At the time of our observations the spectrum of the 22 GHz H2O masers around U Her did not show significant structure, as only a few
features were detected. The averaged cross power spectrum is shown in
an inlay in Fig. 1. We find that the shape of the spectrum is
similar to previous observations, performed with MERLIN, the Very
Large Array (VLA) and with the 100-m radio telescope in Effelsberg
(Baines et al. in preparation; Yates & Cohen 1994; Colomer et al. 2000; Engels et al. 1988). In our spectrum we find that the
strongest feature is located at -15.7 km s-1, and this is the feature
which was used to determine the phase solutions. The stellar velocity
of U Her is
km s-1, which was determined from OH and SiO maser observations (Chapman et al. 1994). As the U Her H2O maser
emission is located between -13 and -20 km s-1, the strongest
feature is not the most blue-shifted feature.
After phase referencing and determining the position of J1628+214 with
respect to the brightest H2O maser feature of U Her, we find that
the position of J1628+214 is shifted by +83 mas in right ascension
and +15 mas in declination with respect to the positions in vL00
(
,
).
A similar process for J1619+2247 results in a position shift of +76 mas in right ascension and +25 mas in declination with respect to
the position in the VLBA calibrator list. As a consistency check we
have also determined the position shift of the brightest maser spot
after phase referencing with respect to J1628+214. We find that the
maser spot is shifted -83 and -11 mas in right ascension and
declination respectively with respect to the a priori assumed
target coordinates.
The formal uncertainty in fitting a Gaussian profile to the reference
source or maser spot is a fraction of the beam width, and depends on
the SNR of the image. For the reference sources the formal position errors
are of the order of 5 mas in each coordinate, for the U Her maser
spot the errors are 1 mas. The best phase connection was
made when phase referencing J1628+214 to the brightest maser spot, so
we assume that the positions as determined with J1628+214 are the most
reliable. The actual phase referencing errors can be estimated from
the difference between the position of the brightest maser spot with
respect to the two reference sources. From this, we conclude that our
positions are accurate to within 10 mas, which is in agreement with
the estimated errors due to the correlator model that are described
above. For the brightest H2O maser spot around U Her we then find a
position of
,
at the time of our observations.
To compare this position to the stellar position of U Her we have
extrapolated the optical position found by the Hipparcos satellite at J1991.25 to our epoch of observation. We have used the proper motion
and parallax determined by monitoring the position of the most
blue-shifted OH maser spot, which was shown to be the stellar
image. The first fit was performed in vL00 for 6 epochs of
observations, a fit including additional epochs was presented in
Vlemmings et al. (2000). As described in vL00, the OH maser proper
motion is entirely consistent with the Hipparcos proper motion. The
error in the transposed position is dominated by the error in proper
motion. At our epoch of observation this error is 6 mas in
each coordinate. Combined with the errors on the parallax and our
position errors, we have been able to compare the radio and optical
position with
18 mas accuracy. Figure 1 shows a map
of the H2O maser features, covering the velocity range indicated in
the spectrum, including the position of the star. Circles indicate the
size of the star and the radio-photosphere. The size of the
radio-photosphere can be estimated from SiO maser observations by
Diamond et al. (1994). Their observations provide an upper limit of
20 mas if, as proposed by Reid & Menten (1997), the
radio-photosphere extends to the edge of the SiO masing region. The
triangle denotes the stellar position as determined when using the
maser positions with J1635+1831 as the reference source.
Although the size of the brightest H2O maser spot (50 mas) is larger than the expected size of the stellar
radio-photosphere, most likely due to the blending of several weaker
maser features, the position of peak intensity matches the predicted
location of the star within the errors. This indicates that the H2O maser spot also coincides with the most blue-shifted OH maser spot
which has been shown to be the amplified stellar image. Thus, also the
brightest H2O maser spot seems to be emission from the stellar
radio-photosphere amplified by the maser medium at the line of sight.
Copyright ESO 2002