Publications of the Korean Astronomical Society pISSN: 1225-153430: 255 ∼ 259, 2015 September eISSN: 2287-6936c©2015. The Korean Astronomical Society. All rights reserved. http://dx.doi.org/10.5303/PKAS.2015.30.2.255

LOW-RESOLUTION SPECTROSCOPIC STUDIES OF GLOBULAR CLUSTERS WITH MULTIPLE POPULATIONS

Dongwook Lim1, Sang-Il Han1, Dong-Goo Roh1,2, and Young-Wook Lee1

1Center for Galaxy Evolution Research and Department of Astronomy, Yonsei University, 50 Yonsei-ro, Seodaemun-gu,Seoul 120-749, Korea

2Korea Astronomy and Space Science Institute, 776 Daedeok-daero, Yuseong, Daejeon 305-348, Korea

E-mail: dwlim@galaxy.yonsei.ac.kr, ywlee2@yonsei.ac.kr

(Received November 30, 2014; Reviced May 31, 2015; Aaccepted June 30, 2015)

ABSTRACT

Recent narrow-band Ca photometry discovered two distinct red giant branch (RGB) populations insome massive globular clusters (GCs) including M22, NGC 1851, and NGC 288. In order to investigatethe differences in light/heavy elements abundances between the two subpopulations, we have performedlow-resolution spectroscopy for stars on the two RGBs in these GCs. We find a significant difference (morethan 4σ) in calcium abundance from the spectroscopic HK′ index for both M22 and NGC 1851. We alsofind a more than 8σ difference in CN band strength between the Ca-strong and Ca-weak subpopulations.For NGC 288, however, we detect the presence of a large difference only in the CN strength. The calciumabundances of the two subpopulations in this GC are identical within errors. We also find interestingdifferences in CN-CH relations among these GCs. While CN and CH indices are correlated in M22, theyshow an anti-correlation in NGC 288. However, NGC 1851 shows no difference in CH between two groupsof stars having different CN strengths. The CN bimodality in these GCs could be explained by pollutionfrom intermediate-mass asymptotic giant branch stars and/or fast-rotating massive stars. For the presenceor absence of calcium bimodality and the differences in CN-CH relations, we suggest these would be bestexplained by how strongly type II supernovae enrichment has contributed to the chemical evolutions ofthese GCs.

Key words: Galaxy: formation — globular clusters: general — globular clusters: individual (M22,NGC 1851, NGC 288) — stars: abundances — techniques: spectroscopic

1. INTRODUCTION

Recent studies of stellar populations in globular clus-ters (GCs) have suggested that most GCs possess twoor more subpopulations. Photometric observations dis-covered, for many GCs, splits from main sequence (MS)to red giant branch (RGB) on the color-magnitude dia-gram (CMD) (e.g., Lee et al., 1999; Siegel et al., 2007).Spectroscopic observations have also found star-to-starvariations in both light and heavy elements abundancesin these GCs (Norris et al., 1996; Da Costa et al.,2009; Carretta et al., 2010; Marino et al., 2011). Abun-dance variations in light elements are suggested to bedue to the chemical pollution and/or enrichment byintermediate-mass asymptotic giant branch (IMAGB)stars (Ventura & D’Antona, 2008) and fast-rotatingmassive stars (FRMSs; Decressin et al., 2007). How-ever, heavy element abundance variations would implythat a later generation of stars was enriched by type IIsupernovae (SNe; Timmes et al., 1995), which would in-dicate a dwarf galaxy connection for the origin of theseGCs. Therefore, searches for these GCs are important asthey can provide additional evidence for building blocks

http://pkas.kas.org

of the Galaxy. In this respect, narrow-band Ca photom-etry, which measures the strength of calcium II H & Klines (Anthony-Twarog et al., 1991), can be a power-ful probe as it can detect even a small spread in calciumabundance. Indeed, recent Ca photometry performed atthe Cerro Tololo Inter-American Observatory (CTIO)has successfully detected splits on the RGB in severalGCs, including M22, NGC 1851, and NGC 288 (J.-W.Lee et al., 2009; Roh et al., 2011). The caveat in thisphotometry, however, is that the adjacent CN band at3883 A can contaminate the measurement if the filtertransmission function intrudes into the CN band (seeLee et al., 2013; Hsyu et al., 2014). In this study, in or-der to investigate the origin of multiple RGBs, we haveperformed low-resolution spectroscopy for stars on thetwo RGBs in M22, NGC 1851, and NGC 288.

