\omega Centauri Abundances

A Large Sample Study of Red Giants in the Globular Cluster Omega Centauri (NGC 5139)

Christian I. Johnson11affiliation: Department of Astronomy, Indiana University, Swain West 319, 727 East Third Street, Bloomington, IN 47405–7105, USA; cijohnson@astro.indiana.edu; catyp@astro.indiana.edu , Catherine A. Pilachowski11affiliation: Department of Astronomy, Indiana University, Swain West 319, 727 East Third Street, Bloomington, IN 47405–7105, USA; cijohnson@astro.indiana.edu; catyp@astro.indiana.edu , R. Michael Rich22affiliation: Department of Physics and Astronomy, UCLA, 430 Portola Plaza, Box 951547, Los Angeles, CA 90095-1547, USA; rmr@astro.ucla.edu , and Jon P. Fulbright33affiliation: Department of Physics and Astronomy, Johns Hopkins University, Baltimore, MD 21218, USA; jfulb@skysrv.pha.jhu.edu 44affiliation: Visiting astronomer, Cerro Tololo Inter–American Observatory, National Optical Astronomy Observatory, which are operated by the Association of Universities for Research in Astronomy, under contract with the National Science Foundation.
Abstract

We present abundances of several light, , Fe–peak, and neutron–capture elements for 66 red giant branch (RGB) stars in the Galactic globular cluster Omega Centauri ( Cen). Our observations lie in the range 12.0V13.5 and focus on the intermediate and metal–rich RGBs. Abundances were determined using equivalent width measurements and spectrum synthesis analyses of moderate resolution (R18,000) spectra obtained with the Blanco 4m telescope and Hydra multifiber spectrograph. Combining these data with previous work, we find that there are at least four peaks in the metallicity distribution function at [Fe/H]=–1.75, –1.45, –1.05, and –0.75, which correspond to about 55, 30, 10, and 5 of our sample, respectively. Additionally, the most metal–rich stars are the most centrally located. Na and Al are correlated despite exhibiting star–to–star dispersions of more than a factor of 10, but the distribution of those elements appears to be metallicity dependent and are divided at [Fe/H]–1.2. About 40–50 of stars with [Fe/H]–1.2 have Na and Al abundances consistent with production solely in Type II supernovae and match observations of disk and halo stars at comparable metallicity. The remaining metal–poor stars are enhanced in Na and Al compared to their disk and halo counterparts and are mostly consistent with predicted yields from 5 M asymptotic giant branch (AGB) stars. At [Fe/H]–1.2, more than 75 of the stars are Na/Al enhanced and may have formed almost exclusively from AGB ejecta. Most of these stars are enhanced in Na by at least 0.2 dex for a given Al abundance than would be expected based on “normal” globular cluster values. All stars in our sample are –rich with [Ca/Fe]=+0.36 (=0.09) and [Ti/Fe]=+0.23 (=0.14). The Fe–peak elements give solar–scaled abundances and similarly small dispersions with [Sc/Fe]=+0.09 (=0.15) and [Ni/Fe]=–0.04 (=0.09). Europium does not vary extensively as a function of metallicity and has [Eu/Fe]=+0.19 (=0.23). However, [La/Fe] varies from about –0.4 to +2 and stars with [Fe/H]–1.5 have [La/Eu] values indicating domination by the s–process. A quarter of our sample have [La/Eu]+1 and may be the result of mass transfer in a binary system. We conclude that the metal–rich population must be at least 1–2 Gyr younger than the metal–poor stars, owing to the long timescales needed for strong s–process enrichment and the development of a large contingent of mass transfer binaries.

stars: abundances, globular clusters: general, globular clusters: individual ( Centauri, NGC 5139). stars: Population II

1 Introduction

Among all of the known Galactic globular clusters, Omega Centauri ( Cen) is unique in the extent of its chemical enrichment. The cluster exhibits huge star–to–star abundance variations that are not limited solely to the light elements, as is the case for most “normal” globular clusters. Instead, Cen stars have [X/Fe]111We make use of the standard spectroscopic notations where [A/B]log(N/N)– log(N/N) and log (A)log(N/N)+12.0 for elements A and B. dispersions of 0.5 to more than 1.0 dex for many elements and span a metallicity range from [Fe/H]–2.2 to nearly –0.5 (e.g., Norris & Da Costa 1995; Suntzeff & Kraft 1996; Smith et al. 2000; Johnson et al. 2008). Additionally, Cen’s red giant branch (RGB) and subgiant branch (SGB) show 4–5 discrete populations in concert with multiple main sequences (Lee et al. 1999; Hilker & Richtler 2000; Pancino et al. 2000; van Leeuwen et al. 2000; Ferraro et al. 2004; Rey et al. 2004; Stanford 2004, 2006; Sollima et al. 2005a; Villanova et al. 2007). These data, along with the apparent age dispersion at the main sequence turnoff, suggest Cen underwent extensive self–enrichment and star formation over 1 Gyr.

While Cen has an estimated mass of 2–710 M (Richer et al. 1991; Meylan et al. 1995; van de Ven et al. 2006), it does not appear to have a particularly deep potential well compared to other lower mass clusters (Gnedin et al. 2002). Combined with the cluster’s retrograde orbit and short disk crossing time (1–210 yrs; Dinescu et al. 1999), it seems unlikely that star formation could have occurred over several Gyrs in the cluster’s current configuration. A proposed scenario is that Cen was once the nucleus of a dwarf spheroidal galaxy that was accreted and stripped apart via gravitational interaction with the Milky Way (e.g., Bekki & Norris 2006). If this is true, then the cluster was probably much more massive in the past.

Until recently, Cen was the only known globular cluster with multiple populations, but new observations (e.g., Piotto 2008) have indicated several of the more massive clusters in the Galaxy host at least two SGBs and/or main sequences despite being monometallic. These anomalous sequences are often interpreted as having large He enhancements ranging from Y0.30–0.38, compared to the canonical He abundance of Y0.25. This assumption applies to the blue main sequence in Cen as well, which is roughly 0.3 dex more metal–rich than the red main sequence and requires Y0.38 to match the observations in this paradigm (Bedin et al. 2004; Norris 2004; Piotto et al. 2005). However, the important caveat remains that while the metallicity difference is measured, the He difference is only inferred. The source of these potential He enhancements remains unknown, but the most likely candidates include: 3–8 M asymptotic giant branch (AGB) stars, super–AGB stars (8–10 M), massive rotating stars, and population III stars (e.g., see Renzini 2008 for a review of this topic). Each of these scenarios poses a unique set of obstacles, but the basic problem is the difficulty in producing a discrete population of He–enriched stars while satisfying other chemical, age, and IMF constraints.

Globular cluster stars appear to have a more complex chemical history than their halo counterparts of similar metallicities, particularly with respect to the light elements oxygen through aluminum. In moderately metal–poor halo stars (–2.0[Fe/H]–1.0), these elements closely mimic the trends predicted for stars forming primarily out of gas polluted by core collapse supernovae (SNe; e.g., Timmes et al. 1995; Samland 1998; Nomoto et al. 2006). That is, the elements remain enhanced at [/Fe]+0.40, but Na and Al, due to their secondary (i.e., metal–dependent) production, slowly increase relative to Fe with increasing metallicity. This is contrasted with the ubiquitous trends observed in globular clusters, which have stars with similar abundance patterns (the so–called “primordial” population) and stars showing signs of varying degrees of high temperature proton–capture processing (the “intermediate” and “extreme” populations; e.g., Kraft 1994; Gratton et al. 2004; Carretta et al. 2008). These tell–tale signs of additional processing are evidenced by the pervasive O–Na and Mg–Al anticorrelations along with the Na–Al correlation observed in all well–studied clusters to date, and are the result of processing in the ON, NeNa, and MgAl cycles (e.g., Gratton et al. 2004). Since these trends are observed in main sequence and turnoff stars (Cannon et al. 1998; Gratton et al. 2001; Cohen et al. 2002; Briley et al. 2004a; 2004b; Boesgaard et al. 2005) as well as RGB stars, it seems likely that the chemical patterns were already imprinted in the gas from which the current generation of stars formed. The source of these abundance patterns is unknown, but intermediate mass AGB stars, which undergo hot bottom burning (HBB) at temperatures exceeding 80–10010 K and experience third dredgeup, are a popular choice because they do not alter [Fe/H], have low velocity ejecta, and produce large quantities of He, thus possibly alleviating some of the He enhancement issues mentioned above. While AGB stars are a qualitatively attractive solution, many problems arise in quantitative analyses and the ejecta yields are highly model dependent (e.g., Denissenkov & Herwig 2003; Fenner et al. 2004; Choi & Yi 2008; Ventura & D’Antona 2008). Other potential polluters include fast rotating massive stars (Decressin et al. 2007) and previous generations of slightly more massive RGB stars (Denissenkov & Weiss 2004); In situ deep mixing may also still play a role in highly evolved RGB stars.

In terms of chemical properties, Cen behaves similarly to Galactic globular cluster populations (aside from the large metallicity spread) in that the various light element relations and enhancements are present in nearly all subpopulations analyzed so far (e.g., Norris & Da Costa 1995; Smith et al. 2000), but the cluster hosts stars of considerable Na/Al enrichment and O depletion. Unlike the field and disk populations that exhibit lower [/Fe] ratios at [Fe/H]–1, presumably due to the contributions of Type Ia SNe, the overwhelming majority of Cen stars at the same metallicity are enhanced. This suggests that Type Ia SNe have played only a minor role in the cluster’s chemical enrichment for all but perhaps the most metal–rich stars (Pancino et al. 2002; but see also Cunha et al. 2002). If Cen is the remnant of a former dwarf spheroidal galaxy then it has evolved much differently than present day dwarf galaxies because they do not show extreme light element enhancements/depletions and often exhibit subsolar [/Fe] abundances (e.g., see review by Geisler et al. 2007). However, Cen does share the stronger s–process component seen in many dwarf spheroidal stars, except that the cluster stars more metal–rich than [Fe/H]–1.5 show s–/r–process ratios indicating complete s–process dominance whereas the dwarf galaxies experienced much weaker s–process enrichment. This is in direct contradiction to globular clusters, which are r–process rich. Lower mass AGB stars (1–4 M), which are thought to produce most of the s–process elements, have therefore had a much more significant effect on Cen’s chemical evolution than is seen in dwarf spheroidals and globular clusters.

In this paper we present spectroscopic analyses of numerous light, , Fe–peak, and neutron–capture elements for 66 stars spanning Cen’s full metallicity range, with an emphasis on the lesser studied intermediate and metal–rich populations. We combine our results with those from the literature and compare Cen to the Galactic thin and thick disk, halo, bulge, other globular clusters, and nearby dwarf spheroidals in an attempt to disentangle the evolution of these very different populations and perhaps isolate chemical signatures that are unique to each subpopulation in Cen.

2 Observations and Reductions

All observations were taken at the Cerro Tololo Inter-American Observatory on 2006 May 26 and 2006 May 27 using the Blanco 4m telescope and Hydra multifiber spectrograph. In each configuration we used the “large” 300 (2) fibers and obtained spectra with two different bench spectrograph setups. The first setup was centered near 6670 Å and spanned approximately 6530–6800 Å while the second setup was centered on 6125 Å and ranged from 6000–6250 Å. Both spectrograph setups employed the 100 slit mask along with the 400 mm Bench Schmidt Camera and 316 line mm echelle grating to achieve a resolving power of R(/)18,000 (0.35 Å FWHM).

Target stars, coordinates, photometry, and membership probabilities were taken from the proper motion study by van Leeuwen et al. (2000). The targets were chosen to be on the upper third of the giant branch and all have V14.0, but priority was given to those with larger B–V indices (i.e., more metal–rich) in the Hydra assignment code. Stars with membership probabilities 70 were excluded from the fiber assignment process. While we did not measure radial velocities for the target stars, cluster members were easily discerned from the field star population because of Cen’s comparatively large radial velocity (V232 km s; Reijns et al. 2006).

We obtained 3, 1800 second exposures for each spectrograph setup with 92 fibers placed on targets. The co–added signal–to–noise (S/N) ratios of the spectra ranged from 25–200, but we only analyzed stars for which the S/N was 50. The final sample includes 66 stars and are shown in Figure 1 along with the data from Johnson et al. (2008) and the full sample of van Leeuwen et al. (2000).

Since Cen exhibits such a large range in metallicity and the various giant branches contain stars in different ratios, selection effects may be more prominent than for typical globular clusters. In Figure 2 we show the observed completion fractions of our current data combined with Johnson et al. (2008) as a function of both V magnitude and B–V color. While there was little increase in the completion fraction for stars with 11.0V12.0, those with 12.0V13.5 increased 5-10 and similar additions are seen in B–V ranging from 1.15 to 1.55. We now have data that are at least uniformly representative across a wide range of temperatures and luminosities; however, the sample is still weighted towards observing more stars in the most metal–poor population. Since the new observations preferentially target stars with metallicities in the range –1.50[Fe/H]–0.50, the increased H opacity and line blocking with increasing metallicity causes these stars to have lower flux in the spectral regions of interest than their more metal–poor counterparts. As a result, stars observed in progressively more metal–rich branches are, on average, more evolved with our magnitude cutoff.

There is some evidence for the presence of a radial metallicity gradient in the cluster (Norris et al. 1996, 1997; Suntzeff & Kraft 1996; Hilker & Richtler 2000; Pancino et al. 2000; Rey et al. 2004; Johnson et al. 2008), and it is important to observe stars at various radii to measure the true metallicity distribution. In Figures 3 and 4 we plot the positions of our program stars and show a normalized cumulative distribution as a function of distance from the cluster center, defined by van Leeuwen et al. (2000) as 132645.9, –472837.0 (J2000). Both figures indicate our combined sample from this study and Johnson et al. (2008) mostly covers stars between 5–15 from the center, which is equivalent to roughly 3.5 to 10.5 core radii where the core radius is 1.40 (Harris 1996; rev. 2003 February). Fiber positioning limitations and increasing stellar densities near the cluster core prevent us from obtaining copious observations inside 1–2 from the center, but we have observed nearly 30–40 of all bright giants inside 10–20.

Data reductions were carried out using various tasks provided in standard IRAF222IRAF is distributed by the National Optical Astronomy Observatories, which are operated by the Association of Universities for Research in Astronomy, Inc., under cooperative agreement with the National Science Foundation. packages. We used ccdproc to trim the overscan region and apply the bias level corrections. Flat–field corrections, ThAr lamp wavelength calibrations, cosmic ray removal, subtraction of scattered light and sky spectra, and extraction of the one–dimensional spectra were performed using the dohydra package. The resultant spectra were then corrected for telluric contamination, continuum flattened, and combined.

3 Analysis

We have derived abundances for nine different elements using local thermodynamic equilibrium (LTE) equivalent width and spectrum synthesis analyses in the combined spectral regions of 6000–6250 Å and 6530–6800 Å. Spectrum synthesis was used for determining all Al abundances because of the potential for CN contamination in metal–rich and CN–strong stars. Model atmosphere parameters including effective temperatures (T), surface gravities (log g), and microturbulence (V) were estimated based on published photometry and the empirical V–T relation given in Johnson et al. (2008).

3.1 Model Stellar Atmospheres

Effective temperatures were estimated via empirical calibrations of V and 2MASS photometry based on the infrared flux method (Blackwell & Shallis 1977). To improve accuracy, we averaged the T values obtained through the color–temperature relations of Alonso et al. (1999; 2001) and Ramírez & Meléndez (2005) for V–J, V–H, and V–K color indices. The photometry was corrected for interstellar reddening and extinction using the corrections recommended by Harris (1996) of E(B–V)=0.12 and McCall (2004) for E(V–J)/E(B–V)=2.25, E(V–H)/E(B–V)=2.55, and E(V–K)/E(B–V)=2.70. Evidence for differential reddening is mainly concentrated near the core (Calamida et al. 2005; van Loon et al. 2007), but the well defined evolutionary sequences seen in Villanova et al. (2007) suggest significant differential reddening is unlikely. Therefore, we have applied a uniform reddening correction to all stars. The temperatures derived from each color index are in very good agreement with an average offset of 21 K (=6 K). Our adopted T values are probably accurate to within 50 K, and are consistent with the star–to–star scatter seen in the calibrations of both studies. Plotting Fe abundance versus excitation potential did not reveal any trends and our adopted photometric temperatures satisfied excitation equilibrium.

Surface gravity was determined by T and absolute bolometric magnitude (M) through the standard relation,

(1)

We applied the bolometric correction to M from Alonso et al. (1999) and used a distance modulus of (m–M)=13.7 (van de Ven et al. 2006). As mentioned in 1, an age spread of 1–4 Gyr is likely present in the cluster, but the difference in mass between the oldest and youngest stars is only of order 0.05 M, which is negligible for surface gravity determinations. Norris et al. (1996) argue that 20–40 of stars on the RGB may be AGB stars with M0.60 M, but this would only lead to abundance uncertainties of order 0.10 dex for species residing in the dominant ionization state (e.g., Fe II). Similar surface gravity and abundance effects may be expected for He–rich stars, which have slightly lower RGB masses due to their shorter lifetimes compared to He–normal stars (e.g., Newsham & Terndrup 2007).

