Summary and Conclusions

In this paper we presented a new set of models for main sequence stars for effective temperatures from $3000\,{\rm K}$ to $10000\,{\rm K}$. We include mostly LTE models, however, NLTE effects become progressively important for effective temperatures larger than about $7000\,{\rm K}$. Therefore, we have included self-consistent NLTE models calculated with a subset of the NLTE species that are available in PHOENIX.

We find that our models agree well with the Kurucz 92 grid for solar type stars ($5000\le \hbox{$\,T_{\rm eff}$}\le 7000\,{\rm K}$) and that our LTE models for stars with effective temperatures up to about $10000\,{\rm K}$ also agree reasonably well with the Kurucz models. For lower temperatures we find some differences between the model grids. The reason for this is that our calculations are designed to include the atmospheres of very low mass stars and brown dwarfs, so that they include a more detailed molecular EOS and a better set of molecular opacity sources than the Kurucz 92 models. For higher effective temperatures, NLTE effects become important. Our results for Vega show that NLTE has an effect on the structure of these models. Therefore, detailed NLTE models are required for effective temperatures higher than about $7000\,{\rm K}$. Furthermore, [Aufdenberg et al.(1997a)Aufdenberg, Hauschildt, Shore, and Baron,Aufdenberg et al.(1997b)Aufdenberg, Hauschildt, Sankrit, and Baron] have shown that the combined effects of line blanketing and spherical geometry are essential ingredients in model atmosphere models for effective temperatures $\mathrel{\hbox{\rlap{\hbox{\lower4pt\hbox{$\sim$}}}\hbox{$\gt$}}}18000\,{\rm K}$. We will present spherical NLTE models for larger effective temperatures in a subsequent paper (Aufdenberg et al, in preparation).

It appears that the assumption of LTE is a useful approximation only in a very narrow range of effective temperatures near that of solar type stars. For cooler and even more so, for hotter stars, NLTE effects are important and should be included in model atmosphere calculations.

We thank Jason Aufdenberg, David Alexander, Robert Kurucz, and Sumner Starrfield for helpful discussions. This work was supported in part by NSF grant AST-9720704, NASA ATP grant NAG 5-3018 and LTSA grant NAG 5-3619 to the University of Georgia, by NSF grant AST-9417242, NASA grant NAG5-3505 and an IBM SUR grant to the University of Oklahoma, and NASA LTSA grant NAG5-3435 to Wichita State University. The NSO/Kitt Peak FTS data of the solar spectrum used in this paper were produced by NSF/NOAO. Some of the calculations presented in this paper were performed on the IBM SP2 and the SGI Origin 2000 of the UGA UCNS, at the San Diego Supercomputer Center (SDSC) and the Cornell Theory Center (CTC), with support from the National Science Foundation, and at the NERSC with support from the DoE. We thank all these institutions for a generous allocation of computer time.