH star owing to its [Fe / H]. TYC5329 seems to showenhancements in the heavy elements, typical of a CEMP-s or -r / sstar, but we could not derive C from this star due to the wave-length coverage. In all dwarfs and subgiants, the 1D LTE abun-dances are larger than the corrected 1D NLTE values, while the3D NLTE corrected abundances in three out of four stars almostbring the C-abundances from the atomic red lines ( > + / Fe] ≤ . / Fe] > . ff erence between the atomic and molecu-lar C abundances are ∼ .
45, with the molecular values alwaysbeing lower. A part of the discrepancy could be due to modellingissues or slightly uncertain stellar parameters, which propagatethrough and a ff ect the atomic and molecular lines di ff erently. Wenote that the NLTE corrected C i agrees with 1D LTE CH within ∼ ± . − . / H], but also on the absolute C abundance (A(C)),where lower A(C) stars are more a ff ected. The 3D A(C) correc-tions to CH range from 0 to − . ff ected. However, in order to get our atomic A(C)in 3D NLTE to agree with the future 3D NLTE molecular values,the future NLTE corrections to CH would need to be large andpositive (0 . . ff set between molecular and atomic C abun-dances.Oxygen behaves like a typically enhanced α − element andfollows the well-known trends as shown in Cayrel et al. (2004),for example. The NLTE corrections clearly lower the averagevalue for O, as seen in Figs. 3–4, and A.1. In the latter figure(A.1), there is a di ff erence in the 1D NLTE abundances of O,which likely arises due to the use of di ff erent codes and treat-ment of collision rates. As shown for Mn in Fig. 10 of Berge-mann et al. (2019), the use of Kaulakis collisions can reduce or enhance the NLTE correction by ∼ ± .
05 dex. The overall NLTEtrend is in agreement with that of Zhao et al. (2016), and we seea hint of a knee, but we also have a few stars with low NLTEO values around [Fe / H] ∼ − .
5. These are discussed further inSect. 6.2.Sodium shows a gradually increasing trend with metallic-ity, except for BD-10_3742, which is very enhanced in Na bya factor of 10 compared to the Sun in LTE which is reduced to[Na / Fe] = / O ra-tio could also indicate a very evolved giant that has undergonesevere stellar mixing events (see, e.g. Placco et al. 2014b). Gen-erally, the star-to-star scatter is fairly large and the NLTE correc-tions clearly reduce the average Na value. The large scatter wasalso shown in Zhao et al. (2016).Both Mg and Si (neutral and ionised) show typical enhancedlevels in metal-poor stars. A few peculiar cases show very high[Mg / Fe]
LTE ∼ . ∼
0) valueshave been derived for TYC5481-00786-1, 2MASS J0023, andHE0420 + + / Fe] of ∼ . / N archival UVESspectra yield a considerably lower Mg abundance by 0 . − . ff au et al. (2017), our trendagrees well with theirs and spans broadly from [S / Fe] ∼ . − . ff ortsto avoid blends, a few cases might su ff er from their e ff ect, return-ing values slightly above literature values (Cayrel et al. 2004;Reggiani et al. 2019). However, below [Fe / H] = −
2, our valuesdrop ([K / Fe] < .
5) and they are in good agreement with thelisted literature studies. In comparison to Zhao et al. (2016), wefind a larger spread and a higher NLTE abundance average for K.This could indicate that unnoticed blends might interfere moreat higher metallicities. The GCE behaviour is clearly di ff erent inLTE and NLTE (see Fig. 5).Calcium shows a clean trend with little scatter as is generallyseen in the literature. We note that one star has a particularlylow [Ca / Fe] (see Sect. 6–Sect. 6.2). Similarly to Ca, Sc showsa flat trend with low star-to-star scatter, which is also the casefor neutral and ionised Ti. This is in agreement with the NLTEresults from Zhao et al. (2016); however, we do not detect anupturn in Sc around [Fe / H] = − ∼ . ii and Ti ii , which is well withinthe uncertainty. We note that the average Ti i and Ti ii di ff er inmost cases, as we did not enforce ionisation equilibrium, unlessthe gravities determined via parallaxes happen to satisfy this.Vanadium was derived in five stars and an additional threeshow upper limits. The trend is very spread with metallicity. Thiswas also noted in Roederer et al. (2014) who studied V i and ii as a function of temperature and [Fe / H]. Similar to them, we findhigher V ii than V i (see ’x’ in Fig. 5). They also note the lack ofatomic data for hyperfine splitting (hfs) for V ii ; however, if wecould implement the missing, hyperfine split oscillator strength,the abundances of V ii would likely decrease. The magnitude ofthe e ff ect also depends on the strength of the lines, which in some Article number, page 11 of 25 & A proofs: manuscript no. 38805corr_arxiv
Fig. 5.
Relative abundances [X / Fe] versus [Fe / H] of element (X) up to Mn. LTE abundances are illustrated by black filled symbols, while NLTEabundances are shown as red symbols, with giants depicted as circles and dwarfs as squares. The ’x’ in the vanadium panel indicates stars forwhich we measured VII rather than VI. LOWESS trends are indicated for LTE (black solid line) and NLTE (red dashed line). In blue, we show theabundances from Cayrel et al. (2004) and Roederer et al. (2014).Article number, page 12 of 25. J. Hansen et al.: PEPSI
Fig. 6.
Relative abundances [X / Fe] vs. [Fe / H] of elements X from Co up to Dy. LTE abundances are illustrated in black filled symbols, whileNLTE abundances are shown as red open symbols, with giants depicted as circles and dwarfs as squares. Locally weighted scatter-plot smoothertrends are indicated for LTE (black solid line) and NLTE (red dashed). In blue, we plot the data from Cayrel et al. (2004), François et al. (2007),and Roederer et al. (2014). Article number, page 13 of 25 & A proofs: manuscript no. 38805corr_arxiv
Fig. 7.