Since we only had a limited number of singly ionized lines available for analysis relative to the number of neutral lines, we relied on the photometric surface gravity estimate instead of ionization equilibrium. In the top panel of Figure 5, we show the differences in derived abundance for neutral and singly ionized species of Fe, Sc, and Ti. For Fe, the average offset between log (Fe I) and log (Fe II) is –0.09 (=0.12), but this is based on a highly discrepant number of lines between the two species and thus may not accurately reflect a systematically low gravity. Sc shares a similar pattern with an average difference of –0.16 dex (=0.22), but these are based on only one line a piece for each species and reliable hyperfine structure information for these transitions is sparse. Ti I and II lines give an average difference of –0.01 dex (=0.18). Overall, the difference between abundances derived from both species is –0.07 dex (=0.17), which is comparable to measurement and model uncertainties. In the bottom panel of Figure 5, we show photometric log g values compared with estimates based on spectroscopic gravity calibrations of globular clusters provided by Kučinskas et al. (2006). The average offset between the two systems is +0.04 dex (=0.17), and is comparable to the uncertainty found by examining ionization equilibrium. This leads us to believe our surface gravity estimates are not in serious error.

We obtained a rough estimate of [Fe/H] using the [Ca/H] calibration based on V and B–V given in van Leeuwen et al. (2000; their equation 15) and assumed [Ca/Fe]+0.30. Likewise, an initial microturbulence value was calculated from the V–T relation given in Johnson et al. (2008) for luminous giants. Our determined T, log g, [Fe/H], and V values were used to generate model atmospheres (without convective overshoot) via interpolation within the ATLAS9333The model atmosphere grids can be downloaded from http://cfaku5.cfa.harvard.edu/grids.html. grid (Castelli et al. 1997). We iteratively adjusted the microturbulence of the model until Fe abundances were independent of line strength following the method described in Magain (1984). Lastly, the model’s metallicity was adjusted to match the derived Fe abundance of each star. A complete list of our adopted model atmosphere parameters along with star identifiers, published photometry, membership probabilities, and S/N estimates for each spectrum are provided in Table 1.

3.2 Derivation of Abundances

3.2.1 Equivalent Width Analysis

For all elements except Al, final abundances were determined using equivalent width analyses and the abfind driver in the LTE line analysis code MOOG (Sneden 1973). Equivalent widths were measured by interactively fitting multiple Gaussians to isolated and blended line profiles. Suitable lines were identified using the solar and Arcturus atlases444The atlases can be downloaded from the NOAO Digital Library at http://www.noao.edu/dpp/library.html.. Our adopted log gf values were determined by measuring equivalent widths in the solar atlas and then modified until all lines yielded abundances consistent with the photospheric values given in Anders & Grevesse (1989). A summary of our line list and measured equivalent widths is given in Tables 2a–2f and the final abundances are in Table 3. Note that all abundance ratios in Table 3 are relative to [Fe/H], which is the average of [Fe I/H] and [Fe II/H] or just [Fe I/H] if Fe II lines are not available. We did not determine [X/Fe] ratios by matching ionization states because many stars did not have reliable Fe II transitions, our typical measured [Fe I/H] values are based on more than 25 lines versus 1 or 2 for [Fe II/H], and [Fe I/H]–[Fe II/H] is approximately equal to the line–to–line dispersion seen in our [Fe I/H] determinations. Giving [X/Fe] ratios relative to [Fe/H] is an attempt to minimize the effects of overionization.

Many line profiles for the odd–Z Fe–peak and neutron capture elements suffer from hyperfine splitting, but the necessary atomic data for several of these transitions are not currently available in the literature. A standard equivalent width analysis will produce an overabundance if this effect is not properly taken into account. Since the error caused by hyperfine splitting increases with line strength, we do not expect our abundances, which are based on unsaturated and generally weak lines, to be strongly affected.

The elements here that may be affected by hyperfine splitting are Sc, La, and Eu. Our Sc abundances are based on the 6210.67 Å Sc I line and 6604.60 Å Sc II line. While hyperfine structure estimates have been produced for Sc II (Prochaska & McWilliam 2000), there is no available information for the Sc I line and neither of these transitions is included in Zhang et al. (2008). Therefore, we have not applied the correction to Sc II, but the offset is probably not too large given that the average equivalent width for the Sc I/II lines is 60 mÅ. Similarly, no hyperfine data exist for the 6774.27 Å La II line and therefore we accept the derived abundances at face value. Europium is slightly more complicated because, in addition to hyperfine splitting, it has two stable, naturally occurring isotopes (Eu and Eu) that are present in nearly equal proportions. For all Eu abundances, we have applied a hyperfine and isotopic line list provided by C. Sneden (2006, private communication). Lacking a priori knowledge of the r–/s–process contributions for La and Eu in Cen, we have assumed a solar mix such that the ratio for La is 25/75 (Sneden et al. 2008) and for Eu 97/3 (Sneden et al. 1996), respectively.

3.2.2 Spectrum Synthesis Analysis

While all other abundances were determined using equivalent width analyses, we derived Al abundances via spectrum synthesis because of the non–trivial contamination from CN lines seen in many of the cooler, more metal–rich stars. For consistency, spectrum synthesis was performed even in cases where CN contamination was not an issue. In Figure 6 we show two sample syntheses for a case with strong (top panel) and weak (bottom panel) CN lines in order to illustrate the line blanketing effects from molecular absorption. For stars where the CN lines were present, we found that a straight–forward equivalent width analysis led to an over estimate of log (Al) by as much as 0.1–0.2 dex compared to spectrum synthesis, but the two methods agreed to within 0.1 dex in spectra without strong CN features. The Al lines are designated by “synth” in Tables 2a–2f.

We created the molecular line list by combining the Kurucz online database555The Kurucz line lists can be accessed at: http://kurucz.harvard.edu/linelists.html. with one provided by B. Plez (private communication, 2007; see also Hill et al. 2002). Since the C, N, and C/C abundances are unknown, we fixed [C/Fe]=–0.5 and C/C=4, which are consistent with Norris & Da Costa (1995) and Smith et al. (2002) for Cen giants. Without accurate C, N, and O data, it is impossible to constrain the molecular equilibrium equations and derive true C and/or N abundances for CN. Therefore, we treated the nitrogen abundance as a free parameter and adjusted it to obtain a best fit to the CN lines. Other metal lines surrounding the Al doublet have excitation potentials 5 eV and are not important contributors in these cool stars.

3.2.3 Abundance Comparison to Other Studies

Although Cen has been the subject of many spectroscopic studies, here we restrict comparison to those measuring abundances using moderately high resolution (R15,000) spectroscopy. The only two previous works for which we have several stars in common are Norris & Da Costa (1995) and Johnson et al. (2008). In the case of Norris & Da Costa, the average difference in measured [Fe/H] for the 7 common stars is –0.02 dex (=0.05), in the sense present minus Norris & Da Costa. The results are similar for most of the other elements with average differences of order 0.10 dex (0.15), and La is the only exception with an average offset of 0.34 (=0.12). This discrepancy is likely due to the difficulty in obtaining accurate La abundances. In comparison to Johnson et al. (2008), the difference in [Fe/H] based on 21 stars in common is –0.10 dex (=0.05) and 0.00 dex (=0.22) for [Al/Fe]. These results are consistent with the small deviations in adopted atmospheric parameters found by Johnson et al. (2008; see their Figures 8–9) comparing that study to other spectroscopic surveys in the literature.

3.3 Abundance Sensitivity to Model Atmosphere Parameters

In Table 4, we show the results of our tests regarding abundance sensitivity to uncertainties in adopted model atmosphere parameters for all elements studied here. We examined how the various log (X) values changed when altering T100 K, log g0.30 cm s, [M/H]0.30 dex, and V0.30 km s. In general, we find that the neutral species tend to be more strongly affected by changes in temperature, but the singly ionized species are influenced by surface gravity changes because of their dependence on electron pressure. However, abundances taken from singly ionized transitions tend to have a larger dependence on T with increasing metallicity while the effects on neutral lines are mitigated. Similarly, only the ionized species have a significant dependence on the model atmosphere’s overall metallicity because their line–to–continuous opacity ratios are more sensitive to changes in the H abundance. For stars with [Fe/H]–1, microturbulence uncertainties have very little influence on the derived abundance because the lines are weak and lie further down the linear portion of the curve of growth, but even in the most metal–rich stars the effects are typically no larger than 0.10–0.15 dex. The lanthanum line is an exception because the more metal–rich Cen stars are very s–process rich and thus the La II lines typically have equivalent widths 75 mÅ. Although each element has a slightly different dependence on these physical parameters, the important point is that the element–to–iron ratio should be mostly invariant. Instead, only the log (X) values should be sensitive to model parameter variations.

As mentioned in 1, it has been argued that several of the intermediate and perhaps metal–rich stars in this cluster may have strong He enhancements extending as large as Y0.38. We do not expect our analysis to be severely altered (see Girardi et al. 2007) and the [X/Fe] ratios should be mostly independent of the adopted He abundance; however, [Fe/H] may be systematically higher in the He–rich stars. To test the effects of He enhancement, we created synthetic spectra using He–normal (Y=0.27) and He–rich (Y=0.35) ATLAS9 models of comparable temperature and metallicity to our stars. We found the line strength differences between the two sets to be much less than 1, with the He–rich model producing stronger lines because of decreased continuous H opacity. This result is consistent with Piotto et al. (2005) and leads us to believe our abundances are robust against possible He variations.

4 Results

4.1 Light elements: Na & Al

The odd–Z elements Na and Al are particularly important because they are among the heaviest elements thought to be produced via proton–capture nucleosynthesis in low and intermediate mass stars. This makes them useful probes for deciphering which processes, in addition to Type II SN production, may have been dominant during various epochs of star formation. Previous large sample, high resolution spectroscopic studies of Cen giants (e.g., Norris & Da Costa 1995; Smith et al. 2000; Johnson et al. 2008) have shown that Na and Al (in addition to C, N, O, and Mg) exhibit very large star–to–star [X/Fe] variations while preserving the Na–Al correlation seen in other Galactic globular clusters. The top two panels of Figure 7 illustrate this point by demonstrating the stark contrast in Na and Al line strengths for stars with similar temperatures and metallicities. Since we can compare stars of similar evolutionary state and metallicity, we can safely assume departures from LTE are not the cause of the observed abundance variations. No NLTE corrections have been applied to our Na and Al results because there are no “standard” values available in the literature and those that are available disagree in magnitude and sign. However, Na and Al abundances determined from the non–resonance, subordinate transitions used here typically have corrections of order 0.20 dex for stars with –2.5[Fe/H]–0.5 (e.g., Gratton et al. 1999; Gehren et al. 2004).

First considering our Na results, we find that there is a general increase in [Na/Fe] as a function of increasing metallicity accompanied by a decrease in the star–to–star scatter. The dominant metallicity group of stars (–1.8[Fe/H]–1.6) have [Na/Fe]=+0.03 (=0.32) with a full range of 1.29 dex while the next population of stars (–1.5[Fe/H]–1.3) have [Na/Fe]=+0.20 (=0.21) and a full range of 0.67 dex, which is smaller by about a factor of 4. However, there is a noticeable change in the distribution of [Na/Fe] for RGB stars in the higher metallicity populations. At [Fe/H]–1.2, 95 (18/19) of the stars are very Na–rich with [Na/Fe]=+0.86 (=0.12). The strong enrichment of this population is in agreement with Norris & Da Costa (1995) who find that at least 75 (6/8) of stars in that metallicity range are Na–rich and at least 50 are O–poor. A two–sample Kolmogorov–Smirnov (K–S) test (Press et al. 1992) shows that the population of stars with [Fe/H]–1.2 is drawn from a different parent population than the [Fe/H]–1.2 group at the 99 level.

By combining our current data with that of Johnson et al. (2008), we have a homogeneous set of [Al/Fe] abundances determined for more than 200 RGB stars. In Figure 8 we show the results of our combined samples for [Al/Fe] and log (Al) as a function of metallicity. Although the sample sizes between Na and Al differ by a factor of 3.5, some key differences standout in the Al data set: (1) there appear to be 2 or 3 different populations of stars, (2) the star–to–star dispersion stays mostly constant until [Fe/H]–1.2, and (3) stars with [Fe/H]–1.2 show a roughly constant log (Al)6.22 (=0.18) as a function of increasing [Fe/H]. However, there are some interesting similarities: (1) the Al data show a clear change in the abundance pattern for stars with [Fe/H]–1.2, (2) the metal–rich RGB stars are predominantly Al–rich, and (3) log (Na)log (Al). It should be noted that despite the large abundance scatter, the Na–Al correlation is present in our data.

We define the three different Al populations as those having [Al/Fe]+0.60, +0.60[Al/Fe]+1.0, and [Al/Fe]+1.0. First considering only stars with [Fe/H]–1.2, the subpopulations break down into [Al/Fe]=+0.34 (=0.14), [Al/Fe]=+0.82 (=0.10), and [Al/Fe]=+1.17 (=0.11), respectively. These represent 50 (83/166), 30 (49/166), and 20 (34/166) of the cluster stars in this metallicity regime. Extending this break down to the entire sample gives a similar distribution of 48 (96/202), 34 (69/202), and 18 (37/202), respectively. This distribution is perhaps tied to the “primordial”, “intermediate”, and “extreme” populations of the O–Na anticorrelation (Carretta et al. 2008). However, only the intermediate Al subpopulation is present at all metallicities. The very enhanced Al stars ([Al/Fe]+1) are only found at [Fe/H]–1.2, and the sequence of low–Al stars ([Al/Fe]+0.60) essentially terminates at about the same metallicity cut–off (this is particularly evident in the bottom panel of Figure 8). A two–sample K–S test confirms the same result as the Na case, which is that the [Al/Fe] distribution for stars with [Fe/H]–1.2 and [Fe/H]–1.2 are different at the 94 level.

These data suggest that Cen’s metal–rich populations may be significantly more chemically homogeneous than the metal–poor (and presumably older) populations, but the gas from which the metal–rich stars formed was enhanced in light elements at a level beyond what is thought to be possible from Type II SNe (e.g., Woosley & Weaver 1995; Chieffi & Limongi 2004). Evidently, at high metallicity it is possible to produce Na in greater quantities than Al.

4.2 elements

The elements are often used to gauge the relative contributions from Type II SNe, which are efficient producers of elements, and Type Ia SNe, which produce mostly Fe–peak elements. Predicted stellar yields from core collapse SNe weighted by a Salpeter initial mass function (IMF; x=1.35) produce [/Fe]+0.30 to +0.50 across a broad range of metallicities (e.g., Chieffi & Limongi 2004). Therefore, values of [/Fe]+0.10 or less suggest Type Ia SNe have contributed some portion of the Fe–peak elements. The most commonly measured elements are O, Mg, Si, Ca, and Ti; however, Ti is more complicated because it has multiple production sources. In globular clusters, a large portion of the stars have had their O and Mg abundances altered by proton–capture nucleosynthesis and therefore these elements cannot be treated as “pure” elements. This restricts discussions regarding enhancement to the heavier elements.

Previous studies of Cen and other globular clusters have shown nearly all stars to be enhanced at [/Fe]+0.40 with very small star–to–star scatter (e.g., see review by Gratton et al. 2004). Our results are consistent with previous work and give [Ca/Fe]=+0.36 (=0.09). Although Ti may straddle being classified as an or Fe–peak element, the stars in our sample are mostly Ti–enhanced with [Ti/Fe]=+0.23 (=0.14). We do not find any stars to be –poor and our lowest derived value is [Ca/Fe]=+0.17, but a handful of –poor stars have been found in this cluster (e.g., Norris & Da Costa 1995; Smith et al. 1995, 2000; Pancino et al. 2002). We do not find any trend in [Ca/Fe] with [Fe/H], which is in contrast to the results of Pancino et al. (2002) who find the most metal–rich stars to be –poor. However, there is real scatter in [/Fe] in this cluster as is evident in the Si and Ca line strength variations seen in the top panel of Figure 7. In any case, larger sample sizes are required to settle this issue, but it seems that the majority of Cen stars are –rich and thus Type Ia SNe have not contributed much to the [X/Fe] ratios. This is a significant problem from a chemical evolution standpoint because either the ejecta from Type Ia SNe were preferentially lost or their presence was suppressed despite a several Gyr timespan in star formation.

4.3 Fe & Fe–peak elements

As mentioned above, Fe–peak elements are produced in both Type II and Type Ia SNe in copious amounts and are the most commonly used tracers of metallicity in stars. These elements are produced in similar conditions during the late stages of stellar evolution and as a result often track together as a function of overall metallicity. Aside from Fe, the other Fe–peak elements analyzed here reproduce this trend with [Sc/Fe]=+0.09 (=0.15) and [Ni/Fe]=–0.04 (=0.09). In both cases, there is no trend in [X/Fe] with [Fe/H]. However, since Cen hosts multiple stellar populations, the behavior of [Fe/H] is not as simple as most monometallic globular clusters.

Large sample spectroscopic and photometric observations of Cen (e.g., Norris & Da Costa 1995; Suntzeff & Kraft 1996; van Leeuwen et al. 2000; Rey et al. 2004; Sollima et al. 2005a; Villanova et al. 2007; Johnson et al. 2008) have shown that the cluster hosts multiple populations of stars with almost no stars being more metal–poor than [Fe/H]=–2, more than half having [Fe/H]–1.7, and the rest forming a high metallicity tail extending to [Fe/H]–0.5. Again combining our new results with those from Johnson et al. (2008), we have a homogeneous set of spectroscopically determined [Fe/H] abundances for 228 RGB stars. In Figure 9 we show a histogram of our combined sample and our results are consistent with the cluster having multiple peaks in the metallicity distribution function at [Fe/H]=–1.75, –1.45, –1.05, and –0.75. These peaks constitute roughly 55, 30, 10, and 5 of our observations, respectively. The percentage of stars contained in each population is nearly identical between our entire sample and a subset having the highest completion fraction (V12.5). This leads us to believe our full sample is representative of the entire cluster population.

