Multiple stellar populations in Globular Clusters A progress

Multiple stellar populations in Globular Clusters A progress report Raffaele Gratton INAF – Osservatorio Astronomico di Padova ITALY

Variations in the strength of CH and CN bands Noticed since early seventies (Osborn 1971) from DDO photometry and spectroscopy Bimodal distribution along the RGB (Norris & Smith 1980 s) NGC 6752 Variations among MS stars in 47 Tuc (Briley et al. 1994)

Most GCs are not homogeneous for other elements O-Na and Mg-Al anti-correlations Cohen, 1978, Ap. J, 223, 487: Scatter in Na abundances in M 13 and M 3 red giants

Signature of the action of high temperature proton capture fusion Survey of bright giants by the Lick. Texas group (Sneden, Kraft, Langer, Ivans et al) Ivans et al. (2001) The Ne. Na cycle which enhances Na, is expected to operate in the same fusion zones in which the ON part of the CNO cycle is fully operative (Denisenkov & Denisenkova 1990; Langer et al. 1993). CNO: T > 20 MK Ne. Na: T~ 35 MK Mg. Al: T> 50 MK 24 Mg depletion: T~70 MK

Li, C, N, O, Na, Al, Mg in unevolved cluster stars NGC 6397 NGC 6752 47 Tuc • negligible convective envelopes • the Al-Mg anticorrelation in TO stars in NGC 6752 requires too high core temperatures Gratton et al. (2001), Bonifacio et al. (2003), Carretta et al. (2004) + Ramirez & Cohen (2002) • dominant H-burning cycle is p-p not CNO Internal mixing is ruled out at least TWO stellar generations!! M 71

The whole star, not only the outer convective envelope, is polluted

First part conclusions (2007) • Globular clusters host different generation of stars • Stars of the later generations formed from material expelled by a fraction of the 1 st gen stars (the polluters) likely diluted by gas with pristine composition (Prantzos & Charbonnel 2006) • In most cases, no detectable production of Fe and heavy elements (some exceptions, mainly massive clusters) • The polluters underwent H-burning at high temperature ( Na-O and Mg-Al anti-correlations) • Various candidates: – Massive AGB stars (M>4 Mo): Ventura et al. 2001 – Fast rotating massive stars (during their main sequence phase): Decressin et al. 2007 – Massive binaries: De Mink et al. 2009 – The combination of these + interacting stars in dense environments (Elmegreen 2017) – Novae: Smith & Kraft 1996, Maccarone et al. 2012

Carretta et al. extensive survey (2004 -15) NGC 483 3 NGC 609 3 NGC 362

Apogee (Meszaros et al. 2015)

Other GCs with spectroscopic evidences of multiple populations NGC 1261: Filler et al. 2012 n n n NGC 2419: Cohen & Kirby 2012 NGC 5053: Boberg et al. 2015 NGC 5634: Carretta et al. 2017 NGC 5694: Mucciarelli et al. 2013 NGC 5824: Da Costa et al. 2014 NGC 5897: Koch & Mc. William 2014 NGC 6266: Yong et al. 2014; Milone 2015 (phot) NGC 6362: Dalessandro et al. 2014 (phot) NGC 6366: Johnson et al. 2013 NGC 6712: Yong et al. 2008 NGC 6723: Gratton et al. 2015

Have all GCs multiple populations? Red: Na-O anticorrelation Crosses: no Na-O anticorrelation Empty: not yet studied Different symbols are for GCs in the MW, LMC or DSph’s Rup 10 6 Pal 3 New definition of GC! Pal 14 NGC 6366

GCs without evidences of multiple populations q q q Pal 12: Cohen et al. 2004 Ter 7: Sbordone et al. 2004 Pal 3: Koch et al. 2009 Pal 14: Caliskan et al. 2012 Rup 106: Villanova et al. 2013 Ter 8: Carretta et al. 2014 They are all small objects

[Fe/H] spread and Mv There seems to be a correlation between [Fe/H] spread and mass: however the correlation is possibly not one-to -one: Ter 5 (Massari et al. 2014) and NGC 5824 (Da Costa et al. 2014) are not very luminous GCs Terzan 5 ω Cen NGC 5824 Garching, October 11, 2012

The extent of the O-Na anticorrelation depends on Mv and concentration

O-Na anticorrelation and HB (Recio-Blanco et al. 2006 , Carretta et al. 2010) ` `

Na-O anticorrelation He HB n n D’Antona et al. 2005: Na-rich stars should be richer in He He-rich stars evolve faster if same mass loss evolved He-rich stars have lower mass they are bluer when on the HB Ex. NGC 2808

Multiple populations in GCs: NGC 2808 (MV=-9. 4): broad band MS colours ΔHe Piotto et al. 2002 Piotto et al. 2007, Ap. JL 661, L 53

Multiple populations in GCs: NGC 2808 (MV=-9. 4): broad band MS colours ΔHe E: Y~0. 38 P: Y~0. 24 I: Y~0. 28 Piotto et al. 2002 Piotto et al. 2007, Ap. JL 661, L 53 Broad band colours (as V-I) are sensitive to He Na-O anticorrelation