Absolute C abundances versus [Fe / H] (LTE). Indications ofCEMP-no (blue line - low A(C)), CEMP-s (red line - high A(C)), and C-normal metal-poor stars (separated by the green line [C / Fe] = + stars amount to ∼
100 mÅ and may cause significant changes inthe abundances if hfs could be included. In both cases, a partof the explanation could be that three out of the four stars withhigh V ii do not show ionisation equilibrium. We note that Ti i and ii also exhibit a similar trend with higher Ti ii values. Thedi ff erence in Ti i and ii can, in part, be explained by the Ti i valuebeing driven by one or two lines.In LTE, both Cr and Mn show subsolar trends ([X / Fe] < / Fe] and [Mn / Fe] are higher. The 3D NLTE abundanceswould be even higher than the abundances derived using the 1DNLTE approach (as also seen in Eitner et al. 2020; Bergemannet al. 2019).Unlike the incomplete Si-burning elements, Co, which isproduced in the complete Si burning, shows slightly enhancedabundances ( ∼ . ff au et al. (2017).The first heavy element ( Z >
30) we analyse is Rb. The abun-dances are based on the 7800 Å line, which is hard to analyseaccurately as it su ff ers from a strong Si-blend and possible tel-luric contamination as well. Few stellar abundances have beenpublished and we add one detection and two upper limits to theliterature, all of which are slightly enhanced.The three elements Sr, Y, and Zr show the typical trendswith Sr and Zr abundances that are higher than those of Y. TheGCE trends are in good agreement with François et al. (2007)and Hansen et al. (2012, 2013). Owing to the lack of observa-tions in the blue-most PEPSI CD1 setting, only four stars havewavelength coverage of the Sr lines. For Zr, we find an excel-lent agreement with the NLTE Zr abundances from Zhao et al.(2016).The heavy elements between Ba and Sm show a large star-to-star scatter (François et al. 2007; Hansen et al. 2012). Wehighlight that Pr and Sm show particularly flat trends with lowerscatter; however, this could be due to our limited samples size.Similarly, Gd–Dy seem to yield a flat trend at a slightly enhanced level. Generally, the heavy elements are derived from the major-ity species and are hence less biased by the LTE approximation.Therefore, the LTE and NLTE abundances of the four heavy el-ements NLTE corrected here (Zr, Ba, Nd, and Eu) show similarGCE behaviours. The NLTE study by Zhao et al. (2016) showsthat Ba was widely spread around zero, which is similar to whatwe find. Also our Eu NLTE abundances are in good agreementwith their trend.Finally, we report upper limits of Pb ([Pb / Fe]
LTE (cid:46) .
8) andTh ([Th / Fe]
LTE (cid:46) .
5) in BD-07_163. Additionally, we note thatthese values are very uncertain and we refrain from using themfurther.
6. Discussion
Here, we discuss stars with peculiar abundance ratios and fo-cus on their heavy element patterns. To explore the origin ofthe neutron-capture elements, we compare their abundances toAGB yield predictions (Cristallo et al. 2011, 2015) and assesss-process contamination at higher metallicities. We explore ther-process using stars with stellar parameters that are represen-tative of our sample with a well-studied r-process pattern. Forthis purpose we use HD20, a metal-poor giant ([Fe / H] = − . (cid:15) ( X ) + log (cid:15) ( X r ), where ther-fraction is taken from Table 5 in Hanke et al. (2020). This en-ables a purer comparison to the r-process at a lower metallicitythan the Sun. Hence, we use the pattern of HD20 as a ’metal-poor Sun’ with a more pure r-process representation, owing toits lower metallicity and r-pattern instead of comparing it to abiased solar scaled r-process pattern.The most metal-rich stars in our sample, [Fe / H] > −
1, areclearly enriched from a highly mixed gas by numerous di ff erentevents (see Sect.6.3). Hence, we only attempted to trace AGBand r-process contributions in stars with [Fe / H] ∼ − . χ between the obser-vations and the yields, using the line-to-line scatter listed in theonline Table as uncertainty. From Fig. 8 and the computed χ , wefind that BD-07_163 and especially TYC5481 are strongly pol-luted by the s-process. None of the models that are contrastedhere provide a good fit to all of the derived abundances. In gen-eral, the heavier elements tend to show an upward trend in Fig. 8.This is somewhat driven by Dy, which from Fig. 6 is also seento be high, so we cannot exclude that we overestimated our Dyabundances a bit. However, based on spectrum synthesis, we es-timate that this is on the 0 . (cid:12) ), fast rotating (V r =
30 km / s), metal-rich([Fe / H] = − .
1) AGB star. Both BD-07_163 and TYC5481 showheavy element contamination that is better fitted by the moremetal-poor AGB star, so the metallicity of BD-07_163 may pre-vent a direct trace of the s-process enrichment and it may bemulti-enriched. However, TYC5481 with an A(C) = / H] ∼ − / Eu] = − .
1, using La as a proxy for Ba and intaking into account that La is an odd element which typicallyyields lower abundances than the even elements (Ba), this staris likely a CEMP-r / s star. This also explains why neither thepure s- nor r-patterns provide a satisfactory fit on their own tothe observations of TYC5481. Meanwhile, the metal-poor gi- Article number, page 14 of 25. J. Hansen et al.: PEPSI
Fig. 8.
Abundance patterns in 1D + LTE, 1D + NLTE compared to thepure r-process representation of HD20 (LTE), and the s-process yieldsfrom the rotating AGB stars with 2 and 5 M (cid:12) and [Fe / H] = − .
1, and − .
1, respectively. ant BD-10_3742 shows a pattern that strongly resembles the r-fraction of HD20, and it likely has predominantly been enrichedin the heavy elements by the r-process. Our metal-poor dwarfstar, BD + Unfortunately, neither Sr nor Eu could be derived from the spec-tra of 2MASS J0023, so classifications adopted in Beers &Christlieb (2005) or Hansen et al. (2019) cannot be applied di-rectly. However, with both [Y / Fe] < − . / Fe] = − . / Fe] (in LTE) and heavy element con-tent of TYC5481 likely mean that it is a CEMP-r / s star; however,the subclassification is more uncertain since we are missing key Fig. 9.
Abundance patterns in 1D + LTE, 1D + NLTE, and 3D + NLTE of2MASS J0023. elements (Ba) for this purpose. TYC5329 could be a CEMP-sstar due to the s-process enhancement in this star; however, wecould not derive C here. Also BD-08_619 is C rich ([C / Fe] = / Fe], and further C corrections are needed tofully understand this star. However, with the higher metallicity,it more likely belongs to the CH group than the CEMP stars.Finally, we note a broad spread in [ α / Fe] with several low- α stars and a few above 0.5. In Sect. 6.2, we explore their kinemat-ics in order to probe if the stars could be accreted into the MilkyWay (see e.g. Nissen & Schuster 2010; Bergemann et al. 2017b;Hansen et al. 2019). Low- α halo stars have likely been accretedfrom dwarf galaxies, which, compared to the MW, o ff er less gasand in turn form a larger population of lower-mass supernovae,thereby explaining the lower α − abundances (see e.g. McWilliamet al. 2013; Reichert et al. 2020, for di ff erent views). For a sub-sample of the chemically most intriguing stars withhigh or low α − content and for the (candidate) CEMP stars(BD + + V indicates retrograde orbits and T denotes the com-bined vertical and radial components relative to the local stan-dard of rest (LSR). As three of the stars show deviations fromthe LSR of more than 210 km s − , we consider them to be typi-cal halo objects (Koppelman et al. 2018). Two of the stars, BD-15_779 and TYC5481, at intermediate LSR velocities between ∼
100 – 200 km s − could be assigned to the metal-poor tail ofthe thick disc (Kordopatis et al. 2013) with their [Fe / H] ∼ − . Z max < ∼ . Article number, page 15 of 25 & A proofs: manuscript no. 38805corr_arxiv -500 -400 -300 -200 -100 0 100 200 300
V [km s -1 ] T [ k m s - ] Fig. 10.
