CEMP stars and AGB models

1
CEMP stars and AGB models
Sara Lucatello INAF-Osservatorio di Padova and
Excellence Cluster Universe, Munich

• Successes, failures and challenges

In collaboration with J. Johnson, T. Masseron,
R. Gratton, E. Carretta, F. Herwig, M. Pignatari
2
Low metallicity surprise

• Carbon Stars (objects rich in C, where the C2
  bands are observed, C/Ogt1) have been recognized
  as a class of astronomical object for more than a
  century.
• They make up for few percent of Pop I and II
  stars
• They have been distinguished between intrisic,
  i.e. C from internal origin, AGB star, and
  extrinsic (classical CH and Ba stars) where C is
  accreted from a now faint and unobservable
  companion.
• In the past 10 years the past and ongoing
  searches for extremely metal poor stars have been
  finding an unexpected high number of C rich stars
  among low metallicity stars
• C-rich (C/Fegt1), metal poor (Fe/Hlt-2.5) are
  called CEMP stars

3
CEMP incidence
HERES over 350 objects at high resolution
R20,000 SDSS sample 5320 lower resolution
(R2,000)

• Fractions of CEMP varies from 9 (Frebel et al
  2007) to 25 (Marsteller et al 2005) depending
  from the authors
• Agreement that it is much higher than at solar
  metallicity (2-3)

4
CEMP with s enrichment

• About 70-80 of CEMP (about 16 of all EMPs) have
  s-process overabundance
• Radial velocity monitoring of CEMP-s are
  consistent with them being all members of
  binaries
• So like classical CH stars they owe their
  composition to mass transfer from a massive
  (1.5-4 Msun)
  companion now a faint white
  
  dwarf
• Excellent chance to study
  AGB
  nucleosynthesis at
  low
  metallicity

5
Model fitting
Models do a good job in general, But there are 2
main discrepancies -N (probably solvable
with CBP) -Eu overabundance
Models Gallino Cristallo Average of CEMP-s
stars 30 objects, pretty robust
6
CEMP-sr possible origins

• At least 50 of CEMP-s has Eu overabundance with
  respect to what expected from s-process
• Stars labelled as CEMP-sr
• Star belongs to a binary which formed from
  r-enriched gas, s-process from AGB
• Binary with massive AGB which explodes as type
  1.5SNe (Zjlstra 2004, Wanajo 2005)
• Binary system with AGB and AIC (e.g. Wanajo 2005)
• AGB transfer only (Johnson Bolte 2004)
• None of the solution proposed so far can account
  for all the observed characheristics (incidence,
  chemical composition, binarity membership etc)

7
Low metallicity AGB and Fluorine

• Multiple sites for 19F production have been
  proposed (WR, SN and AGB), non negligible AGB
  contribution needed in chem. Ev. Models (Renda et
  al. 2004).
• Most models predict that low mass, low
  metallicity AGB stars (1.5-3 Msun) produce a
  large amount of F
• CEMP-s should be highly enhanced in F
• Schuler et al (2007) measured an enormous F
  abundance of F/Fe2.9 dex in log e(F)5.0 dex
  for HE13050132
• Lugaro et al.(2008) compared this result to model
  predictions and concluded that such a F abundance
  should be highly unusual
• We (meJohnson, Masseron, Plez, Pignatari and
  Herwig) obtained CRIRES _at_VLT (R50,000) IR K band
  spectra for 10 more CEMP stars. Large spectral
  coverage so we can measure C,N,O, F and C
  isotopic ratios.

8
AGB and Fluorine
Lugaro et al 2004
Cristallo et al. 2006

• Use CN enhanced model atmospheres by B. Plez
• Measure F abundance from 3 HF lines
• F abundances obtained are much smaller than in
  Schuler et al
• Mostly we can only measure upper limits because
  the HF lines very Teff sensitive
• Comparison to models shows that the F content in
  lower than expected

• Simon Schuler has recently re-determined Fe
  abundance from optical spectrum of HE13050132
  finding a value much higher (1 dex) than assumed
  originally. This would decrease F/Fe but also
  log e (F) as a more metal rich model atmosphere
  would be appropriate.

9
New and unexpected findings

• Are Vrad monitoring really crucial? YES
• CEMP rich in s (including those with
  r-overabundance CEMP-r/s) all in binaries
• Period accurately determined for not many CEMP
• A lot of observations required, relatively easy
  to spot Vrad variations but P determination is
  much more time consuming
• But..It pays off!

10
The story so far

• S-process models very successful in general
• Still several critical issues to solve and/or
  address
• -problems in reproducing some of the
  observed analysis
• -need to take into account effect of close
  companion
• Lots of data will be available in the next few
  years, creating a solid benchmark for model
  testing

We need you!
11
CEMP stars kinds
CEMP-r
130 C-rich stars (95 from HERES, 35 from
literature)
CEMP-no Ba/Felt0 2 CEMP-r 20-30 CEMP are
CEMP-no CEMP-s Ba/Eugt0.5 CEMP-sr 0.0ltBa/Eult
0.5 Over half CEMP-s have Eu overabundance with
respect to what is expected from the pure
s-process
12
The metallicity distribution
CEMP-s Ba/Eugt0.5 CEMP-rs 0.0ltBa/Eult0.5 CEMP
-no Ba/Felt0 CEMP-s and CEMP-rs span same
evolutionary states, have identical metallicity
distributions, have similar binarity fractions
and period distributions
13
CEMP-no possible origins

• Self enrichment (Fujimoto case II) but
  metallicity too high
• Mass transfer from low mass early AGB star before
  s-process
• Born from C enriched material (e.g Ryan et al
  2006)
• Radial velocity monitoring crucial

14
CEMP kinds
94 C-rich stars CEMP-s Ba/Eugt0.5 CEMP-rs 0.0lt
Ba/Eult0.5 CEMP-no Ba/Felt0
15
How do models work?

• See offsets

16
(No Transcript)
17
Successes (and not..)

• Fits lots of elements
• N but CBB
• EugtCEMP-r/s are they really different?

18
(No Transcript)
19
(No Transcript)
20
(No Transcript)
21
HERES and SDSS sample
22
(No Transcript)
23
(No Transcript)
24
(No Transcript)
25
(No Transcript)
26
(No Transcript)
27
New and unexpected findings

• Are Vrad monitoring really crucial? YES
• CEMP rich in s (including those with
  r-overabundance CEMP-r/s) all in binaries
• Period accurately determined for 9 CEMP
• A lot of observations required, relatively easy
  to spot Vrad variations but P determination is
  much more time consuming
• But..It pays off!