Asymmetries in asymptotic giant branch stars and their winds. I. From 3D RHD models to synthetic observables

Context. Asymptotic giant branch (AGB) stars are significant contributors to the metal enrichment of the interstellar medium. They have strong dust-driven winds that have their origin in regions close to the AGB star’s surface, where dense dust clouds form. Aims. In this methods paper, we adapted models from advanced radiation-hydrodynamical (RHD) simulations as input for radiative transfer software to create synthetic observables. A major goal is to describe an AGB star’s non-sphericity and to simulate its effects on the surrounding dusty envelope. Methods. We developed tools in Python to translate models of an AGB star and its dust-driven wind from 3D RHD simulations with CO 5 BOLD into the format used for radiative transfer with RADMC-3D. We preserved the asymmetric shape of the AGB star by including the star as a ‘dust species’ and by using temperature data computed in CO 5 BOLD. The circumstellar silicate dust from the 3D RHD simulation is included using Mg 2 SiO 4 opacity data in RADMC-3D with spatially dependent grain sizes. We compared images and spectral energy distributions (SEDs) created with RADMC-3D of a model snapshot with similar output made with a spherically symmetric stellar atmosphere from the 1D program DARWIN and with a point source star in RADMC-3D. Results. Our CO 5 BOLD model features substantial and clumpy dust formation just above 3.4 au from the grid centre (∼1 R ⋆ above the star), and large-scale structures due to giant convection cells are visible on the stellar surface. With the properties of VLTI as a basis, we have created simple synthetic observables where the dust clouds close to the star and features on the stellar surface are resolved. The flux density and the contrast to the star are high enough that optical interferometers, such as the VLTI, should be able to detect these dust clouds. We find that it is important to include asymmetric stellar models since their irregular shapes, radiation fields, and their dusty envelopes even put their marks on spatially unresolved observables and affect the flux levels and shapes of the SEDs. The effects on flux levels can mostly be linked to the clumpiness of the circumstellar dust. In contrast, the angle-dependent illumination resulting from temperature variations on the stellar surface causes shifts in the wavelengths of the flux maximum, as shown by replacing the asymmetric star with a spherical one. Conclusions. The methods presented here are an important step towards producing realistic synthetic observables and testing predictions of advanced 3D RHD models. With the model used here, we find that optical interferometers should be able to resolve thermal emission from dense clouds in the dust-formation zone close to an AGB star. Taking the angle-dependence of SEDs as a proxy for temporal variations in unresolved data, we conclude that not all variability observed in AGB stars should be interpreted as global changes in the sense of spherical models.