Toomre diagram for a subset of our sample (orange circles) incomparison with the sample of CEMP-no (blue squares) and C-normalmetal-poor halo stars (black diamonds) from Hansen et al. (2015b,2016b,a). Only stars BD + + − , centred on V LSR =
232 km s − . e > R apo <
13 kpc, and R peri > + α − content hinting at a mono-enrichment event (see Sect. 6.3-6.4). At an apocenter distance of 62 kpc and an orbital ellipticityof 0.82, it is an excellent candidate for an outer halo object. Withthe Lindblad diagram in Fig. 11, we attempt to identify a pos-sible progenitor. Here, we show the total specific orbital energyand the specific angular momentum in terms of the azimuthal ac-tion L z = − J φ , both of which are integrals of motion in station-ary, axisymmetric potentials and thus suited to identify coherentgroups of accreted systems (e.g. Gómez et al. 2010; Roedereret al. 2018). Recently, Myeong et al. (2019) identified a majoraccretion event in phase space, dubbed ‘Sequoia’, which broughtin a large number of stars and globular clusters on high-energy,retrograde orbits (see also Massari et al. 2019; Koch & Côté2019). Its location in action space is indicated in Fig. 11. Whilethe star BD + / Fe]
LTE ∼ . / H] LTE ∼ − .
75 are compatible with a metal-poor ex-tension of Sequoia’s metallicity and abundance distribution asshown in Myeong et al. (2019) using LTE abundances. It is worthnoting, however, that our NLTE abundances of this star wouldrather put this view into question. While a unique identificationof this star’s origin in this particular event is therefore pending,an accretion of BD + Here, we explore the question as to whether the most metal-poorstars could have been enriched by a single supernova event, mak-ing them mono-enriched, true second generation stars, or if theyare multi-enriched by several di ff erent events. As suggested byHartwig et al. (2018), this can, at first glance, be observation-ally assessed by inspecting the [Mg / C] ratio as a function of[Fe / H]. This test allows one to asses whether a star, provided -2 -1 0 1 2 3 4 L z [10 kpc km s -1 ] -1.6-1.4-1.2-1-0.8-0.6-0.4-0.20 E t o t [ k m s - ] Fig. 11.
Lindblad diagram for the same stars as in Fig. 10. The green el-lipse [at ( L z , E tot ) ∼ (3200 kpc km s − , − . × ) km s − ] illustrates thelocation of the high-energy, counter-rotating Sequoia merger candidates(Myeong et al. 2019). The same symbols are used as in Fig. 10. Fig. 12. [Mg / C(I)] vs. [Fe / H] in LTE (black) and NLTE (red). Thesymbols show dwarfs and giants, while the enclosed dashed and dot-ted area indicates if the stars are likely to be multi- or mono-enriched(true second generation stars), respectively. The green star representsBD + it is a second-generation star, is likely enriched by one or severalSNe. A mixture of random SNe is far less likely to result in starsthat pass this test than a random individual SN. This is, however,no guarantee that the star actually is a second generation star,and we therefore conducted a constrained comparison to a largegrid of supernova model predictions using a Bayesian approach(for details see Magg et al. 2020, or Section 6.4).Figure 12 shows that there are no obvious candidates, but ourtwo CEMP-no stars (2MASS J0023 and BD + / H] < − . Article number, page 16 of 25. J. Hansen et al.: PEPSI
Table 5.
Constrained and unconstrained best fitting models to the two most metal-poor stars. The reported value is the median, the limits are the2.3 and the 97.7th percentile corresponding to 2 σ lower and upper limits. The limits have been left empty where they are identical to the medianvalue, which indicates that the limit is not well-resolved in the posterior distribution. BD + / H] = − .
75) BD + / H] = − . / Model Mass [M (cid:12) ] Energy [Foe] Mass [M (cid:12) ] Energy [Foe]log (cid:15)
Unconst. constr. Unconstr. constr. Unconstr. constr. Unconstr. constr.LTE 25 . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
10 5 . . A simple comparison to the grid of supernova models fromHeger & Woosley (2010) could indicate, at first glance, thatthe metal-poor stars BD + + (cid:12) . ForBD + χ and the 2 σ limits agree within 1 M (cid:12) and theyyield consistent progenitor masses and a slightly broader rangeof explosion energies. This consistency could misleadingly in-dicate a good match to the model, which is not the case (seeSect. 6.4). Similarly, for BD + ff erent models and adopting di ff erent mixing, dilution, andejected masses. We note that since not all elements can be NLTEcorrected, the NLTE pattern always, or so far, consists of fewerelement abundances than the more complete, but biased, LTEabundance pattern. Hence, we used three sets of patterns, namelya complete set of LTE abundances (’LTE’), a set of LTE abun-dances reduced in number of elements to match the NLTE case(’LTEred’), and finally the NLTE pattern (’NLTE’).Following, we tested the impact of using LTE (red) versusNLTE abundances (see also, e.g. Fig. 13). By blindly adoptingthe best fit model, the LTE versus NLTE case could result in pro-genitor masses di ff ering by up to 14–20 M (cid:12) and energies spreadby a factor of ∼ Here, we use a Bayesian template fitting approach describedin Magg et al. (2020). In this method, the observed abundancepatterns are compared to SNe models from Heger & Woosley(2010). We use this set of yields as it includes around 18000models with a wide range of stellar masses and explosion en-ergies, o ff ering the possibility to fit a wide range of abundancepatterns. While the fitting model could also be used, for exam-ple, with the yields from Ishigaki et al. (2018), its applicability ismore limited in this case because these SNe models are designedto be aspherical. This issue is discussed further in Magg et al.(2020). These yields are based on a grid of 1D simulations ofnon-rotating Pop III core-collapse SNe, spanning a mass rangeof 9 . (cid:12) ≤ M ∗ ≤
100 M (cid:12) and explosion energies (in units of10 erg, abbreviated foe) of 0 . ≤ E ≤
