“Analysis and interpretation of stellar spectra and

nucleosynthesis processes in evolved stars”

D. A. García-Hernández (IAC Support Astronomer)

Instituto de Astrofísica de Canarias, Seminar presentation of IAC postdocs, December 13 2007

Main lines of research

• Chemical abundances of AGB stars and the role of AGB stars in the Early Solar System composition

• Physical study of transition objects between the AGB phase and the Planetary Nebula stage and the spectral analysis (e.g. by using ISO and Spitzer) of their circumstellar dust shells

• CNO isotopic abundances of hydrogen-deficient carbon stars (R Coronae Borealis and HdC stars)

Stellar evolution: AGB stars

Asymptotic Giant Branch: late stage of evolution of low- to intermediate-mass stars (1 M 8 M)

TP phase: strong mass loss enriches the ISM with radionuclides and circumstellar dust grains!

AGB internal structure

AGB stellar nucleosynthesis• Thermal Pulsing phase 12C production, s-element

production (Rb, Zr, Sr, Nd, Ba, Tc, etc.)• 3rd dredge-up very efficient in AGB stars; C/O ratio

increases• Stars eventually turn C-rich and s-process rich

following the M-, MS-, S-, SC-, C-type sequence unless…

• Hot Bottom Burning (if M > 45 M)• When Tbce 2.107 K 12C 13C, 14N (CN-cycle) and

HBB prevents the carbon enrichment (stars remain O-rich)

• 26Al, 7Li production, low 12C/13C & high 17O/16O ratios (Mazzitelli et al. 99; Karakas & Lattanzio 03 )

The s-process in AGB stars

• Free neutrons to form heavier elements (s-elements such as Rb, Zr, Sr, etc.) can be released by 13C(,n) 16O or by 22Ne (,n)25Mg reactions (Busso et al. 99)

• 13C operates during the interpulse period. It is more efficient in 13 M AGB stars. All previous observations of AGB stars are consistent with 13C!

• 22Ne is expected to be efficient in the convective thermal pulse at higher T and Nn. It should become strongly activated in more massive AGB stars (M>45 M). This prediction has never been confirmed by the observations!

The 22Ne neutron source• The operation of the 22Ne neutron source favors the

production of the stable isotope 87Rb (also of 60Fe, 41Ca, 96Zr, 25Mg, 26Mg, etc.) because of the operation of a branching in the s-process path at 85Kr (Beer & Macklin 89)

[Rb/Zr] is a powerful neutron density (and mass) indicator in AGBs!

87Rb is a radioactive isotope (half-life time of ~48.8 Gyr ) and it is frequently used to date moon rocks and meteorites

Massive Galactic O-rich AGBs

• Where are they in our Galaxy?• Theoretical models predicts Rb/Zr in massive

(M>45 M) O-rich AGB stars

• Galactic candidates: OH/IR stars (L, C/O<1, Long Period Variables). Expected to be massive O-rich stars in the final stages of their AGB evolution

Optical observations very difficult due to strong mass-loss (~ 104 106 M

/yr) their chemical composition (Rb, Zr, etc.) is unknown!

Very red stars! Extremely variable stars!

Discovery of Rb-rich AGB stars

• We observed ~100 OH/IR stars in the optical range during 4 observational campaings in La Palma (Spain) and La Silla (Chile)

• We obtained good high-resolution optical echelle spectra for half of the sample

• The other ~50 stars were completely invisible!• We found that these stars are Rb-rich but Zr-poor

22Ne confirmation as predicted theoretically 40 years ago!

• We confirmed for the first time that the [Rb/Zr] ratio can be used as a mass indicator in AGB stars!

Rb-rich AGB stars

• OH/IR stars strong mass loss source of dust • Li-rich HBB 26Al, 13C and 17O producers• Rb-rich but Zr-poor 22Ne important source of 87Rb,

60Fe, 41Ca (but also of 96Zr, 25Mg, 26Mg, etc.) • The more extreme stars are not predicted by the

current models which do not consider the higher mass stars neither the strong mass loss!

