Chemical Compositions of Red Giant Stars from Habitable Zone Planet Finder Spectroscopy
Christopher Sneden, Melike Afsar, Zeynep Bozkurt, Gamze Bocek Topcu, Sergen Ozdemir, Gregory R. Zeimann, Cynthia S. Froning, Suvrath Mahadevan, Joe P. Ninan, Chad F. Bender, Ryan Terrien, Lawrence W. Ramsey, 9 Karin Lind, Gregory N. Mace, Kyle F. Kaplan, Hwihyun Kim, Keith Hawkins, Brendan P. Bowler
DDraft version January 1, 2021
Typeset using L A TEX default style in AASTeX63
CHEMICAL COMPOSITIONS OF RED GIANT STARS FROM HABITABLE ZONE PLANETFINDER SPECTROSCOPY
Christopher Sneden, Melike Afs¸ar,
2, 1
Zeynep Bozkurt, Gamze B¨ocek Topcu, Sergen ¨Ozdemir, Gregory R. Zeimann, Cynthia S. Froning, Suvrath Mahadevan,
8, 9
Joe P. Ninan,
10, 9
Chad F. Bender, Ryan Terrien, Lawrence W. Ramsey,
13, 9
Karin Lind, Gregory N. Mace, Kyle F. Kaplan, Hwihyun Kim, Keith Hawkins, and Brendan P. Bowler Department of Astronomy and McDonald Observatory, The University of Texas, Austin, TX 78712, USA; [email protected] Department of Astronomy and Space Sciences, Ege University, 35100 Bornova, ˙Izmir, Turkey; [email protected] Department of Astronomy and Space Sciences, Ege University, 35100 Bornova, ˙Izmir, Turkey; [email protected] Department of Astronomy and Space Sciences, Ege University, 35100 Bornova, ˙Izmir, Turkey; [email protected] Department of Astronomy and Space Sciences, Ege University, 35100 Bornova, ˙Izmir, Turkey; [email protected] Hobby Eberly Telescope, The University of Texas, Austin, TX 78712, USA; [email protected] Department of Astronomy and McDonald Observatory, The University of Texas, Austin, TX 78712, USA; [email protected] Department of Astronomy & Astrophysics, The Pennsylvania State University, University Park, PA 16803, USA; [email protected] Center for Exoplanets and Habitable Worlds, The Pennsylvania State University, University Park, PA 16803, USA Department of Astronomy & Astrophysics, The Pennsylvania State University, University Park, PA 16803, USA; [email protected] Department of Astronomy and Steward Observatory, University of Azizona, Tucson, AZ 85721, USA; [email protected] Department of Physics and Astronomy, Carleton College, Northfield, MN 55057, USA; [email protected] Department of Astronomy & Astrophysics, The Pennsylvania State University, University Park, PA 16803, USA; [email protected] Department of Astronomy, Stockholm University, AlbaNova University Centre, SE-106 91 Stockholm, Sweden: [email protected] Department of Astronomy and McDonald Observatory, The University of Texas, Austin, TX 78712, USA; [email protected] SOFIA Science Center, NASA Ames Research Center, Mail Stop N232 −
12 P.O. Box 1, Moffett Field, CA 94036, USA;[email protected] Gemini Observatory, Casilla 603, La Serena, Chile; [email protected] Department of Astronomy and McDonald Observatory, The University of Texas, Austin, TX 78712, USA; [email protected] Department of Astronomy and McDonald Observatory, The University of Texas, Austin, TX 78712, USA; [email protected]
ABSTRACTWe have used the Habitable Zone Planet Finder (HPF) to gather high resolution, high signal-to-noise near-infrared spectra of 13 field red horizontal-branch (RHB) stars, one open-cluster giant, andone very metal-poor halo red giant. The HPF spectra cover the 0.81 − µ m wavelength rangeof the zyJ bands, filling in the gap between the optical (0.4 − µ m) and infrared (1.5 − µ m)spectra already available for the program stars. We derive abundances of 17 species from LTE-basedcomputations involving equivalent widths and spectrum syntheses, and estimate abundance correctionsfor the species that are most affected by departures from LTE in RHB stars. Generally good agreementis found between HPF-based metallicities and abundance ratios and those from the optical and infraredspectral regions. Light element transitions dominate the HPF spectra of these red giants, and HPF datacan be used to derive abundances from species with poor or no representation in optical spectra (e.g.,C i , P i , S i , K i ). Attention is drawn to the HPF abundances in two field solar-metallicity RHB starsof special interest: one with an extreme carbon isotope ratio, and one with a rare very large lithiumcontent. The latter star is unique in our sample by exhibiting very strong He i i and S i in HD 122563 are reported, yielding the lowest metallicity determination of[S/Fe] from more than one multiplet. Corresponding author: Christopher [email protected] a r X i v : . [ a s t r o - ph . S R ] D ec Sneden et al.
Keywords: stellar abundances - horizontal branch stars - evolved stars INTRODUCTIONRed horizontal branch (RHB) stars (also called “secondary red clump”; e.g., Girardi et al. 1998, Girardi 1999,Ruiz-Dern et al. 2018) are mostly known for their prominent locations on the Hertzsprung-Russell (HR) diagrams ofglobular clusters. They are evolved stars with double energy sources, burning helium in the core and hydrogen burningin the shell. RHB stars are not easy to identify among field stars, but their cooler counterparts, the red clump (RC)stars, are relatively more numerous and have narrowly constrained temperatures and gravities, making them standout in color-magnitude diagrams. RHB and RC stars have small luminosity ranges, and thus can serve as standardcandles.Accurate chemical compositions of RHB stars along with robust statistics of their occurrence in disk and halopopulations can enhance our knowledge of stellar and Galactic evolution. RHBs are important astrophysical toolsthat can be used to study stellar densities, kinematics, and chemical abundances across the Galactic disk, and allowus to reach out greater distances than the dwarfs or other regular giants would provide (e.g. Girardi 2016). Af¸saret al. (2018a) (here after Af¸sar18a) have studied the detailed chemical compositions and kinematics of a sample of340 candidate field RHB stars in the optical spectral region (3400 − , ar et al. (2018b) (hereafter Af¸sar18b) investigated three RHB stars in detail usinghigh-resolution H- and K-band (1.48 − µ m) spectra obtained with the Immersion Grating Infrared Spectrograph(IGRINS). In that work they argued that in general the IR and optical abundances agree well, and for some elements(e.g., Mg, Si) the IR abundances are more trustworthy than the optical ones. Additionally, that work highlighted theabundances of several elements (e.g., P, S, K) with prominent transitions in the IR whose optical lines are too weakto be detected or too strong to yield reliable abundances.Our work investigates some selected RHB stars from Af¸sar2018a,b in the relatively under-studied zyJ wavelengthregion, using high-resolution spectra obtained with the Habitable Zone Planet Finder (HPF) spectrograph. We derivemetallicities and abundances of 13 RHB stars. We also report on our analyses of the well-known very metal-poor halored giant, HD 122563 (Af¸sar et al. 2016 and references therein, hereafter Af¸sar16), and of one red giant member fromopen cluster NGC 6940 (MMU 105; B¨ocek Topcu et al. 2016, B¨ocek Topcu et al. 2019, here after BT16, and BT19).These two stars will be considered separately from the discussion of the RHB stars.In § §
3. We consider four stars with noteworthy abundance signatures in §
4, and conclude with our summaryin § OBSERVATIONS AND REDUCTIONSThe red giants investigated in this HPF spectroscopic study (Table 1) have been selected from our previous studies(Af¸sar16,18a,b, BT16,19). See those papers for detailed descriptions of the optical and IGRINS observations of ourprogram stars. In general the HPF zyJ wavelength domain has been relatively neglected in stellar abundance studies.In order to sample spectra of red giants in this region, we have selected stars with various chemical compositioncharacteristics. The following sub-sections focus on the details of HPF observations.2.1.
