Accurate Quantitative Spectroscopy of OB Stars H, He III and C IIIIIIV spectrum

1
Accurate Quantitative Spectroscopy of OB StarsH,
He I/II and C II/III/IV spectrum

• María Fernanda Nieva(1,2)
• Norbert Przybilla(1)
• (1) Dr. Remeis Sternwarte Bamberg
  (Germany)
• (2) Observatório Nacional
  (Brazil)

2
Outline

• Introduction/Aims
• Models/Observations
• Our Analysis
• Results/Other Applications
• Summary/Conclusions/Future

3
Introduction

• Carbon
• Early-type OB stars
• Carbon abundance OB-stars

• one of most abundant metals in universe
• created in triple alpha reaction (evolved stars)
• CNO cycle (H ? He massive stars)
• basis of organic chemistry

• luminous/young/massive HR

Sun

• present-day abundances ( ISM)
• stellar evolution
• chemical evolution

Galactic extragalactic
4
(No Transcript)
5
Chemical information (e.g. N/C) tracer for mixing
process (rotation)
Meynet Maeder 2000
Evolution of N/C ratios at surface of rotating
models for vini 300 kms-1
6
Chemical Evolution of the Galaxy Spatial
abundance gradients
Models Alibes et al. (2001)
enriched
OB associations in the MW Disk
primordial
Theory vs. Observations
Daflon et al. 2004
7
Chemical Evolution of the Galaxy Temporal end
point (present-day abundances)
Sun ? sources
Chiappini et al. 2003
? Chemical Evolution Models for the MW
8

• History of e(C) in OB-type stars
  

e(C)log(C/H)12

• Long-standing problem LTE/NLTE
  ( last gt 30 years 1970 )
• (Some C models for NLTE Lennon 1980, Eber
  Butler 1988, Sigut 1996)
• Large non-LTE effects (not well-understood)
• e(C II) strong lines ? e(C II) weak lines
• Strong lines C II 4267, 6578/83 ?
• - very sensitive to non-LTE effects
• - imp. for fast-rotators and low S/N
  (extragalactic)
• e(C II) ? e(C III) no ionization equilibrium

9

• Literature carbon abund. of OB stars
• LTE too low abundances compared to Sun large
  spread
• NLTE still low abundances large spread
• - model atom of Eber Butler 1988
• - Teff from photometric calibrations or Si II/III
  ionization equilibrium (very discrepant !!).

• Solar abundances
• - Old Sun metal-rich compared with vicinity
  (??)

10
Introduction

• Non-LTE e(C) OB III-V in the Solar Vicinity
  (representative sources from the literature)

large scatter
Real from stellar/chemical evolution ?? Artifact
from the analysis ??
sub-solar
Also N, O, etc. Young OB stars metal poor
?? Inconsistent with chemical evolution ?
1 dex !
1.6 dex !
11

• Chemical evolution models
• -instantaneous mixing for rings of 1Kpc no
  scatter in abundance (massive stars)
• -young stars from polluted clouds not metal poor
• Stellar evolution models
• -stars from similar clouds/ young /
  slow-rotators no scatter in abundance

12
Aims of this work

• Accurate Quantitave Spectroscopy
• Solution classical non-LTE problem e(C) OB-stars
• Reliable C II/III/IV model atom calibrated with
  Galactic stars (range OB III-V)
• Special emphasis
• - selection input atomic data
• - accurate atmospheric parameter determination
• Application of model atom to other objects

13
Quantitative Spectroscopy
Our Methodology
Theory synthetic lines
Line fits
Atomic Physics
Model Atoms
Stellar Atmospheres
Chemical Abundances
Observations

• Absolute values (Physics)
• Not relative to other stars or Sun

Spectra

• Non-trivial process (at all !)
• Accuracy depends on all steps of analysis
• Difficult controlling systematic effects

14

• Our Model Calculations
• Model atmospheres LTE hydrostatic
  metal-blanketed in plane parallel geometry
  (ATLAS9 Kurucz 1993)
• NLTE line formation
• Recent versions of
• DETAIL (Giddings, 1981)
• SURFACE (Butler Giddings 1985)
• State-of-the-art model atoms
• C II/III/IV (Nieva Przybilla, 2006, in
  prep.)
• H (Przybilla Butler 2004)
• He I/II (Przybilla 2005)

Hybrid non-LTE approach (Nieva Przybilla
2006, subm. to AA)
radiative transfer statistical equilibrium
for non-LTE calculations
15

• Observations for model calibration
• 6 early-B III-V apparently slow-rotators
• 21500 lt Teff lt 32000 K
• 3.1 lt log g lt 4.3 dex
• randomly distributed in solar vicinity (lt 1 Kpc)
• from associations and field
• Spectra high S/N FEROS data
• Spectra near-IR (FOCES, SUBARU) for 2 stars
• SEDs IUE fluxes Johnson, 2Mass photom.