10. As only elementsup to Zn are included in these yields, we disregard heavier el-ements in the fits. Each model is assigned a likelihood basedon how well its predictions agree with the observationally de-rived abundances. Upper limits on detections are treated as astrict upper limit with a Gaussian error, such that the resultinglikelihood is shaped as a Gaussian error-function. To obtain pos-terior probabilities from the likelihoods, the likelihoods are nor-malised, such that the sum over all likelihoods equals one. Thetreatment of errors and limits is similar to the StarFit tool as usedin Fraser et al. (2017). In such fitting routines, usually only ratios of abundances are fitted, which is commonly realised by an arbi-trary dilution factor that is chosen to optimise the fit likelihood.This dilution factor describes how much metal-free gas the SNejecta mix with before forming a new star and, therefore, changethe overall metallicity of the star. Magg et al. (2020) show thatthese dilution masses are commonly picked too low and are in-compatible with the expected expansion of SN remnants. Fitsthat we mark here as ’constrained’ are fits that were performedwhile enforcing the minimum dilution mass based on the ana-lytic limit on dilution from Magg et al. (2020), whereas fits la-belled as ’unconstrained’ were performed with arbitrary dilutionfactors. The latter are merely shown for reference purposes andthey should not be considered to be viable progenitor scenariosfor the origin of the observed abundance patterns. We note thatour results rely on the SN yields from Heger & Woosley (2010),and di ff erent results may be obtained if a di ff erent set of yields,for example from Ishigaki et al. (2018), is used. Below, we dis-cuss individual, potential mono-enriched, metal-poor stars.In Table 5 we show our comparisons to SN models, indicat-ing the likelihood distributions and limits. We show two cases, asimple ’unconstrained’ case with freely varied dilution and mix-ing, and our second comparison which considers dilution (Magget al. 2020, ’constrained’). Both use a Bayesian inference withpriors, however, the latter ’constrained’ case ensures a lowerlimit of reasonable dilution and we consider it to be better andmore physical.As seen from Table 5, there is little di ff erence in the bestfit model as well as the model ranges between the LTE and thereduced LTE (LTEred). This is likely because there is no di ff er-ence in the abundances listed for each individual element. How-ever, there are fewer elements in the LTEred case, which meansthat there might be a minimum number of elements per patternneeded to provide a decent fit (for BD + Z < , while in LTEred there are only ten). ForBD + ∼
10 elements per star. Furthermore, the combina-tion of elements also matters (see Sect. 6.6), which could explainwhy BD + + α , Cr, and Mn abundances. We also note that this is wherewe see the largest di ff erence between the constrained and uncon-strained fits. This is discussed below. Article number, page 17 of 25 & A proofs: manuscript no. 38805corr_arxiv [ X / H ] unconstrained BD+09_2190 unconstrained BD+09_2190 H H e L i B e B C N O F N e N a M g A l S i P S C l A r K C a S c T i V C r M n F e C o N i C u Z n [ X / H ] constrained BD+09_2190 H H e L i B e B C N O F N e N a M g A l S i P S C l A r K C a S c T i V C r M n F e C o N i C u Z n G a constrained BD+09_21900.0 0.2 0.4 0.6 0.8 1.0 P fit P fit Fig. 13.
Bayesian inference of best fit models to all elements in the range of H to Zn. The column of dashes above each element represents theposterior distribution of abundances in that element as obtained from fitting the abundance patterns to the Heger & Woosley (2010) SN II yields.Top panels show the unconstrained approach, while the two bottom panels show the constrained counter part. Blue indicates NLTE, while redshows LTE. The normalised probability of any model fitting is indicated by the colour bar. Observations are shown as symmetric error bars aroundthe abundance. The limit on Sc applies to the models, not the observations. We note that even with the constrained dilution, almost every abundanceof any individual element could be reproduced. Therefore, the white does not indicate a lack of models in the abundance space, but merely a lackof well-fitting ones.
H H e L i B e B C N O F N e N a M g A l S i P S C l A r K C a S c T i V C r M n F e C o N i C u Z n G a [ X / H ] constrained 2MASS NLTE 0.00.20.40.60.81.0 P f i t Fig. 14.
Constrained model comparison to NLTE abundances of ourMg / C-predicted mono-enriched metal-poor star 2MASS J0023. As inFig. 13, the normalised likelihood is shown by the colour bar.
The CEMP-no candidate 2MASS J0023 poses an interestingcase in this comparison, as CEMP-no stars have been proposedto be bona fide second generation stars. Also Fig. 12 indicates,based on the [Mg / C] ratio, that it is right at the limit, and weseek to find the nature of the progenitor as it could provide im-portant clues as to the formation scenario of CEMP-no stars. Byinspecting Fig. 14, only one model stands out (the one modelin dark blue without likelihood distributions). This means that there is only one or no well-fitting model in the grid of modelsfrom Heger & Woosley (2010). However, the fit is poor; there-fore, we can exclude that 2MASS J0023 is mono-enriched, de-spite its Mg / C-ratio and what simple unconstrained fits wouldlead to. Considering the criteria laid out by Hartwig et al. (2018)that this star is unlikely to form from a combination of multipleSNe, this may indicate that 2MASS J0023 is not a true second-generation star altogether. However, using a di ff erent set of SNmodels, including more exotic kinds of SN, could challenge thisconclusion.We now turn to our most metal-poor star. Unfortunately, wecould not derive C but only place an upper limit and hence wehave a star that is right at the verge of the multi- and mono-enriched area in the [Mg / C] diagram. However, the constrainedfit including all observed elements (in NLTE) shows a perfectcase with statistical likelihood distributions for each element (seeFig. 13). For BD + =
0) would be25.5 M (cid:12) with an explosion energy of 5 foe (10 foe in the reducedLTE case); nevertheless, these values change in NLTE and turnout to be lower (11.2 M (cid:12) and 0.6 foe). In the constrained case,the mass and energy inferred from NLTE and LTE are closer. In
Article number, page 18 of 25. J. Hansen et al.: PEPSI Progenitor mass (M )0.00.20.40.60.81.0 p r o b a b ili t y BD+09_2190 priorLTE unconstr.LTE constr.NLTE unconstr.NLTE constr.LTEred unconstr.LTEred constr. 10 explosion energy(10 erg)BD+09_2190 Fig. 15.
