A Helium-Carbon Correlation on the Extreme Horizontal Branch in ω Centauri
Marilyn Latour, Suzanna K. Randall, Gilles Fontaine, Giuseppe Bono, Annalisa Calamida, Pierre Brassard
aa r X i v : . [ a s t r o - ph . S R ] S e p Accepted for publication in
The Astrophysical Journal
September 12, 2014
Preprint typeset using L A TEX style emulateapj v. 04/17/13
A HELIUM-CARBON CORRELATION ON THE EXTREME HORIZONTAL BRANCH IN ω CENTAURI ∗ M. Latour , S. K. Randall , G. Fontaine , G. Bono , A. Calamida and P. Brassard Accepted for publication in The Astrophysical Journal September 12, 2014
ABSTRACTTaking advantage of a recent FORS2/VLT spectroscopic sample of Extreme Horizontal Branch (EHB)stars in ω Cen, we isolate 38 spectra well suited for detailed atmospheric studies and determine theirfundamental parameters ( T eff , log g , and log N (He)/ N (H)) using NLTE, metal line-blanketed models.We find that our targets can be divided into three groups: 6 stars are hot ( T eff > ∼ T eff < ∼ N (He)/ N (H) > ∼ − ω Cen cluster in a narrow temperature range( ∼ ∼ ω Cen. For the He-rich objects in particular, the clear link between heliumand carbon enhancement points towards a late hot flasher evolutionary history.
Keywords: stars : atmospheres — stars : fundamental parameters — subdwarfs — stars : abundances—Globular Clusters: individual ( ω Centauri) ASTROPHYSICAL CONTEXT
Globular clusters (GCs) are ideal laboratories for con-straining the evolutionary properties of low-mass starsand investigating the formation and kinematic evolu-tion of low–mass stellar systems (Di Cecco et al. 2013;Zocchi et al. 2012). The key advantages in dealing withcluster stellar populations are manifold: cluster starshave the same age and the same iron abundance. More-over, they are located at the same distance and are typi-cally characterized by the same reddening. However, sev-eral of the above assumptions concerning cluster simplestellar populations have been challenged by both spectro-scopic and photometric evidence. It has been shown thatmost GCs host at least two generations of stars differingmainly in light element abundance (Carretta et al. 2010).The second stellar generation is thought to be formedfrom material polluted by the first generation but it isstill not clear exactly how this pollution occurred. Pos-sible polluters might be asymptotic giant branch (AGB)stars during their thermal pulse phase (Ventura et al.2001; Gratton et al. 2004),or fast rotating massive stars(Maeder & Meynet 2006; Decressin et al. 2007). In any ∗ BASED ON OBSERVATIONS COLLECTED AT THE EURO-PEAN ORGANISATION FOR ASTRONOMICAL RESEARCHIN THE SOUTHERN HEMISPHERE, CHILE (PROPOSAL ID386.D-0669 AND 091.D-0791) D´epartement de Physique, Universit´e de Montr´eal, Succ.Centre-Ville, C.P. 6128, Montr´eal, QC H3C 3J7, Canada Dr. Karl Remeis-Observatory & ECAP, Astronomical Insti-tute, Friedrich-Alexander University Erlangen-Nuremberg, Stern-wartstr. 7, 96049 Bamberg, Germany ESO, Karl-Schwarzschild-Str. 2, 85748 Garching bei M¨unchen,Germany Istituto Nazionale de Astrofisica, Osservatorio Astronomico diRoma, via Frascati 33, 00040 Monte Porzio Catone, Italy Universit`a di Roma “Tor Vergata”, Department of Physics,via della Ricerca Scientifica 1, 00133 Rome, Italy Space Telescope Science Institute, 3700 San Martin Drive, Bal-timore, MD 21218 case, the pollution and the subsequent formation of asecond generation of stars happen soon enough in thelife of the cluster (in ∼ -10 yr) that the age spread isusually not detectable at the level of the main sequenceturn off (Gratton et al. 2004).In the specific case of ω Cen, there are atleast three separate stellar populations with a largeundisputed spread in iron abundance (more than1 dex) (Villanova et al. 2007; Calamida et al. 2009;Johnson & Pilachowski 2010). ω Cen also has a dou-ble MS (Anderson 1997; Bedin et al. 2004), the bluersequence being composed of more metal-rich stars thatare also believed to be helium enhanced (Y ∼ .Intriguingly, the occurrence of very blue (thus hot) HBstars in ω Cen (as well as in other massive globular clus-ters with complex populations) cannot be explained bycanonical evolution (D’Cruz et al. 1996). Instead, twocompeting non-canonical scenarios have been proposed.In the first scenario, the blue tail of the HB is explainedby the presence of a He-enhanced second generation ofstars. These stars leave the main sequence with a lowercore mass at a given globular cluster age, resulting in ahigher temperature when they reach the helium burningphase on the HB, thus qualitatively explaining its bluermorphology (D’Antona et al. 2002; Busso et al. 2007). Akey advantage of this scenario is that it can explain the The EHB stars are hot ( T eff > ∼ < ∼ M ⊙ ) that is not massiveenough to sustain significant hydrogen shell burning. M. Latour et al.