HPF Spectra
HPF is a high − resolution ( λ/ ∆ λ ≡ R ∼ − infrared ( zyJ photometric bands, 8100 − The mainHPF goal is to search for exoplanets of cool M dwarf stars by detecting the orbital reflex motions of the parent stars.To achieve high velocity precision ( ∼ − ), the entire HPF instrument is encased in a vacuum chamber cooledto 180 Kelvin, yielding temperature stability of ∼ https://hpf.psu.edu/ PF Chemical Compositions
S/N > zyJ spectral region the telluric molecularblockage ranges from near zero to almost total, varying significantly with wavelength. The dominant telluric contam-inators are H O bands, and their contamination can change substantially on timescales of minutes, usually requiringa hot-star observation to accompany each target star. Some mitigation is provided by the fixed altitude of the HET,which on stable nights can lead to successful use of divisor star spectra acquired some time before or after the targetspectra.HPF observations of target and telluric warm stars were reduced using the HPF pipeline code
Goldilocks . Thispackage processed the raw HPF data by removing bias noise, correcting for nonlinearity, masking cosmic rays, andcalculating the slope/flux and variance image using the algorithms from the pyhxrg module in the tool
HxRGproc (Ninan et al. 2018). The pipeline also processed bright lamp exposures to combine into a master frame to model thetrace of the calibration, sky, and science fibers as well as their respective fiber profiles. The profiles were used asweights in an optimal extraction algorithm (Horne 1986) to produce 1D spectra and their associated propagated errorsfor all 28 orders. The master frame was also used to correct flat fielding features.
Goldilocks uses a single wavelength solution for all observations from the well-calibrated NIR Laser Frequency Comb.HPF’s wavelength solution is extremely stable but does drift on the order of ∼ − throughout the night withlong term drifts over months on the same scale. This amounts to ∼ Goldilocks providesthe drift correction in the observation file headers, it is only important for precision radial velocity studies; we ignoredthis very small effect in our analysis.For each stellar source, the science fiber extractions from
Goldilocks includes light from the target as well as OHemission from the night sky. The sky fiber, located 50.1 (cid:48)(cid:48) away, contains only night sky emission. In most cases, therelative flux ratio between the sky emission and stellar continuum (at the wavelengths of interest) is less than 5%, andthe sky emission has almost no effect on the target spectrum. In those situations we used the spectra of targets asextracted from the science fibers without attempting sky subtraction. For a few small spectral regions in which skysubtraction was critical to accurate representation of the stellar spectra, we performed the sky subtraction manuallyusing the pipeline-extracted star and sky spectra.After these reduction procedures, we performed continuum normalization of the target spectra via the continuum task of IRAF , and we used the telluric task to divide out the telluric contamination from the science target spectra.Later, individual orders were merged into a single continuous spectrum with the scombine task.The Goldilocks pipeline produces rest wavelength solutions that are accurate to levels far beyond what are neededfor our study. To measure the radial velocities (
RV s ) of our stellar targets we made use of rv package in IRAF. Firstwe created synthetic spectra (as outlined in § T eff , log g , and [Fe/H] values to our program stars.Then we used rvcorrect for the heliocentric corrections, and measured the RV values with the cross-correlation task fxcor (Fitzpatrick 1993). We only worked on the wavelength regions that are not affected (or only affected minimally)by the atmospheric telluric lines: 8420 − − − − − RV s from HPF and their standard deviations ( σ ) for each star in Table 2. The σ values were calculatedby taking the average of RV s measured from individual regions listed above. Previous RV measurements of the sametargets (Af¸sar18a and Gaia
DR2) are also listed in Table 2 for comparison purposes.The internal RV accuracies (line-to-line scatters) for individual stars are comparable: (cid:104) σ HP F (cid:105) = 0.18 km s − and (cid:104) σ opt (cid:105) = 0.19 km s − , and the velocity scales agree: (cid:104) RV opt - RV HP F (cid:105) = +0.45 ± − ( σ = 1.96 km s − ). Ourprogram stars also have RV s in the Gaia Data Release 2 Catalog (Arenou et al. 2018),and we list them also in Table 2.The agreement betwen HPF and Gaia is very good: (cid:104) RV opt - RV Gaia (cid:105) = − ± − ( σ = 0.31 km s − ).2.2. Other Spectra https://github.com/grzeimann/Goldilocks Documentation http://iraf.noao.edu/ Sneden et al.
Our program stars have all had previous extensive analyses from optical region spectra. As described in detail byAf¸sar18a, the aHB high-resolution spectra were obtained with the Robert G. Tull Cross-Dispersed Echelle spectrograph(Tull et al. 1995) on the McDonald Observatory 2.7m Harlan J. Smith telescope. The instrumental setup yielded aspectral resolving power of R = 60,000 with effective wavelength coverage for the RHB observations of 4100 − S/N values always exceeded 100. The Tull Spectrograph was also used to obtain a very high
S/N spectrum of the very metal-poor HD 122563. The optical spectrum for open cluster giant NGC 6940 MMU105 was acquired with the High Resolution Spectrograph (Tull 1998) of the Hobby-Eberly Telescope at McDonaldObservatory. It was configured to cover the spectral range 5100 − H and K bands with the ImmersionGrating Infrared Spectrometer (IGRINS; Yuk et al. 2010; Park et al. 2014) when this portable instrument was locatedat the McDonald Observatory 2.7m telescope. The IGRINS spectra for these bright RHB stars covered a wavelengthrange 14,800 − R (cid:39) S/N ≥ ABUNDANCE ANALYSES3.1.
Line Lists
In this exploration of HPF spectra of red giants we first sought to identify all useful atomic and molecular transitionsfor abundances analysis. HPF spectra cover the zyJ wavelength range 8170 − − µ m). In RHB starsthis spectral region contains many atomic lines, nearly all of which arise from light and Fe − group elements. In Figure 1we show one of the HPF spectral orders that has several strong neutral − species lines of Na, Si, Ca, Fe, and Cr, oneof the H i Paschen series lines, Fe ii ii ii line and its multiplet partnersat 10036.66 and 10327.31 ˚Aare zyJ region rarities: ionized-species neutron-capture element ( Z >
30) transitions.Neutron − capture elemental transitions are rarely strong enough in our stars for detection in this wavelength domain.A few Fe − group ionized species lines can be seen in red giant stars, but only one Fe ii multiplet is strong enough forabundance analysis in our program stars.There are also many molecular features in the zyJ region: OH ground state Σ + , CO ground state X Π rovibrational,and CN A Π − X Σ + red system transitions. But in general only CN lines have large absorption strengths. They cansignificantly contaminate many otherwise useful atomic transitions. In practice this problem forces examination ofeach potential atomic line for the presence of blends.With potential line contamination issues, we chose candidate atomic features for abundance analysis by comparingHPF data in each spectral order with synthetic spectra for RHB star HIP 114809. Detailed matching of syntheticand observed spectra was not sought at this point. For all computations of synthetic spectra and predicted single-lineequivalent widths ( EW s) in this paper we used the current version of the LTE plane-parallel line analysis code MOOG(Sneden 1973) . The input atomic and molecular line lists were generated by the auxiliary code linemake . We adoptedthe model atmosphere for HIP 114809 derived for the RHB survey by Af¸sar18a, and whose parameters are given inTable 3. By this process we identified useful transitions of 16 elements; 14 of these have only neutral-species lines. Asingle Ti ii line was detected, along with the very strong Sr ii lines mentioned above.In Table 4 we list the atomic lines chosen for this study. A fundamental limitation of abundance analyses in the zyJ spectral region is the lack of substantial sets of reliable transition probabilities based on laboratory spectroscopy.For this initial HPF abundance study of RHB stars we decided to avoid deriving empirical log( gf ) values from theobserved solar spectrum, as their values are difficult to disentangle from choices of solar model atmosphere, linedamping parameters, and solar microturbulent velocity. The Wisconsin lab atomic physics group has publishedtransition probabilities in the HPF domain for Ti i (Lawler et al. 2013), and Co i (Lawler et al. 2015). For Fe i acollaboration between investigators at Imperial College London Blackett Laboratory and the Wisconsin group (O’Brianet al. 1991; Ruffoni et al. 2014; Den Hartog et al. 2014; Belmonte et al. 2017) has yielded an extensive set of reliable gf values ranging from the U V to the near- IR . We used only lines from those studies in our Ti, Fe, and Co analyses. This ∼ chris/moog.html Available at https://github.com/vmplacco/linemake