Probability from unconstrained and constrained model comparisons, showing preferred mass and energy in the following three cases: LTE(red), LTEred (yellow), and NLTE (blue) for BD + LTE, we obtain 25.5 M (cid:12) and 3.0 foe, while NLTE yields 19.2 M (cid:12) and 1.5 foe, respectively. However, NLTE yields higher valuesthan LTE for BD + χ matching in large samplescould skew the inferred masses and in turn the initial mass func-tion (IMF) at the lowest metallicity of mono-enriched stars in theGalaxy.Similar to the metal-poor stars in Fig. 14 and 15, we com-pared all other stars in our sample to the ‘constrained models’.The obtained fits were poor, their [Mg / C] high, and we can there-fore exclude that all the more metal-rich stars in our sample with[Fe / H] > − . We now explore if some elements provide stronger constraintsthan others on the supernova progenitor when comparing obser-vationally derived stellar abundances to yield predictions. Fig-ure 13 shows the unconstrained (top panels) and the dilution con-strained (bottom panels) predictions compared to the LTE andNLTE abundances of BD + ffi cultto model accurately as nucleosynthesis theory underproduces Ti(see e.g. Kobayashi et al. 2020).The unconstrained (direct) comparison shows clear distribu-tions with peaking likelihoods for all elements, except for O (seeFigs. 13, 15, and 16). These resolved posterior distributions indi-cate that the method is able to recover a range of SN models thatprovide a good fit to the observed abundance patterns. In some cases (e.g. for BD + ff er strong constraints whencombined. This could be explained in part by their di ff erent sen-sitivity to the SN progenitor mass and explosion energy throughtheir odd-even distribution. We emphasise that the di ff erent Naabundance in LTE and NLTE play an important role in constrain-ing the best fit model, and we note that Na NLTE does not presenta good fit in either of the unconstrained or constrained cases.However, to test the element-to-element sensitivity, the pat-tern is fit several times by removing one element at a time. Fig-ure 16 shows such fits to our best mono-enriched candidate,BD + α − elements. However,in NLTE, this becomes even more pronounced. Here, K is alsoseen to play a vital role, and lacking any of these elements couldskew the mass prediction by a factor of ∼ −
5. Again we seethat O, Na, Mg, K, and Mn are important tracers. They representthe odd-even pattern, even in sparsely populated patterns (eightelements), yet they contain Fe-peak elements and α − elementsthat are mass and energy sensitive. The lack of N abundancesin this sense prevents a light odd-even pattern around C-N-O,which would have supported the SN traces and might thereforeimpose a slight bias on our results. However, the current spectrado not allow us to derive N. Future blue spectra in line with what,for example, the CUBES spectrograph (Barbuy et al. 2014 andErnandes et al. 2020 in prep.) will provide, will greatly help ussolidify the true origin of stars, such as BD + Article number, page 19 of 25 & A proofs: manuscript no. 38805corr_arxiv p r o b a b ili t y LTE priorno Ono Nano Mgno Kno Ca no Scno Tino Crno FeAll LTE10 Progenitor mass (M )0.00.20.40.60.81.0 p r o b a b ili t y NLTE 10 explosion energy(10 erg)NLTE Fig. 16.
Constrained fits for the star BD + We note that due to the sparsely populated 3D NLTE pat-terns, we did not attempt find a best fitting (constrained) Pop IIImodel. Nevertheless, we emphasise that it is important to expand3D NLTE patterns and conduct constrained fits to understand thenature of the first population of stars and, in particular, the massspectrum and energetics of the explosion.
7. Conclusion
Stellar archaeology is a powerful tool for exploring and under-standing the chemical evolution of the Galaxy. Despite the lim-ited sample size of our study, the broad range of stellar parame-ters permits one to explore the chemical evolution and the rich-ness of the derived abundance patterns, allowing us to contrastcompeting models in a meaningful way.Our high-resolution PEPSI analysis is an attempt to accu-rately explore the Milky Ways true chemical evolution and thenature of Pop III supernovae. The GCE trends of the 32 studiedelements (34 when including limits) are in agreement with theprevious literature samples and only potassium appears slightlyhigher than anticipated. For the lighter elements, such as C, O,Na, K, Mn, and Co, NLTE corrections play an important role,to the point where the GCE of the MW appears di ff erent in LTEversus NLTE. For Si, Ti, V, and Fe, we derived both neutral andionised abundances; furthermore, in most cases, there is a di ff er-ence or a scatter due to the lack of ionisation equilibrium in thestellar sample. For the heavy elements, where the abundancesare generally derived from the majority species, the NLTE cor- rections are smaller, and the GCE-trends appear consistent inboth LTE and NLTE (see e.g. Zr, Ba, Nd, and Eu).Typically, 1D LTE abundances are used in most studies astheir analysis is faster and more straightforward. However, asshown in this study (and highlighted by Fig. A.1), the overallabundance pattern can look quite di ff erent in LTE and NLTE,even if 1D hydrostatic models atmospheres are used. Yet, thisis not the complete picture as we show by adding 3D (and 3DNLTE) corrected abundances for a small number of elements(C, O, and Fe). The optimal and most physical answer to un-derstanding the chemical evolution of the Galaxy or to pinningdown the nature of the Pop III supernovae can best be answeredwith a pattern of 3D NLTE corrected elements in a statisticalsu ffi cient sample using a dilution constrained method, includingkey elements, such as C and O.The comparison to the yield predictions returned three im-portant results, namely, that the most robust conclusions can beinferred when the stellar pattern consists of ∼ −
10 or moreelements. The conclusions di ff er depending on the physical ba-sis of the radiative transfer models (LTE, NLTE, 1D hydrostaticequilibrium, and 3D hydrodynamics), but also from the yield orSN model side. Here the unconstrained versus the dilution con-strained methods provide di ff erent results. These results are sodi ff erent that they can skew and bias the inferred IMF of theGalaxy if not properly dealt with. In the case of BD + ∼
20 M (cid:12) be-tween NLTE and LTE and unconstrained and constrained meth-ods, respectively. If this would be the outcome of the inferred
Article number, page 20 of 25. J. Hansen et al.: PEPSI mass from numerous mono-enriched second generation stars, themass distribution of the Pop III stars could look very di ff erent.Finally, when comparing observations and theory, it is not onlyimportant to have a su ffi ciently rich, well-sampled pattern, butalso the combination of odd-even elements with Z <
30 (such asC, O, Na, Mg, K, Ca, Ti, and Mn) need to be present to best con-strain the mass and explosion energy of the Pop III progenitor.We highlight that the results also depend on the set of SN modelpredictions that are included in the comparison.Based on α − abundances and kinematics, we traced thechemo-dynamical origin of several chemically peculiar stars andfind strong indications of two fast halo stars (BD + + s star(TYC5481). We tested if the CEMP-no star was a bona fidesecond generation star, as many CEMP-no stars are; however,in following both constrained and unconstrained dilution massmodel comparisons, we could reject the mono-enrichment hy-pothesis for this specific star with the yields we adopted here.The CEMP- s star was likely enriched by an intermediate-massand -metallicity AGB star.We placed constraints on the possible (dominant) AGBdonor at high metallicities, seeing how they add s-process tometal-poor stars, such as TYC5481, while the heavy elementcontent of BD-10_3742 shows a perfect r-process pattern as rep-resented by the ‘metal-poor Sun’ HD20 that serves as a neutron-capture benchmark star with its almost pure r-process pattern.Finally, we find that BD + ∼ . . (cid:12) supernova with a normal explosion energy (3 or 1.5 foe) in LTEor NLTE, respectively. This shows the importance of carefullyassessing the stellar abundances as well as conducting a physi-cal meaningful (constrained) comparison to supernova yields inorder to properly understand the nature of the first stars. Acknowledgements.