2. OBSERVATIONS

Our multi-object spectroscopic observation were per-fomred using the Wide Field Reimaging CCD Camera(WFCCD) mounted on the du Pont 2.5m telescope. Weused the H&K grism with a dispersion of 0.8 A/pixeland a central wavelength of 3700 A. Target stars were

255

256 LIM ET AL.

Figure 1. Spectroscopic target stars identified on the CMDs for M22, NGC 1851, and NGC 288. Blue and red circles indicateselected stars from the bluer and redder RGBs, respectively.

selected from two distinct subpopulations in the (y, hk)CMDs, for M22 and NGC 288, and in the (U , U − I)CMD, for NGC 1851, all obtained from the CTIO 4mtelescope (see Figure 1). In our selection of target stars,a similar number of stars were selected from the two sub-populations. Four or five slit masks, each of which wasmade for stars with similar brightness, were designed us-ing the maskgenx4 code. Typically, each mask includes∼25 slits of 1.2′′ width with slit length longer than 10′′

for obtaining sky spectra. The data reduction was per-formed with IRAF and the modified WFCCD reductionpackage, for which a detailed description can be foundin Prochaska et al. (2006).

In order to compare the strengths of calcium, CN, andCH lines between the two subpopulations, we have mea-sured spectral indices, HK′, S3839, and CH4300 (Lim etal., 2014; Harbeck et al., 2003). The definitions for theseindices are

HK′ = −2.5 logF3916−3985

2F3894−3911 + F3990−4025,

S(3839) = −2.5 logF3861−3884

F3894−3910,

CH4300 = −2.5 logF4285−4315

0.5F4240−4280 + 0.5F4390−4460,

where F3916−3985, for example, is the integrated fluxfrom 3916 to 3985 A. However, spectral indices are af-fected not only by chemical abundance but also by ef-fective temperature (Teff) and surface gravity (log g).Therefore, we calculated δ indices (δHK′, δCN, andδCH) as the difference between the original index foreach star and the least-square fitting line, obtained fromthe full sample, to minimize the effects of temperatureand gravity.

3. RESULTS

Figure 2 shows measured HK′, δHK′, CN, and δCNindices for the RGB stars in M22, NGC 1851, andNGC 288 as functions of magnitude. The differences

in chemical abundance between the two subpopulationsare measured from δHK′ and δCN indices by taking themean values for each subpopulation. In Figure 2, thetwo subpopulations in M22 and NGC 1851 are clearlyseparated by 0.054 and 0.035 in δHK′ index, which are7.5σ and 4.6σ levels, respectively. The mean δHK′

indices of the two subpopulations in NGC 288, how-ever, are almost identical within the standard error, sug-gesting that the difference in δHK′ index is negligible.These results suggest that the separations in M22 andNGC 1851 found in Ca photometry are indeed originat-ing from a difference in calcium abundance, while thereis no evidence for a difference in calcium abundance inNGC 288. In the case of the CN band, all of these threeGCs show clear separations between the two subpop-ulations in δCN index. The differences in δCN indexfor M22, NGC 1851, and NGC 288 are 0.36 (15.9σ),0.42 (12.8σ), and 0.31 (16.6σ), respectively. This alsosuggests that the difference in CN band strength wasmainly responsible for the RGB split in NGC 288.

In addition, we found a systematic difference in theCN-CH correlation among these GCs. In Figure 3, weploted δCN versus δCH of RGB stars in NGC 288,NGC 1851, and M22, respectively, to investigate the cor-relation between CN and CH band strengths. NGC 288shows an anti-correlation between CN and CH indices,which is consistent with the general trend reported inprevious studies. In the case of NGC 1851, we found noapparent relation between the two indices, mostly be-cause the difference in δCH index is negligible betweenthe two subpopulations. On the other hand, M22 ap-parently shows a positive correlation between CN andCH. Note that GCs without a CN-CH anti-correlation(NGC 1851 and M22) are those with calcium abundancevariations. This suggests that the absence of CN-CHanti-correlation is probably due to the effect of SNeenrichment, because only this mechanism can increasethe strengths of every spectral indices, δHK′, δCN, andδCH.