PF Chemical Compositions gf source consistency for these elements in HPF and optical spectra, at the expense of ignoring potentiallyuseful transitions in the HPF region.Of the remaining 13 species, there are 8 with detected lines in our HPF data that have transition probabilities givenin the NIST atomic line database (Kramida et al. 2019) . The NIST atomic line catalog is a critical compilation,and they provide transition probability accuracy assessments. For our lines these quality estimates range from “A+”(uncertainties less than 2%) to “D” (uncertainties less than 50%). We list the NIST line qualities in Table 4. We haveincluded all promising lines without discrimination among NIST quality values, but caution should be observed forthose lines with ratings of “C” and lower.Finally, detected HPF transitions of Ca i , Cr i , Mn i , and Ni i lack recent lab transition work and they do not appearin the NIST database. For these lines we adopted the gf − values given in the Kurucz (2011, 2018) semiempiricalline compendium. Decades of work by Kurucz have yielded a database of millions of transitions. Pragmatic mixingof results from laboratory and theoretical studies along with fresh computations has produced good matches betweenstellar observed and predicted spectra, but individual transition probabilities have widely varying accuracies; thisshould be kept in mind here.S i deserves a more expanded description. There are several lines in the zyJ spectral domain, and they generallyhave gf -values available in the NIST database. Table 1 of Spite et al. (2011) lists the most promising transitions inthe optical and near-IR regions. In our HPF spectra of RHB stars we have all of these lines except the optical tripletat 6757 ˚A (see Costa Silva et al. 2020 for a large survey based on those lines). Detections in our spectra include theground-state [S i ] line at 10821.11 ˚A. Caffau & Ludwig (2007) have explored the use of this transition in solar-typestars. It is an attractive alternative to the high-excitation ( χ (cid:38) i lines that are normally studied, because ithas little T eff sensitivity and is essentially formed in LTE. However, there are some practical problems with this linefor our abundance analyses: (a) the line is very weak; (b) it has an un-identified blend in its red wing (see Figure 1in Caffau & Ludwig); and (c) it is located only 2 ˚A from the edge of its HPF spectral order. For the present HPFexploration we drop the 10821 ˚A line, but it should considered more carefully in future papers.The S i triplet at 9212.9, 9228.1, and 9237.5 ˚A is strong and has been used in several S abundance studies (e.g., Spiteet al. 2011, Koch & Caffau 2011). In our HPF spectra these lines are useful, but with two cautions. First, they aresomewhat saturated in our RHB stars, and the derived S abundances have some dependence on assumed microturbulentvelocity V mic . Second, the spectral region surrounding this S i has a large amount of telluric contamination, which isa potential problem with HPF located at McDonald Observatory. The papers cited here were based on data obtainedat Chilean observatories, with much less water vapor contaminating the stellar spectra.The S i triplet near 10457 ˚A is about half as strong as the 9200 ˚A one, and for metal-rich giants such as ourRHBs these lines are easy to analyze. Finally, Ryde et al. (2019) has identified a line at 10635 ˚A as a S i transition,particularly for evolved stars with T eff (cid:38) EW measurements. About 20 of the lines were either somewhatblended, or very weak, or have hyperfine structure. For these we performed synthetic spectrum calculations.3.2. Equivalent Widths
For the majority of transitions we derived abundances from EW computations. We measured EW s with thespecialized software package SP ECT RE
Fitzpatrick & Sneden (1987) . Interactive fits were made for all lines, withmost of them modeled with Gaussian approximations. For the strongest lines, we used empirically-determined Voigtprofiles. We also set local continua interactively in this procedure, since the IRAF continuum task produce goodgeneral fits to each HPF order blaze function but are inadequate for the small spectral intervals surrounding eachmeasured line. https://physics.nist.gov/PhysRefData/ASD/lines form.html http://kurucz.harvard.edu/linelists.html ∼ chris/spectre.html Sneden et al.
The short wavelength limit for HPF is ∼ ∼ In the parts of this wavelength domain that are relatively free from telluric absorption( ∼ − EW s for RHB stars HIP 33578, HIP 114809, and metal-poor HD 122563 using spectrafrom both instruments. In Figure 2 we show the EW comparisons. The agreement between HPF and TS23 values isat the level of our ability to reliably make EW measurements. Decisions on continuum placement and line profile fitscause EW measurement scatter at the level indicated by the line-to-line scatter. For the strongest lines, assumptionsof appropriate line profile shape (Gaussian or Voigt) also can affect the EW s, but the fractional EW differences inweak and strong lines are similar. The two spectrographs yield essentially identical spectra of red giant stars.3.3. Adopted Model Atmospheres
The model atmospheric parameters of our science targets have been adopted from our previous efforts (Af¸sar16,BT16, Af¸sar18a), in which we describe the derivation of model atmospheres in detail. In summary here, we used asemi-automated version of the MOOG code introduced in § T eff , 0.25 for log g and V mic , and 0.1 for [Fe/H];see Af¸sar18a § Derived Abundances
We determined the abundances of most elements by matching their equivalent widths with computed ones, usingthe stellar atmosphere models (Table 3) and their line parameters (Table 4) described above. Iterations on the outputabundances and re-examinations of the spectra resulted in identification and elimination of aberrant lines in individualstars. In Table 5 we give line-by-line abundances in log (cid:15) units. Program star metallicities [Fe/H] computed fromthe Fe i and Fe ii lines are listed in Table 3. To compute the [Fe/H] values and subsequent abundance ratios [X/Fe]we have assumed the solar abundance set recommended by Asplund et al. (2009).HPF metallicities of RHB stars are in reasonable accord with those derived from optical data (Af¸sar18a). InFigure 3 we illustrate the comparisons. The mean difference for Fe i between the two metallicity sets, defined as∆[Fe i /H] ≡ (cid:104) [Fe i /H] HPF − [Fe i /H] opt (cid:105) = 0.04 ( σ = 0.06). Abundances from both spectral regions are based on thesame Wisconsin/London group transition probability sources. The comparison for Fe ii shows a larger offset, ∆[Fe ii /H] = 0.08 ( σ = 0.05). All of the optical and HPF gf -values for Fe ii are taken from the NIST database, andnearly all of them originate with a critical compilation by Raassen & Uylings (1998), who considered calculated andexperimental transition probabilities in their study. The internal consistency of the Fe ii HPF lines is good, since thetypical line-to-line abundance scatter in this species is small, (cid:104) σ (cid:105) (cid:39) zyJ , and IGRINS HK spectroscopic analyses, extending from ∼ i and Fe ii rangeover (cid:104) [Fe/H] opt (cid:105) = − − (cid:104) [Fe i /H] IGRINS (cid:105) = − σ = 0.04 (Table 6 in Af¸sar18b). The mean Feabundances have similar scatter in other