In addition to Cen being chemically diverse, there is some evidence for a cluster metallicity gradient such that the inner regions contain most of the metal–rich stars (e.g., Suntzeff & Kraft 1996; Norris et al. 1996 Hilker & Richtler 2000; Pancino et al. 2003; Johnson et al. 2008). These results are confirmed in photometric studies (e.g., Rey et al. 2004), which show that the anomalous, metal–rich RGB (RGB–a) is found only near the core of the cluster. In Figure 10, we plot log (Fe) versus distance from the cluster center. We find that the most metal–rich stars are mostly located inside 10, but the metal–poor stars are located uniformly throughout the cluster. The inset plot in Figure 10 shows that the median metallicity stays constant at about log (Fe)=6.0 ([Fe/H]–1.5) at all radii, but the metallicity interquartile and full ranges decrease at large radii. There has been speculation that, in addition to these spatial anomalies, the various cluster populations may exhibit unique kinematic signatures as the result of the cluster formation process (e.g., Norris et al. 1997; Sollima et al. 2005b). However, recent larger sample studies seem to indicate none of Cen’s subpopulations exhibit rotational, proper motion, or radial velocity anomalies (Reijns et al. 2006; Pancino et al. 2007; Johnson et al. 2008; Bellini et al. 2009). These new results seem to negate the merger and background cluster superposition scenarios.

4.4 Neutron–capture elements

Most elements heavier than the Fe–peak are produced via successive neutron captures on seed nuclei, but a limited number of these elements have optical atomic transitions. In the metallicity regime considered here, Ba and La are often the primary tracers of the main s–process component and Eu the primary tracer of the r–process. As previously mentioned, nearly all globular clusters are r–process rich with [Eu/Ba,La]+0.25 (e.g., Gratton et al. 2004), but previous studies have shown that Cen has very strong s–process enhancement, especially at [Fe/H]–1.5 (e.g., Francois et al. 1988; Norris & Da Costa 1995; Smith et al. 1995, 2000). We find in agreement with previous studies that most Cen stars more metal–rich than [Fe/H]–1.7 are strongly s–process enriched based on [La/Eu] ratios approaching and exceeding +0.8. While the most metal–poor stars have [La/Eu]=–0.02, this value rises to [La/Eu]=+0.49 by [Fe/H]–1.4 meaning that [La/Eu] increases by more than a factor of 3 during a span in which [Fe/H] increases by a factor of 2. However, we find that [La/Fe] does not increase appreciably at metallicities exceeding [Fe/H]–1.2 (excluding possible Ba–stars), but all stars in our sample with [Fe/H]–1.2 have [La/Fe]+0.5 and abundance patterns dominated by the s–process. This trend is not shared by Eu, which remains mostly constant at [Eu/Fe]=+0.19 (=0.23) at all metallicities. We have also found that 25 (15/60) of our sample may qualify as Ba–stars with [La/Fe]+1.0. The most extreme case is star LEID 45358, which has [La/Fe]=+2.03 and [La/Eu]=+1.81. For stars in common between the two samples, van Loon et al. (2007) found these to have large Ba4554 indices indicating they are Ba–rich as well.

5 Discussion

5.1 Chemical Enrichment in Cen

Spectroscopic and photometric analyses of Cen stars have revealed a system hosting a complex past. The cluster metallicity apparently increased from [Fe/H]–2.2 to [Fe/H]–0.5 over roughly 2–4 Gyrs (e.g., Stanford et al. 2006) and Cen has experienced a handful of discrete star formation events. The metallicity distribution peaks in our data agree with those found in the literature and correspond to the “MP” ([Fe/H]–1.7), “MINT2” ([Fe/H]–1.4), “MINT3” ([Fe/H]–1.0), and “SGB–a” ([Fe/H]–0.6) populations found on the SGB by Sollima et al. (2005b). However, these classifications are not as well–defined on the main sequence and show considerable complexity (e.g., Bedin et al. 2004; Piotto et al. 2005; Villanova et al. 2007). The apparent kinematic homogeneity of the various stellar populations (e.g. Pancino et al. 2007; Bellini et al. 2009) suggests most, if not all, of the cluster’s chemical enrichment is the result of internal processes rather than a product of multiple merger events. However, the paucity of stars more metal–poor than [Fe/H]–2 means the nascent gas from which the primary population formed was already considerably polluted by massive star ejecta. One of the most striking results discovered so far is that the second most metal–poor stellar population ([Fe/H]–1.4; and perhaps the subsequent more metal–rich stars) may have experienced both a huge increase in He content (dY/dZ70; Piotto et al. 2005) and an equally impressive increase in s–process element abundances compared to the primary population ([Fe/H]–1.7), which contains more than half of all Cen stars. Somehow these events took place while preserving the various light element correlations observed in other globular clusters that do not (in general) have large He and metallicity variations and lack strong s–process signatures. Since the combined Johnson et al. (2008) and current data sets allow us to probe various production sources, we turn now to what our current data add to Cen’s puzzling past.

5.1.1 Supernova Pollution

The majority of elements heavier than Li are produced during various quiescent and explosive nucleosynthetic events in 11 M stars (Woosley & Weaver 1995). These processes, which occur within 2010 years after the onset of star formation, are known to produce an overabundance of elements by about a factor of two relative to the solar /Fe ratio. In addition, massive stars also produce varying amounts of the odd–Z light elements (e.g., C through Al) with metallicity dependent yields of –0.5[X/Fe]+0.3 in the metallicity regime covered by Cen stars (e.g., Woosley & Weaver 1995; Chieffi & Limongi 2004; Nomoto et al. 2006). Although the final abundances of Fe–peak elements are dependent on the explosion energy and mass–cut, they generally track closely to Fe. These stars inevitably produce some neutron capture elements as well, but only the lower mass SNe (8–11 M) are believed to be significant r–process contributors (e.g., Mathews & Cowan 1990; Cowan et al. 1991; Wheeler et al. 1998), while low mass AGB stars (1–4 M) seem the best candidates for s–process production (e.g., Busso et al. 1999).

In contrast, mass transfer Type Ia SNe may take anywhere from 50010 to more than 310 years to evolve (e.g., Yoshii et al. 1996) and could have difficulty forming in low metallicity ([Fe/H]–1) environments (Kobayashi et al. 1998). Nucleosynthesis calculations have shown that these SNe predominantly produce Fe–peak elements and only trace amounts of and light elements (Nomoto et al. 1997). It is estimated that Type Ia SNe have produced at least 50 of the total Fe in the Galaxy and their onset is believed to be the primary cause for the decline in [/Fe] at [Fe/H]–1 in the disk and halo populations. It is for this reason that the [/Fe] ratio is often used as a diagnostic to test the presence of Type Ia SNe in a stellar system.

While there have been some –poor stars found in Cen’s most metal–rich population (Pancino et al. 2002), the general trend of enhancement in the elements suggests a majority of the cluster’s chemical evolution occurred before Type Ia SNe had time to evolve. Determining whether or not Type Ia SNe can form in metal–poor environments could help place additional constraints on Cen’s evolutionary timescale. If the lower limit of [Fe/H]–1 estimated by Kobayashi et al. (1998) is correct and only the most metal–rich population in the cluster is affected by Type Ia SNe ejecta, then this would imply an age difference between the [Fe/H]–1 and [Fe/H]–0.7 groups of 1 Gyr. However, if this limit is at a much lower metallicity, then the cluster would have had to evolve on a much shorter time scale.

In Figures 1113 we show the evolution of all elements measured in this study as a function of [Fe/H] along with those available in the literature for Cen, the Galactic disk, bulge, halo, globular clusters, and nearby dwarf spheroidal galaxies (see Table 5 for data references). From these data we have confirmed: (1) a more than 1.0 dex spread exists for [Na/Fe] and [Al/Fe] and the two elements are correlated, (2) the elements are enhanced by about a factor of two at all metallicities with small star–to–star scatter, (3) there are at least four different metallicity peaks (see also Figure 9) at [Fe/H]=–1.75, –1.45, –1.05, –0.75 with internal dispersions of 0.10 dex in each subpopulation, and (4) there is a large s–process component that manifests itself in the intermediate and metal–rich populations of the cluster. As is the case for other globular clusters, the larger star–to–star variations seen in the light and neutron–capture elements versus the and Fe–peak elements suggest additional production (or destruction) sources other than core collapse SNe. We know that, at least for the light elements, the observed inhomogeneity is not due to incomplete mixing of SN ejecta because the Na/Al enhanced stars are also O–poor (e.g., Norris & Da Costa 1995; Smith et al. 2000).

If massive stars cannot account for all of the observed abundance anomalies in Cen, then how much can they account for? At least in stars with [Fe/H]–1, Type II SNe are responsible for producing nearly all of the and Fe–peak elements. However, IMF weighted theoretical yields of SNe with initial metallicities in the range –2[Fe/H]–0.5 (e.g., Woosley & Weaver 1995; Chieffi & Limongi 2004; Nomoto et al. 2006) produce values roughly consistent with those observed in the disk, halo, and bulge (i.e., [Na/Fe]0; [Al/Fe]+0.3), but Cen (and other globular cluster) stars can reach [Na/Fe]+1.0 and [Al/Fe]+1.4. Using the Al data in Figure 11 to trace the percentage of stars with light element abundance patterns matching those observed in the other Galactic populations at comparable metallicity ([Al/Fe]+0.5), we find 42 (84/202) of our sample fall into this category. It is more difficult to quantify this with the Na data because the sample size is more than a factor of three smaller, but it appears that at least a significant fraction of the stars in Figure 11 with [ Fe/H]–1.2 show [Na/Fe] ratios consistent with the disk, halo, and bulge, but nearly all of the more metal–rich stars are enhanced in Na. This further solidifies the claim that although Type II SNe have had a significant impact on all Cen stars, they are not the only significant nucleosynthesis site. Assuming our data are representative, roughly half of all Cen stars appear to have formed in an environment that was polluted with the ejecta from sources other than Type II SNe.

Further inspection of Figure 11 reveals an interesting trend in Na and Al as a function of [Fe/H]. As noted in 4.2, there is a clear lack of stars showing Na and Al abundances consistent with being polluted solely by Type II SNe at [Fe/H]–1.2. Only 6 (1/17) of Cen giants are “Na–normal” ([Na/Fe]0), and this trend is present in both the Norris & Da Costa (1995) and Smith et al. (2000) data as well. A similar result is observed in the larger sample of Al data in that only 22 (8/36) are “Al–normal” ([Al/Fe]+0.3). While there appears to be a down turn in the maximum [Al/Fe] attained at [Fe/H]–1.2, the rise in [Na/Fe] and [La/Fe] coupled with the stability of [/Fe] and [Eu/Fe] in the same metallicity range indicates this artifact is not the result of Type Ia SNe adding Fe but instead a decrease in the [Al/Fe] ratio being added to the cluster’s ISM by the production source. What is perhaps most intriguing is that despite a (possible) huge increase in He between the [Fe/H]=–1.7 and –1.4 groups, the light element trends are very similar. It would seem that whichever stars are the source of the high He abundances do not produce abnormally large [Na/Fe] and [Al/Fe] ratios because similar enhancements in Na and Al are found in globular clusters that do not show signs of such extreme He variations.

If Eu production can be attributed mostly to 8–10 M stars, then we know from those data alone that chemical enrichment had to have occurred in Cen over more than 20010 years because there are at least four subpopulations with different [Fe/H] and [Eu/Fe] is roughly constant (with some scatter). However, [Eu/Fe] is, on average, consistently at least 0.1–0.2 dex underabundant relative to the other populations shown in Figure 13. The reason for this is not clear, but it could be that the ratio of 8–10 M versus higher mass stars was anomalously low in Cen relative to other systems.

5.1.2 Intermediate Mass AGB Stars

The discovery of significant star–to–star scatter in light elements coupled with the O–Na anticorrelation in stars on the main sequence and subgiant branches of globular clusters seems to indicate that the various relations among the elements O through Al were already imprinted on the gas from which the current generations of stars formed. As discussed in 1, HBB occurring in intermediate mass (5–8 M) AGB stars is currently favored as a likely location for producing the light element trends. These stars have the advantages of preserving their initial [Fe/H] envelope abundances, ejecting enriched material at low velocities, experiencing few third dredge–up episodes (negligible s–process production), and reaching envelope temperatures 7010 K that activate the NeNa and MgAl proton–capture cycles. However, current AGB stellar models are highly sensitive to the adopted treatment of convection and mass loss and it has been pointed out that these scenarios do not explain the role of super–AGB stars (those that ignite core carbon but not neon burning) nor 1–4 M AGB stars (Prantzos & Charbonnel 2006). Models using standard mixing length theory (e.g., Fenner et al. 2004; Karakas & Lattanzio 2007) are unable to reproduce the large O depletions ([O/Fe]–0.6) found in some globular cluster stars (including Cen) and show large enhancements in [C+N+O/Fe], which conflict with observations that the CNO sum is constant (Pilachowski 1988; Dickens et al. 1991; Norris & Da Costa 1995; Smith et al. 1996; Ivans et al. 1999). On the other hand, models adopting the full spectrum of turbulence treatment of convection (e.g., Ventura & D’Antona 2008) show fewer third dredge–up episodes and thus keep the CNO sum roughly constant while explaining some of the C through Al abundance trends seen in globular clusters. Neither case is able to fully explain all light element anomalies, in particular the super O–poor stars and Mg isotopic ratios, which may require a hybrid scenario that includes in situ deep mixing and HBB in 5 M AGB stars (e.g., D’Antona & Ventura 2007) in addition to improvements in key nuclear reaction rates.

Can these stars reproduce what we observe in Cen? Our data have shown that only about half of the stars in our sample are consistent with being formed from gas predominantly polluted by Type II SNe (i.e., the stars are not particularly enhanced in Na and Al compared to disk and halo stars of comparable metallicity). Since 5 M AGB stars likely do not alter the abundances of any elements heavier than Al, we will restrict the discussion to those elements. First turning to the populations with [Fe/H]–1.2, the stars with [Na/Fe]0 and [Al/Fe]+0.5 have envelope material that was likely exposed to high temperature proton–capture processing in an external environment. Although the light element yields are sensitive to both model parameters and nuclear reaction rates, the Ventura & D’Antona (2008) results indicate intermediate metallicity 5–6.5 M AGB stars can produce +0.30[Na/Fe]+0.60 and [Al/Fe]+1.0, while the Fenner et al. (2004) data predict somewhat higher Na and lower Al abundances. These values are consistent with the “intermediate” Al population that has [Al/Fe]=+0.82 (=0.10) and suggest intermediate mass AGB stars could be responsible for the enhancements seen in these stars. However, about 20 of the stars with [Fe/H]–1.2 have [Al/Fe]+1 and [Na/Fe]+0.5. These stars are not accounted for by current AGB models and may have undergone additional in situ deep mixing or require pollution from another unknown source.

As can be seen in the top panel of Figure 14, halo, disk, and bulge stars exhibit a roughly constant [Na/Al]–0.2 from [Fe/H]=–2 to –0.6, while Cen, dwarf spheroidal, and globular cluster stars display a wide range from [Na/Al]=–1 to +0.4 and show a general increase in [Na/Al] with increasing metallicity. Since the final abundances of Na and Al in SN ejecta scale similarly with neutron excess and metallicity (Arnett 1971), the Na/Al ratio is mostly insensitive to metallicity changes and is consistently near [Na/Al]–0.2 (e.g., Woosley & Weaver 1995). The overproduction of Al at low metallicities and underproduction at higher metallicities is consistent with the observed trends in AGB models (e.g., Ventura & D’Antona 2008) due to lower temperatures at the bottom of the convective envelope and shallower mixing in more metal–rich stars. This trend is nearly identically reproduced in globular cluster stars of varying metallicity (bottom panel of Figure 14) and likely indicates the same stars that are responsible for the globular cluster light element anomalies are also prevalent in Cen. Since the same trend is also observed in dwarf spheroidal stars, which are not believed to be strongly enriched in Type II SN ejecta, this may strengthen the case for HBB in intermediate mass AGB stars (or some equivalent H–burning environment) to be the source of light element abundance trends different than those seen in the disk and halo. It is interesting that these two systems share the rise in [Na/Al] versus [Fe/H] with globular clusters because, as their low [/Fe] ratios indicate, star formation has proceeded much differently in dwarf spheroidals despite having comparable main sequence turnoff age ranges with Cen.

The paucity of stars with [Al/Fe]+1 at [Fe/H]–1.2 is also consistent with the predictions of in situ deep mixing at higher metallicities where the increased –gradient is expected to inhibit dredge–up of ON, NeNa, and MgAl cycled material into the stellar envelope via meridional circulation (e.g., Sweigart & Mengel 1979). While the range in Na and Al data track closely to that of other globular clusters at low and intermediate metallicity, the more metal–rich Cen stars show surprising Na enhancements and decreased star–to–star scatter that are not seen in globular clusters of comparable metallicity. This is true even for M4 ([Fe/H]–1.1), which is suspected of having a second, more enriched population without a large spread in Fe (Marino et al. 2008). Although the range of M4’s Al abundances are consistent with the values we find here, the average [Na/Fe] ratio in Cen giants of comparable metallicity is about 0.3 dex larger than the highest [Na/Fe] abundance found by either Ivans et al. (1999) or Marino et al. (2008) in M4. It would seem that there was an additional source of Na in the more metal–rich Cen populations or that hardly any unenriched gas remained to dilute the AGB ejecta. Figure 15 illustrates this point in that the stars with [Fe/H]–1.2 and [Al/Fe]+0.5 have [Na/Fe] ratios that lie above the range expected for a given Al abundance based on typical globular cluster values. The identity of the Na source is only speculative, but if the progenitor AGB population that polluted the gas from which the [Fe/H]–1.2 stars formed was He–rich, the higher temperatures and possible deeper mixing in regions where the NeNa cycle was operating may have contributed to the increased Na abundances. It may also be possible that lower mass, He–rich AGB stars, which evolve more quickly than He–normal stars (and produce more Na and less Al), could have a larger impact than in normal globular clusters. However, 4 M AGB stars are not believed to strongly deplete O and would have to already be O–poor to reproduce the sub–solar [O/Fe] ratios found in many Cen giants with [Fe/H]–1.5.