LOW-RESOLUTION SPECTROSCOPIC STUDIES FOR THE GCs WITH MULTIPLE POPULATIONS 257

Figure 2. Measured spectral indices (HK′, δHK′, CN, and δCN) as functions of magnitude for M22, NGC 1851, and NGC 288,where the blue and red circles are stars in bluer and redder RGBs in Figure 1. The δ indices, δHK′ and δCN, are definedas the height of the original indices, HK′ and CN, above the least square lines (black solid lines). The mean value and theerror of the mean (±1σ) for each subpopulation are denoted by the solid and dashed lines, respectively.

4. DISCUSSION

While the light element variations could be explainedby pollution from IMAGB stars and/or FRMSs (De-cressin et al., 2007; Ventura & D’Antona, 2008), theorigin of the two distinct subpopulations with differ-ent heavy element abundances in some GCs, includingM22 and NGC 1851, is not yet fully understood. Twoscenarios are suggested in the literature, all of whichare apparently possible only in a dwarf galaxy environ-ment. One is that the later generation stars in theseGCs were formed in the metal enriched gas from theearlier generations (hereafter self-enrichment scenario;Timmes et al., 1995, see also Joo & Lee, 2013, and ref-erences therein). The other is that these GCs formedthrough the merging of two proto-galactic GCs withslightly different metallicities (hereafter merger scenario;

Carretta et al., 2010; Bekki & Yong, 2012). The redderRGB stars in M22 and NGC 1851 are enhanced bothin heavy and light elements. In the self-enrichment sce-nario, this would suggest that the redder RGB stars wereaffected by SNe enrichment, together with the contami-nation from IMAGB stars and/or FRMSs. The positiveCN-CH correlation observed in M22 is also naturallyexplained in this scenario as the SNe enrichment wouldhave increased both nitrogen (CN) and carbon (CH)abundances. In the case of NGC 1851, the absence of aCN-CH correlation is most likely because the enhance-ment in carbon abundance by SNe was not enough toovercome the depletion of the same element by IMAGBstars or FRMSs. In this scenario, NGC 288 was not af-fected by SNe, but carbon was only depleted by IMAGBor FRMS. This suggests that the systematic differencesin CN and CH correlation among three GCs is caused by

258 LIM ET AL.

Figure 3. The systematic difference in the CN-CH correlation among RGB stars in NGC 288, NGC 1851, and M22. In eachpanel, a dashed line represents the least-square fitting line obtained from the full sample.

how strongly SNe enrichment has contributed to theirchemical enrichment. The origin of this difference prob-ably has something to do with the difference in theinitial mass of the system as more enriched gas fromSNe would be retained in more massive GCs (see, e.g.,Baumgardt et al., 2008). In the case of the merger sce-nario, the presence of two subpopulations differing inboth heavy and light element abundances can be intu-itively explained by a simple merging of two GCs in theproto-dwarf galaxy environment (Carretta et al., 2010;Bekki & Yong, 2012). A more detailed description ofthis study can be found in Lim et al. (2014).

ACKNOWLEDGMENTS

Support for this work was provided by the Na-tional Research Foundation of Korea to the Center forGalaxy Evolution Research and by the Korea Astron-omy and Space Science Institute under the R&D pro-gram (Project No. 2014-1-600-05) supervised by theMinistry of Science, ICT and future Planning.