program stars.The optical and HPF Fe i abundances have been derived with laboratory transition probabilities (O’Brian et al. 1991;Den Hartog et al. 2014; Ruffoni et al. 2014; Belmonte et al. 2017). Unfortunately, there are very few recent laboratorylog( gf ) values for this species (and for most other species as well) for wavelengths beyond 10,000 ˚A. Therefore the[Fe/H] IGRINS abundances in Af¸sar18b were determined with transition probabilities derived from reverse solar analyses.Additionally, the formal σ -values and the appearance of Figure 4 show that the line-to-line scatter is larger in the HPFabundances than in the optical ones. At least two factors may account for this effect. First, on average the near- IR Fe i lines are much weaker than optical and U V ones. They come from relatively weak branches of multiplets, and The HPF low wavelength limit is defined by the spectrograph setup, whose disperser positions cannot be altered. For TS23 the cross-disperser tilt can be changed to put very long wavelength echelle orders on its CCD detector, but (a) the detector quantum efficiencystrongly declines beyond 9000˚A; (b) the detector suffers increasing amounts of fringing toward 10000 ˚A; and (c) telluric H O absorptionlines are extremely strong in the ∼ − For elements A and B, [A/B] = log ( N A /N B ) (cid:63) – log ( N A /N B ) (cid:12) and log (cid:15) (A) = log ( N A /N H ) + 12.0 . Also, metallicity will be taken tobe the [Fe/H] value. PF Chemical Compositions Second, all of the HPF Fe ii lines occur in the range 8200 − ii abundances with full synthetic spectra, and discarded lines that appear to be severely blended.Probably future abundance surveys can successfully extract reliable abundances from EW analyses of our chosen Fe ii lines, but caution is warranted. 3.5. Abundance Ratios
Mean abundance ratios [X/Fe] are presented in Table 6, and in Figure 5 we show the trends of [X/Fe] with [Fe/H]metallicity. Carbon abundances are not plotted in this figure; they will be discussed in § gf heterogeneity from species to species; see § § − (cid:46) [Fe/H] (cid:46) (cid:39)
0, from HPF spectra as they do in abundance surveys in other spectral regions. Formingaverages of these elements for each star, and defining [Fe-group/Fe] ≡ [ (cid:104) Ti i ,Ti ii ,Cr i ,Mn i ,Co i ,Ni i /Fe (cid:105) ], for thewhole Fe group, the mean for the 13 star RHB set is (cid:104) [Fe-group/Fe] (cid:105) = − σ = 0.04). The Fe-group abundancemeans for each star are shown in the bottom middle panel of Figure 5, showing no trend with metallicity, as expectedfrom previous Galactic disk abundance surveys (e.g., Reddy et al. 2003, 2006; Bensby et al. 2014, Af¸sar18a)The α elements Mg, Si, S, and Ca show a small trend of increasing [X/Fe] values with decreasing [Fe/H]. Defining[ α /Fe] ≡ [ (cid:104) Mg, Si, S, Ca/Fe (cid:105) ], we plot the mean α values for the RHB stars in the bottom right panel of Figure 5.At [Fe/H] (cid:39) α /Fe] (cid:39) (cid:46) − α /Fe] (cid:39) +0.2, again in agreement with the trendsfound in the survey cited above.3.6. Non-Local Thermodynamic Equilibrium Computations
Local thermodynamic equilibrium (LTE) can be an adequate approximation for abundance computations in redgiants for ionized majority species, e.g., the lanthanide ions. However, almost all of our HPF transitions arise fromneutral species. Most of their parent elements are heavily ionized in line-forming atmospheric levels, and their neutraltransitions in the HPF domain arise from high-excitation levels (averaging ∼ P ySM E , the python version of the spectroscopicanalysis code
SM E (Piskunov & Valenti 2017). The computations here are initial estimates of the magnitudes ofnon-LTE effects for each HPF line in a typical RGB star. A more detailed investigation will be carried out in thefuture.For now, we feel more confident with the general magnitude of the non-LTE corrections, and so in Table 7 we presentmean values, standard deviations, and number of lines involved in the calculations for six neutral species for a typicalRHB star and for HD 122563 (to be discussed in § (cid:104) ∆ corr (cid:105) (cid:39) − σ (cid:39) (cid:104) ∆ corr (cid:105) values with green arrows in the panels for Na, Mg, K, and Ca. It is apparent that in each ofthese cases the estimated non-LTE shift significantly reduces the apparent LTE-based overabundances.3.7. The CNO Group This transition probability problem is even more acute in other species, e.g., Fe ii , which has a rich spectrum of lines in the UV ( λ < ii lines were found in our HPF spectra of RHB stars, and there are none in the IGRINS wavelength region. Sneden et al.
Our RHB stars have few CNO strong abundance indicators in the zyJ spectral range. There are some easily-identified C i lines, but this species has played only a very minor role in the extensive literature on C abundances inevolved cool stars. Additionally, there are no strong O i or OH lines in this wavelength domain, thwarting any attemptto derive O abundances. Two prominent CN red-system (0 −
0) bands are seen in the metal-rich stars of Figure 1: theR-branch head at (cid:39) (cid:39) mean /Fe] values are based on multiple C indicators: the CH G-band in the4280 − Swan bands with bandheads near 5160 and 5630 ˚A. Abundances from C i are determinedand tabulated by Bozkurt et al. but they do not participate in the C abundance means. The optical N abundanceshave been determined using C mean abundances derived earlier in concert with the O abundances (from the [O i ] line).Carbon isotopic ratios come mostly from the 8004 ˚A CN feature that is the basis for most C/ C estimates in redgiants.Abundances from HPF C i transitions in Table 8 are repeated from Table 6. Mean values for the stellar sam-ple have not been computed because star-to-star abundance differences for C, N, and C/ C are natural inevolved giants. The three sets of C abundances in Table 8 are offset from each other but they do correlatewell: (cid:104) [C CI,opt /Fe] − [C mean,opt /F e ] (cid:105) = +0.10, σ = 0.08; (cid:104) [C CI,HP F /Fe] − [C CI,opt /F e ] (cid:105) = +0.19, σ = 0.08; and (cid:104) [C CI,HP F /Fe] − [C mean,opt /F e ] (cid:105) = +0.30, σ = 0.12. The C i HPF lines clearly yield aberrantly high abundancescompared to the other values. The uncertainties in the mean differences are consistent with uncertainties in theobservations and abundance computations involving C i lines. This issue should be investigated in the future, withattention to accuracy and consistency of transition probability sources, stellar atmosphere parameter dependences,and non-LTE sensitivities of these high-excitation ( χ (cid:38) mean abundances of Table 6. Formally the optical and HPF N abundances agree very well: (cid:104) [N CN,HP F /Fe] − [N CN,opt /F e ] (cid:105) = +0.01, σ = 0.05. However, the CN transition data used here all come from onerecent extensive laboratory study (Brooke et al. 2014), and Sneden et al. (2014) have shown that various CN bandsfrom blue through the near- IR yield internally consistent abundance results. The red system (0-0) bands in the HPFregion are much stronger than the (2-0) bands near 8000 ˚A that have dominated CN studies in the optical spectralregion. Probably this will enable CN detection in warmer RHB stars than those included in this study.Finally, we attempted to estimate C/ C ratios from the CN (0-0) Q-branch bandhead that occurs at 10923 ˚A,almost 3 ˚A blueward of the CN bandhead. However, in our RHB stars the CN lines are very weak, and this smallspectral region is contaminated by telluric lines that are difficult to remove accurately. Therefore Table 8 lists onlynine isotopic ratios, and a couple of these have large estimated uncertainties. For most stars the best conclusion is thatthe HPF C/ C values are consistent with, but probably inferior to those derived from the optical (2-0) transitions. STARS OF SPECIAL INTERESTAf¸sar18a identified about 150 true field RHB giants out of their original candidate sample of ∼
340 stars. Our RHBtargets to observe with HPF included two that Af¸sar18b analyzed with IGRINS H- and K-band spectra, 12 chosen atrandom from the Af¸sar18a list, and two that seemed to be worth special attention. Here we discuss these two RHBfield stars, the one open cluster giant, and metal-poor HD 122563.4.1.
