OH 231.8 is a bipolar nebula (e.g. Reipurth 1987) with a high-velocity outflow seen in OH and CO, and it has a rich molecular spectrum (Morris et al. 1982, 1987). SiO maser emission has been detected by Barvainis & Clemens (1984), Morris et al. (1987), Jewell et al. (1991), and in this work. The central star is a long-period variable of spectral class M with a period of about 680 days (Feast et al. 1983; Cohen et al. 1985). Cohen et al. demonstrated the presence of a blue companion, and the bipolar outflow is therefore most likely caused by mass transfer between the central stars. The SiO maser is probably associated with the long-period variable.
OH 42.3 belongs to a class of objects where the velocity width of the maser emission is larger than that of the OH emission and it has been
suggested that they are in the early stages of becoming high-velocity
outflow sources (Gomez et al. 1994).
SiO maser emission was detected by Jewell et al. (1991), and in this
work. The OH maser emission is variable with a period of more than 2000
days.
Other objects where the velocity width of the
maser emission is larger than that of the OH emission are
and
(Gomez et al. 1994). OH 37.1 is a star with non-variable OH
emission (Herman & Habing 1985).
It was tentatively detected in SiO by Jewell et
al. (1991), but not by Gomez et al. (1990),
Nyman et al. (1993a), or in this work, and we therefore regard it as
a non-detection. No SiO maser has been found toward OH 12.8 (Jewell
et al. 1991).
OH 19.2 was not included in our sample, but it has been detected in SiO emission by Jewell et al. (1991). It has a peculiar OH maser spectrum with at least four peaks, and maps of the OH emission has been interpreted in terms of a bipolar outflow (Chapman 1988). The OH emission is variable with a period of about 600 days (Chapman 1988).
OH 15.7 is a non-variable OH/IR star (Herman & Habing 1985)
with a double peaked energy distribution.
It was tentatively detected by Jewell et al. (1991) and in this work,
and regarded as detected by Nyman et al. (1993a) but with a low
signal to noise ratio. The
possible maser components cover a similar velocity range
in the different observations. No SiO masers have been found towards other
objects of this type, e.g. and
(this
work; Jewell et al. 1991).
The SiO masers in OH 231.8, OH 42.3, OH 19.2, and the tentatively detected
OH 15.7 all have properties similar to those of
SiO masers seen towards Mira variables and OH/IR stars: line widths of
and center velocities close to the mean velocity of
the OH emission.
OH 19.2 is observed to be bipolar and the OH and
maser properties of OH 42.3 can be modeled
in terms of a bipolar outflow (Gomez et al. 1994). Their SiO
maser emission can therefore be explained in the same way as for OH 231.8
(Sect. 5.1): the central objects are binary systems where one of the
components is a variable red giant presently undergoing mass loss. The
bipolar nebula is shaped by the mass transfer between the two stars and the
SiO maser is associated with the variable star. Schwarz et al. (1995)
made a survey of SiO masers in dusty symbiotic systems, and emission was
found only in systems with wide orbits. Their observations show that SiO
masers may exist near variable stars also in binary systems.
OH 15.7 has probably very recently left the AGB. Its maser
emission has been very irregular over the past years and recently
disappeared (Engels 1997), indicating that the mass loss rate is
irregular and decreasing.
Only two SiO masers were clearly detected in the objects classified as PPNe in our sample, and one, OH 15.7+0.8, was tentatively detected. This is considerably fewer than towards the OH/IR stars. The low detection rate of SiO emission towards PPNe could be explained if these sources are situated at larger distances than OH/IR objects. However, distance estimates for the various types of objects indicate that they, on the average, are located at similar distances. It is also possible to estimate whether the non-detections are significant by comparing the SiO maser fluxes with the total energy fluxes, since it is likely that the luminosities of OH/IR stars are similar to those of PPNe. In Fig. 2 (click here) we have plotted the SiO (v=2, J=1-0) energy flux versus the far infrared energy flux (calculated from the IRAS intensities). For the OH/IR stars the far infrared energy flux should be a good approximation to the total energy flux, since they mainly radiate in the far infrared, but for the PPNe and PNe it is only a lower limit. Many PPNe have far infrared energy fluxes comparable to those of the detected OH/IR stars, indicating that the absence of SiO emission is real and not caused by a distance effect.
Figure 2: SiO (v=2, J=1-0) integrated flux densities plotted as
function of the far infrared energy flux. Detections, as well as upper
limits, are included. The upper limits to the SiO integrated flux densities
were calculated in the same way as in Nyman et al.
(1993a). The objects are divided into OH/IR stars, PPNe, PNe, and
unclassified sources ("?") according to Table 2 (click here). Note that the FIR
energy flux is only an upper limit to the total energy flux of the PPNe and
PNe. OH 15.7 was not detected in the v=2 transition and only tentatively
detected in the v=1 transition. (Nyman et al.
1993a, reported an integrated flux density of in the v=1 transition)