Evolved star water maser cloud size determined by star size

Abstract

Cool, evolved stars undergo copious mass loss but the details of how the matter is returned to the ISM are still under debate. We investigated the structure and evolution of the wind at 5 to 50 stellar radii from Asymptotic Giant Branch and Red Supergiant stars. 22-GHz water masers around seven evolved stars were imaged using MERLIN, at sub-AU resolution. Each source was observed at between 2 and 7 epochs (several stellar periods). We compared our results with long-term Pushchino single dish monitoring. The 22-GHz emission is located in ~spherical, thick, unevenly filled shells. The outflow velocity doubles between the inner and outer shell limits. Water maser clumps could be matched at successive epochs separated by <2 years for AGB stars, or at least 5 years for RSG. This is much shorter than the decades taken for the wind to cross the maser shell, and comparison with spectral monitoring shows that some features fade and reappear. In 5 sources, most of the matched features brighten or dim in concert from one epoch to the next. One cloud in W Hya was caught in the act of passing in front of a background cloud leading to 50-fold, transient amplification. The masing clouds are 1-2 orders of magnitude denser than the wind average and contain a substantial fraction of the mass loss in this region, with a filling factor <1%. The RSG clouds are ~10x bigger than those round the AGB stars. Proper motions are dominated by expansion, with no systematic rotation. The maser clouds survive for decades (the shell crossing time) but the masers are not always beamed in our direction. Radiative effects cause changes in flux density throughout the maser shells on short timescales. Cloud size is proportional to parent star size; clouds have a similar radius to the star in the 22-GHz maser shell. Stellar properties such as convection cells must determine the clumping scale.