Multi-dimensional radiative transfer in the expanding atmospheres of massive stars

点击翻译按钮获取中文摘要

Massive stars are critical agents in galactic evolution, both in the present epoch as well as in the early cosmos, and they are key tools to interpret our Universe. Their high luminosities drive energetic outflows, the stellar winds, that are fundamentally important for the conditions in the neigboring ISM, for triggering star formation, and for massive star evolution. During recent years, the "basic" mass-loss behavior of hot luminous stars became understood, though, e.g., effects from wind-inhomogeneities potentially imply lower mass-loss rates than predicted. However, present limitations in our atmospheric models restrict our diagnostic capabilities to test more advanced predictions. These refer particularly to the effects of fast rotation, magnetic fields, and neighborhood to the Eddington-limit/super-Eddington conditions, requiring a multi-dimensional treatment, not only with respect to hydrodynamics, but also regarding radiative transfer and occupation numbers.Though many groups have invested considerable effort in developing their own multi-D radiative transfer codes designed for specific purposes, none of these codes has been routinely applied to "realistic" wind conditions accounting for non-LTE effects. Here, we suggest a project with the "ultimate" goal (beyond the present proposal) to test recent and upcoming theoretical models predicting the impact of (rapid) rotation, magnetic-fields, and inhomogeneities on the stellar winds and photospheres of massive stars, by coupling 3-D radiative transfer with corresponding NLTE-calculations based on our fast-performing atmosphere code FASTWIND.Our "immediate" objective of this proposal is to develop, in a first step, adequate tools for the solution of 3-D radiative transfer (continuum and line radiation) in asymmetric, rapidly expanding atmospheres with non-monotonic velocity fields. We plan to use a Cartesian grid, and to test the accuracy and performance of two methods, namely a spatially implicit finite volume method as suggested by Adam (1990), and a (more conventional) short characteristics method, both amended by an Accelerated Lambda Iteration. After comparing the two methods, we will decide on our final approach. Since parallelization is of prime importance, we will consider and test suitable strategies for various architectures. As first applications, we will study the formation of UV-resonance lines and Balmer_alpha (the most important mass-loss diagnostics) in models of rapidly rotating and magnetically confined winds, and compare with observations. Already these investigations will allow for preliminary constraints on such models.The techniques developed in this and future projects will enable to investigate many more multi-D problems in massive star atmospheres. To speed up progress, we plan to make the developed tools available to the community.