2025A&A...699A.326B

Context. We carry out decades-long single-dish monitoring of the variation in water-maser emission associated with the circumstellar envelopes (CSEs) of different types of evolved stars.
Aims. We follow the variation in the maser emission over long time intervals (multiple optical periods) to understand the structure and kinematics of the stellar wind that drives the maser clouds away from the star. We also determine how this may depend on the stellar properties.
Methods. We carried out monitoring campaigns with single-dish telescopes of water-maser emission at 22 GHz in the CSEs of four stars: o Cet, R Leo, χ Cyg, and R Cas. The observations took place with some interruptions between 1987 and 2023. The exact time interval differed from one star to the next, but no star was monitored for fewer than 15 years.
Results. The variability in integrated flux in the masers in R Cas and o Cet followed the variability in the optical with the same period, but with a lag of about one-third in phase. R Leo was too often below our sensitivity threshold for us to determine a radio period. Remarkably, no maser at all was detected in χ Cyg. The variability in the masers in R Cas has a distinctive pattern. The total flux, modulated by the pulsations of the star, gradually increases to a maximum, which is followed by a similar decrease. This takes about 20 years. The pattern is repeated after an interval of quiescence of several years. Our observations have covered about one and a half cycle of this pattern so far. During its decline from a maximum, the variation in the flux resembles a damped harmonic oscillator. There are two dominant emission components that move almost tangentially on either side of the star with respect to the observer. The redshifted component likely originates from a single cloud and seems to be falling back towards the star. The blue component, moving in the CSE hemisphere nearest to us, has no drift in the line-of-sight velocity and appears to originate in a time series of short-living clouds with line-of-sight velocities within ∼1 km s^–1 of each other. No systematic velocity drifts are found in R Leo and o Cet. A few bursts of emission were detected at infrequent times in R Cas and R Leo that lasted about a year and caused an increase in the flux density by 1-2 orders of magnitude. The velocity range of the maser emission is ≲10 kms^–1, which is narrower than the majority of the Miras and semi-regular variables we studied so far. In particular, in the stars with a low bolometric luminosity, o Cet and R Leo, only the brightest maser components from a limited part of the CSE are visible.
Conclusions. The existence of a zone in the CSE with favourable conditions for maser excitation is confirmed most clearly in the case of R Cas through the unique pattern of its maser variability. The bolometric luminosity of a star and the velocity range of its water-maser emission, that is, the number of emission components in the spectra, are clearly correlated. Most maser components in the CSEs we studied originated in clouds that move almost perpendicular to the line of sight. The redshifted emission in R Cas is consistent with an origin in a single cloud that lived for at least about eight years.