EHB population assuming a standard mass-loss efficiencyalong the RGB branch. In contrast, the second evolution-ary scenario predicts that some stars experience a he-lium core flash after having evolved away from the RGB.This “hot flasher” scenario has been modeled by severalgroups (e.g., Castellani & Castellani 1993; D’Cruz et al.1996; Brown et al. 2001) and has indeed been shown toproduce helium burning stars that settle at the very hotend of the EHB. Different “flavors” of hot flashers mayoccur depending on the evolutionary stage of the star atthe time of the helium flash. If the ignition of heliumhappens before the star reaches the white dwarf coolingcurve, the hydrogen-burning shell forms a barrier thatprevents the inner convection zone from reaching the en-velope of the star. This situation is often referred toas an “early hot flash” and results in a hot EHB starwith a H-rich atmosphere. If on the other hand theflash does not occur until the star has settled onto thewhite dwarf cooling sequence, the reduced entropy of themuch weaker hydrogen-burning shell allows the convec-tion zone to extend out to the surface, mixing the heliumand carbon from the core with the hydrogen present inthe atmosphere. Depending on the degree of hydrogenshell burning at the time of the flash, the mixing effi-ciency may vary. In the most common “late hot flasher”case, the deep mixing is likely to burn most of the hydro-gen carried into the interior and the resulting star willthus arrive at the blue end of the EHB with an atmo-sphere dominated by helium. The surface compositionpredicted from models of late flashers is around 95 to 96% helium by mass and 3 to 4 % carbon (Brown et al.2001; Cassisi et al. 2003). Evolutionary paths for dif-ferent types of hot flashers can be found in Figure 4 ofBrown et al. (2001). Note that in all these cases the starmust have lost a large amount of its hydrogen envelopeon the RGB via some mechanism.The hot flasher scenario can also be invoked to ex-plain EHB stars (spectral types sdO and sdB) among theGalactic field population, in particular the He-rich sdBs(Lanz et al. 2004; Miller Bertolami et al. 2008). He-poorsdB stars in the field can be modeled in terms of a canon-ical evolutionary scenario, where the He-flash occurs atthe tip of the RGB (Dorman et al. 1993). The neces-sary mass loss can also be explained in terms of binaryinteractions such as Roche lobe overflow and commonenvelope evolution, mergers involving at least one Hewhite dwarf, or even planet ingestion (see Heber 2009for a review of formation mechanisms). Complicationsappear with the helium-enriched subdwarfs, whose ex-istence cannot be explained by canonical evolution. Inthese cases, alternate scenarios, such as the late heliumflash discussed above or the merger of two white dwarfs(Saio & Jeffery 2000), are invoked. However, accordingto these scenarios, He-rich hot subdwarfs should quicklysettle on the zero age helium main sequence (ZAHEMS),which does not appear to be the case from observations(Stroeer et al. 2007; Heber et al. 2006). Understandingthe formation of EHB stars, both in the field and in glob-ular clusters, therefore remains a challenge.In this context, it remains essential to characterizeas many hot subdwarfs as possible. Globular clustersEHB stars in particular have been studied less thantheir field counterparts due to the obvious observationaldifficulties. However, for ω Cen there have been sev- eral surveys of HB and EHB stars aimed at gaining in-sight on the formation mechanism and evolutionary sta-tus of these objects (see, e.g., Moehler et al. 2002, 2007,2011, and Moni Bidin et al. 2012). These studies com-bined spectroscopic observations and model atmospheretechniques to derive atmospheric parameters for severalhot subdwarf stars in that cluster. Quite interestingly,Moehler et al. (2011) found preliminary evidence for acorrelation between carbon and helium enhancement.This would point toward the hot flasher scenario as theorigin of helium-enriched stars, but does not rule out thepossibility of the He-enhanced scenario also playing a role(Cassisi et al. 2009). Shedding light on the EHB evolu-tion in globular clusters is indeed of great importance,not only for the understanding of the late evolutionarystages of low-mass stars, but also for the interpretationof the multiple populations observed in some GCs.Another development concerning ω Cen and its popu-lation of hot subdwarfs has been the recent discovery ofshort-period EHB pulsators as reported by Randall et al.(2009, 2011). Contrary to initial expectations, these vari-ables turned out not to be the analogs of field sdB pul-sators discovered almost two decades ago (Kilkenny et al.1997), but members of a new family of H-rich sdO pulsat-ing stars with effective temperatures clustering around50,000 K. Interestingly, and despite extensive searches(Johnson et al. 2013), no field counterparts to those sdOpulsators have been found . In an effort to map this new ω Cen instability strip in the log g - T eff diagram, FORSspectroscopy was obtained at the VLT for a sample of60 EHB star candidates. Preliminary results for 19 starswere reported in Randall et al. (2013) and the completestudy will be presented elsewhere (Randall et al. 2014, inpreparation). For the purposes of the present study, weselected 38 stars whose spectra showed no signs of pol-lution from nearby stars for more detailed spectroscopicmodeling. During the course of this we found that themajority of the sample have He-rich atmospheres and,moreover, that carbon features can be seen in 25 of them,thus opening up the possibility of investigating the rela-tionship between helium richness and carbon abundancein hot subdwarfs in a globular cluster environment. Inlight of the currently raging debate on EHB star evolu-tion we felt that this would be a most worthwhile en-deavour to pursue. In Section 2 we thus present ourobservational material, followed by the resulting funda-mental parameters and carbon abundances measured forour stars in Section 3. Finally, in Section 4 we discussour results and compare them to similar results for fieldstar samples and theoretical predictions. OBSERVATIONAL MATERIAL
The spectra used here were taken from the initial sam-ple of 60 ω Cen EHB star candidates as described indetail by Randall et al. 2014 (in preparation). The spec-troscopic targets had been selected as EHB star candi-dates based on their brightness and colour in the ω CenWFI/ACS catalogue (Castellani et al. 2007). Since theprimary aim of the observations was to map the sdO in-stability strip, the color cut favors the hotter part of the Pulsating hot subdwarfs have also been discovered quite re-cently in the globular cluster NGC 2808 as reported by Brown et al.(2013) but fundamental parameters still have to be determined forsome of them. elium-Carbon Correlation on the EHB in ω Centauri Figure 1.
Spectra of the 38 stars that make up our sample of EHB stars in ω Cen. The spectra are displayed over the 3980 − − , the accepted value for ω Cen (Harris1996). The spectra are presented in the same order as in Table 1, and the color used to indicate their names makes it easy to associatethem with their respective group (see Section 3). Panels a and b.
M. Latour et al.
Figure 1.
Continued. Panels c and d. elium-Carbon Correlation on the EHB in ω Centauri Figure 1.
Continued. Panels e and f.
M. Latour et al.
Figure 1.