5.1.3 Low Mass AGB Stars

Lower mass, thermally pulsing AGB stars (1–4 M), which evolve over 15010 to 2.510years (Schaller et al. 1992), are thought to be the primary producers of the main s–process component in the Galaxy at metallicities found in Cen (e.g., Busso et al. 2001). Smith et al. (2000) showed that the [Rb/Zr] ratio in Cen was consistent with the s–process being produced in 1.5–3.0 M AGB stars, implying a monotonic, total evolutionary timescale of 2–3 Gyrs. This is consistent with most other estimates (e.g., Stanford et al. 2006; but see also Villanova et al. 2007). Since these stars have the longest formation timescale, the presence of their chemical signatures sets a lower limit on relative age estimates.

The halo and disk populations are known to exhibit a steady rise in the contribution of s–process elements at [Fe/H]–2.5 (e.g., Simmerer et al. 2004), but globular cluster heavy element abundances are dominated by the r–process (e.g., Gratton et al. 2004) and are indicative of the rather rapid chemical evolution timescales of normal globular clusters compared to the disk and halo. Interestingly, dwarf spheroidal stars tend to have a stronger s–process component than any of the Galactic populations (e.g., Geisler et al. 2007), but one that is much smaller than that seen in Cen. This, along with the evidence for Type Ia SN pollution, implies dwarf spheroidal galaxies evolve much differently than most other Galactic stellar systems and do so with a rather subdued star formation rate (e.g., Mateo 2008).

However, the Galactic bulge is believed to have formed rapidly, as constrained by turnoff photometry (Ortolani et al. 1995; Kuijken & Rich 2002; Zoccali et al. 2003; Clarkson et al. 2008) as well as by measured high [/Fe] (e.g., McWilliam & Rich 1994; Fulbright et al. 2006; Lecureur et al. 2007). Theoretical studies argue for timescales significantly less than 10 yrs (e.g., Elmegreen 1999; 2008; Ballero et al. 2007). Yet despite a metallicity that is high compared to the halo and Cen, (e.g., Fulbright et al. 2006, Zoccali et al. 2008) the s–process elements are seen to exhibit Solar [X/Fe] ratios (McWilliam & Rich 1994) that would appear to require low and intermediate mass stars to have provided significant input to the bulge’s chemical evolution.

In Figure 13 we show the evolution of [La/Fe], [Eu/Fe], and [La/Eu] as a function of [Fe/H] for Cen and other stellar populations. Since all but the most metal–poor group of Cen stars show significant enhancement in the s–process element La (and the [La/Eu] ratio), we find in agreement with previous studies that at least 10 years had to have passed between the formation of the primary population at [Fe/H]=–1.7 and the final population at [Fe/H]=–0.7 to allow the low mass progenitor populations enough time to evolve. A significant percentage (25) of stars in our sample have [La/Fe]+1.0 and may be the result of binary mass transfer from a 4 M AGB companion. However, none of these stars are present in the dominant, most metal–poor population but are found at [Fe/H]–1.5 with most being present at [Fe/H]–1. It is unknown whether the prevalence of such stars at higher metallicities is a result of the longer formation timescales needed for one of the companions to evolve, an anomalous increase in the binary fraction at higher metallicity, or a sample selection effect. If the result is not a selection effect, then this may be a clear indication that the more metal–rich stars are at least 1–210 years older than the metal–poor population.

Figures 15 and 16 show [Na/Fe] and [La/Fe] versus [Al/Fe], which could be a useful indicator regarding the relative importance of low versus intermediate mass AGB stars. In most globular clusters there is little evidence of light elements showing any correlation with heavy neutron–capture elements on top of the correlations seen among the various light elements (e.g., Smith 2008), which implies the elements lighter than Al are produced in a different astrophysical site over different timescales than those produced via the s–process and r–process. This may mean that the current generation of globular cluster stars have abundance signatures strongly weighted towards pollution from more massive AGB stars compared to those 4 M. On the contrary, Cen exhibits a mild correlation between La and Al (as well as Na) and as stated above shows [La/Fe] ratios well in excess of the roughly [La/Fe]+0.5 maximum found in globular cluster stars, especially at [Fe/H]–1.5. Current AGB nucleosynthesis models (e.g., Ventura & D’Antona 2008) suggest that this correlation is unlikely to be the result of 1–4 M stars dominating the chemical enrichment of Cen because AGB stars in that mass range are shown to produce Na without significantly depleting O, which contradicts the O–Na anticorrelation observed in the cluster giants and prevalence of O–poor stars at higher metallicities (e.g., Norris & Da Costa 1995). For the other populations shown in Figure 16, only the bulge data show any hint of a Na–Al correlation, but that is not believed to be the result of the same mechanism at work in globular clusters (Lecureur et al. 2007). However, the current lack of heavy element data in the bulge makes it difficult to draw conclusions regarding the impact of low mass AGB stars in that environment. Since none of the other populations show any correlation between La and Al (or Na), it appears that Cen is (as always) a special case where both low and intermediate mass AGB stars have had significant influence on the cluster’s chemical evolution.

In this paper and previous studies, it has been shown that Cen is an extremely complex object with an intriguing formation history. Nearly all aspects of its past remain a mystery and although it has been shown that the cluster experienced multiple star formation episodes (and probably significant mass loss), there is evidence both for and against simple monotonic chemical enrichment (i.e., metal–poor stars are older than more metal–rich stars). It appears that Cen shares many chemical characteristics with a variety of systems that formed under widely different conditions and the cluster exhibits signs of both rapid and extended star formation. One of the interesting issues raised by our data is the significance of the apparent transition in light element abundance trends at [Fe/H]–1.2. It seems as if the stars with [Fe/H]–1.2 were made almost entirely out of AGB ejecta, but the populations with [Fe/H]–1.2 contain groups of stars that likely formed both with and without the presence of AGB pollution in nearly equal proportions. The lack of –poor stars in all but perhaps the most metal–rich population poses a serious problem and Cen’s enrichment history challenges the paradigm of chemical evolution that for timescales 1 Gyr, Type Ia SNe contribute Fe–peak and –poor material that drive down the [/Fe] ratio to near solar composition. It may be that the cluster lost too much mass before the onset of Type Ia SNe or the ejecta were located too far outside the core to be retained. This may be corroborated by evidence that there is no radial preference in the location of X–ray binaries in Cen due to a lack of mass segregation (e.g., Gendre et al. 2003). While the observation of large numbers of RGB, SGB, and main sequence stars are needed to understand the full picture of Cen’s evolution, the large fluctuations in light element abundances such as Na and Mg, which are often used as metallicity tracers, make low resolution or integrated light studies difficult to decipher. However, future large sample, high resolution studies spanning both the giant branches and main sequences should help further isolate the chemical signatures of each subpopulation and allow more quantitative analyses.

6 Summary

We have determined abundances of several light, , Fe–peak, and neutron–capture elements for 66 RGB stars in the globular cluster Cen using moderate resolution (R18,000) spectra. Two different Hydra spectrograph setups were employed spanning 6000–6250 Å and 6530–6800 Å, yielding co–added S/N ratios of about 50–200. The observations covered the full cluster metallicity regime with an emphasis on the intermediate and metal–rich populations. The elemental abundances were determined using either equivalent width analyses or spectrum synthesis, with the addition of hyperfine structure data when available.

The light elements Na and Al show large abundance inhomogeneities that span more than a factor of 10 and the elements are correlated. The Al data set was supplemented with that from Johnson et al. (2008) and yielded [Fe/H] and [Al/Fe] abundances for more than 200 RGB stars. From these data we find evidence for the existence of possibly three different populations of stars with distinct [Al/Fe] patterns. The three sequences segment into those with [Al/Fe]=+0.34 (=0.14), [Al/Fe]=+0.82 (=0.10), and [Al/Fe]=+1.17 (=0.11) and represent 48, 34, and 18 of our sample, respectively. These may be inherently tied to the “primordial,” “intermediate,” and “extreme” populations found in normal globular clusters that exhibit varying degrees of O depletion and Na enhancement. However, there appears to be a break in the distribution of both Na and Al at [Fe/H]–1.2. Stars with [Fe/H]–1.2 have abundances in the range –0.1[Al/Fe]+1.4 and –0.5[Na/Fe]+0.6 with at least half of the stars exhibiting light element abundances consistent with the disk and halo populations, but more than 75 of stars with [Fe/H]–1.2 are enhanced in Na and Al with values exceeding those found in the disk, halo, and even some globular clusters. None of the stars with [Al/Fe]+1.0 are found at [Fe/H]–1.2. A two–sided K–S test reveals the Na and Al abundances on either side of the [Fe/H]=–1.2 cutoff to have a 90 probability of being drawn from different parent populations.

All of our program stars are enhanced in elements with [Ca/Fe]=+0.36 (=0.09) and [Ti/Fe]=+0.23 (=0.14), despite showing a range of more than a factor of 30 in [Fe/H]. The Fe–peak elements share the same small range in star–to–star scatter but give roughly solar–scaled values of [Sc/Fe]=+0.09 (=0.15) and [Ni/Fe]=–0.04 (=0.09). Our results are in agreement with previous studies as we find multiple peaks in the metallicity distribution function at [Fe/H]=–1.75, –1.45, –1.05, and –0.75 and few stars with [Fe/H]–1.8. These populations represent about 55, 30, 10, and 5 of our sample, respectively. Additionally, we find evidence supporting the idea that the most metal–rich stars are more centrally concentrated, and there appears to be a decrease in the star–to–star metallicity dispersion as a function of increasing distance from the cluster core.

The neutron–capture elements La and Eu yield abundances indicative of strong s–process enrichment in all but the most metal–poor stars. We find that nearly all Cen stars with [Fe/H]–1.5 have [La/Eu]+0.5, which contradicts the generally r–process dominated nature of normal globular cluster stars that have [La/Eu]–0.25. Despite the sharp rise in [La/Fe], the Eu abundance remains fairly constant across all metallicities with [Eu/Fe]=+0.19 (=0.23). However, 25 of our sample contains stars with [La/Fe]+1.0 that are possibly the result of mass transfer in a binary system. These stars are also known to have large Ba4554 indices and are predominantly found at [Fe/H]–1.3.

Comparing these results with the abundance trends observed in the Galactic halo, disk, bulge, globular clusters, and nearby dwarf spheroidal galaxies indicates the current generation of Cen stars share many chemical characteristics found in each of those populations but contain key differences. The elevated [/Fe] and solar–scaled Fe–peak abundances suggest that Type II SNe have dominated the production of metals in the cluster with almost no contribution from Type Ia SNe. However, we find that at least 40–50 of stars in our sample have [Na/Fe] and [Al/Fe] ratios that exceed the yields expected from moderately metal–poor SNe. Previous studies have shown that the Na and Al enhanced stars are also O–poor, which implies that these stars were polluted by material that has been exposed to high temperature proton–capture burning. This is corroborated by examining the behavior of [Na/Al] as a function of metallicity. Type II SNe are expected to produce a nearly metallicity independent yield of [Na/Al]–0.2 over –2[Fe/H]–0.5, which matches observations of disk and halo stars, but Cen, normal globular cluster, and dwarf spheroidal stars span a range of –1[Na/Al]+0.4. Therefore, our data strongly support the idea of an additional source of light elements in these environments.

HBB occurring in intermediate mass AGB stars is a favored location for producing Na and Al while destroying O. Current AGB nucleosynthesis models predict our observed trends, that more Al is produced at low metallicity and more Na produced at high metallicity, and may explain stars with +0.5[Al/Fe]+1.0. However, they may not be adequate to reproduce the 20 of metal–poor stars with [Al/Fe]+1, which may require some other source (e.g., in situ mixing or massive rotating stars). What is perhaps most intriguing is that we find evidence for two different subpopulations separated as being either more metal–poor or metal–rich than [Fe/H]–1.2. Most of the stars with [Fe/H]–1.2 appear to have formed almost entirely out of AGB ejecta and have [Na/Fe] and [Al/Fe] abundances well above those found in the disk and halo at similar metallicity, while those at [Fe/H]–1.2 show more of a continuum between strong SN pollution and AGB pollution. Since we did not choose targets based on known chemical properties (e.g., CN strength), it seems that the prevalence of Na and Al enhanced stars at higher metallicity is likely not a selection effect. Interestingly, although all Cen giants exhibit the same Na–Al correlation found in other globular clusters, the Cen stars with [Fe/H]–1.2 have more Na for a given Al abundance by 0.2 dex compared to what is expected based on the trend seen in normal globular clusters. There is also a mild correlation between La and both Na and Al, but it is unclear how La relates to these elements. The decreasing maximum value of [Al/Fe] at [Fe/H]–1.2 is not shared by Na and La and suggests a decrease in the [Al/Fe] abundance being added to the cluster’s ISM rather than an increase in Fe due to Type Ia SNe.

The sharp increase in the abundance of [La/Fe] and [La/Eu] with increasing metallicity coupled with the relatively long lifetimes of stars thought to produce most of the s–process elements is consistent with the generally adopted chemical evolution timescale of 2–4 Gyr. However, other stellar systems that evolved over 1 Gyr exhibit the characteristic downturn in [/Fe], but this trend is mostly absent in Cen stars. Even though it is highly probable that Cen did not evolve as a closed box, the apparent preferential retention of Type II versus Type Ia SN ejecta or even the suppression of Type Ia SNe at [Fe/H]–1 at timescales exceeding 1–2 Gyrs remains an important problem.

This publication makes use of data products from the Two Micron All Sky Survey, which is a joint project of the University of Massachusetts and the Infrared Processing and Analysis Center/California Institute of Technology, funded by the National Aeronautics and Space Administration and the National Science Foundation. This research has made use of NASA’s Astrophysics Data System Bibliographic Services. RMR acknowledges support from grant AST–0709479 from the National Science Foundation. Support of the College of Arts and Sciences and the Daniel Kirkwood fund at Indiana University Bloomington for CIJ is gratefully acknowledged. We would like to thank the referee for his/her thoughtful comments that led to improvement of the manuscript. Facilities: CTIO