HIP 33578: Undiluted Interior CN − Cycle Products on the Surface
HIP 33578 is an RHB field star that was an unremarkable member of the Af¸sar18a survey. But that study con-centrated on kinematics, Fe metallicities, and limited Fe-group and α element abundance ratios of its RHB sample.Bozkurt et al. (in preparation) will provide much more extensive abundance information on the LiCNO elements thatare sensitive to stellar interior fusion cycles and envelope mixing, along with results many other element groups. In thiswork we have noticed that HIP 33578 is one of the rare metal-rich field giants with a very low carbon isotopic ratio.Fortunately we were also able to obtain an IGRINS H- and K-band spectrum of this star. The optical, IGRINS, andnow HPF spectra present a solid observational case for the extreme C/ C value, as we illustrate in Figure 6. For all
PF Chemical Compositions < CN and CO in each small spectral interval. The best estimate from synthetic/observedspectrum matches from each of these molecular bands is C/ C = 3 ±
1. This ratio at the surface of HIP 33578is lower than, but consistent within the observational and theoretical uncertainties, of the interior CN-cycle C/ Cvalue presented first discussed in detail by Caughlan & Fowler (1962).Further support comes from the C and N abundances. From Table 8 the mean values for RHB stars excludingHIP 33578 are (cid:104) [C/Fe] (cid:105) = − σ = 0.14) and (cid:104) [N/Fe] (cid:105) = 0.57 ( σ = 0.14). The optical results for HIP 33578 are[C/Fe] = − C/ C of this star.The LiCNO abundances and carbon isotopic ratio of HIP 33578 are in the domain of the “weak G-Band” (wkG)disk giants (Sneden et al. 1978; Adamczak & Lambert 2013; Palacios et al. 2016). For HIP 33578 Bozkurt et al. (2021,in preparation), supported by our HPF results, have derived log (cid:15) (Li) < − C/ C = 3, while approximate mean values from Adamczak & Lambert (2013) and Palacios et al. (2016)are (cid:104) log (cid:15) (Li) (cid:105) = 3 to 10, (cid:104) [C/Fe] (cid:105) (cid:39) − (cid:104) [N/Fe] (cid:105) (cid:39) (cid:104) [O/Fe] (cid:105) = 0.0, and (cid:104) C/ C (cid:105) (cid:39) → N conversion. This star’s Li abundance is very much smaller than the typical wkG giant, but thesamples of both Adamczak & Lambert (2013) and Palacios et al. (2016) have stars with no obvious Li enhancement.Our single star cannot provide much new insight into the evolutionary history of this small red giant subclass. However,the RHB status of HIP 33578 is shared by many wkG stars. From its T eff = 5118 K and its photometry and parallaxgiven in Af¸sar et al. (2018a) we suggest that log(L/L (cid:12) (cid:39) HIP 99789: a rare Red Horizontal Branch Star with Extremely High Lithium
In contrast to HIP 33578, the RHB star HIP 99789 has ordinary CNO abundances and C/ C for our RHBsample (Table 8) but an extremely high Li content: log (cid:15) (Li) (cid:39) i absorption line at10830 ˚A. In Figure 7 we show a montage of our RHB spectra in the wavelength region surrounding the 10830 ˚A line.This transition arises from a very high-excitation (19.8 eV) metastable level of He i . It has been used extensivelyin past studies to trace solar/stellar chromospheric activity and wind outflows, and has gained recent attention as apotential indicator of outflows from exoplanet atmospheres (e.g., Ninan et al. 2020). Inspection of Figure 7 suggeststhat more than half of our RHB sample, 8 out of 13 stars, have extremely little or no detectable He i absorption at10830 ˚A. Another three stars, HIP 476, HIP 33578, and HIP113610, have modest He i lines. But HIP 29962 andHIP 99789 have strong 10830 ˚A absorptions, suggesting substantial chromospheric activity in their atmospheres. Wecannot identify any other distinguishing spectroscopic signatures in HIP 99789, but the possible connection betweenhigh Li abundance and He i activity is intriguing and should be pursued with a dedicated set of HPF observations ofa larger sample of high-Li stars.Bozkurt et al. (in preparation) will show that HIP 99789 is the most Li-rich member of the Af¸sar18a sample. However,its appearance among the RHBs may match a recent discovery about evolved stars with high Li. Singh et al. (2019)have determined Li abundances in red giants whose evolutionary status can be assigned to either the first-ascent RGor the red clump on the basis of their asteroseismological properties determined from the Kepler satellite (Mathuret al. 2017). Significant Li abundances can only be found in the He core-burning red clump stars (see their Figure 4).The standard red clump is cooler than the RHB; the highest T eff reported by Mathur et al. for their Li-rich stars is4999 K, and almost all other stars are cooler than 4900 K. Most of our RHBs have T eff (cid:38) T eff = 5054 K and log g = 2.41 (Table 3). The temperature is not radically different than those of theMathur et al. Li-rich sample, and the gravity is in the middle of their log g distribution. HIP 99789 seems to supporttheir assertion that Li enhancement is strongly associated with the He-core burning phase of late stellar evolution.However, many decades ago Alexander (1967) speculated that red giants could temporarily enhance their Li contentsby ingesting companion planets in their bloated envelopes. Some evidence connecting Li-rich giants to planetarysystems has accumulated recently, e.g., Adam´ow et al. (2018) and references therein. HIP 99789 may have one ormore stellar or planetary companions, as indicated by its abnormally large apparent spatial acceleration. Close binarycompanions can influence unresolved spectra by diluting line depths of the primary star, adding in new lines, andproducing anomalous RV offsets. None of our targets show signs of being double-line spectroscopic binaries from their0 Sneden et al.
RV cross-correlation functions. But low mass ratio or longer-period companions could still be present in these systems.Recently, Brandt et al. (2019) cross-matched astrometry from
Hipparcos and
Gaia
DR2 to produce the
Hipparcos-GaiaCatalog of Accelerations (HGCA), which provides systematics-corrected proper motions spanning a ≈ Gaia and
Hipparcos-Gaia scaled positional difference propermotions to calculate accelerations in our sample. Among the 17 stars we observed, all except for NGC6940 MMU 105have
Hipparcos astrometry and therefore have entries in the HGCA catalog. Four out of these 16 stars have significantastrometric accelerations: HIP 29962 (149 ±
22 m s − yr − ), HIP 99789 (670 ±
19 m s − yr − ), HIP 113610 (255 ±
18 m s − yr − ), and HIP 114809 (76 ± − yr − ). Interestingly, HIP 29962, HIP 99789, and HIP 114809 also havethe largest discrepancies ( ≈ − ) between the RVs measured from the optical Tull data and our HPF spectra,offering further evidence that these stars have companions based on radial accelerations. The limited astrometric andRV sampling makes it challenging to constrain the nature of the companion, but the amplitude of these accelerationssuggest that the companions are most likely low-mass stars or perhaps white dwarfs, both of which are not expectedto meaningfully impact the spectra and abundances we derive in this study. While no solid connection can be arguedbetween the large Li of HIP 99789 and its abnormally large apparent acceleration, it should be studied in the future.4.3. NGC 6940 MMU 105: The First Open Cluster Star With Extremely Broad High Resolution Spectral Coverage
The cool red giant MMU 105 is a member of the open cluster NGC 6940, and we have previously reported on itschemical composition from optical (BT16) and IGRINS (BT19) spectra. The instrumental setups and observationswere like those discussed in § − σ = 0.07) is somewhat smaller than the cluster mean [Fe/H] = − σ = 0.06). Note that MMU105 is the coolest (4765 K) of the 12 RGs. The IGRINS H- and K-band iron abundance is larger, (cid:104) [Fe/H] (cid:105) = − α elements in this cluster have theabundances of slightly metal-poor disk stars. For elements with 21 ≤ Z ≤ (cid:104) [Fe-group/Fe] (cid:105) = 0.01 ± (cid:104) [ α /Fe] (cid:105) = 0.17 ± >
30) appear to be slightly overabundant byan average of about 0.2 dex. From IGRINS data, (cid:104) [Fe-group/Fe] (cid:105) = 0.15 ± α /Fe] = [ (cid:104) Mg, Si, S, Ca/Fe (cid:105) ] =0.11 ± C/ C optical = 15 +3 − , C/ C IGRINS = 23 ±
3, and now we can add C/ C HP F = 15 +5 − . Both optical and IGRINSdata reveal underabundances of C and overabundances of N. The HPF N abundances from the (0-0) bandheads agreewell with the previous values, and the HPF C i abundances are in accord with the optical and IGRINS abundancesafter accounting for the systematic C i offset discussed above. The isochrones discussed in BT16 and BT19 suggestthat all NGC 6940 RGs are core He-burning red clump stars with mostly undergone canonical first dredge-up. TheHPF C, N, and C/ C values should not be the primary data for assessing mixing in NGC 6940 MMU 105, butclearly they are in agreement with our previous optical and IGRINS studies.4.4.
HD 122563: The Brightest Very Metal − Poor Halo Giant
The red giant HD 122563 (HIP 068594) is the only very low metallicity member of the Yale Catalog of Bright Stars(Hoffleit & Warren 1995). It was first noted as a high proper motion star (Roman 1955), and was one of the first verymetal − poor stars subject to detailed abundance analysis (Wallerstein et al. 1963). Among many noteworthy chemicalcomposition features, HD 122563 has a very large relative oxygen abundance ([O/Fe] (cid:39) +0.6; Lambert et al. 1974),very low carbon isotopic ratio ( C/ C = 5 ±
2; Lambert & Sneden 1977), and weak neutron-capture abundances inan r -process pattern (Sneden & Parthasarathy 1983; Honda et al. 2006).The HD 122563 spectrum portion displayed in Figure 1 has identifications of three Si i lines and one each of H i ,Ca i , Sr ii . Neglecting the H i Paschen line, these five transitions represent a significant fraction (5/34) of the total As renormalized to the solar abundances of Asplund et al. (2009); see BT19.