Continued. Panels g and h. elium-Carbon Correlation on the EHB in ω Centauri T eff > ∼ i triplet (5167, 5172, and 5183˚A), Na i doublet (5890 and 5896 ˚A), and Ca ii K line.Of course some of these lines originating from the inter-stellar medium can be seen in the spectra of our uncon-taminated sample, but unlike pollution by a companion,the interstellar lines are not redshifted at the cluster’svelocity ( ∼
232 km s − , Harris 1996).The spectra were obtained in March 2011 and April2013 using the MXU mode of FORS2 mounted at theVLT on Cerro Paranal, Chile. Each spectrum is basedon the combination of two 2750 s exposures obtainedwith the 600B grating and a slit width of 0.7 ′′ , and hasa wavelength resolution of ∼ B magnitude of ∼ iii complex near 4650˚A for each spectrum and compared this to the derived Heabundance as obtained in a preliminary spectral analysis.The results of this operation are summarized in Figure2, which shows that the C iii feature is detectable in25 of our sample of 38 stars. Most interestingly, how-ever, Figure 2 suggests a clear correlation between theHe abundance and the strength of that C feature. Thisfinding provided the main incentive to push further andattempt a quantitative measurement of the carbon abun-dance through detailed atmosphere modeling. SPECTROSCOPIC ANALYSIS
Fundamental Parameters
Given that the spectra appeared to span a significantrange in effective temperature and helium abundance, wedecided to build dedicated grids of NLTE model atmo-spheres in order to estimate the fundamental parametersof our sample stars in a homogeneous way. Because mostof the stars were expected to be hotter than typical sdBs(He ii lines are present in most spectra) and also richer inhelium, we adopted solar abundances for carbon, nitro- gen, and oxygen in our models. This proxy metallicitywas adopted simply because these elements are the mostimportant perturbators of the atmospheric structure ofa hot star at their normal abundances. Note that wedid not add iron in our computations because a solarabundance would not have significantly changed the at-mospheric structure in the presence of CNO in solar pro-portions (Haas et al. 1996; Latour et al. 2011), and theextra computation time needed outweighed the limitedbenefits. Our model atmospheres and synthetic spec-tra were computed with the public codes TLUSTY andSYNSPEC and include the following ions (besides thoseof H and He) : C ii to C v , N ii to N vi and O ii toO vii . Note that, as usual with TLUSTY, the highestionization stage of each element is taken as a one-levelatom. Additional information on the model atoms can befound on TLUSTY’s Web site and in Lanz & Hubeny(2003; 2007). The grid we computed included modelswith T eff between 26,000 K and 58,000 K in steps of2,000 K, log g between 5.2 and 6.4 in steps of 0.2 dex,and log N (He)/ N (H) from − β upto and including H11) as well as all the strong heliumlines of both ionization stages (present between λ χ minimization procedure sim-ilar to that of Saffer et al. (1994). After applying the ω Cen radial velocity shift correction of 232 km s − tothe spectra, the minimization procedure automaticallycorrect for any residual discrepancies by matching thecore of the observed lines with the modeled ones. How-ever, a few observed spectra did not include the He ii Figure 2.
Correlation between the He abundance and the equiva-lent width of the CIII 4650 complex detected formally in 25 of our38 sample stars. The equivalent width is evaluated in arbitraryunits. http://nova.astro.umd.edu/ M. Latour et al.
Figure 3.
Resulting fits for four stars of our sample. Top panels (a-b) feature Group 2 stars with a helium abundance of around solar.The He i and ii lines are well reproduced but some residuals can be seen in the fit to the lower hydrogen lines. The bottom-left panel (c)shows fit for a helium enriched star, and the bottom-right (d) one illustrates the fit for a hotter, helium depleted star (this is the pulsatingstar identified as V1 in Randall et al. 2011).Note that H ǫ , traced by a dashed line, was not included in the fitting procedure because it ispolluted by a fairly strong H component of the Ca ii doublet (the K component is also seen in the observed spectrum). line at 5412 ˚A, and a few others were cut in the blue dueto their positions on the CCD chip so that the higherBalmer lines (between H8 and H11) could not alwaysbe included. In addition, the H ǫ line was explicitly disre-garded because of the interstellar pollution caused by theH line of Ca ii . Our derived parameters ( T eff , log g , andlog N (He)/ N (H)), as well as the abundances by massfraction are listed in Table 1, and Figure 3a to Figure 3ddisplay representative fits for four stars. Given the rela-tive faintness of the targets, the results we achieved arequite satisfactory in terms of simultaneoulsy fitting allof the available lines. This suggests that the derived at-mospheric parameters are reliable. Note, however, thatthe quoted uncertainties refer only to the formal errorsof the fits; the true uncertainties will certainly be larger.A cursory inspection of our results suggests that ourtarget stars can naturally be divided into three groups,and this is indicated in Table 1. The seven coolest ob-jects in the table form our Group 1 and are typicalH-rich sdB stars. Our Group 2 is constituted of the25 following stars, which are He-rich subdwarfs with log N (He)/ N (H) > ∼ − low the imposed limit. Our Group 3 is made of the 6hottest objects in our sample and is a collection of hotH-rich sdO subdwarfs, including 4 pulsators (5034421,177238, 154681, 281063).The natural separation between the three groups ofstars is easily seen in Figure 4, which depicts the heliumnumber abundance (relative to hydrogen) as a functionof effective temperature for the 38 stars of our sample.The coolest and hottest stars show significant under-abundances of helium, while among the He-rich objects(Group 2, in red) a positive correlation is seen betweenthe two parameters. This is reminiscent of the relationsfound by Edelmann et al. (2003) for field sdB stars (seetheir figure 5). Figure 5 shows our sample in the log N (He)/ N (H)-log g plane, where the six coolest starsstand out with their low helium abundances and sur-face gravities. Combined with the lower effective tem-peratures inferred, this implies they are typical heliumcore burning sdB stars. In contrast, the Group 3 starsmust certainly be the analogs of post-EHB H-rich hotsdO subdwarfs on their way to the white dwarf regime.It is worth mentioning that the Group 3 objects show aHe abundance lower than the He richest stars of Group2 by ∼ Helium and Carbon Abundances elium-Carbon Correlation on the EHB in ω Centauri Table 1
