Forum Springer Astron. Astrophys.
Forum Whats New Search Orders

Astron. Astrophys. 319, L1-L4 (1997)

Previous Section Next Section Title Page Table of Contents

3. Results and discussion

RT Vir is a semiregular variable (SRb) at the estimated distance between 120 pc (Szymczak & Engels 1995) and 460 pc (Menten & Melnick 1991). RT Vir is surrounded by strong water masers with flux density steadily greater than 100 Jy.

The distribution of the water masers at epoch April 6, 1995 is shown in Fig. 1. The figure demonstrates that the distribution size of water masers is about 160 mas in right ascention, which is roughly twice as large as those previously obtained by BCJ, BJ and YC (80-110 mas). It is interesting to investigate the relation in time variation between the distribution size of the water masers and the phase of the optical light curve. Therefore, we analyzed unpublished data of RT Vir in American Association of Variable Star Observation (AAVSO) with the phase dispersion minimization method to correctly estimate the light phase in the previous (BCJ on January 1985, YC on February 1985, and BJ on December 1988) and the present observations (January-May 1995). The analysis showed that the pulsation period of RT Vir is 375 days and that the previous observations were near the light maxima and the present observations are vice versa. Thus, the distribution size of the water masers are likely to be correlated with the optical light curve in such a way that the distribution size is largest near the optical minimum. The similar variation of spatial structure of stellar water masers correlated with the stellar light curve was first discovered in a Mira variable R Aql by BJ, and our case of RT Vir is the second for the late type stars after R Aql and the first for the semiregular variables.

[FIGURE] Fig. 1. Spatial distribution of detected water masers around RT Vir on April 6, 1995. The shown position is that of the cross-power flux peak of each individual maser spot. The relative position errors are 5-15 mas in right ascention and 10-50 mas in declination. The stellar velocity is estimated to be 17-18 km s-1 (e.g., BJ; YC).

Fig. 1 also shows that the blue-shifted and red-shifted maser components with respective to the stellar velocity (17-18 km s-1, e.g., BJ; YC) are clearly separated in the celestial plane. The direction of the line of separation is roughly east-west and consistent with those reported in the previous results by BCJ, BJ and YC.

Fig. 2 shows measured shifts of line-of-sight velocities in time for seven water maser spots which are strong enough and well determined in spatial position. The estimated rates of the velocity shifts range from -3.55 to 1.76 km s-1 yr-1 and are listed in Table 3. We could not reliably measure the velocity shifts of other maser spots because of the difficulty in identifying the same maser spot among successive observing sessions due to the limited positional accuracy in our wide-field mapping.

[FIGURE] Fig. 2. Shifts of line-of-sight velocities in time of detected strong water masers around RT Vir during four months in 1995. The component [FORMULA] = 12.9 km s-1 at the first epoch is assumed to be a blend of two maser spots.


Table 3. Rates of shifts in line-of-sight velocities in time of strong water masers around RT Vir

Fig. 1. reminds more or less simple stream pattern like bipolar flow or rotation. However, it is evident from Fig. 2 that there is no clear one-to-one correspondence between acceleration/deceleration and blue-shift/red-shift of the maser spots. It is therefore difficult to reconcile the observed results to the simplest kinematical models like rotating disk or bipolar flow. Perhaps, the real mass-loss process involves a fairly complicated streaming pattern in the circumstellar space. The present results mostly preclude a rotating disk model of the masing gas. In fact, even if we disregard the mismatch between the observed and expected velocity-acceleration patterns, we still need material ejected from the surface of the slowly rotating star to acquire the rotational velocity of order 10 km s-1 to explain the observed velocities and accelerations in terms of rotation. Greenhill et al. (1995) also suggest that there is little rotational motion of material with SiO masers nearby the stellar surface. It is premature to attempt to draw any further physical interpretation from the newly detected velocity shifts, in view of the small number of maser spots for which the velocity shifts are measured. Furthermore, we cannot reject yet the possibility that the shifts of line-of-sight velocities are caused only by changes in the masing region within the circumstellar clouds and might not directly reflect the real gas motions. Sullivan III (1971, 1973) claimed that the changes in line-of-sight velocities of water masers are likely to be due to changes in line strength of the three transitions within the fine structure. Therefore, more systematic and sensitive monitoring of the stellar water masers using VLBI is highly desirable for obtaining unambiguous conclusions on the newly discovered phenomena and on the mass-loss process of the evolved star. Then it is indispensable that the monitoring is performed with short time intervals from several days to one month between successive epochs because the spatial distribution of the water masers may change within the time intervals. We expect that it will be possible to measure similar velocity shifts for dozens of maser spots around RT Vir in more sensitive future VLBI observations with better mapping accuracy. Also, the high resolution VLBI monitoring will enable us to investigate transverse velocities of the maser spots and their accelerations or decelerations on the basis of proper motions of the maser spots and their time variations, which will further constrain the mass-loss dynamics of the evolved star.

Previous Section Next Section Title Page Table of Contents

© European Southern Observatory (ESO) 1997

Online publication: July 3, 1998