PF Chemical Compositions § i HP F /H] = − σ = 0.07), is consistent with optical value: [Fe i opt /H] = − σ = 0.12), Of the other species entered in Table 10, we draw attention to Si i , S i , and Sr ii .The abundances from Si i lines in all three spectral domains are displayed in Figure 8. The transition probabilitiesfor all the transitions were taken from the NIST database. Si abundances derived from HPF lines perhaps are the mostreliable set. Certainly, as discussed in § RW ≡ log( EW / λ ) (cid:46) − zyJ Si i lines in the HPF spectrum yield amean Si abundance about 0.1 dex higher than do the lines in the IGRINS H- and K-band region. Resolution of thissmall discrepancy should begin with a comprehensive new transition probability study.We have detected S i lines in HD 122563, as we illustrate in Figure 9. The triplet of lines near 10457 ˚A is strong inmetal-rich giants such as HIP 114809 (top panel), and the 10455 ˚A line is blended enough that its abundance mustbe derived by synthetic spectrum computations. The two strongest lines of the triplet are clearly seen in HD 122563(bottom panel). We have derived an S abundance from two of these lines and all three of the 9220 ˚A lines. Our LTEabundance is [S/Fe] = +0.75 ( σ = 0.26), and if we apply the non-LTE correction of − CONCLUSIONSWe have used the Habitable Zone Planet Finder to explore the spectra of red giant stars in the zyJ (8100 − C/ C ratios. It ispossibly a member of the weak G-band subclass. HIP 99789 is a (perhaps not so) rare Li-rich RHB star. The uniqueHPF contribution is our discovery that this star has a strong He i i triplet,yielding an abundance in one of the lowest-metallicity cases.HPF is dedicated to measuring high-precision radial velocities in M-dwarf stars, but we have shown in this paperthat it can be an attractive instrument for investigating the chemical compositions of warmer red giants. Most usefultransitions arise from light and Fe-group neutral species. For some elements, e.g., Mg, Si, and S, the abundances fromthe HPF zyJ spectral region may prove to be superior to those from the optical region.2 Sneden et al.
ACKNOWLEDGMENTSWe thank Noriyuki Matsunaga and George Preston for helpful comments on the manuscript. These results are basedon observations obtained with the Habitable-zone Planet Finder Spectrograph on the Hobby-Eberly Telescope. Wethank the Resident Astronomers and Telescope Operators at the HET for the skillful execution of our observationsof our observations with HPF. The Hobby-Eberly Telescope is a joint project of the University of Texas at Austin,the Pennsylvania State University, Ludwig-Maximilians-Universit¨at M¨unchen, and Georg-August Universit¨at Gottin-gen. The HET is named in honor of its principal benefactors, William P. Hobby and Robert E. Eberly. The HETcollaboration acknowledges the support and resources from the Texas Advanced Computing Center. We acknowledgesupport from NSF grants AST-1006676, AST-1126413, AST-1310885, AST-1517592, AST-1310875, AST-1910954,AST-1907622, AST-1909506, and support from the Heising-Simons foundation in our pursuit of precision spectroscopyin the NIR.This work also used the Immersion Grating Infrared Spectrometer (IGRINS) that was developed under a collabo-ration between the University of Texas at Austin and the Korea Astronomy and Space Science Institute (KASI) withthe financial support of the US National Science Foundation under grants AST-1229522 and AST-1702267, of theMcDonald Observatory of the University of Texas at Austin, and of the Korean GMT Project of KASI. Support forthis study was also provided by National Science Foundation grant AST-1616040.
Software: linemake (https://github.com/vmplacco/linemake), MOOG (Sneden 1973), IRAF (Tody 1986, Tody1993), SPECTRE (Fitzpatrick & Sneden 1987), Goldilocks (https://github.com/grzeimann/Goldilocks Documentation),PySME (Wehrhahn 2019; 10.5281/zenodo.3520617)
PF Chemical Compositions
Adamczak, J., & Lambert, D. L. 2013, ApJ, 765, 155Adam´ow, M., Niedzielski, A., Kowalik, K., et al. 2018,A&A, 613, A47Af¸sar, M., Bozkurt, Z., B¨ocek Topcu, G., et al. 2018a, AJ,155, 240Af¸sar, M., Sneden, C., Frebel, A., et al. 2016, ApJ, 819, 103Afs , ar, M., Sneden, C., Wood, M. P., et al. 2018b, ApJ, 865,44Alexander, J. B. 1967, The Observatory, 87, 238Amarsi, A. M., Lind, K., Osorio, Y., et al. 2020, A&A, 642,A62Arenou, F., Luri, X., Babusiaux, C., et al. 2018, A&A, 616,A17Asplund, M., Grevesse, N., Sauval, A. J., & Scott, P. 2009,ARA&A, 47, 481Belmonte, M. T., Pickering, J. C., Ruffoni, M. P., et al.2017, ApJ, 848, 125Bensby, T., Feltzing, S., & Oey, M. S. 2014, A&A, 562, A71B¨ocek Topcu, G., Af¸sar, M., Schaeuble, M., & Sneden, C.2015, MNRAS, 446, 3562B¨ocek Topcu, G., Af¸sar, M., & Sneden, C. 2016, MNRAS,463, 580B¨ocek Topcu, G., Af¸sar, M., Sneden, C., et al. 2019,MNRAS, 485, 4625Bowler, B. P., Cochran, W. D., Endl, M., et al. 2020, arXive-prints, arXiv:2012.04847Brandt, T. D., Dupuy, T. J., & Bowler, B. P. 2019, AJ,158, 140Brooke, J. S. A., Ram, R. S., Western, C. M., et al. 2014,ApJS, 210, 23Caffau, E., & Ludwig, H. G. 2007, A&A, 467, L11Caughlan, G. R., & Fowler, W. A. 1962, ApJ, 136, 453Costa Silva, A. R., Delgado Mena, E., & Tsantaki, M. 2020,A&A, 634, A136De Silva, G. M., Freeman, K. C., Bland-Hawthorn, J., et al.2015, MNRAS, 449, 2604Den Hartog, E. A., Lawler, J. E., Sneden, C., Cowan, J. J.,& Brukhovesky, A. 2019, ApJS, 243, 33Den Hartog, E. A., Ruffoni, M. P., Lawler, J. E., et al.2014, ApJS, 215, 23Fitzpatrick, M. J. 1993, in Astronomical Society of thePacific Conference Series, Vol. 52, Astronomical DataAnalysis Software and Systems II, ed. R. J. Hanisch,R. J. V. Brissenden, & J. Barnes, 472Fitzpatrick, M. J., & Sneden, C. 1987, in Bulletin of theAmerican Astronomical Society, Vol. 19, Bulletin of theAmerican Astronomical Society, 1129Gaia Collaboration, Brown, A. G. A., Vallenari, A., et al.2018, A&A, 616, A1 Girardi, L. 1999, MNRAS, 308, 818—. 2016, ARA&A, 54, 95Girardi, L., Groenewegen, M. A. T., Weiss, A., & Salaris,M. 1998, MNRAS, 301, 149Halverson, S., Mahadevan, S., Ramsey, L., et al. 2014, inSociety of Photo-Optical Instrumentation Engineers(SPIE) Conference Series, Vol. 9147, Ground-based andAirborne Instrumentation for Astronomy V, 91477ZHoffleit, D., & Warren, W. H., J. 1995, VizieR Online DataCatalog, V/50Honda, S., Aoki, W., Ishimaru, Y., Wanajo, S., & Ryan,S. G. 2006, ApJ, 643, 1180Horne, K. 1986, PASP, 98, 609Koch, A., & Caffau, E. 2011, A&A, 534, A52Kramida, A., Yu. Ralchenko, Reader, J., & and NIST ASDTeam. 2019, NIST Atomic Spectra Database (version5.7.1), [Online]. Available: https://physics.nist.gov/asd [Aug 11 2020] NationalInstitute of Standards and Technology, Gaithersburg,MD.Kurucz, R. L. 2011, Canadian Journal of Physics, 89, 417Kurucz, R. L. 2018, in Astronomical Society of the PacificConference Series, Vol. 515, Workshop on AstrophysicalOpacities, ed. C. Mendoza, S. Turck-Chi´eze, & J. Colgan,47Lambert, D. L., & Sneden, C. 1977, ApJ, 215, 597Lambert, D. L., Sneden, C., & Ries, L. M. 1974, ApJ, 188,97Lawler, J. E., Guzman, A., Wood, M. P., Sneden, C., &Cowan, J. J. 2013, ApJS, 205, 11Lawler, J. E., Sneden, C., & Cowan, J. J. 2015, ApJS, 220,13Mahadevan, S., Ramsey, L., Bender, C., et al. 2012, inSociety of Photo-Optical Instrumentation Engineers(SPIE) Conference Series, Vol. 8446, Ground-based andAirborne Instrumentation for Astronomy IV, 84461SMahadevan, S., Ramsey, L. W., Terrien, R., et al. 2014, inSociety of Photo-Optical Instrumentation Engineers(SPIE) Conference Series, Vol. 9147, Ground-based andAirborne Instrumentation for Astronomy V, 91471GMathur, S., Huber, D., Batalha, N. M., et al. 2017, ApJS,229, 30Metcalf, A. J., Anderson, T., Bender, C. F., et al. 2019,Optica, 6, 233Ninan, J. P., Bender, C. F., Mahadevan, S., et al. 2018, inSociety of Photo-Optical Instrumentation Engineers(SPIE) Conference Series, Vol. 10709, High Energy,Optical, and Infrared Detectors for Astronomy VIII,107092U Sneden et al.