Conclusion (2012) n n n n The presence of multiple populations is the rule for massive GCs They are not present in low mass GCs Mv is not the mass at the epoch of formation of the GCs: this was likely larger than the present mass The limiting mass for the onset of the phenomenon is not well clear, but it is at least ~2 x 105 Mo Other parameters might play a role (metallicity, concentration, epoch of formation) He abundance variations may be present There is a strong correlation with the second parameter effect on the HB

Multiple populations from HST photometry (Milone, Piotto et al. ) From Dotter et al. 2014 Near UV bands are sensitive to N Na-O anticorrelation

Multiple populations from HST photometry (Milone, Piotto et al. ) n n Possibility to derive simultaneously information on Na-O anticorrelation and ΔHe if accurate (HST) photometry is available over a wide spectral range Milone et al. (2012)

HST legacy program: 58 GCs (Piotto et al. 2015)

Definition of MPs (Milone et al. 2017)

General trends (Milone et al. 2017)

Results (2017) Multiple populations (variations of N, likely correlated with Na) are present in all GCs n They can often be separated in a few discrete populations n The separation between MP sequences are function of metallicity and Mv n The fraction of FG stars decreases with Mv n Large variations in He are rare: in most GCs the most He-rich stars have Y~0. 3 or less n

Concerns: Fornax and WLM Stellar debris is only a small fraction of the total stellar mass for a normal IMF n The missing 90 -95% of the FG stars had to escape n These escaping stars presumably comprise the halo of the MW (Prantzos & Charbonnel 2006) n BUT: no evidence of these stars in dwarf galaxies Fornax and WLM (Larsen et al. 2012, 2014; Elmegreen et al. 2012) n

Concerns: YMC n n Idea: observe very massive clusters when multiple generations are occurring Various studies by Bastian and co-workers n n n No evidence of on-going star formation in YMO (Bastian et al. 2013; Cabrera-Ziri et al. 2014) A number of exposed local YMC exist (M>=106 M⊙, age<=15 Myr) even high-mass clusters are able to clear any natal gas within them within a few Myr after formation (Longmore et al. 2014) the clusters in M 83 are all exposed (no longer embedded) by <4 Myr (Hollyhead et al. 2015) Difficulty to reproduce He, Na and O production with a single polluter (Bastian et al. 2015) Conclusion: no satisfactory mechanism known (see also Renzini et al. 2015)

The EMSTO quarrel Many populous clusters in the LMS/SMC with age ~1 Gyr show an extended MS TO that can be interpreted as a spread in age of several hundreds Myr (Bertelli et al. 2003) possible evidence of MPs (Mackey et al. 2008) In some cases dual red clump (Girardi et al. 2009) Bertelli et al. 2003 Keller et al. 2011

Controversy n Other effects may mimic an age spread: Binaries (Bertelli et al. 2003) n Spread in stellar rotation (Bastian & De Mink 2009; D’Antona et al. 2015) n Interacting binaries (Yang et al. 2011) n Variability (Salinas et al. 2016) n n Age spread still controversial; see e. g: Pros’: Milone et al. (2017) n Cons’: Bastian et al. (2017) n

Polluters n The polluters for Intermediate stars: Is active in all GCs with M>2 x 105 Mo; the fraction of stars is 50 -75% n Produce moderately extended Na-O anticorrelation (Na enhancement by ~0. 6 -0. 8 dex, O depletion by ~0. 4 -0. 6 dex) n In some clusters produce a moderate Mg-Al anticorrelation n Produce a small He enhancement (ΔY≤ 0. 04) n Need to be a Li producer (log n(Li)~2. 5) n this pattern well matches massive AGB stars undergoing HBB (lifetime 30 -100 Myr)

Polluters n The polluters for Extreme stars: n n n Is active only in most massive GCs (M>106 Mo); the fraction of stars is at most ~25% of total Produce stars with Na/Al enhanced by ~0. 4 -1. 0 dex and O/Mg depleted by ~0. 5 -1. 0 dex In extreme cases even an Ar-K cycle (Cohen & Kirby 2012) Produce a large He enhancement (ΔY~0. 10) Do not need to be a Li producer He production (but not Mg-Al and Ar-K cycle) matches fast rotating massive stars (Decressin et al. lifetime a few Myr) Alternatively, massive AGB stars may also be considered (D’Antona et al. 2016)

D’Antona et al. 2016 scenario (NGC 2808)

D’Antona et al. 2016 scenario

n n n n The Elmegreen (2017) scenario Interactions between massive stars and massive star binaries can shred stellar envelopes Massive stars should also scatter low-mass stars out of the cluster Agreement with observations for a cloud core mass of ~4 x 106 Mo and density of 2 x 106 cm-3 Mechanism efficient at early epochs when pressure was higher Ratio of gas consumption over lifetime of massive stars should be low (~10%) Efficiency of scattering low-mass stars per unit dynamical time should be large (a few per cent) Velocity dispersion of embedded GC comparable to gas dispersions of galaxies stellar ejection of multi-star