Why are massive Orich AGB stars in our Galaxy not Sstars

1
Why are massive O-rich AGB stars in our Galaxy
not S-stars?

• D. A. García-Hernández (IDC-ESAC, Madrid, Spain)
• In collaboration with P. García-Lario (IDC-ESAC),
  B. Plez (GRAAL, France), A. Manchado (IAC,
  Spain), F. DAntona (OAR, Italy), J. Lub H.
  Habing (Sterrewacht Leiden, The Netherlands)

Gdansk, June 29 2005
2
AGB stellar nucleosynthesis
1

• Main processes during the Thermal Pulsing phase
  ? 12C, s-element production (Rb, Zr, Ba, Tc, Nd,
  etc.) (3rd dredge-up)
• 3rd dredge-up increases C/O ratio forming M-,
  MS-, S-, SC-, C-type stars
• Hot Bottom Burning (Mgt4 M?)
• When Tbce? 2.107 K ? 12C ? 13C, 14N (CN-cycle)
  and HBB prevents the carbon star formation
• 7Li production and low 12C/13C ratios (Sackman
  Boothroyd 1992 Mazzitelli et al. 1999 )

D.A. García-Hernández
3
Previous works (MCs)
2

• HBB activation in massive AGB stars in the
  Magellanic Clouds (MCs) (e.g. Smith Lambert
  1989 Plez et al. 1993 Smith et al. 1995)
• Characteristics ?7 ? Mbol ? ?6 ( M 4 ? 8 M?
  )
• log
  ?(Li) ? ( 1 ? 4 dex)
• 
  C/O lt 1 ( 0.5 )
• 12C/
  13C ? ( lt 10 )
• ?(s-process) ? (
  s/Fe gt 0.5 dex)

D.A. García-Hernández
4
Previous works (Milky Way)
3

• Li-rich AGBs not so luminous (?6 ? Mbol ? ?3.5)
• ? S-, SC-, C-type stars with low mass (e.g. Abia
  Isern 96, 97, 00 Abia Wallerstein 98) ? not
  yet well understood!!
• HBB models predicts log ?(Li) ? in massive (Mgt4
  M?) O-rich AGB stars (e.g. Mazzitelli et al. 99)
• Galactic candidates OH/IR stars (L?, C/Olt1,
  Long Period Variables)
• ? Optical observations very difficult due to
  strong mass-loss ( 10?4 ? 10?6 M? /yr) ? at
  present ?(Li) and ?(s-process) are unknown!

D.A. García-Hernández
5
Massive Galactic O-rich AGBs
4
Selection of the sample (102 OH/IR stars)

• Long Period Variables (P 300 ? 1000 days)
• Large amplitude variability (8 ? 10 mag in V)
• Late-type stars (gtM5)
• OH maser emission emitters (Vexp(OH) lt 25 km s-1)
• Comparison stars plus 9 C-rich stars (18 objects)
• Members of the galactic disk population with
  strong IR excesses detected by IRAS

D.A. García-Hernández
6
Observations and data reduction
5

• Echelle spectra with UES (WHT, La Palma) and
  CASPEC (ESO 3.6m) in 1996-1997 at R 50,000 (4
  runs)
• Spectral range 5000 ? 9000 Å. We were mainly
  interested in the Li I 6708 Å region
• Exposures times of 10 ? 30 minutes with S/Ngt100
  in the Li I region
• Data reduction with the ECHELLE software package
  in IRAF

D.A. García-Hernández
7
UES echelle spectra
6
Blue example

Red example
D.A. García-Hernández
8
Overview
7

• 25 stars detected in the Li I 6708 Å line
• 32 stars non-detected in the Li I 6708 Å line
• 45 stars too red at 6708 Å or without OPC
• Extremely red spectra dominated by TiO bands
• Absence of molecular bands of ZrO (YO, LaO, etc.)
• From Vdoppler Li I, Ca I, TiO (stellar) are
  formed deeper than K I, Rb I (very probably of
  circumstellar origin)
• Some stars also show H? emission (shock-waves)

ZrO 6474 Å region
Li I 6708 Å region
D.A. García-Hernández
9
Progenitor masses
8

• Period and Vexp(OH) as distance-independent mass
  indicators (e.g. Chen 2001 Jiménez-Esteban 2004)
• Different sources (masses) depending on P and
  Vexp(OH)

IRAS
Galaxy
D.A. García-Hernández
10
Chemical analysis
9

• Classical model atmospheres (HE, LTE, etc.) for
  cool stars (MARCS) and the TURBOSPECTRUM
  spectral synthesis code (Plez et al. 1992)
• TiO, ZrO are included and atomic lines from
  VALD-2
• The whole machinery was tested on the high
  resolution spectrum of the Sun and Arcturus
• Spectral regions of interest (60 Å) Li I 6708
  Å
• 
  ZrO 6474 Å
• 
  K I 7699 Å Rb I 7800 Å

D.A. García-Hernández
11
Overall strategy
10

• Initial range for Teff and log g from the VK
  photometry
• Further constraints on the set of stellar
  parameters (M, Teff, C/O, log g, ?, z, ?(Zr),
  CNO, 12C/13C) using spectral synthesis

Model vs. observations
M2 M? C/O0.5 log g?0.5
?3 km s-1 (z, CNO,12C/13C)?