Ninan, J. P., Stefansson, G., Mahadevan, S., et al. 2020,ApJ, 894, 97O’Brian, T. R., Wickliffe, M. E., Lawler, J. E., Whaling,W., & Brault, J. W. 1991, Journal of the Optical Societyof America B Optical Physics, 8, 1185Palacios, A., Jasniewicz, G., Masseron, T., et al. 2016,A&A, 587, A42Park, C., Jaffe, D. T., Yuk, I.-S., et al. 2014, in Society ofPhoto-Optical Instrumentation Engineers (SPIE)Conference Series, Vol. 9147, Society of Photo-OpticalInstrumentation Engineers (SPIE) Conference Series, 1Piskunov, N., & Valenti, J. A. 2017, A&A, 597, A16Raassen, A. J. J., & Uylings, P. H. M. 1998, Journal ofPhysics B Atomic Molecular Physics, 31, 3137Reddy, B. E., Lambert, D. L., & Allende Prieto, C. 2006,MNRAS, 367, 1329Reddy, B. E., Tomkin, J., Lambert, D. L., & AllendePrieto, C. 2003, MNRAS, 340, 304Roman, N. G. 1955, ApJS, 2, 195Roy, A., Halverson, S., Mahadevan, S., & Ramsey, L. W.2014, in Society of Photo-Optical InstrumentationEngineers (SPIE) Conference Series, Vol. 9147,Ground-based and Airborne Instrumentation forAstronomy V, 91476BRuffoni, M. P., Den Hartog, E. A., Lawler, J. E., et al.2014, MNRAS, 441, 3127 Ruiz-Dern, L., Babusiaux, C., Arenou, F., Turon, C., &Lallement, R. 2018, A&A, 609, A116Ryde, N., Hartman, H., Oliva, E., et al. 2019, A&A, 631, L3Shetrone, M., Cornell, M. E., Fowler, J. R., et al. 2007,PASP, 119, 556Singh, R., Reddy, B. E., Bharat Kumar, Y., & Antia,H. M. 2019, ApJL, 878, L21Sneden, C. 1973, ApJ, 184, 839Sneden, C., Lambert, D. L., Tomkin, J., & Peterson, R. C.1978, ApJ, 222, 585Sneden, C., Lucatello, S., Ram, R. S., Brooke, J. S. A., &Bernath, P. 2014, ApJS, 214, 26Sneden, C., & Parthasarathy, M. 1983, ApJ, 267, 757Spite, M., Caffau, E., Andrievsky, S. M., et al. 2011, A&A,528, A9Tull, R. G. 1998, in Proc. SPIE, Vol. 3355, OpticalAstronomical Instrumentation, ed. S. D’Odorico, 387–398Tull, R. G., MacQueen, P. J., Sneden, C., & Lambert, D. L.1995, PASP, 107, 251Wallerstein, G., Greenstein, J. L., Parker, R., Helfer, H. L.,& Aller, L. H. 1963, ApJ, 137, 280Yuk, I.-S., Jaffe, D. T., Barnes, S., et al. 2010, in Society ofPhoto-Optical Instrumentation Engineers (SPIE)Conference Series, Vol. 7735, Society of Photo-OpticalInstrumentation Engineers (SPIE) Conference Series, 1
PF Chemical Compositions Figure 1.
Spectra in an HPF echelle order that has only moderate telluric line contamination (removed in the reduction processdescribed in § − temperature open clusterred clump giant, and the bottom panel has the very metal − poor field giant HD 122563. In the top panel some prominent atomic lines arelabeled. All identified lines except Fe ii and Sr ii are neutral-species transitions. The CN (0 −
0) R-branch bandhead is also indicated. Sneden et al.
Figure 2.
Comparison of equivalent widths measured with the McDonald Observatory TS23 optical echelle spectrograph (Tull et al.1995) and the HPF. The dashed line represents the mean difference, (cid:104) ∆( EW ) (cid:105) = − σ = 1.6 m˚A. Three stars, identified in the figure legend, were used for this test. See 3.2 for details. Figure 3.
Comparison of Fe i and Fe ii metallicities with those from optical spectra Af¸sar18a after adjustment to the solar abundancesof Asplund et al. (2009). The dashed line represents the mean offset between the two Fe i sets, and the dotted lines represent the standarddeviation. PF Chemical Compositions Figure 4.
Abundances of Fe lines plotted as a function of wavelength. The wavelength axis is plotted logarithmically due to its largerange. Results from optical TS23, HPF, and IGRINS spectra are shown with different colors. The lines denote statistics of the HPF Fe i results; the solid line is for (cid:104) [Fe i /H] (cid:105) HPF , and the two dashed lines represent the 1- σ standard deviation for the HPF abundances. Sneden et al.
Figure 5.
Abundance ratios [X/Fe] plotted as functions of the [Fe/H] metallicity for most elements in this study. The solar valuesof these quantities, [Fe/H] ≡ ≡ ii lines are depicted with open symbols. In the Fe-group (bottom middle) panel, the points are colored blueto emphasize that they are simple means of Ti, Cr, Mn, Co and Ni abundances. In the α -group (bottom right) panel, the points have redcolors; they are simple means of Mg, Si, S, and Ca abundances. PF Chemical Compositions Figure 6.
Prominent C molecular features in three wavelength domains of HIP 33578. The dots represent the observed spectra, andthe lines represent synthetic spectra with C/ C values defined in the figure legend by the line colors. Sneden et al.
Figure 7.
Spectra of the RHB stars surrounding the He i i transition. The relative flux scale for HIP 476 is correct, andvertical offsets of 0.4 have been added in succession to the spectra of the other 14 RHBs. Three prominent photospheric absorption lineshave been identified, and there are smaller atomic and CN lines that can bee seen in the plot. PF Chemical Compositions Figure 8.
Si abundances from optical (TS23), HPF, and IGRINS data in HD 122563. The solid and dashed horizontal lines show themean and standard deviation values for the HPF results.
Figure 9. S i in the RHB star HIP 114809 and in the very metal-poor giant HD 122563. In each panel the observed spectra are shownwith open circles. The red line represents the synthetic spectrum with no contribution from S i , the black line is for the best overall Sabundance, and the blue and green lines are for S abundances that are 0.3 dex smaller and larger than the best abundance. Sneden et al.