REFERENCES

Anthony-Twarog, B. J., Twarog, B. A., Laird, J. B., &Payne, D., 1991, Ca II H and K Filter Photometry onthe UVBY System. I - The Standard System, AJ, 101,1902

Baumgardt, H., Kroupa, P., & Parmentier, G., 2008, TheInfluence of Residual Gas Expulsion on the Evolution ofthe Galactic Globular Cluster System and the Origin ofthe Population II halo, MNRAS, 384, 1231

Bekki, K. & Yong, D., 2012, On the Origin of the Stellar Haloand Multiple Stellar Populations in the Globular ClusterNGC 1851, MNRAS, 419, 2063

Carretta, E., Gratton, R. G., & Lucatello, S., et al.,2010, Abundances for a Large Sample of Red Giants inNGC 1851: Hints for a Merger of Two Clusters?, ApJ,722, L1

Da Costa, G. S., Held, E. V., Saviane, I., & Gullieuszik, M.,2009, M22: An [Fe/H] Abundance Range Revealed, ApJ,705, 1481

Decressin, T., Meynet, G., Charbonnel, C., Prantzos, N., &

Ekstrom, S., 2007, Fast Rotating Massive Stars and theOrigin of the Abundance Patterns in Galactic GlobularClusters, A&A, 464, 1029

Harbeck, D., Smith, G. H., & Grebel, E. K., 2003, CN Abun-dance Variations on the Main Sequence of 47 Tucanae, AJ,125, 197

Hsyu, T., Johnson, C. I., Lee, Y.-W., & Rich, R. M.,2014, Light-Element Chemistry and the Double Red Gi-ant Branch in the Galactic Globular Cluster NGC 288,PASP, 126, 597

Joo, S. -J. & Lee, Y. -W., 2013, Star Formation Histories ofGlobular Clusters with Multiple Populations. I. ω CEN,M22, and NGC 1851, ApJ, 762, 36

Lee, J. -W., Kang, Y. -W., Lee, J., & Lee, Y. -W., 2009, En-richment by Supernovae in Globular Clusters with Multi-ple Populations, Nature, 462, 480

Lee, Y. -W., Han, S. -I., & Joo, S. -J., et al., 2013, TwoDistinct Red Giant Branch Populations in the GlobularCluster NGC 2419 as Tracers of a Merger Event in theMilky Way, ApJ, 778, L13

Lee, Y. -W., Joo, J. -M., & Sohn, Y. -J., et al., 1999, MultipleStellar Populations in the Globular Cluster ω Centauri astracers of a merger event, Nature, 402, 55

Lim, D., Han, S. -I., & Lee, Y. -W., et al., 2014, Low-Resolution Spectroscopy for the Globular Clusters withMultiple Red Giant Branches: M22, NGC 1851 andNGC 288, ApJS, submitted

Marino, A. F., Milone, A. P., & Piotto, G., et al., 2011,Sodium-Oxygen Anticorrelation and Neutron-capture El-ements in Omega Centauri Stellar Populations, ApJ, 731,64

Norris, J. E., Freeman, K. C., & Mighell, K. J., 1996, TheGiant Branch of omega Centauri. V. The Calcium Abun-dance Distribution, ApJ, 462, 241

Prochaska, J. X., Weiner, B. J., Chen, H. -W., & Mulchaey,J. S., 2006, Probing the Intergalactic Medium-GalaxyConnection toward PKS 0405-123. III. The Galaxy Surveyand Correlations with O VI Absorbers, ApJ, 643, 680

Roh, D. -G., Lee, Y. -W., & Joo, S. -J., et al., 2011, ApJ,Two Distinct Red Giant Branches in the Globular ClusterNGC 288, 733, L45

Siegel, M. H., Dotter, A., & Majewski, S. R., et al., 2007,The ACS Survey of Galactic Globular Clusters: M54 and

LOW-RESOLUTION SPECTROSCOPIC STUDIES FOR THE GCs WITH MULTIPLE POPULATIONS 259

Young Populations in the Sagittarius Dwarf SpheroidalGalaxy, ApJ, 667, L57

Timmes, F. X., Woosley, S. E., & Weaver, T. A., 1995,Galactic chemical evolution: Hydrogen through zinc,ApJS, 98, 617

Ventura, P. & D’Antona, F., 2008, Self-enrichment byAsymptotic Giant Branch Stars in Globular Clusters:Comparison Between Intermediate and High Metallicities,MNRAS, 385, 2034