Dynamic atmosphere and wind models of C-type asymptotic giant branch stars

Context. Mass loss through stellar winds governs the evolution of stars on the asymptotic giant branch (AGB). In the case of carbonrich AGB stars, the wind is believed to be driven by radiation pressure on amorphous carbon (amC) dust forming in the atmosphere. The structural complexity of amC is evident from the diversity of laboratory optical data that are available in the literature. Consequently, the choice of dust optical data will have a significant impact on atmosphere and wind models of AGB stars. Aims.We compare two commonly used optical data sets of amC and investigate how the wind characteristics and photometric properties resulting from dynamical models of carbon-rich AGB stars are influenced by the micro-physical properties of dust grains. Methods. We computed two extensive grids of carbon star atmosphere and wind models with the DARWIN 1D radiation-hydrodynamical code. A defining feature of these models is a self-regulating feedback between the time-dependent dynamics, grain growth, and dust optical properties. Thus, they are able to predict combinations of mass-loss rates, wind velocities, and grain sizes for given stellar parameters and micro-physical data. Each of the two grids uses a different amC optical data set. The stellar parameters of the models were varied in terms of the effective temperature, luminosity, stellar mass, carbon excess, and pulsation amplitude to cover a wide range of possible combinations. A posteriori radiative transfer calculations were performed for a sub-set of the models, resulting in photometric fluxes and colours. Results. We find small, but systematic differences in the predicted mass-loss rates for the two grids. The grain sizes and photometric properties are affected by the different dust optical data sets. Higher absorption efficiency leads to the formation of a greater number of grains, which are smaller. Models that are obscured by dust exhibit differences in terms of the covered colour range compared to observations, depending on the dust optical data used. Conclusions. An important motivation for this study was to investigate how strongly the predicted mass-loss rates depend on the choice of dust optical data, as these mass-loss values are more frequently used in stellar evolution models. Based on the current results, we conclude that mass-loss rates may typically differ by about a factor of two for DARWIN models of C-type AGB stars for commonly used dust optical data sets.