CJH acknowledges support from the Max Planck Societyand from the ChETEC COST Action (CA16117), supported by COST (Euro-pean Cooperation in Science and Technology). CJH is grateful to A. Amarsifor providing 3D and NLTE computations for C, O, and Fe, and also to S.Cristallo and H. Reggiani for fruitful dialogues and helpful input. A.K., M.B.,and R.S.K. gratefully acknowledge funding by the Deutsche Forschungsgemein-schaft (DFG, German Research Foundation) – Project-ID 138713538 – SFB 881("The Milky Way System"), subprojects A03, A05, A10, A11, as well as subpro-jects B01, B02, and B08. EC gratefully acknowledge support from the FrenchNational Research Agency (ANR) funded project “Pristine” (ANR-18-CE31-0017). H.W.Z. acknowledges the National Natural Science Foundation of ChinaNo. 11973001 and National Key R&D Program of China No. 2019YFA0405504.MM was supported by the Max-Planck-Gesellschaft via the fellowship of the In-ternational Max Planck Research School for Astronomy and Cosmic Physics atthe University of Heidelberg (IMPRS-HD). R.S.K. also acknowledges supportfrom Germany’s Excellence Strategy in framework of the Heidelberg Clusterof Excellence STRUCTURES (grant EXC-2181 / / / / // pepsi.aip.de / ). This workhas made use of data from the European Space Agency (ESA) mission Gaia ( ), processed by the Gaia
Data Process- ing and Analysis Consortium (DPAC, ). Funding for the DPAC has been provided by na-tional institutions, in particular the institutions participating in the
Gaia
Mul-tilateral Agreement. Finally, we would like to thank the anonymous referee foruseful comments.
References
Aguado, D. S., Allende Prieto, C., González Hernández, J. I., & Rebolo, R. 2018,ApJ, 854, L34Alexeeva, S., Pakhomov, Y., & Mashonkina, L. 2014, Astronomy Letters, 40,406Alexeeva, S. A. & Mashonkina, L. I. 2015, MNRAS, 453, 1619Amarsi, A. M., Asplund, M., Collet, R., & Leenaarts, J. 2016a, MNRAS, 455,3735Amarsi, A. M., Barklem, P. S., Collet, R., Grevesse, N., & Asplund, M. 2019a,A&A, 624, A111Amarsi, A. M., Lind, K., Asplund, M., Barklem, P. S., & Collet, R. 2016b, MN-RAS, 463, 1518Amarsi, A. M., Nissen, P. E., Asplund, M., Lind, K., & Barklem, P. S. 2019b,A&A, 622, L4Amarsi, A. M., Nissen, P. E., & Skúladóttir, Á. 2019c, A&A, 630, A104Aoki, W., Beers, T. C., Christlieb, N., et al. 2007, ApJ, 655, 492Asplund, M., Grevesse, N., Sauval, A. J., & Scott, P. 2009, ARA&A, 47, 481Bailer-Jones, C. A. L., Rybizki, J., Fouesneau, M., Mantelet, G., & Andrae, R.2018, AJ, 156, 58Barbuy, B., Bawden Macanhan, V., Bristow, P., et al. 2014, Ap&SS, 354, 191Barklem, P. S. 2018, A&A, 610, A57Barklem, P. S., Belyaev, A. K., Dickinson, A. S., & Gadéa, F. X. 2010, A&A,519, A20Barklem, P. S., Belyaev, A. K., Spielfiedel, A., Guitou, M., & Feautrier, N. 2012,A&A, 541, A80Beers, T. C. & Christlieb, N. 2005, ARA&A, 43, 531Belyaev, A. K. & Voronov, Y. V. 2017, A&A, 606, A106Belyaev, A. K., Voronov, Y. V., Mitrushchenkov, A., Guitou, M., & Feautrier, N.2019, MNRAS, 487, 5097Belyaev, A. K., Voronov, Y. V., Yakovleva, S. A., et al. 2017, ApJ, 851, 59Belyaev, A. K. & Yakovleva, S. A. 2017, A&A, 608, A33Belyaev, A. K. & Yakovleva, S. A. 2018, MNRAS, 478, 3952Belyaev, A. K., Yakovleva, S. A., & Barklem, P. S. 2014, A&A, 572, A103Bensby, T., Feltzing, S., & Lundström, I. 2003, A&A, 410, 527Bergemann, M. 2011, MNRAS, 413, 2184Bergemann, M. & Cescutti, G. 2010, A&A, 522, A9Bergemann, M., Collet, R., Amarsi, A. M., et al. 2017a, ApJ, 847, 15Bergemann, M., Collet, R., Schönrich, R., et al. 2017b, ApJ, 847, 16Bergemann, M., Gallagher, A. J., Eitner, P., et al. 2019, A&A, 631, A80Bergemann, M., Hansen, C. J., Bautista, M., & Ruchti, G. 2012, A&A, 546, A90Bergemann, M. & Nordlander, T. 2014, arXiv e-prints, arXiv:1403.3088Bergemann, M., Pickering, J. C., & Gehren, T. 2010, MNRAS, 401, 1334Bonifacio, P., Ca ff au, E., Spite, M., et al. 2015, A&A, 579, A28Bonifacio, P., Spite, M., Cayrel, R., et al. 2009, A&A, 501, 519Bruls, J. H. M. J., Rutten, R. J., & Shchukina, N. G. 1992, A&A, 265, 237Busso, M., Gallino, R., & Wasserburg, G. J. 1999, ARA&A, 37, 239Butler, K. & Giddings, J. 1985, Newsletter on the analysis of astronomical spec-tra, No. 9, University of LondonCa ff au, E., Bonifacio, P., François, P., et al. 2011, Nature, 477, 67Carlsson, M. 1986, Uppsala Astronomical Observatory Reports, 33Carretta, E., Bragaglia, A., Gratton, R., D’Orazi, V., & Lucatello, S. 2009, A&A,508, 695Cayrel, R., Depagne, E., Spite, M., et al. 2004, A&A, 416, 1117Chan, C., Müller, B., & Heger, A. 2020, MNRAS, 495, 3751Chen, K.