16
Our Analysis
Step 2 Carbon
Step 1 H, He I/II
No grids ! Detailed analysis for each star !
Fine tunning for atmospheric parameters
Solve the atmosphere

Calibration of model atom
17
Step 2 Carbon
Our Analysis
Step 1 H, He I/II
I
Initial Teff , log g
I
Initial Teff , log g, x, z, e(He), v sin i
No grids ! Detailed analysis for each star !
NLTE C populations
Stellar atmosphere
Synthetic H/He/C profiles
NLTE H/He populations
Comparison with observed spectra
Synthetic H/He profiles
Variables
Comparison with observed spectra
Parameter verification
Teff , log g, x, z, e(C), v sin i
200 levels gt 1300 radiat. transitions gt 5300
collis. transitions
Empirical calibration of C model atom
Parameter verification
M
Modified Teff , log g, x, z, e(He), v sin i
New set of atomic data
M
Modified Teff , log g
MI ?
no
MI ? se(C) min?
yes
no
Final Teff , log g, x, z, e(He), v sin i

• Accurate Stellar Parameters
• Calibrated C model for 1 Star
• Accurate Carbon Abundance

yes
To Step 2
Verify Step 1
18
Step 2 Carbon
Our Analysis
Step 1 H, He I/II
I
Initial Teff , log g
I
Initial Teff , log g, x, z, e(He), v sin i
No grids ! Detailed analysis for each star !
NLTE C populations
Stellar atmosphere
Synthetic H/He/C profiles
NLTE H/He populations
Comparison with observed spectra
Synthetic H/He profiles
Variables
Comparison with observed spectra
Parameter verification
Teff , log g, x, z, e(C), v sin i
200 levels gt 1300 radiat. transitions gt 5300
collis. transitions
Empirical calibration of C model atom
Parameter verification
M
Modified Teff , log g, x, z, e(He), v sin i
New set of atomic data
Same procedure for
M
Modified Teff , log g

• 6 early-B III-V stars
• 21500 lt Teff lt 32000 K
• 3.1 lt log g lt 4.3 dex

MI ?
no
MI ? se(C) min?
yes
no
Final Teff , log g, x, z, e(He), v sin i

• Accurate Stellar Parameters
• Calibrated C model for 1 Star
• Accurate Carbon Abundance

yes
To Step 2
Verify Step 1
19
Near-IR
Simultaneous fits to all mesurable H/He lines
Visual
H Balmer
H Paschen Data FOCES,
Calar Alto, Spain
He I
He I K-Band Data Subaru, Hawaii
He II
Nieva Przybilla, 2006b (subm. to AA)
Data FEROS, ESO
HR 3055
20
Spectral Energy Distributions
parameters H, He fits
Data IUE fluxes Johnson 2Mass photometry
Nieva Przybilla, 2006a, ApJ, 639, L39
21
Carbon quality of spectra/fits
up to 40 lines !!
Spectra FEROS, ESO
4267
Nieva Przybilla, 2006c (in prep.)
Abundances from line profile fits (similar values)
22
New 2 C II emission lines (not reproduced by LTE)
LTE only absorption lines radiation through
cooler medium
Teff
C II NLTE
calibration stars
NLTE coupling radiation field atomic
transitions emission upper levels overpopulated
Nieva Przybilla, 2006a, ApJ, 639, L39
Carbon abundance similar rest of lines
23
Sensitivity of Carbon Abundance to Atmospheric
Parameters
Our solution for HR 3055 Teff 31200200 K
log g 3.950.05 dex x 8 1 km s-1
Typical systematic discrepancies from the
literature in Teff , log g and x gtgt our
uncertainties
C IV up to 1.1 dex! C III up to 0.35 dex! C II
up to -0.35 dex!
4267
6583
Teff -2000 K
6578
log g 0.2 dex
Discrepancies increase with ? atomic data !!
x 5 km s-1
24
Different Sets of Atomic Data oscillator
strengths
FFT Froeser Fischer Tachiev 2004 OP Opacity
Project N02 Nahar 2002
25
Different Sets of Atomic Data s photoionization
OP Opacity Project N02a Nahar 2002a
26
Sensitivity of Carbon to Atomic Data
To photoionization cross-sections from ab-initio
calculations
Lines very sensitive to non-LTE effects
De(C) up to 0.8 dex
Empirical selection of input atomic data
(simultaneously all lines / all stars)
Our C II model n 9, l 3 OP 5 n 9, 3 l
8 N02a
27
Sensitivity of Carbon to Atomic Data
To collisional ionization cross-sections from
empirical approximations (literature) Approx. OK?
Lines very sensitive to non-LTE effects
Empirical selection of input atomic data
(simultaneously all lines / all stars)
CBF threshold values of 6f 2F0 6g 2G Our
model 1 order mag gt N02a
28
Results Non-LTE C abundances from individual
lines
Non-LTE vs. LTE