Atmospheric and Other Parameters for the 38 Stars of our SampleNumber T eff (K) log g log N (He)/ N (H) X(H) X(He)5238307 25711 ±
400 5.35 ± − ± ± ± ±
468 5.48 ± − ± ± ± ±
602 5.53 ± − ± ± ± ±
454 5.38 ± − ± ± ± ±
280 5.48 ± − ± ± ± ±
281 5.41 ± − ± ± ± ±
317 5.75 ± − ± ± ± ±
392 5.67 ± − ± ± ± ±
327 5.73 ± − ± ± ± ±
403 5.77 ± − ± ± ± ±
316 5.91 ± − ± ± ± ±
307 5.71 ± − ± ± ± ±
428 5.84 ± − ± ± ± ±
335 5.54 ± − ± ± ± ±
328 5.71 ± − ± ± ± ±
401 5.76 ± − ± ± ± ±
402 5.72 ± − ± ± ± ±
506 5.59 ± − ± ± ± ±
387 5.75 ± − ± ± ± ±
408 5.70 ± − ± ± ± ±
428 5.71 ± − ± ± ± ±
327 5.70 ± − ± ± ± ±
433 5.72 ± − ± ± ± ±
368 5.82 ± − ± ± ± ±
863 5.90 ± − ± ± ± ±
599 5.93 ± ± ± ± ±
530 5.97 ± − ± ± ± ±
340 5.60 ± − ± ± ± ±
549 5.69 ± − ± ± ± ±
371 5.66 ± − ± ± ± ±
523 6.06 ± ± ± ± ±
362 6.01 ± ± ± ± ±
637 5.88 ± − ± ± ± ±
824 5.89 ± − ± ± ± ±
877 6.07 ± − ± ± ± ±
758 5.89 ± − ± ± ± ± ± − ± ± ± ± ± − ± ± ± As mentioned earlier, our spectra show a correlationbetween the presence (and strength) of carbon lines andthose of helium. This is the same phenomenon as de-scribed by Stroeer et al. (2007), who demonstrated a linkbetween helium enrichment and the presence of carbonand/or nitrogen lines in field sdO stars. Given that ourspectra of ω Cen stars are rather limited in resolutionand sensitivity, they are not optimally suited for study-ing weak metal lines in the optical domain. Neverthe-less, carbon lines were easily found in our spectra, evenfor the less helium-rich stars, thanks to the strong C iii complexes around 4070 ˚A and 4650 ˚A (see Fig. 1). Forthe weaker nitrogen lines, it was possible to associate fea-tures in the spectra with N ii and iii only in the mosthelium-rich stars. Therefore, we decided to focus our ef-forts on the carbon lines and specifically on quantifyingthe amount of carbon present in the atmosphere of thesestars.In order to accurately derive the carbon abundances,we built a small grid of model atmospheres for each starin our sample, keeping the fundamental parameters ofthe models fixed at the values given in Table 1, but vary-ing the carbon abundance. Looking back at the spectro-scopic fits we obtained for our Group 1 stars, it becameclear that the solar abundance initially assumed for C,N, and O was far too high, yielding strong metal lines inthe synthetic spectra that were not recovered in the ob- servations. In fact, with the exception of a few very weaklines in the hottest star, none of the Group 1 stars showany metal lines whatsoever. Adding to this the annoyingtendency of our strongest C lines to blend with the O ii lines (see below), we decided to include only carbon as ametal in our small model grids for Group 1 stars. For theother helium-poor stars in Group 3, the higher effectivetemperatures wipe out most of the metal lines even whenincluded at solar abundance, so we were able to use mod-els with the original solar amount of oxygen and nitrogen.For both of these H-rich groups we varied the carbonabundance of our models from log N (C)/ N (H)= − − N (C)/ N (H)= − λλ iii complexes with theO ii lines became problematic, because at the low reso-lution of our observations (2.6 ˚A) a solar amount of oxy-0 M. Latour et al.
Figure 4.
Helium abundance versus effective temperature for the38 stars of our sample. Group 1 stars are found at lower temper-atures and are illustrated in purple. Group 2 stars are depictedin red and are generally He-rich objects. A clear trend of increas-ing helium abundance with effective temperature can be noticedamong them. Finally, the hottest stars forming Group 3 are shownin blue. The error bars include only the formal uncertainties ofthe fitting procedure and should be regarded as lower limits. Thedotted line indicates the solar helium abundance.
Figure 5.
Similar to Fig. 4, but showing helium abundance versussurface gravity. Though no clear systematic trends are seen, eachgroup of stars is well-defined in this diagram. The position ofGroup 1 stars (in purple) is consistent with typical EHB subdwarfB stars. gen in a model without carbon produces small featuresmimicking weak C iii lines. Therefore, we reduced theamount of oxygen to log N (O)/ N (H) = − λλ iii complexes around 4070 ˚A and4650 ˚A, the C iii lines at 4162.9 and 4186.9 ˚A, the C ii lines at 4267.3, 4516.8 and 4619.0 ˚A, and C iv at 4658.3˚A. The abundances obtained from these five spectral re-gions are listed in Table 2 together with the weightedmean adopted as the final abundance for each star andthe equivalent mass fraction. In a few spectra, the Clines were visible only marginally or not at all; in thesecases we deduced an upper limit on the abundance usingthe strongest C line (usually λ ii lines require a C abundance higher than themean in order to be well reproduced (thus they appeartoo weak in our models), and the C iii line at 4070 ˚A of-ten gives a lower abundance than its counterpart at 4650˚A. The C ii doublet around 4074.5 ˚A also often appearstoo strong when using the abundance needed to repro-duce the neighbouring C iii complex. An example of thelatter behaviour is illustrated for star 5138707 (see Fig.6). The reason for these systematic differences in derivedC abundance based on different lines remains unclear.Figure 7 highlights one of the main results of our spec-troscopic study of EHB stars in ω Cen: we find a strongcorrelation between the helium and carbon abundances.The different groups of stars are also easily distinguish-able in this plot: the most He-rich stars of Group 2form the carbon-enriched population, the rest of Group2 forms a clump around the solar abundance of carbon,and the hot stars of Group 3 are found at the lower abun-dance end of the correlation. Finally, for the six coolestGroup 1 stars we can place only an upper limit on thecarbon abundance. The Group 1 star 5180753 is a bitpeculiar in that it falls into the low abundance regionpopulated by the Group 3 stars. Indeed, it is the onlyGroup 1 star with a measured carbon abundance. Interms of its (low) effective temperature, this star belongsto Group 1, but its carbon and helium abundances, andalso to a lesser extend its surface gravity, are more simi-lar to the Group 3 stars. This explains its odd position in elium-Carbon Correlation on the EHB in ω Centauri Figure 6.