References

  • Alonso et al. (1999) Alonso, A., Arribas, S., & Martínez-Roger, C. 1999, A&AS, 140, 261
  • Alonso et al. (2001) Alonso, A., Arribas, S., & Martínez-Roger, C. 2001, A&A, 376, 1039
  • Anders & Grevesse (1989) Anders, E., & Grevesse, N. 1989, Geochim. Cosmochim. Acta, 53, 197
  • Arnett (1971) Arnett, W. D. 1971, ApJ, 166, 153
  • Ballero et al. (2007) Ballero, S. K., Matteucci, F., Origlia, L., & Rich, R. M. 2007, A&A, 467, 123
  • Bedin et al. (2004) Bedin, L. R., Piotto, G., Anderson, J., Cassisi, S., King, I. R., Momany, Y., & Carraro, G. 2004, ApJ, 605, L125
  • Bekki & Norris (2006) Bekki, K., & Norris, J. E. 2006, ApJ, 637, L109
  • Bellini et al. (2009) Bellini, A., et al. 2009, A&A, 493, 959
  • Bensby et al. (2003) Bensby, T., Feltzing, S., & Lundström, I. 2003, A&A, 410, 527
  • Bensby et al. (2005) Bensby, T., Feltzing, S., Lundström, I., & Ilyin, I. 2005, A&A, 433, 185
  • Blackwell & Shallis (1977) Blackwell, D. E., & Shallis, M. J. 1977, MNRAS, 180, 177
  • Boesgaard et al. (2005) Boesgaard, A. M., King, J. R., Cody, A. M., Stephens, A., & Deliyannis, C. P. 2005, ApJ, 629, 832
  • Briley et al. (2004) Briley, M. M., Cohen, J. G., & Stetson, P. B. 2004a, AJ, 127, 1579
  • Briley et al. (2004) Briley, M. M., Harbeck, D., Smith, G. H., & Grebel, E. K. 2004b, AJ, 127, 1588
  • Busso et al. (1999) Busso, M., Gallino, R., & Wasserburg, G. J. 1999, ARA&A, 37, 239
  • Busso et al. (2001) Busso, M., Gallino, R., Lambert, D. L., Travaglio, C., & Smith, V. V. 2001, ApJ, 557, 802
  • Calamida et al. (2005) Calamida, A., et al. 2005, ApJ, 634, L69
  • Cannon et al. (1998) Cannon, R. D., Croke, B. F. W., Bell, R. A., Hesser, J. E., & Stathakis, R. A. 1998, MNRAS, 298, 601
  • Carretta et al. (2004) Carretta, E., Gratton, R. G., Bragaglia, A., Bonifacio, P., & Pasquini, L. 2004, A&A, 416, 925
  • Carretta et al. (2008) Carretta, E., Bragaglia, A., Gratton, R. G., & Lucatello, S. 2008, arXiv:0811.3591
  • Castelli et al. (1997) Castelli, F., Gratton, R. G., & Kurucz, R. L. 1997, A&A, 318, 841
  • Cavallo et al. (2004) Cavallo, R. M., Suntzeff, N. B., & Pilachowski, C. A. 2004, AJ, 127, 3411
  • Chieffi & Limongi (2004) Chieffi, A., & Limongi, M. 2004, ApJ, 608, 405
  • Choi & Yi (2008) Choi, E., & Yi, S. K. 2008, MNRAS, 386, 1332
  • Clarkson et al. (2008) Clarkson, W., et al. 2008, ApJ, 684, 1110
  • Cohen et al. (2002) Cohen, J. G., Briley, M. M., & Stetson, P. B. 2002, AJ, 123, 2525
  • Cowan et al. (1991) Cowan, J. J., Thielemann, F.-K., & Truran, J. W. 1991, Phys. Rep., 208, 267
  • Cunha et al. (2002) Cunha, K., Smith, V. V., Suntzeff, N. B., Norris, J. E., Da Costa, G. S., & Plez, B. 2002, AJ, 124, 379
  • D’Antona & Ventura (2007) D’Antona, F., & Ventura, P. 2007, MNRAS, 379, 1431
  • Decressin et al. (2007) Decressin, T., Meynet, G., Charbonnel, C., Prantzos, N., & Ekström, S. 2007, A&A, 464, 1029
  • Denissenkov & Herwig (2003) Denissenkov, P. A., & Herwig, F. 2003, ApJ, 590, L99
  • Denissenkov & Weiss (2004) Denissenkov, P. A., & Weiss, A. 2004, ApJ, 603, 119
  • Dickens et al. (1991) Dickens, R. J., Croke, B. F. W., Cannon, R. D., & Bell, R. A. 1991, Nature, 351, 212
  • Dinescu et al. (1999) Dinescu, D. I., Girard, T. M., & van Altena, W. F. 1999, AJ, 117, 1792
  • Elmegreen (1999) Elmegreen, B. G. 1999, ApJ, 517, 103
  • Elmegreen et al. (2008) Elmegreen, B. G., Bournaud, F., & Elmegreen, D. M. 2008, ApJ, 688, 67
  • Fenner et al. (2004) Fenner, Y., Campbell, S., Karakas, A. I., Lattanzio, J. C., & Gibson, B. K. 2004, MNRAS, 353, 789
  • Ferraro et al. (2004) Ferraro, F. R., Sollima, A., Pancino, E., Bellazzini, M., Straniero, O., Origlia, L., & Cool, A. M. 2004, ApJ, 603, L81
  • Francois et al. (1988) Francois, P., Spite, M., & Spite, F. 1988, A&A, 191, 267
  • Fulbright (2000) Fulbright, J. P. 2000, AJ, 120, 1841
  • Fulbright et al. (2006) Fulbright, J. P., McWilliam, A., & Rich, R. M. 2006, ApJ, 636, 821
  • Fulbright et al. (2007) Fulbright, J. P., McWilliam, A., & Rich, R. M. 2007, ApJ, 661, 1152
  • Gehren et al. (2004) Gehren, T., Liang, Y. C., Shi, J. R., Zhang, H. W., & Zhao, G. 2004, A&A, 413, 1045
  • Geisler et al. (2007) Geisler, D., Wallerstein, G., Smith, V. V., & Casetti-Dinescu, D. I. 2007, PASP, 119, 939
  • Gendre et al. (2003) Gendre, B., Barret, D., & Webb, N. A. 2003, A&A, 400, 521
  • Girardi et al. (2007) Girardi, L., Castelli, F., Bertelli, G., & Nasi, E. 2007, A&A, 468, 657
  • Gnedin et al. (2002) Gnedin, O. Y., Zhao, H., Pringle, J. E., Fall, S. M., Livio, M., & Meylan, G. 2002, ApJ, 568, L23
  • Gratton et al. (1999) Gratton, R. G., Carretta, E., Eriksson, K., & Gustafsson, B. 1999, A&A, 350, 955
  • Gratton et al. (2001) Gratton, R. G., et al. 2001, A&A, 369, 87
  • Gratton et al. (2004) Gratton, R., Sneden, C., & Carretta, E. 2004, ARA&A, 42, 385
  • Harris (1996) Harris, W. E. 1996, AJ, 112, 1487
  • Hilker & Richtler (2000) Hilker, M., & Richtler, T. 2000, A&A, 362, 895
  • Hill et al. (2002) Hill, V., et al. 2002, A&A, 387, 560
  • Ivans et al. (1999) Ivans, I. I., Sneden, C., Kraft, R. P., Suntzeff, N. B., Smith, V. V., Langer, G. E., & Fulbright, J. P. 1999, AJ, 118, 1273
  • Ivans et al. (2001) Ivans, I. I., Kraft, R. P., Sneden, C., Smith, G. H., Rich, R. M., & Shetrone, M. 2001, AJ, 122, 1438
  • James et al. (2004) James, G., François, P., Bonifacio, P., Carretta, E., Gratton, R. G., & Spite, F. 2004, A&A, 427, 825
  • Johnson et al. (2005) Johnson, C. I., Kraft, R. P., Pilachowski, C. A., Sneden, C., Ivans, I. I., & Benman, G. 2005, PASP, 117, 1308
  • Johnson et al. (2008) Johnson, C. I., Pilachowski, C. A., Simmerer, J., & Schwenk, D. 2008, ApJ, 681, 1505
  • Karakas & Lattanzio (2007) Karakas, A., & Lattanzio, J. C. 2007, Publications of the Astronomical Society of Australia, 24, 103
  • Kobayashi et al. (1998) Kobayashi, C., Tsujimoto, T., Nomoto, K., Hachisu, I., & Kato, M. 1998, ApJ, 503, L155
  • Kraft (1994) Kraft, R. P. 1994, PASP, 106, 553
  • Kuijken & Rich (2002) Kuijken, K., & Rich, R. M. 2002, AJ, 124, 2054
  • Kučinskas et al. (2006) Kučinskas, A., Hauschildt, P. H., Brott, I., Vansevičius, V., Lindegren, L., Tanabé, T., & Allard, F. 2006, A&A, 452, 1021
  • Lecureur et al. (2007) Lecureur, A., Hill, V., Zoccali, M., Barbuy, B., Gómez, A., Minniti, D., Ortolani, S., & Renzini, A. 2007, A&A, 465, 799
  • Lee et al. (2005) Lee, J.-W., Carney, B. W., & Habgood, M. J. 2005, AJ, 129, 251
  • Magain (1984) Magain, P. 1984, A&A, 134, 189
  • Marino et al. (2008) Marino, A. F., Villanova, S., Piotto, G., Milone, A. P., Momany, Y., Bedin, L. R., & Medling, A. M. 2008, A&A, 490, 625
  • Mateo (2008) Mateo, M. 2008, The Messenger, 134, 3
  • Mathews & Cowan (1990) Mathews, G. J., & Cowan, J. J. 1990, Nature, 345, 491
  • McCall (2004) McCall, M. L. 2004, AJ, 128, 2144
  • McWilliam (1997) McWilliam, A. 1997, ARA&A, 35, 503
  • McWilliam & Rich (1994) McWilliam, A., & Rich, R. M. 1994, ApJS, 91, 749
  • Meylan et al. (1995) Meylan, G., Mayor, M., Duquennoy, A., & Dubath, P. 1995, A&A, 303, 761
  • Newsham & Terndrup (2007) Newsham, G., & Terndrup, D. M. 2007, ApJ, 664, 332
  • Nomoto et al. (2006) Nomoto, K., Tominaga, N., Umeda, H., Kobayashi, C., & Maeda, K. 2006, Nuclear Physics A, 777, 424
  • Norris & Da Costa (1995) Norris, J. E., & Da Costa, G. S. 1995, ApJ, 447, 680
  • Norris et al. (1996) Norris, J. E., Freeman, K. C., & Mighell, K. J. 1996, ApJ, 462, 241
  • Norris et al. (1997) Norris, J. E., Freeman, K. C., Mayor, M., & Seitzer, P. 1997, ApJ, 487, L187
  • Ortolani et al. (1995) Ortolani, S., Renzini, A., Gilmozzi, R., Marconi, G., Barbuy, B., Bica, E., & Rich, R. M. 1995, Nature, 377, 701
  • Pancino et al. (2000) Pancino, E., Ferraro, F. R., Bellazzini, M., Piotto, G., & Zoccali, M. 2000, ApJ, 534, L83
  • Pancino et al. (2002) Pancino, E., Pasquini, L., Hill, V., Ferraro, F. R., & Bellazzini, M. 2002, ApJ, 568, L101
  • Pancino et al. (2003) Pancino, E., Seleznev, A., Ferraro, F. R., Bellazzini, M., & Piotto, G. 2003, MNRAS, 345, 683
  • Pancino et al. (2007) Pancino, E., Galfo, A., Ferraro, F. R., & Bellazzini, M. 2007, ApJ, 661, L155
  • Pilachowski (1988) Pilachowski, C. A. 1988, ApJ, 326, L57
  • Piotto et al. (2005) Piotto, G., et al. 2005, ApJ, 621, 777
  • Piotto (2008) Piotto, G. 2008, Memorie della Societa Astronomica Italiana, 79, 334
  • Prantzos & Charbonnel (2006) Prantzos, N., & Charbonnel, C. 2006, A&A, 458, 135
  • Press et al. (1992) Press, W. H., Teukolsky, S. A., Vetterling, W. T., & Flannery, B. R. 1992, Numerical Recipes in FORTRAN: The Art of Scientific Computing (2nd ed.; Cambridge: Cambridge Univ. Press)
  • Prochaska & McWilliam (2000) Prochaska, J. X., & McWilliam, A. 2000, ApJ, 537, L57
  • Ramírez & Cohen (2002) Ramírez, S. V., & Cohen, J. G. 2002, AJ, 123, 3277
  • Ramírez & Meléndez (2005) Ramírez, I., & Meléndez, J. 2005, ApJ, 626, 465
  • Reddy et al. (2003) Reddy, B. E., Tomkin, J., Lambert, D. L., & Allende Prieto, C. 2003, MNRAS, 340, 304
  • Reddy et al. (2006) Reddy, B. E., Lambert, D. L., & Allende Prieto, C. 2006, MNRAS, 367, 1329
  • Reijns et al. (2006) Reijns, R. A., Seitzer, P., Arnold, R., Freeman, K. C., Ingerson, T., van den Bosch, R. C. E., van de Ven, G., & de Zeeuw, P. T. 2006, A&A, 445, 503
  • Renzini (2008) Renzini, A. 2008, MNRAS, 391, 354
  • Rey et al. (2004) Rey, S.-C., Lee, Y.-W., Ree, C. H., Joo, J.-M., Sohn, Y.-J., & Walker, A. R. 2004, AJ, 127, 958
  • Richer et al. (1991) Richer, H. B., Fahlman, G. G., Buonanno, R., Fusi Pecci, F., Searle, L., & Thompson, I. B. 1991, ApJ, 381, 147
  • Samland (1998) Samland, M. 1998, ApJ, 496, 155
  • Sbordone et al. (2007) Sbordone, L., Bonifacio, P., Buonanno, R., Marconi, G., Monaco, L., & Zaggia, S. 2007, A&A, 465, 815
  • Schaller et al. (1992) Schaller, G., Schaerer, D., Meynet, G., & Maeder, A. 1992, A&AS, 96, 269
  • Shetrone & Keane (2000) Shetrone, M. D., & Keane, M. J. 2000, AJ, 119, 840
  • Shetrone et al. (2001) Shetrone, M. D., Côté, P., & Sargent, W. L. W. 2001, ApJ, 548, 592
  • Shetrone et al. (2003) Shetrone, M., Venn, K. A., Tolstoy, E., Primas, F., Hill, V., & Kaufer, A. 2003, AJ, 125, 684
  • Simmerer et al. (2004) Simmerer, J., Sneden, C., Cowan, J. J., Collier, J., Woolf, V. M., & Lawler, J. E. 2004, ApJ, 617, 1091
  • Smith et al. (1995) Smith, V. V., Cunha, K., & Lambert, D. L. 1995, AJ, 110, 2827
  • Smith et al. (1996) Smith, G. H., Shetrone, M. D., Bell, R. A., Churchill, C. W., & Briley, M. M. 1996, AJ, 112, 1511
  • Smith et al. (2000) Smith, V. V., Suntzeff, N. B., Cunha, K., Gallino, R., Busso, M., Lambert, D. L., & Straniero, O. 2000, AJ, 119, 1239
  • Smith (2008) Smith, G. H. 2008, PASP, 120, 952
  • Sneden (1973) Sneden, C. 1973, ApJ, 184, 839
  • Sneden et al. (1996) Sneden, C., McWilliam, A., Preston, G. W., Cowan, J. J., Burris, D. L., & Armosky, B. J. 1996, ApJ, 467, 819
  • Sneden et al. (1997) Sneden, C., Kraft, R. P., Shetrone, M. D., Smith, G. H., Langer, G. E., & Prosser, C. F. 1997, AJ, 114, 1964
  • Sneden et al. (2004) Sneden, C., Kraft, R. P., Guhathakurta, P., Peterson, R. C., & Fulbright, J. P. 2004, AJ, 127, 2162
  • Sneden et al. (2008) Sneden, C., Cowan, J. J., & Gallino, R. 2008, ARA&A, 46, 241
  • Sollima et al. (2005) Sollima, A., Ferraro, F. R., Pancino, E., & Bellazzini, M. 2005a, MNRAS, 357, 265
  • Sollima et al. (2005) Sollima, A., Pancino, E., Ferraro, F. R., Bellazzini, M., Straniero, O., & Pasquini, L. 2005b, ApJ, 634, 332
  • Stanford et al. (2004) Stanford, L. M., Da Costa, G. S., Norris, J. E., & Cannon, R. D. 2004, Memorie della Societa Astronomica Italiana, 75, 290
  • Stanford et al. (2006) Stanford, L. M., Da Costa, G. S., Norris, J. E., & Cannon, R. D. 2006, ApJ, 647, 1075
  • Suntzeff & Kraft (1996) Suntzeff, N. B., & Kraft, R. P. 1996, AJ, 111, 1913
  • Sweigart & Mengel (1979) Sweigart, A. V., & Mengel, J. G. 1979, ApJ, 229, 624
  • Timmes et al. (1995) Timmes, F. X., Woosley, S. E., & Weaver, T. A. 1995, ApJS, 98, 617
  • van de Ven et al. (2006) van de Ven, G., van den Bosch, R. C. E., Verolme, E. K., & de Zeeuw, P. T. 2006, A&A, 445, 513