-J., Heger, A., Whalen, D. J., et al. 2017, MNRAS, 467, 4731Choplin, A., Tominaga, N., & Meyer, B. S. 2020, A&A, 639, A126Collet, R., Asplund, M., & Trampedach, R. 2007, A&A, 469, 687Cooper, A. P., D’Souza, R., Kau ff mann, G., et al. 2013, MNRAS, 434, 3348Côté, B., Eichler, M., Arcones, A., et al. 2019, ApJ, 875, 106Cristallo, S., Abia, C., Straniero, O., & Piersanti, L. 2015, ApJ, 801, 53Cristallo, S., Piersanti, L., Straniero, O., et al. 2011, ApJS, 197, 17Dehnen, W. & Binney, J. 1998, MNRAS, 294, 429Dobrovolskas, V., Kuˇcinskas, A., Ste ff en, M., et al. 2013, A&A, 559, A102Du ff au, S., Ca ff au, E., Sbordone, L., et al. 2017, A&A, 604, A128Edvardsson, B., Andersen, J., Gustafsson, B., et al. 1993, A&AS, 102, 603Eitner, P., Bergemann, M., Hansen, C. J., et al. 2020, A&A, 635, A38Ezzeddine, R., Frebel, A., Roederer, I. U., et al. 2019, ApJ, 876, 97Fellhauer, M., Wilkinson, M. I., Evans, N. W., et al. 2008, MNRAS, 385, 1095François, P., Depagne, E., Hill, V., et al. 2007, A&A, 476, 935Fraser, M., Casey, A. R., Gilmore, G., Heger, A., & Chan, C. 2017, MNRAS,468, 418 Article number, page 21 of 25 & A proofs: manuscript no. 38805corr_arxiv
Frebel, A., Christlieb, N., Norris, J. E., et al. 2006, ApJ, 652, 1585Frebel, A., Ji, A. P., Ezzeddine, R., et al. 2019, ApJ, 871, 146Freytag, B., Ste ff en, M., Ludwig, H.-G., et al. 2012, Journal of ComputationalPhysics, 231, 919Gaia Collaboration, Brown, A. G. A., Vallenari, A., et al. 2018, A&A, 616, A1Gallagher, A. J., Bergemann, M., Collet, R., et al. 2020, A&A, 634, A55Gallagher, A. J., Ca ff au, E., Bonifacio, P., et al. 2016, A&A, 593, A48Gómez, F. A., Helmi, A., Brown, A. G. A., & Li, Y.-S. 2010, MNRAS, 408, 935Gustafsson, B., Edvardsson, B., Eriksson, K., et al. 2008, A&A, 486, 951Hanke, M., Hansen, C. J., Ludwig, H.-G., et al. 2020, A&A, 635, A104Hansen, C. J., Bergemann, M., Cescutti, G., et al. 2013, A&A, 551, A57Hansen, C. J., Hansen, T. T., Koch, A., et al. 2019, A&A, 623, A128Hansen, C. J., Ludwig, H.-G., Seifert, W., et al. 2015a, AstronomischeNachrichten, 336, 665Hansen, C. J., Nordström, B., Bonifacio, P., et al. 2011, A&A, 527, A65Hansen, C. J., Nordström, B., Hansen, T. T., et al. 2016a, A&A, 588, A37Hansen, C. J. & Primas, F. 2011, A&A, 525, L5Hansen, C. J., Primas, F., Hartman, H., et al. 2012, A&A, 545, A31Hansen, T., Hansen, C. J., Christlieb, N., et al. 2015b, ApJ, 807, 173Hansen, T. T., Andersen, J., Nordström, B., et al. 2016b, A&A, 586, A160Hartwig, T., Yoshida, N., Magg, M., et al. 2018, MNRAS, 478, 1795Heger, A. & Woosley, S. E. 2002, ApJ, 567, 532Heger, A. & Woosley, S. E. 2010, ApJ, 724, 341Hollek, J. K., Frebel, A., Placco, V. M., et al. 2015, ApJ, 814, 121Horowitz, C. J., Arcones, A., Côté, B., et al. 2019, Journal of Physics G NuclearPhysics, 46, 083001Ilyin, I. V. 2000, PhD thesis, Astronomy Division Department of Physical Sci-ences P.O.Box 3000 FIN-90014 University of Oulu FinlandIshigaki, M. N., Tominaga, N., Kobayashi, C., & Nomoto, K. 2018, ApJ, 857, 46Ito, H., Aoki, W., Honda, S., & Beers, T. C. 2009, ApJ, 698, L37Käppeler, F., Gallino, R., Bisterzo, S., & Aoki, W. 2011, Review of ModernPhysics, 83, 157Kaulakys, B. 1991, Journal of Physics B Atomic Molecular Physics, 24, L127Keller, S. C., Bessell, M. S., Frebel, A., et al. 2014, Nature, 506, 463Kiselman, D. 1991, A&A, 245, L9Kobayashi, C., Karakas, A. I., & Lugaro, M. 2020, arXiv e-prints,arXiv:2008.04660Kobayashi, C., Tominaga, N., & Nomoto, K. 2011, ApJ, 730, L14Kobayashi, C., Umeda, H., Nomoto, K., Tominaga, N., & Ohkubo, T. 2006, ApJ,653, 1145Koch, A. & Côté, P. 2019, A&A, 632, A55Koppelman, H., Helmi, A., & Veljanoski, J. 2018, ApJ, 860, L11Kordopatis, G., Gilmore, G., Wyse, R. F. G., et al. 2013, MNRAS, 436, 3231Lee, Y. S., Beers, T. C., Masseron, T., et al. 2013, AJ, 146, 132Lind, K., Bergemann, M., & Asplund, M. 2012, MNRAS, 427, 50Magg, M., Nordlander, T., Glover, S. C. O., et al. 2020, arXiv e-prints,arXiv:2006.12517Magic, Z., Collet, R., Asplund, M., et al. 2013, A&A, 557, A26Mashonkina, L. 2020, MNRAS, 493, 6095Mashonkina, L. & Gehren, T. 2000, A&A, 364, 249Mashonkina, L., Gehren, T., Shi, J.