• 6 early-B III-Vstars
• 21500 lt Teff lt 32000 K
• 3.1 lt log g lt 4.3 dex

Solution to C II/III/IV and C II/III ionization
equilibrium even for lines very sensitive to
non-LTE effects !!
29
Carbon Abundances in the Solar Vicinity
(literature)
Old spread artifact ?
1.6 dex !
30
Accurate Pristine Carbon Abundances in the Solar
Vicinity
Our results vs. literature
e(C)OB e(C)Sun (accurate results)
e(C)OB 8.33 0.03 (only uncertainties)
avoids systematic errors in atmospheric
parameters and atomic data
e(C)Sun 8.39 0.05 (total error)
31
Observational Constraints for Stellar Evolution
BA-Supergiants Przybilla, Firnstein,
Schiller (Sternwarte Bamberg ) Stellar Evolution
after MS
Chemical information (e.g. N/C) tracer for mixing
process
Evolutionary tracks Meynet Maeder (2003)
This work progenitors of Supergiants
Starting point for stellar evolution models
32
Other Applications

• Non-LTE chemical abundances in
• Fast-rotators in LMC (old C model sample to be
  re-analysed) (Korn et al. 2005)
• Extreme Helium Stars (Przybilla et al. 2006a)
• Sub-dwarfs B Stars (Przybilla et al. 2006b)
• Hypervelocity Star HVS8 (Nieva et al. 2006, in
  prep.)

With Calibrated C model lower variables
33
Summary

• State-of-the-art C II/III/IV model atom
  empirically calibrated (for NLTE calculations)
• Critically selected input atomic data
• Highly accurate stellar parameters/C abundances
  free of systematic errors
• Applications to different kind of objects
• Solution to C II/III/IV ionization equilibrium
• Solution to (unreported) C II emission lines
• Solution to classical C II ll 4267/6578-82 ?
  problem and now possible applications to
  fast-rotators and extragalactic objects
• Highly uniform C abundance for 6 early B-type
  stars (nearly solar !)

Conclusions
34

• Message
• Atmospheric parameter determination not trivial
  in chemical abundance analysis !!
• Atomic data has to be checked (not only f !)
• Both are basis of
• Accurate Quantitative Spectroscopy

• Basic checklist
• Stars in HRD over ZAMS?
• All possible indicators for parameters agree?
• Abundance vs. EW (individual lines)

35
Future

• Improve on statistics in MW (larger sample),
  data FOCES, Calar Alto (Spain)
• LMC/SMC objects high S/N spectra of
  slow-rotators (still not available)

Benchmark abundances
Goal different metallicities, empirical
calibration of stellar evolution and
galactochemical evolution models
-high accuracy -free of systematic errors
36
References
We Thank

• Uli Heber (support of the project)
• Katia Cunha (support of the project)

• Martin Altmann (FEROS data)
• Joachim Puls (Subaru data)

• DAAD (Ph.D. Scholarship)

• Korn, A., Nieva, M.F., Daflon, S., Cunha, K.,
  2005, ApJ, 633, 899
• Nieva, M.F. Przybilla, N., 2006a, ApJ, 639, L39
• Nieva, M.F. Przybilla, N., 2006b, subm. to AA
• Nieva, M.F. Przybilla, N., 2006c, in prep.
• Nieva, M.F., Heber, U., Edelmann, H., Przybilla,
  N., 2006, in prep.
• Przybilla, N., 2005, AA, 443, 293
• Przybilla, N. Butler, K., 2004, ApJ, 609, 1181
• Przybilla, N., Nieva, M.F., Edelmann, H., 2006,
• Przybilla, N., Nieva, M.F., Heber, U., Edelmann,
  H., 2006