  • van Leeuwen et al. (2000) van Leeuwen, F., Le Poole, R. S., Reijns, R. A., Freeman, K. C., & de Zeeuw, P. T. 2000, A&A, 360, 472
  • van Loon et al. (2007) van Loon, J. T., van Leeuwen, F., Smalley, B., Smith, A. W., Lyons, N. A., McDonald, I., & Boyer, M. L. 2007, MNRAS, 382, 1353
  • Ventura & D’Antona (2008) Ventura, P., & D’Antona, F. 2008, MNRAS, 385, 2034
  • Villanova et al. (2007) Villanova, S., et al. 2007, ApJ, 663, 296
  • Wheeler et al. (1998) Wheeler, J. C., Cowan, J. J., & Hillebrandt, W. 1998, ApJ, 493, L101
  • Woolley (1966) Woolley, R. R. 1966, Royal Observatory Annals, 2, 1
  • Woosley & Weaver (1995) Woosley, S. E., & Weaver, T. A. 1995, ApJS, 101, 181
  • Yoshii et al. (1996) Yoshii, Y., Tsujimoto, T., & Nomoto, K. 1996, ApJ, 462, 266
  • Zhang et al. (2008) Zhang, H. W., Gehren, T., & Zhao, G. 2008, A&A, 481, 489
  • Zoccali et al. (2003) Zoccali, M., et al. 2003, A&A, 399, 931
  • Zoccali et al. (2008) Zoccali, M., Hill, V., Lecureur, A., Barbuy, B., Renzini, A., Minniti, D., Gómez, A., & Ortolani, S. 2008, A&A, 486, 177
Figure 1: Color–magnitude diagram of Cen’s RGB. The filled red circles represent the stars observed for this study and the filled blue squares show the stars observed for Johnson et al. (2008). There are 22 stars which overlap with Johnson et al. and those are also indicated by filled red circles. The complete sample, including stars not observed here, are from van Leeuwen et al. (2000).
Figure 2: Histogram showing the observed completion fraction of this study combined with the data of Johnson et al. (2008). The top panel shows the completion fraction as a function of V magnitude and the bottom panel shows the completion fraction as a function of B–V color.
Figure 3: Program stars are shown in terms of position in the field. The symbols are the same as in Figure 1. The cross indicates the field center at 201.691, –47.4769 (J2000) (132645.9, –472837.0) and the ellipses indicate 1, 5, and 10 times the core radius of 1.40.
Figure 4: Normalized cumulative distribution for our combined sample as a function of distance from the cluster center. This plot shows the fraction of our total sample observed inside a given radius. The cluster center reference is the same as in Figure 3.
Figure 5: The top panel shows a plot of the difference in abundance as derived from both neutral and singly ionized species as a function of [Fe/H]. The filled circles represent Fe, the filled boxes are Sc, and the filled triangles are Ti. The bottom panel shows the log g values adopted from photometry versus the calibrated T–log g relation from Kučinskas et al. (2006). In both panels the straight line indicates perfect agreement.
Figure 6: Sample spectrum syntheses are shown for two stars of varying T, [Fe/H], and CN strength, but similar [Al/Fe] ratios. The relative intensity scales are the same in both figures. The solid line shows the best fit to the observed spectrum, the dotted lines illustrate deviations 0.30 dex, and the dashed line indicates how the spectrum would appear if Al were absent.
Figure 7: Sample spectra are shown in three different wavelength regions to highlight the line strength differences seen in Na, Al, and La. Each panel contains stars of roughly the same T and [Fe/H]. Note the differences seen in both the Si and Ca features compared to Fe in the top panel.
Figure 8: The top panel shows [Al/Fe] plotted as a function of [Fe/H] and the bottom panel shows log (Al) plotted as a function of log (Fe). The symbols are the same as those in Figure 1.
Figure 9: Histogram of derived [Fe/H] values for the combined sample of this study and Johnson et al. (2008) with bin sizes of 0.10 dex. The dashed line histogram shows the results from Johnson et al. (2008).
Figure 10: Fe is plotted as a function of distance from the cluster center. The points show the data from both this study and Johnson et al. (2008). We have averaged the Fe abundances for stars observed in both studies. The inset plot shows the mean and quartile distributions in 5 bins. The vertical lines represent the full data range (except outliers) and open circles indicate mild outliers between 1.5 and 3.0 times the interquartile range.
Figure 11: Plots of [Na/Fe], [Al/Fe], and [Ca/Fe] versus [Fe/H] are shown with data from this study and the literature. The filled circles are values from the combined sample of this study and Johnson et al. (2008), the open circles are from Norris & Da Costa (1995), the filled squares are from Smith et al. (2000), the open squares are from Francois et al. (1988), and the stars are from Smith et al. (1995). Literature values are provided for the thin/thick disk (open black boxes), halo (open blue boxes), bulge (open green boxes), dwarf spheroidals (filled cyan boxes), globular clusters with 1 bars (filled magenta boxes), and the Sagittarius dwarf spheroidal (open cyan triangles). References are given in Table 5.
Figure 12: Plots of [Sc/Fe], [Ti/Fe], and [Ni/Fe] versus [Fe/H] are shown with data from this study and the literature. The symbols and [X/Fe] scales are the same as in Figure 11.
Figure 13: Plots of [La/Fe], [Eu/Fe], and [La/Eu] versus [Fe/H] are shown with data from this study and the literature. The symbols are the same as in Figure 11. The dashed lines indicating pure s–process and r–process abundance ratios are taken from McWilliam (1997).
Figure 14: [Na/Al] ratios as a function of metallicity are shown for a variety of populations. The symbols in the top panel are the same as those in Figure 11 and the blue points in the bottom panel represent individual globular cluster stars.
Figure 15: The top panel shows [Na/Fe] versus [Al/Fe] and compares Cen data to results from individual globular cluster stars. The bottom panel shows the same set of stars but plots [La/Fe] versus [Al/Fe]. The symbols are the same as those in Figure 14.
Figure 16: The top panel shows [Na/Fe] versus [Al/Fe] with data from this study and the literature. The bottom panel shows [La,Ba/Fe] versus [Al/Fe] where the Ba data are used as a tracer for the s–process in the disk and halo while La is used in all other cases. The symbols are the same as those in Figure 11.
Star Alt. V BV J H K M Mem. T log g [Fe/H] V S/N S/N
LEIDaaIdentifier from van Leeuwen et al. (2000). ROAbbIdentifier from Woolley (1966). Prob.ccMembership probability from van Leeuwen et al. (2000). (K) (cm s) Avg. (km s) 6125 Å 6670 Å
9 370 12.529 1.250 10.382 9.755 9.627 1.543 99 4505 1.20 1.41 2.05 100 100
6017 240 12.233 1.420 9.717 8.982 8.808 1.839 98 4145 0.85 1.22 2.00 125 100
12013 394 12.579 1.319 10.242 9.560 9.402 1.493 98 4305 1.10 1.31 1.85 100 125
15023 234 12.182 1.166 9.964 9.352 9.231 1.890 100 4455 1.05 1.80 2.05 125 150
16015 213 12.127 1.122 9.979 9.373 9.210 1.945 100 4510 1.05 1.80 1.60 125 125
17015 325 12.430 1.156 10.235 9.610 9.497 1.642 100 4470 1.15 1.82 1.65 150 100
19062 464 12.803 1.144 10.601 10.001 9.872 1.269 98 4470 1.30 1.75 1.50 100 100
22037 307 12.339 1.186 10.178 9.559 9.402 1.733 100 4485 1.10 1.76 1.90 150 150
23061 296 12.337 1.188 10.158 9.472 9.390 1.735 100 4460 1.10 1.66 1.65 150 125
24027 5969 13.013 1.099 10.952 10.344 10.226 1.059 100 4600 1.45 1.57 1.80 125 75
24056 364 12.474 1.145 10.363 9.708 9.584 1.598 100 4520 1.20 1.80 1.65 150 125
25026 569 12.875 1.067 10.807 10.238 10.137 1.197 100 4630 1.40 1.83 1.65 75 125
26022 5788 13.095 1.071 11.071 10.486 10.380 0.977 100 4660 1.50 1.75 1.60 75 100
26086 295 12.787 1.313 10.330 9.650 9.464 1.285 98 4205 1.10 1.13 1.90 100 75
28092 5896 12.521 1.207 10.301 9.649 9.522 1.551 100 4425 1.15 1.61 1.70 50 50
30013 540 12.895 1.249 10.737 10.116 9.955 1.177 100 4485 1.35 1.33 1.80 100 75
32169 5510 13.331 1.173 10.975 10.289 10.128 0.741 100 4285 1.40 0.91 1.70 100 100
32171 251 12.189 1.383 9.897 9.176 9.051 1.883 100 4325 0.95 1.48 1.70 100 100
33099 175 12.100 1.483 9.615 8.848 8.691 1.972 100 4155 0.80 1.18 2.05 125 125
34029 243 12.107 1.452 9.635 8.878 8.719 1.965 99 4170 0.80 1.26 2.00 200 150
34225 557 13.017 1.229 10.608 9.932 9.820 1.055 100 4270 1.25 1.20 2.15 75
35172 237 12.414 1.399 10.043 9.310 9.127 1.658 100 4245 1.00 1.22 1.85 100 75
35201 263 12.530 1.360 10.268 9.530 9.389 1.542 100 4335 1.10 1.20 1.90 100 100
36282 290 12.351 1.155 10.179 9.575 9.449 1.721 100 4500 1.15 1.82 1.85 100 100
38198 12.474 1.374 10.122 9.425 9.242 1.598 100 4275 1.05 1.53 1.95 150 150
38232 12.236 1.449 9.724 9.021 8.834 1.836 100 4160 0.85 1.45 2.00 175 175
39026 287 12.333 1.373 9.943 9.208 9.059 1.739 100 4240 0.95 1.49 1.95 175 175
39048 451 12.887 1.420 10.106 9.335 9.114 1.185 99 3965 0.95 0.98 2.00 100 125
39129 12.843 1.361 10.639 9.982 9.833 1.229 100 4430 1.30 1.32 1.85 75
39392 4579 13.413 1.194 11.108 10.440 10.265 0.659 100 4330 1.45 0.91 1.90 100 100
41033 463 12.900 1.258 10.257 9.503 9.330 1.172 99 4060 1.05 1.09 2.10 100 125
42508 600 13.041 1.137 10.783 10.144 9.990 1.031 100 4390 1.35 1.58 1.90 100 100
43061 357 12.602 1.431 9.744 8.973 8.780 1.470 100 3930 0.80 0.93 2.15 125 150
43389 12.856 1.301 10.387 9.686 9.500 1.216 100 4190 1.10 1.40 1.65 100 100
45358 13.067 1.449 10.552 9.798 9.637 1.005 99 4145 1.15 1.03 1.95 125 150
45485 3804 13.391 1.146 11.056 10.351 10.253 0.681 99 4315 1.45 1.06 1.75 100 100
46121 12.891 1.400 10.101 9.400 9.175 1.181 100 3985 0.95 0.96 1.85 125 125
47215 13.491 1.300 10.894 10.108 9.994 0.581 100 4095 1.30 0.77 1.85 100 125
48116 12.847 1.474 10.129 9.342 9.109 1.225 100 3990 0.95 0.86 1.75 150
48323 500 13.081 1.461 10.286 9.458 9.273 0.991 100 3945 1.00 0.75 1.80 100
49013 312 12.325 1.299 10.046 9.399 9.231 1.747 99 4365 1.05 1.61 2.00 150 150
49037 509 12.864 0.994 10.839 10.313 10.175 1.208 100 4690 1.45 1.77 1.50 100 100
51021 171 11.984 1.470 9.391 8.633 8.424 2.088 100 4075 0.70 1.41 1.90 150 150
51080 236 12.317 1.428 9.809 9.058 8.900 1.755 100 4150 0.85 1.48 2.00 125 150
51132 421 12.874 1.365 10.267 9.503 9.375 1.198 100 4090 1.05 1.11 2.10 125 100
54022 2594 13.360 1.412 10.825 10.069 9.910 0.712 100 4135 1.25 0.66 1.90 100 100
54105 386 12.771 1.293 10.194 9.473 9.277 1.301 100 4105 1.00 1.18 2.05 100 100
55101 480 12.923 1.252 10.505 9.832 9.650 1.149 100 4240 1.20 1.00 1.90 100 100
55111 182 11.969 1.476 9.502 8.804 8.591 2.103 100 4185 0.75 1.44 2.00 150 200
55142 367 12.442 1.445 10.008 9.257 9.086 1.630 100 4195 0.95 1.12 2.10 100 125
56040 204 12.379 1.162 10.204 9.606 9.423 1.693 100 4475 1.15 1.76 2.05 150 175
57010 207 12.154 1.412 9.744 9.001 8.864 1.918 99 4225 0.85 1.62 2.00 175 175
59036 289 12.396 1.182 10.209 9.554 9.433 1.676 100 4455 1.10 1.75 1.90 175 150
59090 271 12.316 1.255 10.081 9.451 9.267 1.756 100 4405 1.05 1.64 1.85 100 100
59094 164 12.174 1.157 10.091 9.527 9.386 1.898 97 4600 1.10 1.85 1.50 75
60066 2118 13.086 1.253 11.067 10.470 10.330 0.986 100 4640 1.50 1.29 2.00 50 50
63021 1878 13.169 1.181 11.031 10.389 10.293 0.903 100 4515 1.45 1.42 1.60 75 75
65046 601 12.904 1.098 10.873 10.312 10.193 1.168 100 4665 1.45 1.69 1.45 100
66047 472 12.704 1.351 10.452 9.779 9.628 1.368 100 4380 1.20 1.26 1.75 75 75
66054 232 12.111 1.290 9.908 9.226 9.123 1.961 100 4430 1.00 1.71 2.00 75
69027 1471 13.365 1.399 10.944 10.229 10.050 0.707 100 4225 1.35 0.89 2.00 75 75
76038 316 12.535 1.411 10.009 9.238 9.049 1.537 98 4120 0.95 1.29 2.05 125 100
77025 194 12.197 1.339 9.861 9.171 9.008 1.875 99 4300 0.95 1.68 2.05 100 100
80029 218 12.273 1.293 9.949 9.248 9.111 1.799 97 4310 1.00 1.66 1.85 100 125
81018 217 12.282 1.174 10.015 9.377 9.244 1.790 100 4395 1.05 1.87 1.85 50 50
85027 264 12.370 1.260 10.035 9.358 9.217 1.702 99 4315 1.00 1.62 1.95 75 75
Table 1: Photometry, Membership, and Model Atmosphere Parameters
Element E.P. log gf 9 6017 12013 15023 16015 17015 19062 22037 23061 24027 24056
(Å) (eV)
6003.01 Fe I 3.88 1.07 40 46
6007.32 Ni I 1.68 3.35 70 21 22
6007.97 Fe I 4.65 0.70 53 19
6008.57 Fe I 3.88 0.94 99
6015.25 Fe I 2.22 4.66 17 12
6024.07 Fe I 4.55 0.06 84 107 89 59 52 59 61 58 60 49