?
Teff, FWHM Li and Zr (s-elements)
chemical abundances ( log ?(Li), log ?(Zr) )
?2 test
?
?
fixed parameters
D.A. García-Hernández
12
Best fit in the Li I region
11
Teff3000 K, log ?(Li) 1.3 are needed to fit
the observations!
Zoom
D.A. García-Hernández
13
Best fit in the ZrO 6474 Å region
12
Zr/Fe0.0 is needed to fit the
observations! Comparison with a galactic S-star

D.A. García-Hernández
14
IRAS 10436 a galactic S-star
13
Zr/Fe1.0 is needed to fit the observations!
D.A. García-Hernández
15
Li and Zr abundances
14

• Li detected stars show log ?(Li) 1 ? 3 dex
• Li non-detected stars show log ?(Li) lt 0.0 dex
• Uncertainty of ?log ?(Li) 0.4 ? 0.6 dex
  (sensitivity to the atmosphere parameters)
• All stars show upper limits to the Zr abundance
  consistent with no s-element overabundance
  

Zr/Fe lt 0.0 ? 0.25 dex for Teff gt 3000 K
Zr/Fe lt 0.25 ? 0.5 dex for Teff lt 3000 K
D.A. García-Hernández
16
P and Vexp(OH) vs. HBB
15
No clear correlation between log ?(Li) and P,
Vexp(OH) But no Li-rich stars with P lt 400 days
and Vexp(OH) lt 6 km s-1 Half of the stars with
higher P and Vexp(OH) are Li-rich
D.A. García-Hernández
17
Theory vs. observations
16

• - Stars with Plt400 days and Vexp(OH)lt6 km s-1
  are non-HBB stars (3 M? lt M lt 4 M?) ? non Li-rich
  
• - Stars with higher P and Vexp(OH) ? are HBB
  stars (M gt 4 M?) ? Li-rich but why only half of
  them are Li-rich?
• - Both type of stars experience strong mass loss
  and only a few thermal pulses (and less efficient
  because of the high metallicity) ? no s-process
  enhancement
• - The obscured stars must also be HBB stars and
  they represent the more massive AGB stars in the
  Galaxy
• ? This scenario is consistent with the strong IR
  excess detected by IRAS and the HBB and
  nucleosynthesis model predictions!

D.A. García-Hernández
18
Galaxy vs. Magellanic Clouds
17

• Massive O-rich AGB stars in the MCs are S-stars
  and 80 of them are also Li-rich ? HBB stars
• Why are these stars s-element enriched?
• Metallicity effect!
• - Theoretical models predict a higher efficiency
  of the dredge-up in low metallicity environments
  (e.g. Busso et al. 1988 2001 Straniero et al.
  1995 2000 Lugaro et al. 2003 Herwig 2004)
• - Lower metallicity ? lower dust production (van
  Loon 00) ? less efficient mass loss ? longer AGB
  lifetime in the MCs compared to the Galaxy!

D.A. García-Hernández
19
Conclusions
18

• - 25 stars detected in the Li I 6708 Å line, 32
  stars non-detected and 45 stars too red (or no
  OPC)
• The chemical analysis revealed that half of the
  stars with useful optical spectra are Li-enriched
  ? HBB
• - All stars in the sample are considerably
  massive
• (M gt 3 M?) but only the more massive ones (M
  gt 4 M?) experience HBB. The lack of lithium in
  some HBB stars is a consequence of the timescale
  of the Li production phase (104 years)

D.A. García-Hernández
20
Conclusions
19

• As a consequence of the different metallicity,
  massive galactic O-rich AGB stars are not
  s-process enriched in strong contrast to
  Magellanic Cloud massive AGB stars ?
  Observational evidence that the chemical
  evolution during the AGB is strongly modulated by
  the metallicity!!
• Need of extending the analysis to other Galaxies
  in the Local Group with a wide variety of
  metallicities

D.A. García-Hernández
21
Li I 6708 Å region
D.A. García-Hernández
22
ZrO 6474 Å region
D.A. García-Hernández
23
IRAS vs. P and Vexp
D.A. García-Hernández
24
Galactic latitude vs. Vexp(OH)
D.A. García-Hernández
25
Zoom in the Li I region
log ?(Li)1.3 is needed to fit the
observations!
D.A. García-Hernández
26
Other possible hypotheses?

• Are they more massive stars (M gt 4 M?)?
• - This is not consistent with the non-detection
  of Li in any of them!
• Are they lower mass stars (M lt 1.5 M?)?
• - This is not consistent with the lack of
  s-process elements. Other low-mass stars of S-
  and C-type show strong s-process element
  enrichment
• - A early stage as AGB stars is also not
  consistent with the strong IR excess observed by
  IRAS

D.A. García-Hernández
27
Timescale of the Li production
HBB models (Mazzitelli et al. 1999) explain the
lack of lithium in half of the massive O-rich AGB
stars where the HBB is active! The Li-rich phase
is of the order of the interpulse time (104
years)!
D.A. García-Hernández
28
Li production at low metallicity
HBB models (Mazzitelli et al. 1999) explain the
higher detection rate of Li-rich stars in the MCs
because they predict a lower mass limit of only
3.0?3.7 M? (in the LMC) for the HBB activation
and a faster lithium production
D.A. García-Hernández