Issue |
A&A
Volume 528, April 2011
|
|
---|---|---|
Article Number | A32 | |
Number of page(s) | 12 | |
Section | Stellar atmospheres | |
DOI | https://doi.org/10.1051/0004-6361/201016151 | |
Published online | 23 February 2011 |
Three-dimensional surface convection simulations of metal-poor stars
The effect of scattering on the photospheric temperature stratification
1
Max-Planck-Institut für Astrophysik,
Karl-Schwarzschild-Str. 1,
85741
Garching b. München,
Germany
e-mail: remo@mpa-garching.mpg.de; hayek@mpa-garching.mpg.de; asplund@mpa-garching.mpg.de
2
Research School of Astronomy & Astrophysics,
Cotter Road,
Weston
ACT
2611,
Australia
3
Institute for Theoretical Astrophysics,
Sem Sælands Vei,
0315
Oslo,
Norway
e-mail: b.v.gudiksen@astro.uio.no
4
Niels Bohr Institute, University of Copenhagen,
Juliane Maries Vej 30,
2100, Copenhagen,
Denmark
e-mail: aake@nbi.dk
5
JILA, University of Colorado, 440 UCB, Boulder, CO
80309-0440,
USA
e-mail: trampeda@lcd.colorado.edu
Received: 15 November 2010
Accepted: 8 January 2011
Context. Three-dimensional (3D) radiative hydrodynamic model atmospheres of metal-poor late-type stars are characterized by cooler upper photospheric layers than their one-dimensional counterparts. This property of 3D model atmospheres can dramatically affect the determination of elemental abundances from temperature-sensitive spectral features, with profound consequences on galactic chemical evolution studies.
Aims. We investigate whether the cool surface temperatures predicted by 3D model atmospheres of metal-poor stars can be ascribed to approximations in the treatment of scattering during the modelling phase.
Methods. We use the Bifrost code to construct 3D model atmospheres of metal-poor stars and test three different ways to handle scattering in the radiative transfer equation. As a first approach, we solve iteratively the radiative transfer equation for the general case of a source function with a coherent scattering term, treating scattering in a correct and consistent way. As a second approach, we solve the radiative transfer equation in local thermodynamic equilibrium approximation, neglecting altogether the contribution of continuum scattering to extinction in the optically thin layers; this has been the default mode in our previous 3D modelling as well as in present Stagger-Code models. As our third and final approach, we treat continuum scattering as pure absorption everywhere, which is the standard case in the 3D modelling by the CO5BOLD collaboration.
Results. For all simulations, we find that the second approach produces temperature structures with cool upper photospheric layers very similar to the case in which scattering is treated correctly. In contrast, treating scattering as pure absorption leads instead to significantly hotter and shallower temperature stratifications. The main differences in temperature structure between our published models computed with the Stagger- and Bifrost codes and those generated with the CO5BOLD code can be traced to the different treatments of scattering.
Conclusions. Neglecting the contribution of continuum scattering to extinction in optically thin layers provides a good approximation to the full, iterative solution of the radiative transfer equation in metal-poor stellar surface convection simulations, and at a much lower computational cost. Our results also demonstrate that the cool temperature stratifications predicted for metal-poor late-type stars by previous models by our collaboration are not an artifact of the approximated treatment of scattering.
Key words: hydrodynamics / convection / radiative transfer / scattering / stars: late-type / stars: atmospheres
© ESO, 2011
Current usage metrics show cumulative count of Article Views (full-text article views including HTML views, PDF and ePub downloads, according to the available data) and Abstracts Views on Vision4Press platform.
Data correspond to usage on the plateform after 2015. The current usage metrics is available 48-96 hours after online publication and is updated daily on week days.
Initial download of the metrics may take a while.