The winds of cool luminous AGB stars are commonly assumed to be driven by radiative acceleration of dust grains which form in the extended atmospheres produced by pulsation-induced shock waves. The dust particles gain momentum by absorption or scatte
ring of stellar photons, and they drag along the surrounding gas particles through collisions, triggering an outflow. This scenario, here referred to as Pulsation-Enhanced Dust-DRiven Outflow (PEDDRO), has passed a range of critical observational tests as models have developed from empirical and qualitative to increasingly self-consistent and quantitative. A reliable theory of mass loss is an essential piece in the bigger picture of stellar and galactic chemical evolution, and central for determining the contribution of AGB stars to the dust budget of galaxies. In this review, I discuss the current understanding of wind acceleration and indicate areas where further efforts by theorists and observers are needed.
The evolution and spectral properties of stars on the AGB are significantly affected by mass loss through dusty stellar winds. Dynamic atmosphere and wind models are an essential tool for studying these evolved stars, both individually and as members
of stellar populations, to understand their contribution to the integrated light and chemical evolution of galaxies. This paper is part of a series testing state-of-the-art atmosphere and wind models of carbon stars against observations, and making them available for use in various theoretical and observational studies. We have computed low-resolution spectra and photometry (in the wavelength range 0.35-25 mu) for a grid of 540 dynamic models with stellar parameters typical of solar-metallicity C-rich AGB stars and with a range of pulsation amplitudes. The models cover the dynamic atmosphere and dusty outflow (if present), assuming spherical symmetry, and taking opacities of gas-phase species and dust grains consistently into account. To characterize the time-dependent dynamic and photometric behaviour of the models in a concise way we defined a number of classes for models with and without winds. Comparisons with observed data in general show a quite good agreement for example regarding mass-loss rates vs. (J-K) colours or K magnitudes vs. (J-K) colours. Some exceptions from the good overall agreement, however, are found and attributed to the range of input parameters (e.g. relatively high carbon excesses) or intrinsic model assumptions (e.g. small particle limit for grain opacities). While current results indicate that some changes in model assumptions and parameter ranges should be made in the future to bring certain synthetic observables into better agreement with observations, it seems unlikely that these pending improvements will significantly affect the mass-loss rates of the models.
We study the circumstellar environment of the M-type AGB star RT Vir using mid-infrared high spatial resolution observations from the ESO-VLTI focal instrument MIDI. The aim of this study is to provide observational constraints on theoretical predict
ion that the winds of M-type AGB objects can be driven by photon scattering on iron-free silicate grains located in the close environment (about 2 to 3 stellar radii) of the star. We interpreted spectro-interferometric data, first using wavelength-dependent geometric models. We then used a self-consistent dynamic model atmosphere containing a time-dependent description of grain growth for pure forsterite dust particles to reproduce the photometric, spectrometric, and interferometric measurements of RT Vir. Since the hydrodynamic computation needs stellar parameters as input, a considerable effort was first made to determine these parameters. MIDI differential phases reveal the presence of an asymmetry in the stellar vicinity. Results from the geometrical modeling give us clues to the presence of aluminum and silicate dust in the close circumstellar environment (< ~5 stellar radii). Comparison between spectro-interferometric data and a self-consistent dust-driven wind model reveals that silicate dust has to be present in the region between 2 to 3 stellar radii to reproduce the 59 and 63 m baseline visibility measurements around 9.8 micron. This gives additional observational evidence in favor of winds driven by photon scattering on iron-free silicate grains located in the close vicinity of an M-type star. However, other sources of opacity are clearly missing to reproduce the 10-13 micron visibility measurements for all baselines. This study is a first attempt to understand the wind mechanism of M-type AGB stars by comparing photometric, spectrometric, and interferometric measurements with state-of-the-art, self-consistent dust-driven wind models. The agreement of the dynamic model atmosphere with interferometric measurements in the 8-10 micron spectral region gives additional observational evidence that the winds of M-type stars can be driven by photon scattering on iron-free silicate grains. Finally, a larger statistical study and progress in advanced self-consistent 3D modeling are still required to solve the remaining problems.