Results of our fitting procedure for five regions containing carbon lines for the four stars presented in Fig. 3a to Fig. 3d, withtheir mean log N (C)/ N (H) indicated. The carbon abundance in the model is that listed in Table 2 for the respective region. If a regionwas not fit, the mean abundance was used in the model. some of the previous plots. It is worth mentioning that inspite of the differences in derived C abundance depend-ing on the line used, the same correlation is seen for eachof the five lines fit. This correlation can be describedby a linear regression between the mean abundance ofcarbon and the helium abundance as evaluated from the30 stars for which carbon was detected. It is shown inFigure 7 by the straight black line and can be describedwith the equationlog N(C) / N(H) = 1 . ± . × log N(He) / N(H) − . ± . . (1)A similar positive correlation was found by N´emeth et al.(2012) in their study of field subdwarfs, however theirslope is steeper.We mentioned at the beginning of this section thatthe nitrogen lines are relatively weak in our spectra andare only discernible in the most helium rich stars. Nev-ertheless, we attempted to roughly estimate the amountof nitrogen present. For the seven most carbon richstars, we computed model atmospheres with nitrogenabundances of 10 and 100 times the solar value (theoriginal models computed for fitting these stars had asolar abundance of nitrogen) and compared, by eye, theresulting synthetic spectra with those observed. Theonly useful line for this exercise turned out to be N iii at 4640 ˚A. We estimate that two out of these seven stars may have nitrogen abundances as high as 100 timessolar. These two stars are 165237 and 5039935, whichare also the most helium and carbon enriched. For theremaining five carbon rich stars the abundances seemto vary between 10 and 100 times solar. For the rest ofGroup 2, no nitrogen lines are detected above the noise,allowing us to place an upper abundance limit of roughlysolar, since this produces lines comparable to the noiselevel of the observations. So, while the quality andresolution of our spectra prevent us from formally fittingand deriving firm abundances for the nitrogen lines, ourinspection indicates that N enrichment appears to gohand in hand with helium and carbon enrichment. DISCUSSION
Comparison with field sdB-sdO stars
The distribution of our sample of stars in the log g - T eff diagram is illustrated in Figure 8. Here, the sizeof each circle is proportional to the logarithmic he-lium abundance, filled circles denoting He-poor stars (log N (He)/ N (H) < ∼ − ω Cen (Z=0.002, solidlines, and Z=0.0003, dashed lines) and normal helium2
M. Latour et al.
Table 2
Inferred Carbon Abundances (log N (C)/ N (H))Name C iii iii iii ii ii < − < − × − < − < − × − < − < − × − < − < − × − < − < − × − < − < − × − < − − ± − ± × − − ± − ± − ± × − − ± − ± − ± − ± − ± × − − ± − ± − ± − ± × − − ± − ± − ± − ± − ± × − − ± − ± − ± − ± × − − ± − ± − ± − ± − ± × − − ± − ± − ± − ± − ± × − − ± − ± − ± − ± × − − ± − ± − ± × − − ± − ± − ± − ± − ± × − − ± − ± − ± × − − ± − ± − ± − ± − ± × − − ± − ± − ± − ± × − − ± − ± − ± − ± × − − ± − ± − ± − ± − ± − ± × − − ± − ± − ± − ± − ± × − − ± − ± − ± − ± × − − ± − ± − ± − ± − ± × − − ± − ± − ± − ± − ± − ± × − − ± − ± − ± − ± − ± − ± × − − ± − ± − ± − ± − ± × − − ± − ± − ± − ± − ± − ± × − − ± − ± − ± − ± − ± − ± × − − ± − ± − ± − ± − ± − ± × − − ± − ± − ± − ± − ± − ± × − − ± − ± − ± − ± × − − ± − ± − ± × − − ± − ± − ± × − − ± − ± × − < − < − × − < − < − × − content (Y ∼ T eff ∼ ω Cen sample into perspec-tive, Figure 9 shows an equivalent picture for hot sub-dwarfs in the Galactic field. These data are the resultof many years of efforts first presented by Green et al.(2008) through the Arizona-Montr´eal Spectroscopic Pro-gram and updated recently by Fontaine et al. (2014) onthe basis of model atmospheres comparable to those usedby us. In Figure 9, the ZAEHB, TAEHB and ZAHEMSwere computed with our local hot subdwarf evolutionarycode at Universit´e de Montr´eal. Here, the models have ametallicity of Z = 0.02, representative of the field stars depicted. Note, in particular, how most of the H-richobjects fall in between the ZAEHB and TAEHB as ex-pected, while the other H-rich stars can be interpretedas evolved, post-EHB objects on their way to the whitedwarf cooling domain. This is similar to what we findfor ω Cen, with the slight difference that the field H-rich sdOs are predominantly found at lower log g thanthose in our sample. But it is the location of the bulkof He-rich stars in the field in relation to their locationin ω Cen that is particularly intriguing. While the He-rich stars in the field cluster around 45,000 K and areclearly hotter than the He-core burning EHB, the ma-jority of their cluster counterparts are distinctly coolerand in line with the predicted EHB. Note that the fielddistribution of He-sdOs as described in the independentstudies of Stroeer et al. (2007) and N´emeth et al. (2012)is very similar to the results shown in Figure 9.In view of our findings we checked the indepen-dent spectroscopic study of Moehler et al. (2011) which,among other things, led to the characterization of 17 He-rich subdwarfs in ω Cen . A look at their Table 4 shows Note that the majority of their helium-rich subsample con-sists of different stars from ours; only three stars are present in elium-Carbon Correlation on the EHB in ω Centauri Figure 7.
