Spectroscopic analysis of DA white dwarfs with 3D model atmospheres^⋆

¹ Zentrum für Astronomie der Universität Heidelberg, Landessternwarte, Königstuhl 12, 69117 Heidelberg, Germany
e-mail: ptremblay@lsw.uni-heidelberg.de; hludwig@lsw.uni-heidelberg.de
² Leibniz-Institut für Astrophysik Potsdam, An der Sternwarte 16, 14482 Potsdam, Germany
e-mail: msteffen@aip.de
³ Centre de Recherche Astrophysique de Lyon, UMR 5574 CNRS, Université de Lyon, École Normale Supérieure de Lyon, 46 allée d’Italie, 69364 Lyon Cedex 07, France
e-mail: Bernd.Freytag@ens-lyon.fr

Received: 18 July 2013
Accepted: 28 August 2013

Abstract

We present the first grid of mean three-dimensional (3D) spectra for pure-hydrogen (DA) white dwarfs based on 3D model atmospheres. We use CO⁵BOLD radiation-hydrodynamics 3D simulations instead of the mixing-length theory for the treatment of convection. The simulations cover the effective temperature range of 6000 < T_eff (K) < 15 000 and the surface gravity range of 7 < log g < 9 where the large majority of DAs with a convective atmosphere are located. We rely on horizontally averaged 3D structures (over constant Rosseland optical depth) to compute ⟨3D⟩ spectra. It is demonstrated that our ⟨3D⟩ spectra can be smoothly connected to their 1D counterparts at higher and lower T_eff where the 3D effects are small. Analytical functions are provided in order to convert spectroscopically determined 1D effective temperatures and surface gravities to 3D atmospheric parameters. We apply our improved models to well studied spectroscopic data sets from the Sloan Digital Sky Survey and the White Dwarf Catalog. We confirm that the so-called high-log g problem is not present when employing ⟨3D⟩ spectra and that the issue was caused by inaccuracies in the 1D mixing-length approach. The white dwarfs with a radiative and a convective atmosphere have derived mean masses that are the same within ~0.01 M_⊙, in much better agreement with our understanding of stellar evolution. Furthermore, the 3D atmospheric parameters are in better agreement with independent T_eff and log g values from photometric and parallax measurements.