ABUNDANCE VARIATIONS IN GLOBULAR CLUSTERS:

from light to heavy elements

Collaborators: Raffaele Gratton, Sara Lucatello (INAF Padova), Angela Bragaglia, Eugenio Carretta (INAF Bologna), Anna F. Marino (Max Planck Institute, Heidelberg), Chris Sneden (The University of Texas at Austin), Simon W. Campbell, Maria Lugaro, John Lattanzio, George Angelou (Monash University) Thomas Masseron (Universite de Brussels) Inese Ivans (University of Utah) Marco Pignatari (University of Basel)

Valentina D’Orazi

Dept. of Physics and Astronomy, Macquarie UniversityMonash Centre for Astrophysics, Monash University

“A Simple Stellar Population is defined as an assembly of coeval, initially chemically homogeneous, single stars ..

Four main parameters are required to describe a SSP, namely its age, composition (Y,Z), and the initial mass

function

..In nature the best example of SSPs are stellar clusters” (Renzini & Buzzoni 1986).

Globular Clusters for many years considered as ideal

benchmarks for studying stellar evolution

& synthesis population models

THIS TRADITIONAL PERSPECTIVE IS NOW PROVEN TO BE TOO SIMPLISTIC….

Globular Clusters ARE NOT Simple Stellar Populations Photometry

Piotto et al. (2007)

ω Cen

Lee et al. (1999)

Pancino et al. (2000)

NGC 2808

NGC 1851

Bedin et al. (2004)

Milone et al. (2008)

Spectroscopy

Lick-Texas group (from Ivans et al. 2001)

Since ’70s anti-correlations between light elements

(C, N, O, Na, Mg, Al) the abundances of C, O, Mg are depleted where those of N, Na, Al are enhanced

Cohen (1978); Peterson (1980); Norris (1981)

Marino et al.(2008, 2009)

M4

M22

A PREVIOUS GENERATION of stars which synthesized in their interiors p-capture elements are RESPONSIBLE for these

chemical signatures in GC stars

HOT hydrogen burning, where the ON, NeNa, and MgAl chains are operating - the ON reduces O, the NeNa increases Na

(T ~ 30 million K), while the MgAl produces Al (T~65 million K)

IM-AGB stars (4 – 8 M) experiencing Hot Bottom Burning

(e.g. Ventura & D’Antona 2009)

Winds of Fast Rotating Massive Stars

(e.g. Decressin et al. 2007)

Still debated……

OUR SURVEYNa-O anticorrelation and HB in 19 GCs

FLAMES@VLT (Giraffe+UVES),>100 hrs

Carretta et

al. (2009a)

P=primordial FGI=Intermediate SGE=Extreme SG

Fe-peak, Na, O, Mg, Al abundances derived for ~1200 stars

All the GCs show the Na-O anti-correlation

the second generation is always PRESENT

The shape of Na-O distribution changes from

cluster to cluster POLLUTER’S MASS is

varying: this change is driven by both Luminosity

(~mass) & Metallicity

The Mg-Al anticorrelation

This can also be explained through high-temperature (T~ 65 million K) proton capture nucleosynthesis, via the MgAl chain (Mg depleted, Al enhanced).

Kraft et al. 1997

Yong et al. 2003 (NGC 6752)

Carretta et al. (2009b)

The Mg-Al anticorrelation is not present in ALL the GCs (

POLLUTER’S MASS)

Why Lithium ??

Among the light elements Li has a special role.