Helium abundance versus the mean carbon abundance,still using the same color coding as in Fig 4. The upper limits onthe carbon abundance inferred for eight stars are indicated by ar-rows instead of error bars. This diagram shows an obvious relationbetween the abundances of the two elements, which is illustratedby the linear regression (black solid line) based on the 30 stars forwhich a carbon abundance was obtained. Dotted lines indicate thesolar helium and carbon abundances. that they also infer effective temperatures around or be-low 40,000 K for most of the He-rich stars. In fact, theirhelium rich subsample is even found at slightly lowertemperatures than ours. It is not clear if this small dif-ference is a systematic effect related to the way in whichthe atmospheric parameters were estimated or the con-sequence of a different color cutoff in the sample favour-ing hotter stars for our study. Interestingly, the sampleof EHB stars analysed in NGC 2808 by Moehler et al.(2004) comprises a modest number of He-enriched starsthat are found at similar temperatures as those in ω Cen.The difference in temperature between He-rich EHBstars in ω Cen and the field was already reported ina review by Heber (2009), together with the findingthat the field population contains a higher fraction ofstars with strongly enriched helium abundances. In-deed, only a few stars in our and Moehler et al.’s sam-ple have log N (He)/ N (H) > ω Cen. These differences point towardsfundamental differences between the helium-enrichedEHB star population in the field and in ω Cen, and arelikely related to the fact that sdB and sdO stars in GCshave older (12 −
13 Gyr) and typically metal poorer pro-genitors than their field counterparts.The hot subdwarf stars found in the Sloan Digital SkySurvey (SDSS) might shed some light on these differ- both samples: 53945, 75981, and 5142999 (164808 in Moehler et al.2011) ences. Hirsch (2009) carried out a spectral analysis of asample of hot subdwarf spectra selected from the SDSScatalog (Data Release 7) and found a population of starsquite similar to those in ω Cen in terms of the T eff − log N (He)/ N (H) distribution (see his Figure 7.1 andalso Heber et al. 2006). These stars have helium abun-dances between solar and log N (He)/ N (H)= 0.0, andeffective temperatures mostly below 40,000 K. A pre-liminary spectral analysis of hot subdwarfs from SDSSDR10 (P. N´emeth, priv. comm.) reveals a similar pop-ulation. The space distribution and kinematics of theHirsch (2009) sample suggest that most of the stars be-long to the halo population, which is likely similar to thatof ω Cen in that it is relatively old and metal-poor. Incontrast, most surveys of bright field stars favour metal-rich population I stars, which might explain why the at-mospheric properties derived are different.
Comparison with evolutionary theory
The different characteristics and properties of the starsin our sample are likely to bear traces of their evolution-ary history. The question is how were these stars formed?As already indicated in the Introduction, the answer tothis is by no means straightforward.For the helium-poor stars in our sample (Group 1 and3) the evolutionary status at least is easier to pin downsince their atmospheric parameters are consistent withthem being typical EHB stars (Group 1) and hotterevolved post-EHB stars (Group 3). Let us remember thatcanonical evolution identifies EHB stars as the progenyof red giant stars that are subject to important mass lossbefore or at the helium flash, leaving behind a heliumcore burning star stripped of most of its hydrogen en-velope. According to Figure 8a, our modest sample ofhydrogen-rich sdBs sits right on the EHB as predictedfor models with a canonical He-abundance (Y ∼ ω Cen (Y ∼ ∼ M. Latour et al.
Figure 8.
Distribution of our sample of ω Cen EHB stars in the log g - T eff plane. The size of a given circle is a logarithmic measure of theHe abundance relative to that of H. He-poor and He-rich stars are represented by filled and open circles respectively. Left - the ZAEHB andTAEHB are plotted for two different metallicities, Z=0.002 (solid line) and Z=0.0003 (dashed line), and a normal helium content, Y ∼ the hydrogen and leave behind an atmosphere composedof approximately 96% helium by mass and 3 to 4% car-bon. But in our sample only the three most helium-richstars have a mass fraction of helium higher than 90%,with a maximum carbon mass fraction of 1.5%. These Figure 9.
Similar to Fig. 8, but depicting the distribution ofhot subdwarfs in the field (Fontaine et al. 2014). The ZAHEMS,as well as ZAEHB and TAEHB (assuming a core mass of 0.47 M ⊙ )now refer to models with a metallicity of Z =0.02. three stars could still potentially fit within the frameworkof a late-flash event, but the bulk of our Group 2 starshave substantially lower helium abundances and cannotbe reproduced by this scenario. We should mention thatan intermediate type of flash mixing (known as “shallowmixing” as opposed to “deep mixing”) was proposed andstudied by Lanz et al. (2004) and Miller Bertolami et al.(2008). During a shallow mixing event, the inner andouter parts of the star are not mixed as efficiently; inparticular, hydrogen is only diluted (not burned) withthe helium and carbon-rich material dredged up fromthe core. Therefore, the amount of hydrogen remainingin the envelope is higher in a shallow mixing case. Whilethis scenario seems to fit our measured abundances bet-ter than deep mixing event, the latter is the more usualoutcome of a hot flasher event.Examples of late flasher evolutionary tracks(Miller Bertolami et al. 2008) are shown in Figure10 overplotted with our sample. It is obvious thatthe model tracks do not reproduce the observationsvery well and in particular predict higher values of log g on the EHB (where the stars spend most of theirHe-burning life) than measured. This may be a conse-quence of the fact that neither the shallow nor the deepmixing models consider gravitational settling. Indeedcalculations indicate that the inclusion of gravitationalsettling moves the HB tracks towards lower effectivetemperatures and surface gravities (Moehler et al. 2004;Michaud et al. 2011). Moreover, the late flasher modelsdo not account for the mean molecular barrier whendealing with the convective zones, which may also affectthe ZAHB location of the models.When modeling the evolution of stars as compact asthose on the EHB it is essential to take into consid- elium-Carbon Correlation on the EHB in ω Centauri Figure 10.