6027.06 Fe I 4.07 1.14 47 68 63 25 15 24 35 22 28 57 20
6035.35 Fe I 4.29 2.56
6056.01 Fe I 4.73 0.44 52 51 43 23 22 18 36 16 26 30
6064.63 Ti I 1.05 1.92 22 60 59 12 8 21 7 11 8 8
6065.49 Fe I 2.61 1.51 152 189 169 132 114 127 116 119 131 117 111
6078.50 Fe I 4.79 0.34 41 52 59 21 22 20 29 22 28 33 22
6079.02 Fe I 4.65 1.06 30 7 9 9 13
6082.72 Fe I 2.22 3.61 52 87 72 33 23 30 21 26 33 22
6084.11 Fe II 3.20 3.87 23 18 9 13 9 14 9
6086.29 Ni I 4.26 0.54 19 31 25 8 7 23
6089.57 Fe I 5.02 0.92 13 9 8 8
6091.18 Ti I 2.27 0.42 52 44 19
6094.38 Fe I 4.65 1.66 7 8
6096.67 Fe I 3.98 1.88 15 38 28 13 7 15 7 11
6102.18 Fe I 4.83 0.37 36 72 63 24 19 22 28 26 25 28 24
6102.73 Ca I 1.88 0.81 158 173 166 113 106 116 117 116 116 120 109
6108.13 Ni I 1.68 2.51 95 135 112 65 51 67 68 59 75 71 60
6111.08 Ni I 4.09 0.89 13 25 8 13 17
6120.25 Fe I 0.91 5.95 12 47 38 4 11 11 10
6121.01 Ti I 1.88 1.32 18 13
6122.23 Ca I 1.89 0.26 181 231 204 146 129 152 143 138 144 151 141
6126.22 Ti I 1.07 1.42 48 105 85 19 19 22 26 25 24 27 18
6128.98 Ni I 1.68 3.36 42 79 64 22 21 15 27 21 31 16
6146.24 Ti I 1.87 1.51 12
6149.25 Fe II 3.89 2.78 28 34 28 26 19 14 24 22
6151.62 Fe I 2.18 3.33 67 95 86 43 40 44 44 41 48 52 45
6154.23 Na I 2.10 1.57 50 34 20
6157.73 Fe I 4.07 1.22 51 74 27 22 23 25 26 24 36 25
6160.75 Na I 2.10 1.27 19 76 62 3 3 12 8 7 26 12
6161.30 Ca I 2.52 1.28 58 108 99 25 25 29 38 31 37 60 32
6162.18 Ca I 1.90 0.07 205 248 226 167 141 155 167 163 163 170 154
6165.36 Fe I 4.14 1.51 27 35 39 11 15 23 18 18 17 13
6166.44 Ca I 2.52 1.11 53 96 100 39 42 43 49 36 44 53 40
6169.04 Ca I 2.52 0.69 84 123 118 50 59 57 67 54 60 67 62
6169.56 Ca I 2.52 0.42 103 163 137 73 72 66 83 76 77 104 77
6173.34 Fe I 2.22 2.89 103 129 120 71 57 76 71 73 75 82 70
6175.37 Ni I 4.09 0.55 32 36 35 13 10 24 11 11 10 16
6176.82 Ni I 4.09 0.42 34 45 55 26 15 21 26 16 17 18
6177.25 Ni I 1.83 3.53 20 47 35 10 11 11 14 14
6180.21 Fe I 2.73 2.66 56 93 94 36 38 31 40 32 48 43 36
6186.72 Ni I 4.11 0.96 10 16 8 8 9
6188.00 Fe I 3.94 1.69 23 54 43 11 19 18 9 17 20 12
6200.32 Fe I 2.61 2.41 100 119 118 67 60 59 74 62 71 73 64
6210.67 Sc I 0.00 1.53 11 31 10 7
6219.29 Fe I 2.20 2.42 130 152 107 84 99 106 104 105 96 91
6226.74 Fe I 3.88 2.19 21 8
6229.23 Fe I 2.84 3.00 44 62 56 21 18 18 21 12 28 17
6232.65 Fe I 3.65 1.23 80 122 85 48 46 41 59 46 49 59 48
6546.24 Fe I 2.76 1.65 131 131 107 91 87 94 98 106 103 98
6551.68 Fe I 0.99 5.77 17 15
6554.23 Ti I 1.44 1.16 21 76 71
6556.07 Ti I 1.46 1.10 58 56 12 10
6559.57 Ti II 2.05 2.30 38 39 39 51 36 35 42 38 32 41
6574.25 Fe I 0.99 5.02 68 96 88 41 21 33 43 40 47 35 30
6586.31 Ni I 1.95 2.81 57 77 76 24 34 25 28 34 34
6592.92 Fe I 2.73 1.52 155 106 119 102 123 123 114
6593.88 Fe I 2.43 2.42 154 145 93 78 96 83 92 93 98 83
6597.57 Fe I 4.79 0.95 9 15 14 12 19 16
6604.60 Sc II 1.36 1.48 61 41 50 40 42 45
6606.97 Ti II 2.06 2.79 12 32 16 18 11 13 17
6608.04 Fe I 2.28 3.96 52 38 16 23
6609.12 Fe I 2.56 2.69 78 104 97 50 50 59 62 62 66 61
6625.02 Fe I 1.01 5.37 40 91 80 26 21 31 20 17 21
6627.54 Fe I 4.55 1.58 13 13 23 9
6633.75 Fe I 4.79 0.80
6643.63 Ni I 1.68 2.01 132 155 115 81 102 116 110 111 99 110
6645.12 Eu II 1.37 0.20 4 12 4 4 9 5 10 12 9 10 16
6646.96 Fe I 2.61 3.96 24 8
6648.12 Fe I 1.01 5.92 4 9 8 14
6677.99 Fe I 2.69 1.35 160 185 162 130 115 110 125 112 127 129 118
6696.03 Al I 3.14 1.57 synth synth synth synth synth synth
6698.66 Al I 3.14 1.89 synth synth synth synth synth synth
6703.57 Fe I 2.76 3.01 36 80 56 20 16 27 20 17
6710.32 Fe I 1.48 4.83 38 91 68 18 16 17 26 22
6717.68 Ca I 2.71 0.61 100 64 61 78 82 72 78 89 64
6726.67 Fe I 4.61 1.07 22 30 14 9
6733.15 Fe I 4.64 1.48 21 11 8
6739.52 Fe I 1.56 4.79 42 10 28 17
6743.12 Ti I 0.90 1.65 31 100 83 12 15 14 18 19 17
6774.27 La II 0.13 1.75 7 92 33 9 13 8 25 5 7 18 6
Table 2a: Linelist and Equivalent Widths
Element E.P. log gf 25026 26022 26086 28092 30013 32169 32171 33099 34029 34225 35172
(Å) (eV)
6003.01 Fe I 3.88 1.07 88
6007.32 Ni I 1.68 3.35 18 67 16 50 57
6007.97 Fe I 4.65 0.70 67 34 48
6008.57 Fe I 3.88 0.94 40 121
6015.25 Fe I 2.22 4.66 26 17
6024.07 Fe I 4.55 0.06 56 45 113 66 89 126 110 97 113
6027.06 Fe I 4.07 1.14 80 73 49 90 75 75
6035.35 Fe I 4.29 2.56
6056.01 Fe I 4.73 0.44 16 21 61 33 66 32 52
6064.63 Ti I 1.05 1.92 14 96 31 79 24 85 75 66
6065.49 Fe I 2.61 1.51 100 91 189 145 179 142 190 181 174
6078.50 Fe I 4.79 0.34 15 64 40 77 49 61 54
6079.02 Fe I 4.65 1.06 57 28 57 55 30 43
6082.72 Fe I 2.22 3.61 14 17 76 66 84 86 93 78
6084.11 Fe II 3.20 3.87 19 22
6086.29 Ni I 4.26 0.54 8 32 30 49 30 30 26 26
6089.57 Fe I 5.02 0.92 41 14 33 30
6091.18 Ti I 2.27 0.42 11 67 21 47 19 61 53 50
6094.38 Fe I 4.65 1.66
6096.67 Fe I 3.98 1.88 38 42 19 34
6102.18 Fe I 4.83 0.37 21 23 81 23 48 87 39 81 56 65
6102.73 Ca I 1.88 0.81 98 95 213 115 156 183 143 185 184 189
6108.13 Ni I 1.68 2.51 35 58 129 76 102 122 105 139 125
6111.08 Ni I 4.09 0.89 11 33 17 48 20 37 37
6120.25 Fe I 0.91 5.95 54 16 46 52 43 42
6121.01 Ti I 1.88 1.32 45 40 33 28 30
6122.23 Ca I 1.89 0.26 145 134 245 150 175 248 172 233 233 228
6126.22 Ti I 1.07 1.42 16 112 30 70 114 72 114 115 97
6128.98 Ni I 1.68 3.36 13 61 44 75 48 74 72 69
6146.24 Ti I 1.87 1.51 43 27
6149.25 Fe II 3.89 2.78 15 28
6151.62 Fe I 2.18 3.33 20 39 109 78 76 65 102 98 95
6154.23 Na I 2.10 1.57 5 112 108 30 74 50 68
6157.73 Fe I 4.07 1.22 14 28
6160.75 Na I 2.10 1.27 17 123 18 62 110 43 92 68 92
6161.30 Ca I 2.52 1.28 17 136 58 90 132 74 119 117 117
6162.18 Ca I 1.90 0.07 134 147 265 183 202 253 199 257 247 248
6165.36 Fe I 4.14 1.51 67 38 71 33 50 49 48
6166.44 Ca I 2.52 1.11 29 42 114 33 82 122 76 118 107 104
6169.04 Ca I 2.52 0.69 50 63 153 65 98 149 94 139 129 135
6169.56 Ca I 2.52 0.42 64 82 174 74 104 161 108 156 152 156
6173.34 Fe I 2.22 2.89 53 49 145 84 116 134 103 153 128 143
6175.37 Ni I 4.09 0.55 7 19 36 43 18 44 34 36
6176.82 Ni I 4.09 0.42 10 63 14 35 50 36 44 47 56
6177.25 Ni I 1.83 3.53 50 32 68 20 56 44 37
6180.21 Fe I 2.73 2.66 22 28 109 52 64 113 70 100 102 90
6186.72 Ni I 4.11 0.96 16 17 20
6188.00 Fe I 3.94 1.69 75 42 71 31 62 54 50
6200.32 Fe I 2.61 2.41 39 60 120 73 86 125 96 127 123 120
6210.67 Sc I 0.00 1.53 66 59 24 67 49 52
6219.29 Fe I 2.20 2.42 83 83 118 117 161 123 160 160 152
6226.74 Fe I 3.88 2.19 7 32 15 54 30
6229.23 Fe I 2.84 3.00 69 43 78 33 98 63 61
6232.65 Fe I 3.65 1.23 37 121 82 96 110 82 127 98 104
6546.24 Fe I 2.76 1.65 73 77 134 126
6551.68 Fe I 0.99 5.77 51
6554.23 Ti I 1.44 1.16 17 97 36 79 69
6556.07 Ti I 1.46 1.10 27 28 96 62
6559.57 Ti II 2.05 2.30 28 35 30 29 72 49
6574.25 Fe I 0.99 5.02 30 19 70 103 72 118 112 115
6586.31 Ni I 1.95 2.81 34 20 77 63 51 66 50 82 81 65 81
6592.92 Fe I 2.73 1.52 104 108 186 115 151 139 210 177 182
6593.88 Fe I 2.43 2.42 76 64 146 98 146 124 174 158 156 164
6597.57 Fe I 4.79 0.95 17 32
6604.60 Sc II 1.36 1.48 61 85 73 66
6606.97 Ti II 2.06 2.79 13
6608.04 Fe I 2.28 3.96 63 48 24 59 58
6609.12 Fe I 2.56 2.69 44 67 77 125 73 114 121 107
6625.02 Fe I 1.01 5.37 17 17 106 52 120 94
6627.54 Fe I 4.55 1.58 36 10 21 20 14 27
6633.75 Fe I 4.79 0.80
6643.63 Ni I 1.68 2.01 97 90 188 116 143 145 140 189 156 149 162
6645.12 Eu II 1.37 0.20 11 11 10 54 10 28 22 34 20 36 16
6646.96 Fe I 2.61 3.96 45 15 26 10 37 28 34 42
6648.12 Fe I 1.01 5.92 66
6677.99 Fe I 2.69 1.35 103 111 200 147 150 232 192 191
6696.03 Al I 3.14 1.57 synth synth synth synth synth synth synth synth synth synth
6698.66 Al I 3.14 1.89 synth synth synth synth synth synth synth synth synth
6703.57 Fe I 2.76 3.01 25 74 29 54 74 77 76
6710.32 Fe I 1.48 4.83 31 61 56 84
6717.68 Ca I 2.71 0.61 62 67 99 106 152
6726.67 Fe I 4.61 1.07 10 30
6733.15 Fe I 4.64 1.48 19 29 27 16 26
6739.52 Fe I 1.56 4.79 13 30 33 62 61
6743.12 Ti I 0.90 1.65 111 65 97 47 129 123 88 99
6774.27 La II 0.13 1.75 7 7 73 51 86 75 112 91 68 90 93
Table 2b: Linelist and Equivalent Widths
Element E.P. log gf 35201 36282 38198 38232 39026 39048 39129 39392 41033 42508 43061
(Å) (eV)
6003.01 Fe I 3.88 1.07 73
6007.32 Ni I 1.68 3.35 96
6007.97 Fe I 4.65 0.70 58 66
6008.57 Fe I 3.88 0.94 131
6015.25 Fe I 2.22 4.66 23 11 36 28 34
6024.07 Fe I 4.55 0.06 108 61 89 90 87 141 130 134 91
6027.06 Fe I 4.07 1.14 71 51 74 53 111 93 96 33
6035.35 Fe I 4.29 2.56 19
6056.01 Fe I 4.73 0.44 19 46 45 50 65 66 67 30 72
6064.63 Ti I 1.05 1.92 57 29 65 57 140 83 112 15 146
6065.49 Fe I 2.61 1.51 177 117 155 167 162 207 196 203 135 224
6078.50 Fe I 4.79 0.34 60 29 49 55 50 77 77 80 38 84
6079.02 Fe I 4.65 1.06 48 16 29 27 58 62 17 66
6082.72 Fe I 2.22 3.61 83 21 67 83 69 76 38 122
6084.11 Fe II 3.20 3.87 19 18 14
6086.29 Ni I 4.26 0.54 39 14 28 28 22 38 39 37 19
6089.57 Fe I 5.02 0.92 22 43
6091.18 Ti I 2.27 0.42 43 26 43 34 117 64 82 105
6094.38 Fe I 4.65 1.66 12 9 9 29 22 23 7 28
6096.67 Fe I 3.98 1.88 15 34 24 58 53 51 15
6102.18 Fe I 4.83 0.37 63 18 48 58 48 80 40
6102.73 Ca I 1.88 0.81 168 114 163 177 162 248 204 220 130 265
6108.13 Ni I 1.68 2.51 121 62 108 119 109 155 128 148 89 150
6111.08 Ni I 4.09 0.89 32 21 23 15 40 44 10 48
6120.25 Fe I 0.91 5.95 34 22 32 34 87 56 76 10 90
6121.01 Ti I 1.88 1.32 19 36 56 95
6122.23 Ca I 1.89 0.26 210 149 196 210 211 253 283 172 355
6126.22 Ti I 1.07 1.42 97 19 83 106 87 183 116 156 48 168
6128.98 Ni I 1.68 3.36 64 55 60 59 93 87 83 28 102
6146.24 Ti I 1.87 1.51 17 42 56 50
6149.25 Fe II 3.89 2.78 20 17 22
6151.62 Fe I 2.18 3.33 84 38 83 105 89 123 104 117 70 132
6154.23 Na I 2.10 1.57 70 28 14 17 149 108 121 7 127
6157.73 Fe I 4.07 1.22 18 74 78 66 56
6160.75 Na I 2.10 1.27 87 14 43 36 35 137 129 127 29 154
6161.30 Ca I 2.52 1.28 107 24 87 88 88 127 61
6162.18 Ca I 1.90 0.07 233 153 205 236 219 362 277 307 182 366
6165.36 Fe I 4.14 1.51 41 39 37 33 65 74 27 68
6166.44 Ca I 2.52 1.11 98 34 84 97 91 170 120 142 60 156
6169.04 Ca I 2.52 0.69 124 48 101 121 116 145 168 84 185
6169.56 Ca I 2.52 0.42 134 122 138 125 209 158 185 108
6173.34 Fe I 2.22 2.89 120 69 109 128 109 152 148 161 96 175
6175.37 Ni I 4.09 0.55 32 16 31 31 54 47 53 17 50
6176.82 Ni I 4.09 0.42 43 39 42 48 61 51 72 62
6177.25 Ni I 1.83 3.53 36 34 44 26 69 60 64
6180.21 Fe I 2.73 2.66 86 37 78 93 84 115 111 108 62 131
6186.72 Ni I 4.11 0.96 15 12 31 24 47 34
6188.00 Fe I 3.94 1.69 45 11 34 41 42 73 68 69 18 76
6200.32 Fe I 2.61 2.41 118 63 109 114 98 141 121 134 70 149
6210.67 Sc I 0.00 1.53 39 18 31 28 161 69 100 143
6219.29 Fe I 2.20 2.42 156 102 125 153 148 174 121 189
6226.74 Fe I 3.88 2.19 20 18 51 40 28 13 43
6229.23 Fe I 2.84 3.00 62 19 46 63 57 88 93 30 106
6232.65 Fe I 3.65 1.23 109 44 97 100 93 135 106 125 67 128
6546.24 Fe I 2.76 1.65 155 117
6551.68 Fe I 0.99 5.77 32 28 52
6554.23 Ti I 1.44 1.16 66 59 153 19 145
6556.07 Ti I 1.46 1.10 59 49 75 73 155
6559.57 Ti II 2.05 2.30 46 49 60 28 47
6574.25 Fe I 0.99 5.02 94 106 86 146 66 119 138 51 162
6586.31 Ni I 1.95 2.81 77 65 72 73 52 77 86 100
6592.92 Fe I 2.73 1.52 108 189 168 208 162 196 205 132
6593.88 Fe I 2.43 2.42 150 95 133 140 137 193 130 159 187 97 199
6597.57 Fe I 4.79 0.95 33 17 48
6604.60 Sc II 1.36 1.48 64 37 72 82 57
6606.97 Ti II 2.06 2.79 33
6608.04 Fe I 2.28 3.96 44 67 42 86 25 58 70
6609.12 Fe I 2.56 2.69 100 58 98 102 86 81 111 129 77 173
6625.02 Fe I 1.01 5.37 87 63 97 66 55 108
6627.54 Fe I 4.55 1.58 25 9 16 13 32 28 12 29
6633.75 Fe I 4.79 0.80
6643.63 Ni I 1.68 2.01 162 92 151 161 155 186 136 150 174 133 188
6645.12 Eu II 1.37 0.20 32 7 24 16 12 39 12 34 16 12 28
6646.96 Fe I 2.61 3.96 29 16 22 21 56 19 36 39 55
6648.12 Fe I 1.01 5.92 10 37 45 40 105 109
6677.99 Fe I 2.69 1.35 196 129 180 172 231 164 209 215 156 247
6696.03 Al I 3.14 1.57 synth synth synth synth synth synth synth synth synth synth synth
6698.66 Al I 3.14 1.89 synth synth synth synth synth synth synth
6703.57 Fe I 2.76 3.01 34 48 60 59 95 57 81 25 123
6710.32 Fe I 1.48 4.83 52 68 55 127 61 32 139
6717.68 Ca I 2.71 0.61 57 103 120 103
6726.67 Fe I 4.61 1.07 22 22 32 11
6733.15 Fe I 4.64 1.48 8 13 6 27 18 40 24 30
6739.52 Fe I 1.56 4.79 10 41 51 95
6743.12 Ti I 0.90 1.65 89 56 88 187 119 153 44 171
6774.27 La II 0.13 1.75 71 12 71 47 49 155 37 52 87 30 98
Table 2c: Linelist and Equivalent Widths
