Introduction

As part of our ongoing effort to calculate accurate line blanketed stellar atmospheres we present the results our ``NextGen'' model atmosphere grid for ``high'' effective temperatures ($3000\le \hbox{$\,T_{\rm eff}$}\le 10000\,{\rm K}$). The results for very low mass stars (VLMS) and Brown Dwarfs involve the complication of dust condensation and are discussed in Allard et al, (in preparation). All the NextGen models are computed with the same assumptions to allow internally consistent analysis of, e.g., HR diagrams of globular clusters or population synthesis models.

Most of the models have been calculated under the assumption of local thermodynamic equilibrium (LTE) for the ionization/dissociation equilibria and the level populations. Thomson and Rayleigh scattering are included and cause the source function to deviate from the Planck function. The LTE assumption has been made mainly to be consistent with the VLMS part of the NextGen grid. The LTE models are being used as starting points for a large NLTE model grid that is currently being constructed and that will replace the LTE model grid (Hauschildt et al, in preparation). The NLTE models that are currently available in the NextGen model grid feature only a small subset of NLTE species and will be replaced by full NLTE models. We will discuss briefly how NLTE models change the results for one example.

[Kurucz(1992)Kurucz] has calculated an extensive grid of LTE model atmospheres with his ATLAS9 code using Opacity Distribution Functions (ODF) to include line blanketing. This model grid has set the standard for a large number spectral analyses over a wide range of effective temperatures. [Castelli & Kurucz(1994)Castelli and Kurucz] have used the Kurucz 92 grid and the new opacity sampling version of Kurucz's model atmosphere code ATLAS12 [Kurucz(1993a)Kurucz] to compute model atmospheres for Vega. They find very good agreement over a wide range in wavelengths, improving significantly upon earlier results. This is mostly due to Kurucz's update to his line lists, resulting in an increase in the total line opacity resulting in better fits to the ultraviolet spectra.

We use the Kurucz 92 grid to test our calculations at effective temperatures larger than about $5000\,{\rm K}$. At lower temperatures, larger differences are expected due to differences between the equations of state (EOS) and the treatment of molecular opacities in ATLAS12 and PHOENIX. For higher effective temperatures, when molecules and molecular opacities become less important, the synthetic spectra should be in good agreement with our LTE models because we use the Kurucz atomic line list [Kurucz(1994)Kurucz]; small differences are expected due to different bound-free and free-free opacity sources and different numerical methods used.