• These stars, if present at the ESS, are more important at the early stages a highly variable Rb/Sr ratio and 87Rb/87Sr ages should be taken with caution

(García-Hernández et al. 06, 07)

ESS radionuclides inventory

• SN scenario may explain 60Fe but not other radionuclides such as 26Al, 41Ca, 107Pd

• Low-mass AGBs reproduce the other radionuclides but do not explain 60Fe. 22Ne is needed! (Wasserburg et al.06)

• Both models do not completely explain many of the radionuclide ESS concentrations. In particular, they cannot explain the 87Rb anomalies detected in CAIs

• We cannot discard SN and low-mass AGBs in the ESS, but massive AGBs probably also played an important role, as evidenced by important Rb/Sr variations in CAIs (Podosek et al. 91; McKeegan and Davis 03) !

CAIs evidence

• 87Rb anomalies are present in CAIs (as deduced from 87Sr/86Sr variations) (e.g. Podosek et al. 91)

• 41Ca and 26Al (also 25Mg, 26Mg) also present • CAIs display important 60Fe concentrations and it

has been found that 60Fe excesses are correlated with 96Zr 22Ne! (Quitté et al. 07)

• It is a mere coincidence that CAIs show all the

chemical anomalies expected in massive AGB stars? New AGB stellar nucleosynthesismodels explain these anomalies,giving a self-consistent solution to theESS radionuclides (Trigo-Rodríguez et al. 08, submitted to MAPS)

CS dust shells AGB-PN

• The dust sequence from AGB to PNe as seen by ISO

- A massive O-rich star

- A very low-mass O-rich star

• Spitzer/IRS survey of heavily obscured PN precursors

- Characterization of the IR spectral properties

- Study of the total obscuration phase

O-rich transition sources

O-rich PN precursors where the transition from amorphous to crystalline dust structure istaking place

The crystalline silicate features are detected in emission inside the amorphous silicate absorptions at 9.7 & 18 m

Young IR PNe

IR PNe as indicated by the detection of nebular emission lines (e.g. [Ar II], [Ar III], [Ne II], [S III])

O-rich PNe showing amorphous and crystalline silicates probably massive PNe

R CrB and HdC stars

• Hydrogen-deficient (~105) luminous stars

• Their origins have remained a puzzle for decades

• Two scenarios have survived theoretical and observational scrutiny:

- DD scenario (merger of a He and C-O WD)

- FF scenario (final pAGB He flash in a CSPN)

• CNO isotopic abundances are a powerful tool for discriminating between the DD and FF scenarios

CNO isotopic abundances

Gemini/PHOENIXnear-IR (R~50,000) spectra of RCB & HdCstars

12C/13C, 14N/15N16O/17O/18O ratioscan be derived from CO & CN transitions

in the K-band (2.3 m)

18O-rich stars

• HdC stars and some RCBs are strongly enriched in 18O (16O/18O~0.24) but C and N are in the form of 12C and 14N, respectively

• This suggests that HdC and RCB stars are related objects and that they probably formed from a WD merger. FF scenario cannot produce 18O-rich stars

• Calculations of the nucleosynthesis achieved during the merger in the DD scenario need to be developed! (merger process in a few days with acretion rates of

150 M yr-1, Clayton et al. 07)

A very red massive O-rich AGB

Visual Red Infrared

Extremely variable stars

Visually Bright! Not found!

Optical echelle spectra“Blue example”

“Red example”

A wide variety of Rb abundances are needed to fit the observations!

First detection of strong Rubidium overabundances in massive AGB stars!

Teff=3000 K

[Rb/Fe]=+0.1

[Rb/Fe]=+0.9

[Rb/Fe]=+1.6

[Rb/Fe]=+2.3

Rb abundances vs. Vexp(OH) Vexp(OH) can be taken as a distance-independent mass indicator in OH/IR stars(e.g. Jiménez-Esteban et al 05)

The more extreme stars are not predicted by the current modelswhich do not consider the higher mass stars neither the strong mass loss![Rb/Zr] >[Rb/Fe] 0.5 in these stars