-R., Korn, A. J., & Grupp, F. 2011, A&A,528, A87Mashonkina, L., Jablonka, P., Pakhomov, Y., Sitnova, T., & North, P. 2017a,A&A, 604, A129Mashonkina, L., Korn, A. J., & Przybilla, N. 2007, A&A, 461, 261Mashonkina, L., Ryabchikova, T., & Ryabtsev, A. 2005, A&A, 441, 309Mashonkina, L., Ryabtsev, A., & Frebel, A. 2012, A&A, 540, A98Mashonkina, L., Sitnova, T., & Belyaev, A. K. 2017b, A&A, 605, A53Mashonkina, L., Sitnova, T., Yakovleva, S. A., & Belyaev, A. K. 2019, A&A,631, A43Massari, D., Koppelman, H. H., & Helmi, A. 2019, A&A, 630, L4McWilliam, A., Wallerstein, G., & Mottini, M. 2013, ApJ, 778, 149Meynet, G., Ekström, S., & Maeder, A. 2006, A&A, 447, 623Myeong, G. C., Vasiliev, E., Iorio, G., Evans, N. W., & Belokurov, V. 2019,MNRAS, 488, 1235Nissen, P. E. & Schuster, W. J. 2010, A&A, 511, L10Nordlander, T., Amarsi, A. M., Lind, K., et al. 2017, A&A, 597, A6Pillepich, A., Madau, P., & Mayer, L. 2015, ApJ, 799, 184Placco, V. M., Beers, T. C., Roederer, I. U., et al. 2014a, ApJ, 790, 34Placco, V. M., Frebel, A., Beers, T. C., & Stancli ff e, R. J. 2014b, ApJ, 797, 21Placco, V. M., Frebel, A., Beers, T. C., et al. 2016, ApJ, 833, 21Przybilla, N., Butler, K., Becker, S. R., Kudritzki, R. P., & Venn, K. A. 2000,A&A, 359, 1085Reggiani, H., Amarsi, A. M., Lind, K., et al. 2019, A&A, 627, A177Reichert, M., Hansen, C. J., Hanke, M., et al. 2020, A&A, in press(arXiv:2004.01195)Roederer, I. U., Hattori, K., & Valluri, M. 2018, AJ, 156, 179Roederer, I. U., Preston, G. W., Thompson, I. B., et al. 2014, AJ, 147, 136Ruchti, G. R., Bergemann, M., Serenelli, A., Casagrande, L., & Lind, K. 2013,MNRAS, 429, 126 Sbordone, L., Bonifacio, P., Ca ff au, E., et al. 2010, A&A, 522, A26Schlafly, E. F. & Finkbeiner, D. P. 2011, ApJ, 737, 103Schörck, T., Christlieb, N., Cohen, J. G., et al. 2009, A&A, 507, 817Sitnova, T. M. 2016, Astronomy Letters, 42, 734Sitnova, T. M. & Mashonkina, L. I. 2018, Astronomy Letters, 44, 411Sitnova, T. M., Mashonkina, L. I., & Ryabchikova, T. A. 2016, MNRAS, 461,1000Sitnova, T. M., Yakovleva, S. A., Belyaev, A. K., & Mashonkina, L. I. 2020,Astronomy Letters, 46, 120Sneden, C. A. 1973, PhD thesis, The University of Texas at Austin.Steenbock, W. & Holweger, H. 1984, A&A, 130, 319Strassmeier, K. G., Ilyin, I., Järvinen, A., et al. 2015, AstronomischeNachrichten, 336, 324Strassmeier, K. G., Ilyin, I., & Ste ff en, M. 2018, A&A, 612, A44Takahashi, K., Umeda, H., & Yoshida, T. 2014, ApJ, 794, 40Tinsley, B. M. 1979, ApJ, 229, 1046Tominaga, N. 2009, ApJ, 690, 526Velichko, A. B., Mashonkina, L. I., & Nilsson, H. 2010, Astronomy Letters, 36,664Watson, D., Hansen, C. J., Selsing, J., et al. 2019, Nature, 574, 497Winteler, C., Käppeli, R., Perego, A., et al. 2012, ApJ, 750, L22Yakovleva, S. A., Belyaev, A. K., & Kraemer, W. P. 2018, Chemical Physics,515, 369Yakovleva, S. A., Belyaev, A. K., & Kraemer, W. P. 2019, MNRAS, 483, 5105Youakim, K., Starkenburg, E., Martin, N. F., et al. 2020, MNRAS, 492, 4986Zhang, H. W., Gehren, T., & Zhao, G. 2008, A&A, 481, 489Zhao, G., Mashonkina, L., Yan, H. L., et al. 2016, ApJ, 833, 225 Article number, page 22 of 25. J. Hansen et al.: PEPSI
Appendix A: Online material
Online material available on CDS and in the published paper (open access).
Article number, page 23 of 25 & A proofs: manuscript no. 38805corr_arxivArticle number, page 24 of 25. J. Hansen et al.: PEPSI
Fig. A.1.
Abundance patterns in 1D LTE (yellow squares), 1D NLTE (red triangles with corrections as listed in Table 3 with the exception of C,which in the red triangles are based on Alexeeva & Mashonkina (2015)), and 1D NLTE and 3D NLTE for C and O (in black), while 3D LTE isshown for Fe (corrections from Amarsi et al. 2019b,c, as described in Sect. 4.3). Progenitor mass (M )0.00.20.40.60.81.0 p r o b a b ili t y BD+24_1676 10 explosion energy(10 erg)BD+24_1676priorLTE unconstr.LTE constr.NLTE unconstr.NLTE constr.LTEred unconstr.LTEred constr. Fig. A.2.
Probability from unconstrained and constrained model comparisons, showing preferred mass and energy in the following three cases:LTE (red), LTEred (yellow), and NLTE (blue) for BD + Appendix A.1: Testing mono versus multi enrichment at low-metallicity
For our second most metal-poor dwarf, BD + ffi ciently sampled parameter space. We, therefore, conclude thatthis star is likely multi-enriched (or not a true second-generation star), which is in good agreement with the r + s mixed heavy elementpattern.s mixed heavy elementpattern.