Element E.P. log gf 43389 45358 45485 46121 47215 48116 48323 49013 49037 51021 51080
(Å) (eV)
6003.01 Fe I 3.88 1.07 82
6007.32 Ni I 1.68 3.35 91
6007.97 Fe I 4.65 0.70 37 84 60 49
6008.57 Fe I 3.88 0.94 89 90
6015.25 Fe I 2.22 4.66 22 32 8
6024.07 Fe I 4.55 0.06 101 117 114 105 144 129 134 71 50 102
6027.06 Fe I 4.07 1.14 64 74 85 90 93 95 53 19 62 79
6035.35 Fe I 4.29 2.56
6056.01 Fe I 4.73 0.44 45 67 59 65 80 66 71 40 14 50 58
6064.63 Ti I 1.05 1.92 66 106 67 130 121 140 146 23 8 69 68
6065.49 Fe I 2.61 1.51 157 192 175 198 204 200 206 142 164 170
6078.50 Fe I 4.79 0.34 56 76 69 82 76 92 42 21 42 49
6079.02 Fe I 4.65 1.06 28 42 47 67 15 28 31
6082.72 Fe I 2.22 3.61 74 96 89 120 117 122 50 17 81 81
6084.11 Fe II 3.20 3.87 18 20
6086.29 Ni I 4.26 0.54 19 37 34 43 53 47 50 22 22
6089.57 Fe I 5.02 0.92
6091.18 Ti I 2.27 0.42 52 82 69 86 110 106 113 16 47 53
6094.38 Fe I 4.65 1.66 9 16 19 22 31 27
6096.67 Fe I 3.98 1.88 34 45 48 54 42 64 12 7 28 30
6102.18 Fe I 4.83 0.37 55 89 82 102 48 21 56 59
6102.73 Ca I 1.88 0.81 162 219 195 216 237 256 257 137 98 184 175
6108.13 Ni I 1.68 2.51 111 137 120 142 141 140 152 92 44 124 120
6111.08 Ni I 4.09 0.89 20 48 37 44 55 51 52 12 22 26
6120.25 Fe I 0.91 5.95 37 58 37 75 78 84 93 16 52 39
6121.01 Ti I 1.88 1.32 33 49 20 95 99 14 32 31
6122.23 Ca I 1.89 0.26 206 285 318 338 397 409 159 137 236 227
6126.22 Ti I 1.07 1.42 98 97 161 157 181 185 38 118 109
6128.98 Ni I 1.68 3.36 55 85 80 90 100 101 109 43 11 71 67
6146.24 Ti I 1.87 1.51 65 81 19
6149.25 Fe II 3.89 2.78 13 22 16 24
6151.62 Fe I 2.18 3.33 96 119 104 114 119 114 74 29 104 95
6154.23 Na I 2.10 1.57 30 107 81 68 126 137 137 24 21
6157.73 Fe I 4.07 1.22 37 17 77
6160.75 Na I 2.10 1.27 44 130 113 84 128 151 139 15 4 33 35
6161.30 Ca I 2.52 1.28 101 112 62 20 98 103
6162.18 Ca I 1.90 0.07 224 311 239 314 336 383 378 188 150 233 234
6165.36 Fe I 4.14 1.51 47 62 47 87 89 30 10 38 39
6166.44 Ca I 2.52 1.11 90 141 105 144 132 179 163 66 29 101 107
6169.04 Ca I 2.52 0.69 117 153 131 149 173 179 177 87 45 132 124
6169.56 Ca I 2.52 0.42 140 189 160 175 196 194 102 78 132 143
6173.34 Fe I 2.22 2.89 117 144 132 150 154 159 103 42 128 125
6175.37 Ni I 4.09 0.55 27 63 49 60 46 70 52 31 33
6176.82 Ni I 4.09 0.42 53 71 69 77 82 54 66 28 17 51 49
6177.25 Ni I 1.83 3.53 34 54 56 61 66 83 67 32 11 46 33
6180.21 Fe I 2.73 2.66 83 122 118 123 125 116 61 20 104 93
6186.72 Ni I 4.11 0.96 28 46 29 41 51 17 21 16
6188.00 Fe I 3.94 1.69 40 65 55 80 66 79 79 36 55 48
6200.32 Fe I 2.61 2.41 106 128 116 133 148 92 52 120 121
6210.67 Sc I 0.00 1.53 88 59 107 147 150 17 44
6219.29 Fe I 2.20 2.42 134 131 77 155 150
6226.74 Fe I 3.88 2.19 21 28 54 47 50 12 19 23
6229.23 Fe I 2.84 3.00 61 88 57 87 91 41 63 58
6232.65 Fe I 3.65 1.23 93 133 116 110 122 115 131 88 35 96
6546.24 Fe I 2.76 1.65
6551.68 Fe I 0.99 5.77 86 50
6554.23 Ti I 1.44 1.16 62 147 24 16 81 62
6556.07 Ti I 1.46 1.10 78 142 103 88
6559.57 Ti II 2.05 2.30 41 42 49 50 61 47
6574.25 Fe I 0.99 5.02 89 123 95 132 131 59 122
6586.31 Ni I 1.95 2.81 54 98 70 88 62 95 98
6592.92 Fe I 2.73 1.52 163 166 170
6593.88 Fe I 2.43 2.42 176 140 112 67 153 153
6597.57 Fe I 4.79 0.95 34 68 11 26
6604.60 Sc II 1.36 1.48 45 77 62
6606.97 Ti II 2.06 2.79 56 33
6608.04 Fe I 2.28 3.96 44 70 46 80 66 52
6609.12 Fe I 2.56 2.69 89 131 109 131 91 50 132 111
6625.02 Fe I 1.01 5.37 94 120 86 57 86 83
6627.54 Fe I 4.55 1.58 19 32 27 25 38 21 17
6633.75 Fe I 4.79 0.80
6643.63 Ni I 1.68 2.01 159 161 190 164 134 75 159 151
6645.12 Eu II 1.37 0.20 21 23 9 24 14 7 22 21
6646.96 Fe I 2.61 3.96 43 43 61 9 24 18
6648.12 Fe I 1.01 5.92 73 82 83 21 52 51
6677.99 Fe I 2.69 1.35 171 202 177 208 164 180
6696.03 Al I 3.14 1.57 synth synth synth synth synth synth synth synth synth
6698.66 Al I 3.14 1.89 synth synth synth synth synth synth
6703.57 Fe I 2.76 3.01 63 64 80 93 34 17 67 75
6710.32 Fe I 1.48 4.83 59 104 82 107 122 42 13 65 64
6717.68 Ca I 2.71 0.61 81 60
6726.67 Fe I 4.61 1.07 26 34 24
6733.15 Fe I 4.64 1.48 22 21
6739.52 Fe I 1.56 4.79 38 25 57 48
6743.12 Ti I 0.90 1.65 82 183 164 40 121 100
6774.27 La II 0.13 1.75 64 161 75 91 101 18 10 41 55
Table 2d: Linelist and Equivalent Widths
Element E.P. log gf 51132 54022 54105 55101 55111 55142 56040 57010 59036 59090 59094
(Å) (eV)
6003.01 Fe I 3.88 1.07 101 49
6007.32 Ni I 1.68 3.35
6007.97 Fe I 4.65 0.70 53
6008.57 Fe I 3.88 0.94 114
6015.25 Fe I 2.22 4.66 22 25 7
6024.07 Fe I 4.55 0.06 123 135 120 119 91 56 82 74 65 39
6027.06 Fe I 4.07 1.14 85 112 84 88 54 31 51 27 45
6035.35 Fe I 4.29 2.56
6056.01 Fe I 4.73 0.44 75 84 58 62 45 25 48 26 32
6064.63 Ti I 1.05 1.92 111 112 100 78 54 103 9 47 13 13
6065.49 Fe I 2.61 1.51 202 200 191 168 203 122 164 132 131 92
6078.50 Fe I 4.79 0.34 73 73 46 73 23 39 33
6079.02 Fe I 4.65 1.06 58 71 58 62 26 9 22 13
6082.72 Fe I 2.22 3.61 110 122 110 88 77 102 21 67 29 36
6084.11 Fe II 3.20 3.87 17
6086.29 Ni I 4.26 0.54 40 51 34 41 18 47 7 22 15 13
6089.57 Fe I 5.02 0.92 40
6091.18 Ti I 2.27 0.42 88 108 68 75 30 67 30 12
6094.38 Fe I 4.65 1.66 15 29 23 23 20 8
6096.67 Fe I 3.98 1.88 46 59 51 38 28 40 26 10 9
6102.18 Fe I 4.83 0.37 92 104 81 77 63 84 27 52 34 32
6102.73 Ca I 1.88 0.81 221 248 219 199 110 156 131 136 81
6108.13 Ni I 1.68 2.51 149 152 135 128 119 148 63 106 73 73
6111.08 Ni I 4.09 0.89 52 40 18 11 14
6120.25 Fe I 0.91 5.95 56 75 59 38 64 36 12 16
6121.01 Ti I 1.88 1.32 64 99 32 34 17 44 16 9
6122.23 Ca I 1.89 0.26 304 364 260 245 216 264 142 194 162 159 115
6126.22 Ti I 1.07 1.42 162 173 126 116 98 138 23 81 26 35
6128.98 Ni I 1.68 3.36 86 107 92 80 65 94 24 59 23 34
6146.24 Ti I 1.87 1.51 45 46
6149.25 Fe II 3.89 2.78 28 26 30 23 25
6151.62 Fe I 2.18 3.33 124 120 112 94 111 43 91 57 58 29
6154.23 Na I 2.10 1.57 118 153 104 108 20 106 12 16 15
6157.73 Fe I 4.07 1.22 71 28 72 45 42
6160.75 Na I 2.10 1.27 143 143 110 121 30 121 29 28 16
6161.30 Ca I 2.52 1.28 84 33 79 46 51
6162.18 Ca I 1.90 0.07 316 344 267 257 224 286 157 212 189 166 136
6165.36 Fe I 4.14 1.51 75 90 72 66 49 59 10 28 15 19
6166.44 Ca I 2.52 1.11 147 155 130 122 97 136 34 79 46 45 26
6169.04 Ca I 2.52 0.69 159 171 141 158 111 152 53 105 68 73 39
6169.56 Ca I 2.52 0.42 190 205 180 166 134 169 83 121 86 95 60
6173.34 Fe I 2.22 2.89 133 155 149 144 128 149 80 113 84 83 44
6175.37 Ni I 4.09 0.55 58 63 53 41 29 49 20 34 13 22
6176.82 Ni I 4.09 0.42 68 76 54 53 46 57 17 40 27 25
6177.25 Ni I 1.83 3.53 61 72 51 45 30 60 15 32 15 13
6180.21 Fe I 2.73 2.66 131 119 110 106 78 107 42 78 45 51 28
6186.72 Ni I 4.11 0.96 22 30 16 33 20 17 7
6188.00 Fe I 3.94 1.69 69 88 59 77 30 79 14 38 16 16 15
6200.32 Fe I 2.61 2.41 150 149 124 129 114 135 69 111 68 83 50
6210.67 Sc I 0.00 1.53 103 137 78 56 80 27 11 14
6219.29 Fe I 2.20 2.42 174 168 156 95 140 106 112 81
6226.74 Fe I 3.88 2.19 34 22
6229.23 Fe I 2.84 3.00 88 102 78 77 58 87 41 24 25
6232.65 Fe I 3.65 1.23 129 155 113 128 90 134 54 83 59 59
6546.24 Fe I 2.76 1.65 135 112 121
6551.68 Fe I 0.99 5.77 38
6554.23 Ti I 1.44 1.16 161 92 59 44 28 29
6556.07 Ti I 1.46 1.10 117 127 80 106 54 16
6559.57 Ti II 2.05 2.30 44 50 59 32 50 35 28
6574.25 Fe I 0.99 5.02 132 115 113 99 120 43 92 44 42
6586.31 Ni I 1.95 2.81 83 95 78 84 70 35 63 38
6592.92 Fe I 2.73 1.52 200 225 200 203 166 204 143 128
6593.88 Fe I 2.43 2.42 179 169 169 141 157 96 138 100
6597.57 Fe I 4.79 0.95 21 11
6604.60 Sc II 1.36 1.48 66 44 67 44
6606.97 Ti II 2.06 2.79 14 26
6608.04 Fe I 2.28 3.96 73 77 48 45
6609.12 Fe I 2.56 2.69 152 128 122 117 62 112 59 74
6625.02 Fe I 1.01 5.37 90 70 37
6627.54 Fe I 4.55 1.58 31 39 19 13 7
6633.75 Fe I 4.79 0.80
6643.63 Ni I 1.68 2.01 170 193 166 156 150 161 105 151 111 131
6645.12 Eu II 1.37 0.20 26 34 18 18 8 42 9 12 12 8
6646.96 Fe I 2.61 3.96 67 36 38 22 20
6648.12 Fe I 1.01 5.92 86 51 32 10
6677.99 Fe I 2.69 1.35 209 228 228 196 182 127 179 132 137
6696.03 Al I 3.14 1.57 synth synth synth synth synth synth synth synth synth
6698.66 Al I 3.14 1.89 synth synth synth synth synth synth synth synth
6703.57 Fe I 2.76 3.01 80 87 96 95 60 97 17 51
6710.32 Fe I 1.48 4.83 112 114 63 45 25
6717.68 Ca I 2.71 0.61 73 118 92 83
6726.67 Fe I 4.61 1.07 7 15
6733.15 Fe I 4.64 1.48 19 28 25 29 12 23
6739.52 Fe I 1.56 4.79 89 44 17 35
6743.12 Ti I 0.90 1.65 141 158 138 116 91 128 21 65 26 34
6774.27 La II 0.13 1.75 155 161 88 90 44 111 14 62 10 23
Table 2e: Linelist and Equivalent Widths
Element E.P. log gf 60066 63021 65046 66047 66054 69027 76038 77025 80029 81018 85027
(Å) (eV)
6003.01 Fe I 3.88 1.07 81
6007.32 Ni I 1.68 3.35
6007.97 Fe I 4.65 0.70
6008.57 Fe I 3.88 0.94
6015.25 Fe I 2.22 4.66 13 11
6024.07 Fe I 4.55 0.06 74 71 39 89 106 72 76 62
6027.06 Fe I 4.07 1.14 61 56 81 43 42 45
6035.35 Fe I 4.29 2.56
6056.01 Fe I 4.73 0.44 40 54 48 39 31 21 32
6064.63 Ti I 1.05 1.92 26 24 52 121 84 29 14 13 20
6065.49 Fe I 2.61 1.51 132 98 156 186 181 149 138 121 153
6078.50 Fe I 4.79 0.34 34 30 57 74 55 46 36 31 36
6079.02 Fe I 4.65 1.06 18 17 22 79 41 22 13
6082.72 Fe I 2.22 3.61 73 95 105 60 52 22 54
6084.11 Fe II 3.20 3.87 10 15
6086.29 Ni I 4.26 0.54 15 17 33 33 15 14 13
6089.57 Fe I 5.02 0.92 11 14 15 10
6091.18 Ti I 2.27 0.42 25 18 31 102 69 15 12
6094.38 Fe I 4.65 1.66
6096.67 Fe I 3.98 1.88 18 46 33 22 22
6102.18 Fe I 4.83 0.37 42 41 60 105 76 43 44 19 41
6102.73 Ca I 1.88 0.81 134 122 95 167 219 195 145 134 137 152
6108.13 Ni I 1.68 2.51 95 96 40 103 164 147 94 90 100
6111.08 Ni I 4.09 0.89 17 30 9 14
6120.25 Fe I 0.91 5.95 17 32 68 57 18 15 13
6121.01 Ti I 1.88 1.32 15 58 62 8 14 13
6122.23 Ca I 1.89 0.26 167 161 124 211 288 267 176 169 179
6126.22 Ti I 1.07 1.42 62 44 164 144 59 49 21 58
6128.98 Ni I 1.68 3.36 45 44 10 59 88 85 44 48 28 46
6146.24 Ti I 1.87 1.51 39
6149.25 Fe II 3.89 2.78 27 25 25 31 13 19 21
6151.62 Fe I 2.18 3.33 55 82 137 115 76 72 43 68
6154.23 Na I 2.10 1.57 33 131 37 10 13 18 20
6157.73 Fe I 4.07 1.22 55 24 54 51 58
6160.75 Na I 2.10 1.27 82 36 45 125 50 17 10 19
6161.30 Ca I 2.52 1.28 89 62 20 94 122 58 61 35 65
6162.18 Ca I 1.90 0.07 217 199 145 225 341 281 185 176 179 194
6165.36 Fe I 4.14 1.51 51 20 19 42 67 59 20 23 27
6166.44 Ca I 2.52 1.11 78 73 39 98 153 130 63 70 66 69
6169.04 Ca I 2.52 0.69 115 95 44 133 173 141 89 86 66 95
6169.56 Ca I 2.52 0.42 108 109 80 133 201 173 108 106 90 111
6173.34 Fe I 2.22 2.89 113 91 110 172 129 105 100 72 92
6175.37 Ni I 4.09 0.55 28 21 49 19 20 16
6176.82 Ni I 4.09 0.42 50 36 18 78 70 27 34 31 34
6177.25 Ni I 1.83 3.53 40 62 60 21 24 16
6180.21 Fe I 2.73 2.66 90 61 40 91 109 66 59 65 65
6186.72 Ni I 4.11 0.96 16 24 33
6188.00 Fe I 3.94 1.69 39 34 43 79 61 30 29
6200.32 Fe I 2.61 2.41 99 95 57 106 138 125 105 89 61 88
6210.67 Sc I 0.00 1.53 18 32 68 18
6219.29 Fe I 2.20 2.42 109 83 165 165 129 124 118 133
6226.74 Fe I 3.88 2.19 32 48 31 21 14
6229.23 Fe I 2.84 3.00 36 65 77 81 33 31 45
6232.65 Fe I 3.65 1.23 75 36 126 116 75 70 48 74
6546.24 Fe I 2.76 1.65 114 158 162 129 137 98 144
6551.68 Fe I 0.99 5.77 23
6554.23 Ti I 1.44 1.16 96 32
6556.07 Ti I 1.46 1.10 51 108 38
6559.57 Ti II 2.05 2.30 45 42 58 38 34
6574.25 Fe I 0.99 5.02 51 52 53 123 81 67 90
6586.31 Ni I 1.95 2.81 75 94 104 65 49 74
6592.92 Fe I 2.73 1.52 121 164 141 186 199 154 152
6593.88 Fe I 2.43 2.42 130 99 160 132 118 85 123
6597.57 Fe I 4.79 0.95 23 12
6604.60 Sc II 1.36 1.48 46 46 58 55
6606.97 Ti II 2.06 2.79 22 21 43 21
6608.04 Fe I 2.28 3.96 57 24 25 34 34
6609.12 Fe I 2.56 2.69 89 61 137 122 89 51 74
6625.02 Fe I 1.01 5.37 44 24 49 41 49
6627.54 Fe I 4.55 1.58 31
6633.75 Fe I 4.79 0.80
6643.63 Ni I 1.68 2.01 135 113 180 185 142 137 119 142
6645.12 Eu II 1.37 0.20 20 13 14 9 16 18 10 7 17 11
6646.96 Fe I 2.61 3.96 13 25 33 11
6648.12 Fe I 1.01 5.92 14 36 49 22 19 11 25
6677.99 Fe I 2.69 1.35 141 138 229 175 125 165
6696.03 Al I 3.14 1.57 synth synth synth synth synth synth synth synth synth synth
6698.66 Al I 3.14 1.89 synth synth synth synth synth synth synth synth synth
6703.57 Fe I 2.76 3.01 44 53 100 73 40 39 47
6710.32 Fe I 1.48 4.83 40 27 74 38 33 35 42
6717.68 Ca I 2.71 0.61 101 101
6726.67 Fe I 4.61 1.07 13 35 24 15
6733.15 Fe I 4.64 1.48 27 26 13
6739.52 Fe I 1.56 4.79 24