Table 1 . Basic Parameters of Target Stars and Observation Dates
Star Name RA(2000) a DEC(2000) a B a V a K a d b Date observed(h m s) ( ◦ (cid:48) (cid:48)(cid:48) ) (mag) (mag) (mag) (pc)Observed TargetsHIP 476 00 05 41.96 +13 23 46.5 6.43 5.55 3.77 122 2018-10-11HIP 19740 04 13 56.38 +09 15 49.7 5.71 4.89 2.97 128 2018-10-10HIP 27091 05 44 41.43 +12 24 55.1 7.89 6.94 4.76 291 2018-11-04HIP 28417 06 00 06.04 +27 16 19.8 7.64 6.62 4.21 145 2018-10-04HIP 29962 06 18 26.38 +49 11 09.1 8.63 7.74 5.54 199 2018-10-09HIP 32431 06 46 10.37 +23 22 16.4 7.34 6.49 4.32 221 2018-10-09HIP 33578 06 58 36.73 −
07 11 08.9 7.89 7.13 5.14 145 2018-11-19HIP 57748 11 50 36.25 +43 39 32.2 8.66 7.90 6.01 210 2019-05-18HIP 99789 20 14 46.98 +26 47 32.6 9.20 8.33 6.15 470 2018-11-21HIP 106775 21 37 41.66 +15 12 57.5 8.37 7.44 5.24 358 2018-11-16HIP 113610 23 00 37.91 +00 11 09.1 7.10 6.23 4.22 183 2018-10-08HIP 114809 23 15 23.04 +25 40 20.1 7.64 6.81 4.76 209 2018-11-12HIP 116053 23 30 54.85 +07 59 48.9 9.04 8.10 5.89 243 2018-11-19NGC6940 MMU 105 20 34 25.46 +28 05 05.6 11.90 10.66 7.82 1021 2019-06-19HD 122563 14 02 31.84 +09 41 09.9 7.10 6.19 3.73 290 2018-07-17Telluric StandardsHR 26 00 10 02.20 +11 08 44.9 5.46 5.53 5.70 93 2018-10-11HR 1307 04 13 34.56 +10 12 44.9 6.27 6.25 6.03 145 2018-10-10HR 1808 05 27 10.09 +17 57 43.9 5.30 5.40 5.59 203 2018-11-04HR 2084 05 57 59.65 +25 57 14.1 4.76 4.82 5.03 476 2018-10-04HR 2257 06 22 03.55 +59 22 19.5 6.21 6.07 5.51 152 2018-10-09HR 2529 06 51 33.05 +21 45 40.1 5.26 5.26 5.17 129 2018-10-09HR 2648 07 02 54.77 −
04 14 21.2 4.80 5.00 5.50 373 2018-11-19HIP 55182 11 17 55.15 +40 50 14.4 8.98 8.93 8.88 345 2019-05-18HR 7734 20 14 04.88 +36 36 17.5 6.45 6.46 6.36 283 2018-11-21HR 8258 21 35 27.03 +24 27 07.7 6.35 6.22 5.79 198 2018-11-16HR 8773 23 03 52.61 +03 49 12.1 4.40 4.52 4.75 125 2018-10-08HIP 115579 23 24 43.38 +36 21 43.8 6.89 7.02 7.43 413 2018-11-12HR 8963 23 37 56.80 +18 24 02.4 5.48 5.48 5.42 74 2018-11-19HR 7958 20 46 38.59 +46 31 54.1 6.36 6.30 6.04 151 2019-06-19HIP 68209 13 57 52.12 +16 12 07.5 7.65 7.58 7.41 256 2018-07-17 a SIMBAD Database b Adopted from Gaia DR2 (Gaia Collaboration et al. 2018)
PF Chemical Compositions Table 2.
Radial Velocities
Star RV a σ RV σ RV σ δ RV δ RVHPF HPF opt opt Gaia Gaia opt − HPF Gaia − HPFHIP 476 2.06 0.17 2.31 0.12 1.90 0.12 0.25 − − − − − − − − − − − − − − − − − − − − − − − − − − − − − − − − − − − − − − − − − − a All velocities are in km s − Table 3.
Program Star Model Atmospheres
Star Name T eff log g V mic [Fe i ] σ [Fe ii ] σ [Fe i ] σ ii ] σ − opt a opt opt opt opt opt opt HPF HPF HPF HPF HPF HPFField RHB StarsHIP 476 5109 2.68 1.36 − − − − − − − − − − − − − − − − − − − − − − − − − − − − − − − − − − − − − − − − − − − · · · · · · · · · HIP 114809 5139 2.59 1.31 − − − − − − − − − − − − − · · · · · · · · · a The optical values are taken from high-resolution optical spectroscopy obtained with the McDonald Observatory 2.7m TS23echelle spectrograph (Af¸sar16, BT16, Af¸sar18a). Sneden et al.
Table 4.
Atomic Lines a λ Species χ log( gf ) Source b (˚A) (eV)8335.15 C i − i − i i i − i − i i − i i − a The full version of this table in ascii form is avail-able on − line b Sources of the transition probabilities: NIST= the NIST Atomic Spectra Database,https://physics.nist.gov/asd, with nota-tion of their assessment of log( gf ) quality;LAWLER13 = Lawler et al. (2013); KURUCZ= the Kurucz (2011, 2018) compendium,http://kurucz.harvard.edu/linelists.html; DEN-HARTOG14 = Den Hartog et al. (2014);RUFFONI14 = Ruffoni et al. (2014); OBRIAN91= O’Brian et al. (1991); LAWLER15 = Lawleret al. (2015). Table 5.
Line Abundances a Species λ (1) b (2) (3) (4) (5) (6) (7) (8) (9) (10) (11) (12) (13) (14) (15)(˚A) m˚A m˚A m˚A m˚A m˚A m˚A m˚A m˚A m˚A m˚A m˚A m˚A m˚A m˚A m˚AC i · · · · · · C i · · · · · · C i · · · · · · C i · · · · · · · · · · · · · · · · · · · · · · · · C i · · · · · · · · · C i · · · · · · · · · · · · C i · · · · · · C i · · · · · · C i · · · · · · C i · · · · · · a The full version of this table in ascii form is available on − line b Star identifications: (1) HIP 476; (2) HIP 19740, (3) HIP 27091; (4) HIP 28417; (5) HIP 29962: (6) HIP 32431; (7) HIP 33578;(8) HIP 57748; (9) HIP 99789; (10) HIP 106775; (11) HIP 113610; (12) HIP 114809; (13) HIP 116043; (14) NGC 6940 MMU105; (15) HD122563
PF Chemical Compositions Table 6.
Mean Abundances
Star C i a Na i Mg i Si i P i S i K i Ca i Ti i Ti ii Cr i Mn i Co i Ni i Sr i HIP 476 b − − − − − − − − − − − − − · · · − − · · · − − · · · − · · · · · · − − − · · · − − · · · − − − · · · − · · · − · · · − · · · − · · · · · · − · · · − · · · − − − · · · − · · · − − − − − − · · · − − · · · − − · · · − − · · · − − − · · · − · · · − − · · · · · · − · · · − − · · · − − − − − − − · · · − · · · − · · · · · · − − − · · · − − · · · − a The C i mean abundances are tabulated here, but discussion of carbon is referred to 3.7 b For each star there are three rows, giving in order the mean abundances, the standard deviations σ , and the number of lines contributing to themeans. Sneden et al.
Table 7.
Mean Non-LTE Abundance Corrections
Species (cid:104) ∆ corr (cid:105) a σ (cid:104) ∆ corr (cid:105) a σ b HD122 b HD122 b C i − − i − · · · · · · · · · Mg i − − · · · i − · · · · · · · · · K i − · · · · · · · · · Ca i − − b ∆ corr is the computed shift for a line need to correct its LTE abundancefor non-LTE effects b HD122 = HD 122563
Table 8.
LiCNO Abundances
Star log (cid:15) (Li) [C CI /Fe] [C mean /Fe] a [N CN /Fe] b [O OI /Fe] C/ Coptical spectraHIP 476 0.72 − − − · · · HIP 19740 < − − − − − < − − < − − < − − < − − < − − − − − − < − − < − − − < − − · · · HIP 116053 0.42 − − − < − − − · · · − ± · · · ± · · · > · · · − ± · · · ± · · · · · · HIP 27091 · · · − ± · · · ± · · · +5 − HIP 28417 · · · ± · · · ± · · · +2 − HIP 29962 · · · − ± · · · ± · · · · · · HIP 32431 · · · − ± · · · ± · · · +5 − HIP 33578 · · · − ± · · · ± · · · +2 − HIP 57748 · · · ± · · · ± · · · · · · HIP 99789 · · · − ± · · · ± · · · · · · HIP 106775 · · · − ± · · · ± · · · +4 − HIP 113610 · · · − ± · · · ± · · · · · · HIP 114809 · · · ± · · · ± · · · +10 − HIP 116053 · · · − ± · · · ± · · · +10 − N6940 MMU 105 · · · − ± · · · ± · · · +5 − a The optical mean C abundances are based on CH and C features. b Both optical and HPF N abundances assume the C and O abundances determined from optical [O i ] transitions. PF Chemical Compositions Table 9.
HD 122563 Line Abundances a Wavelength Species χ log( gf ) log (cid:15) Source b ˚A eV10683.08 C i i i − i − i − i − i i i i − a The full version of this table in ascii form is available on − line b As defined in Table 5
Table 10.
HD 122563 MeanAbundances
Species (cid:104) [X/Fe] (cid:105) σ i − i · · · i i i · · · ii i i − a ii a For Fe ii