Comparison of our sample with late-flasher evolu-tionary tracks by Miller Bertolami et al. (2008). The tracks arefor a metallicity Z=0.001 and respectively refer to a deep-mixingevent (M=0.48150 M ⊙ , dotted line) and a shallow mixing event(M=0.49145 M ⊙ , dashed-dotted line). Points at 5 Myr intervalsare shown on the first track, to give an idea of the evolutionarytimescale in the different regions. eration the diffusion processes occuring in their atmo-spheres. Assuming that He-rich subdwarfs are born withmore or less the atmospheric composition predicted bythe late flasher scenario, the heavier elements will quicklydissipate from the atmosphere due to gravitational set-tling if there are no competing mechanisms to slow downthis process. Indeed, the time needed after the pri-mary helium flash for the star to settle on the ZAEHBis of the order of ∼ yr (Miller Bertolami et al. 2008;Brown et al. 2001), while diffusion if left unimpeded willtransform the initially He-rich atmospheric compositionto one dominated by H on a timescale of only ∼ yr.This is illustrated in Figure 11, where we show the evo-lution of the surface abundances of H, He, and C (inmass fraction) for a 0.47 M ⊙ subdwarf with an initialatmospheric composition as predicted by the late flasherscenario. Hence, diffusion must be slowed down in theHe-rich subdwarfs if they are to be detected as such. Themost likely mechanisms for this are stellar winds or in-ternal turbulence (Hu et al. 2011). Note that radiativelevitation is not a dominant contributor to the slowingdown of gravitational settling, since it can only maintaina subsolar amount of carbon and helium in the atmo-sphere of a subdwarf such as the one modeled in Figure11.The relation we find between the carbon and the he-lium abundance (Figure 7) for EHB stars in ω Cen isalso observed in field EHB stars (N´emeth et al. 2012),and can largely be interpreted as the signature of diffu-sion effects. The fact that the correlation between the Cand He abundances is positive and that the slope (1.36)is larger than 1 in a log-log C vs He abundance plotis a strong indication that chemical separation is going
Figure 11.
Mass fraction of H, He and C as a function of agefor an evolutionary model including gravitational settling. Afterroughly 10 yr, the initial helium and carbon rich composition ofthe atmosphere has become hydrogen dominated. The amount ofHe and C that can be supported by radiative levitation is indicatedby dotted lines. on in these stars, albeit slowed down by a competingagent, with carbon sinking faster than helium as can beexpected. Interestingly, in this plot stars belonging toGroup 2 and 3 follow the same relation, thus suggest-ing a possible evolutionary link between them. The hotsdOs could be post late flashers rapidly evolving towardsthe white dwarf cooling sequence. Miller Bertolami et al.(2008) suggested that due to diffusion effects, the He-rich late flashers could turn into hydrogen-rich objectsbefore approaching the white dwarf regime. A commonorigin for these two groups of stars could offer an expla-nation for their relatively high surface gravity. Canoni-cal post-EHB evolutionary tracks usually predict risingluminosities after core He-exhaustion and can accountrather well for the lower surface gravities measured forthe hotter field sdOs (log g . ω Cen EHB stars. Such a rela-tion has been suspected for many years (Moehler et al.2002, 2007, 2011), but has never been quantified untilnow. By measuring the carbon abundances in our spec-tra using dedicated models we were able to confirm thathelium enhancement is directly linked with carbon abun-dances being around or above solar. The linearity of theHe-C relation (in the logarithmic plane) likely bears thesignature of diffusion effects, but also requires a forma-tion mechanism that enriches the surface of the star notonly with helium but also with carbon. Therefore, ourresults strongly favor a late flasher history over the He-enhanced scenario for our helium-enriched stars. In fact,three quarters of our sample could fit within this frame-6
M. Latour et al. work . Unfortunately, current evolutionary models forlate hot-flashers cannot fully explain the characteristicsof our objects; their position in the log g - T eff diagramis not correctly reproduced and diffusion certainly needsto be taken into account to recover the chemical compo-sitions measured. We are hopeful that the results pre-sented here will trigger the development of more sophisti-cated models fine-tuned to the EHB star population of ω Cen and eventually help solve the evolutionary mystery.This work was supported in part by the NSERCCanada through a doctoral fellowship awarded to M.L. and through a research grant awarded to G. F. Thelatter also acknowledges the contribution of the CanadaResearch Chair Program. M. L. also acknowledges fund-ing by the Deutsches Zentrum fr Luft- und Raumfahrt(grant 50 OR 1315). This work was partially supportedby PRIN–INAF 2011 ”Tracing the formation and evolu-tion of the Galactic halo with VST” (P.I.: M. Marconi)and by PRIN–MIUR (2010LY5N2T) ”Chemical and dy-namical evolution of the Milky Way and Local Groupgalaxies” (P.I.: F. Matteucci).REFERENCES