Li is produced in Big Bang nucleosynthesis

It is enriched during the galaxy evolution,and

destroyed in the stellar interior(Tburn starting @ 2.5 MK)

WMAP logn(Li)=2.72 ± 0.06 (Cyburt et al. 2008) Li-plateau logn(Li)=2.1 – 2.3 (halo stars)

..solution still far…

Lithium & p-capture elements (1)

Li has a fundamental role thanks to the fact that it is very easily destroyed in stellar interiors:

It is expected that at CNO/NeNa cycle temperatures NO Li is left

Polluting material (ejected from the first generation stars) has Li ~ 0

Na-poor, O/Li-rich stars are the FIRST POPULATION born in the cluster share the same chemical composition of field stars

Na-rich, O/Li-poor stars, i.e. the SECOND GENERATION, formed from gas progressively enriched by the ejecta of first population

IF PRISTINE AND POLLUTING MATERIAL ARE MIXED IN DIFFERENT PROPORTIONS THEN LITHIUM AND OXYGEN ARE EXPECTED TO BE

CORRELATED, AND LITHIUM AND SODIUM ANTICORRELATED

Lithium & p-capture elements (2)

While Fast Rotating Massive Stars can only destroy Li,

IM-AGB stars can also produce it

THE CAMERON-FOWLER MECHANISM

(“7Be transport” mechanism, Cameron & Fowler 1971)

Any production of Lithium tends to erase the

Li–O(Na) (anti–)correlation

WHICH ARE THE POLLUTERS ??

History (1): NGC 6752Pasquini et al. (2005)

[Fe/H] = –1.5 (Carretta et al. 2009c)(m-M)=13.13

Basing on 9 TO stars, Pasquini et al. found a Li depletion up to ~ 1 dex below the Spite plateau

Li-Na anticorrelation

Li-N anticorrelation

Li-O correlation

History (2): NGC 6397

[Fe/H] = – 1.99 (Carretta et al. 2009c)

(m-M) = 12.50

Bonifacio et al. (2002) on only 4 stars:

NO Li variation

Lind et al. (2009) the first large sample of Li, Na determinations in TO and early SGB stars, i.e. ~100 stars

“a limited number of Na-enhanced and Li-deficient stars strongly contribute to forming a significant anti-correlation between the abundances of Na and Li.” (Lind et al. 2009)

(1) Li, Na, O in GC dwarfs: the case of 47 Tuc

~100 TO starsFLAMES Giraffe spectra

HR15n (Li I)HR19A (Na I @8183-8194 Å

O I @7771-7775Å)

Na-O distributions in dwarfs and giants are identical evolutionary effects acting during the RGB phase (D’Antona for M13) can be ruled out –at least for this cluster-

The largest database of this kind available so far

Red solid line dilution model(**)

(**) [X]=log[(1-dil) x10[XO] + dil x10[Xp]],

where [XO] and [Xp] are logarithmic

abundances of

original and processed material

Prantzos & Charbonnel (2006)

D ’Orazi et al. (2010a)

Li-O are weakly positively correlated within a large scatter Stars with low O have also low Li content, but at higher oxygen, Li can assume all values, ranging from 1.54±0.06 to 2.78±0.08

Li-Na show NO anti-correlation

A simple dilution model fails in reproducing both Li-O and Li-Na distributions (maybe this model is just the upper envelope)

Different behaviour with respect to NGC 6397 ([Fe/H]=-1.99)

The scatter is reminiscent (a Pop. II analog?) of what found in the OC M67 (e.g. Randich et al. 2000) and in general in cool (Teff~5800K) disk stars (Ryan et al. 2001)

(2) Li-Na anticorrelation in NGC 6121 (M 4)

FLAMES UVES (R~50000, setup 580) spectra for ~90 giant stars from Marino et al. (2008)

Stellar parameters + abundances for Fe, Na, O Marino et al. (2008)

Na-O anticorrelation

Red stars: V > Vbump

A depletion of Li of a factor of ~20 is predicted at 1 DUP

(at the bump luminosity Li 0, thermohaline mixing, see Charbonnel & Zahn 2007)

D ’Orazi & Marino (2010)

NO Li-Na

anticorrelation

Na-rich and

Na-poor stars

have the SAME

Li content, BUT

the scatter is larger

for the first group

V > Vbump

1st generation 2nd generation

Any Lithium production tends to erase the Li-Na anticorrelation

WHICH ARE THE MODEL PREDICTIONS ?

