The moving atmospheres of red giants

Research project P 14365 The moving atmospheres red giants Josef HRON 08.05.2000 Variability is one of the main characteristics of stars on the asymptotic giant branch (AGB). Pulsation is clearly identified as the cause of the light and radial velocity variations observed for Miras and large amplitude Semiregular variables (SRVs). The large pulsation amplitudes and the cool photospheric temperatures of Miras lead to a very complex structure and temporal behaviour of the atmospheres. This is well known from spectroscopy and stars to be qualitatively reproduced by current (dynamic) model atmospheres. However, detailed quantitative compatisons of high resolution spectra with dynamic models are still lacking. The mechanism causing the light (and velocity) variability in AGB stars with small visual amplitudes is much less understood than for Miras. Besides pulsation various other mechanisms, have been proposed in the literature: large convective cells, starspots and active regions or the formation of dust. In the proposed project we will focus on two open questions with regard to AGB variables: (1) Identification of the causes of light and velocity variations in small amplitude variables: Models of convection in solar type stars predict characteristic line asymmetries and shift which are also seen in the observed line profiles. Evidence for such line shifts exists now for AGB stars. Therefore we will analyse new quasisimultaneous (IR-) spectroscopy and optical photometry for periodicities, for line profile asymmetries and changes and for correlations between light and velocity changes. The results will the be compared with synthetic line profiles and spectra computed on the basis of (i) hydrostatic model atmospheres with assumed surface structures in temperature and velocity and (ii) dynamic model atmospheres in spherical symmetry. The surface structure will be mainly adopted from existing models and scenarios for convection in late type stars. This give constraints on the AGB instability strip as well as the characteristics of convection for red giants. (2) The "warm molecular envelope": Low excitation CO(dv=2) lines of some Miras and a few small amplitude AGB variables show evidence for a component with a temperature around K and an almost zero velocity relative to the stellar systemic velocity. This layer, the so called "warm molecular envelope", is claimed to be common in evolved red giants in general. However, so far ist presence has been deduced only through an excess absorption relative to spectra based on hydrostatic atmospheres. We will therefore compare (mostly available) high resolution infrared of Miras and SRVs with synthetic spectra derived from dynamic model atmospheres. This will give important clues on the outer regions of AGB star atmospheres where the stellar wind originates and on the physical "completeness" of current dynamic model atmospheres.

Red giants represent the late stage in the evolution of most stars, that is also of our sun. In this stage the stars reach such low temperatures in their outer layers (less than 3000 degrees Celsius) that molecules and cosmic dust are formed. At the same time the stars are very large (such a star would extend out to Jupiters orbit) and this makes them very unstable to global oscillations and to more local motions in their atmospheres. Dust, molecules and atmospheric motions together cause a stellar wind, carrying several earth masses of gas and dust per year into interstellar space. This material is then available for the formation of new stars and planets. The project investigated two specific questions connected to the atmospheric motions of red giants: the possible causes for observed small brightness changes and the properties and origin of observed layers in the atmosphere apparently not moving relative to the star. It could be shown that the likely origin of the small brightness variations are global oscillations although motions related to the mixing of matter on the surface (convection) may have some influence, too. Furthermore it was demonstrated that non-moving layers are a natural consequence of stellar models taking into account the global oscillations and that such models can explain the observed vertical and temporal changes of gas-velocity in the stellar atmosphere rather well. To achieve the goals of the project new observations were obtained and analysed and extensive computer simulations were carried out. The observational data consisted of brightness measurements in the visible light partly obtained with two small robotic telescopes operated by the University of Vienna in the USA. For studying the atmospheric motions, spectroscopic observations in the infrared were mostly used, carried out at large telescopes in the USA. For the simulations computer codes were adopted and developed which calculate the motion, temperature and density of gas, the formation of molecules and dust and the transport of radiation under the influence of global stellar oscillations.