Anderson, J. 1997, PhD thesis, University of California atBerkeleyBedin, L. R., Piotto, G., Anderson, J., Cassisi, S., King, I. R.,Momany, Y., & Carraro, G. 2004, ApJ, 605, L125Brown, T. M., Landsman, W. B., Randall, S. K., Sweigart, A. V.,& Lanz, T. 2013, ApJ, 777, L22Brown, T. M., Sweigart, A. V., Lanz, T., Landsman, W. B., &Hubeny, I. 2001, ApJ, 562, 368Busso, G., Cassisi, S., Piotto, G., Castellani, M., Romaniello, M.,Catelan, M., Djorgovski, S. G., Recio Blanco, A., Renzini, A.,Rich, M. R., Sweigart, A. V., & Zoccali, M. 2007, A&A, 474,105Calamida, A., Bono, G., Stetson, P. B., Freyhammer, L. M.,Piersimoni, A. M., Buonanno, R., Caputo, F., Cassisi, S.,Castellani, M., Corsi, C. E., Dall’Ora, M., Degl’Innocenti, S.,Ferraro, I., Grundahl, F., Hilker, M., Iannicola, G., Monelli, M.,Nonino, M., Patat, N., Pietrinferni, A., Prada Moroni, P. G.,Primas, F., Pulone, L., Richtler, T., Romaniello, M., Storm, J.,& Walker, A. R. 2009, ApJ, 706, 1277Carretta, E., Bragaglia, A., Gratton, R. G., Recio-Blanco, A.,Lucatello, S., D’Orazi, V., & Cassisi, S. 2010, A&A, 516, A55Cassisi, S., Salaris, M., Anderson, J., Piotto, G., Pietrinferni, A.,Milone, A., Bellini, A., & Bedin, L. R. 2009, ApJ, 702, 1530Cassisi, S., Schlattl, H., Salaris, M., & Weiss, A. 2003, ApJ, 582,L43Castellani, M. & Castellani, V. 1993, ApJ, 407, 649Castellani, V., Calamida, A., Bono, G., Stetson, P. B.,Freyhammer, L. M., Degl’Innocenti, S., Moroni, P. P., Monelli,M., Corsi, C. E., Nonino, M., Buonanno, R., Caputo, F.,Castellani, M., Dall’Ora, M., Del Principe, M., Ferraro, I.,Iannicola, G., Piersimoni, A. M., Pulone, L., & Vuerli, C. 2007,ApJ, 663, 1021D’Antona, F., Caloi, V., Montalb´an, J., Ventura, P., & Gratton,R. 2002, A&A, 395, 69D’Antona, F., Caloi, V., & Ventura, P. 2010, MNRAS, 405, 2295D’Cruz, N. L., Dorman, B., Rood, R. T., & O’Connell, R. W.1996, ApJ, 466, 359Decressin, T., Meynet, G., Charbonnel, C., Prantzos, N., &Ekstr¨om, S. 2007, A&A, 464, 1029Di Cecco, A., Zocchi, A., Varri, A. L., Monelli, M., Bertin, G.,Bono, G., Stetson, P. B., Nonino, M., Buonanno, R., Ferraro, I.,Iannicola, G., Kunder, A., & Walker, A. R. 2013, AJ, 145, 103Dorman, B., Rood, R. T., & O’Connell, R. W. 1993, ApJ, 419,596 Edelmann, H., Heber, U., Hagen, H.-J., Lemke, M., Dreizler, S.,Napiwotzki, R., & Engels, D. 2003, A&A, 400, 939Fontaine, G., Green, E. M., Brassard, P., Latour, M., & Chayer,P. 2014, in Astronomical Society of the Pacific ConferenceSeries, Vol. 481, 6th Meeting on Hot Subdwarf Stars andRelated Objects, ed. V. Van Grootel, E. M. Green,G. Fontaine, & S. Charpinet, 83Gratton, R., Sneden, C., & Carretta, E. 2004, ARA&A, 42, 385Green, E. M., Fontaine, G., Hyde, E. A., For, B.-Q., & Chayer, P.2008, in Astronomical Society of the Pacific Conference Series,Vol. 392, Hot Subdwarf Stars and Related Objects, ed.U. Heber, C. S. Jeffery, & R. Napiwotzki, 75Haas, S., Dreizler, S., Heber, U., Jeffery, S., & Werner, K. 1996,A&A, 311, 669Harris, W. E. 1996, AJ, 112, 1487Heber, U. 2009, ARA&A, 47, 211Heber, U., Hirsch, H., Str¨oer, A., O’Toole, S., Haas, S., &Dreizler, S. 2006, Baltic Astronomy, 15, 91Hirsch, H. 2009, PhD thesis, Friedrich-Alexander-Universit¨atErlangen-N¨urnbergHu, H., Tout, C. A., Glebbeek, E., & Dupret, M.-A. 2011, ArXive-printsJohnson, C. B., Green, E. M., Wallace, S., O’Malley, C. J.,Amaya, H., Biddle, L., & Fontaine, G. 2013, ArXiv e-printsJohnson, C. I. & Pilachowski, C. A. 2010, ApJ, 722, 1373Kilkenny, D., Koen, C., O’Donoghue, D., & Stobie, R. S. 1997,MNRAS, 285, 640Lanz, T., Brown, T. M., Sweigart, A. V., Hubeny, I., &Landsman, W. B. 2004, ApJ, 602, 342Lanz, T. & Hubeny, I. 2003, ApJS, 146, 417—. 2007, ApJS, 169, 83Latour, M., Fontaine, G., Brassard, P., Green, E. M., Chayer, P.,& Randall, S. K. 2011, ApJ, 733, 100Maeder, A. & Meynet, G. 2006, A&A, 448, L37Michaud, G., Richer, J., & Richard, O. 2011, A&A, 529, A60+Miller Bertolami, M. M., Althaus, L. G., Unglaub, K., & Weiss,A. 2008, A&A, 491, 253Moehler, S., Dreizler, S., Lanz, T., Bono, G., Sweigart, A. V.,Calamida, A., Monelli, M., & Nonino, M. 2007, A&A, 475, L5Moehler, S., Dreizler, S., Lanz, T., Bono, G., Sweigart, A. V.,Calamida, A., & Nonino, M. 2011, A&A, 526, A136Moehler, S., Sweigart, A. V., Landsman, W. B., & Dreizler, S.2002, A&A, 395, 37Moehler, S., Sweigart, A. V., Landsman, W. B., Hammer, N. J.,& Dreizler, S. 2004, A&A, 415, 313Moni Bidin, C., Villanova, S., Piotto, G., Moehler, S., Cassisi, S.,& Momany, Y. 2012, A&A, 547, A109N´emeth, P., Kawka, A., & Vennes, S. 2012, MNRAS, 427, 2180Pietrinferni, A., Cassisi, S., Salaris, M., & Castelli, F. 2006, ApJ,642, 797Piotto, G., Villanova, S., Bedin, L. R., Gratton, R., Cassisi, S.,Momany, Y., Recio-Blanco, A., Lucatello, S., Anderson, J.,King, I. R., Pietrinferni, A., & Carraro, G. 2005, ApJ, 621, 777Randall, S. K., Calamida, A., & Bono, G. 2009, A&A, 494, 1053Randall, S. K., Calamida, A., Fontaine, G., Bono, G., &Brassard, P. 2011, ApJ, 737, L27Randall, S. K., Calamida, A., Fontaine, G., Green, E. M.,Monelli, M., Alonso, M. L., Catelan, M., Bono, G., Dhillon,V. S., & Marsh, T. R. 2013, in European Physical Journal Webof Conferences, Vol. 43, European Physical Journal Web ofConferences, 4006Saffer, R. A., Bergeron, P., Koester, D., & Liebert, J. 1994, ApJ,432, 351Saio, H. & Jeffery, C. S. 2000, MNRAS, 313, 671Stroeer, A., Heber, U., Lisker, T., Napiwotzki, R., Dreizler, S.,Christlieb, N., & Reimers, D. 2007, A&A, 462, 269Ventura, P., D’Antona, F., Mazzitelli, I., & Gratton, R. 2001,ApJ, 550, L65Villanova, S., Piotto, G., King, I. R., Anderson, J., Bedin, L. R.,Gratton, R. G., Cassisi, S., Momany, Y., Bellini, A., Cool,A. M., Recio-Blanco, A., & Renzini, A. 2007, ApJ, 663, 296Zocchi, A., Bertin, G., & Varri, A. L. 2012, A&A, 539, A6512