D ’Antona & Ventura (2010)

M 4

•Low mass AGB polluters (~4M) moderate Li production (≈Plateau value)

•The “vertical” Na-O anti-correlation in M4 confirms very low depletion of O

•No Al variations (no MgAl cycle)

From light to heavy elements

Barium abundances in 15 Globular ClustersFrom Giraffe spectra

INTERMEDIATE AGB STARS (4 – 8 M) AS CANDIDATE POLLUTERS

IS THERE ALSO THE CONTRIBUTION OF

LOW MASS AGBs (s-process variation and CNO NOT constant) ??

1. Barium vs. [Na/Fe]/[O/Fe]D’Orazi et al. (2010b)

Barium and Na-O anticorrelation

There is NO segregation along the NaO anticorrelation between Ba-rich and Ba-poor stars

Barium stars

Quite LARGE UNCERTAINTIES but STATISTICS

5 Ba stars on a total of 1205 ~0.4 %

FIELD STARS ~2 %

[Fe/H] Mv

47 TUC -0.76 -9.42

NGC 288 -1.32 -6.74

NGC 6254 -1.57 -7.48

NGC 6397 -1.99 -6.63

NGC 6752 -1.55 -7.73

4 of 5 Ba-stars are P: between P stars, the fraction of Ba stars reaches ~2% CLUSTER ENVIRONMENT

What’s next?? (1)

Run of Lithium with p-capture reaction elements in:

NGC 6218, NGC 3201, NGC 5904

ESO P87 30 h with FLAMES@VLT (PI VD) ~100 RGB stars per GC

Preliminary results in NGC 6218 indicate an M4-like behaviour:

Li is CONSTANT between First and Second Generation stars (D’Orazi+ 2012, in prep.)

``The chemical composition of nearby young

clusters/associations

Collaborators: Silvano Desidera Raffaele Gratton (INAF Padova) Katia Biazzo , Elvira Covino (INAF Napoli), Sergei M. Andrievsky (Odessa National Observatory/GEPI Paris) Gayandhi De Silva (AAO) Claudio Melo (ESO Chile) Sofia Randich (INAF Arcetri) Carlos Torres (Laboratorio Nacional/MCT, Brazil)

D’Orazi et al. (2009)

Anticorrelation between [Ba/Fe] ratio and cluster age

Galactic chemical evolution model only assuming a higher Ba yield from low-mass AGB stars (i.e. 1-1.5 Msun) than that previously predicted(confirmed from other s-process elements, Maiorca + 2011)

Is the enhancement in the Ba content shared by the majority (totality?) of young clusters

populating the solar neighbourhood?

Are the nearby young clusters characterised by a unique [Ba/Fe] value?

Do they show any intrinsic internal dispersion?

Do the other s-process elements follow the enhancement in Ba?

While a chemical evolution model with enhanced Ba production can account for the observed raising trend up to ~500 Myr, it dramatically fails in reproducing the

young stellar clusters.

S-process elements in AB Doradus (~70 Myr), Carina-Near (~200 Myr) and Ursa Major (~500 Myr)

D’Orazi et al. 2012

We find that while the s-process elements Y, Zr, La, and Ce exhibit solar ratios in all three associations, Ba is over-abundant by 0.2 dex.

Run of s-process elements with OC ages

None of the current models can account for such a trend in Ba, without bearing similar enhancement in other s-process elements

Chromospheric effects????

(i) CaII H&K chromospheric emission, (logR_HK)(ii) coronal emission (X-ray luminosity)(iii) rotational velocity (vsini)

Due to the presence of a hot chromosphere, one would expect a T(tau) function less steep compared to that of old stars (the outer atmosphere should be heated at a certain extent by the upper chromosphere levels).

NLTE effects of the 5853 Å line

Over-ionisation

Although no correlation between [Ba/Fe] and several activity indicators seems to be present, we conclude that different effects

could be at work which may (directly or indirectly) be related to the presence of

hot stellar chromospheres.

need for a large, homogeneous investigation of s-process abundances in

clusters younger than the Hyades to draw final conclusions on this issue and provide observational constraints to new

theoretical models.

..Stay tuned…