Please Note: the e-mail address(es) and any
external links in this paper were correct when it was written in 1995,
but may no longer be valid.
Go to: Front page, Participants, Group Photograph, Preface, Contributions, Acknowledgements
MERLIN Observations of the Radio Jets of SS433
Frederick H.
Jowett and
Ralph E.
Spencer
e-mail: fhj@jb.man.ac.uk
Nuffield Radio Astronomy Laboratories, University of Manchester,
Jodrell Bank, Lower Withington, SK11 9DL, UNITED KINGDOM
Abstract:
The unusual binary star SS433 was observed on 5 epochs in December, 1991 and
January, 1992 using the United Kingdom's Multi-Element Radio Linked
Interferometer Network (MERLIN) operating at a frequency of 5 GHz. The
maps produced from this data show the characteristic ``S-shaped'' morphology of
SS433's precessing jets. These jets were demonstrated to behave, in general, as
dictated by the ``Kinematic Model''. Several knots of radio emission in the
jets were identified and their motion could be followed between the images. By
studying their evolution, these knots were shown to be ballistic, not slowing
down noticeably before fading from view and by comparing their apparent
velocity with the predictions of the Kinematic Model the distance to SS433 was
measured at 
. The brightness evolution of the knots was
found to be consistent with exponential decay of time constant 16 days. Two
anomalous knots were identified and believed to be caused by drastic variation
of the jet ejection velocity.
The canonical model of SS433 describes a binary system comprising a neutron
star and OB or Wolf-Rayet companion. Matter from the companion is transferred
onto the compact star via an accretion disc and a collimated jet of material is
ejected along the axis of the disc at an angle of 
to the
line-of-sight. By some unknown mechanism, the accretion disc, and hence the
ejection vector of the jet, precesses around a cone of half-angle 
every 162.5 days. This precession results in optical emission lines
which ``move'' in frequency periodically, due to varying doppler shift
determined by the projected ejection velocity of the jet material. This
behaviour is described well by the ``Kinematic Model'' which assumes a constant
magnitude of the ejection velocity of 
(Abell & Margon (1979),Margon & Anderson (1989)).
On radio maps, imaging the regions from the central engine of scales 
down to 
, the precessing jet manifests itself in a
characteristic ``S-shaped'' morphology of synchrotron-radiating material.
Previous observations, particularly with MERLIN and VLBI (e.g.
Spencer (1984),Vermeulen et al. (1993)), have shown knots of radio emission which move
ballistically outwards from the core, without showing any appreciable
deceleration over distances out to 5000 AU. By comparing the apparent velocity
of the knots with the Kinematic Model, previous workers have derived distances
to SS433 between about 4.5 and 5.0 kpc.
VLBI observations of the
inner 500 AU of the jets have discovered that knots are often, but not always,
ejected symmetrically from the core, emerging from permanent features termed
``core wings'' (Vermeulen et al. (1993)). These knots then brighten by a factor of
two as they move away from the core, as they enter a region known as the
``Brightening Zone'' which has an extent of about 250 AU. After this, the knots
begin to fade. This fading, presumably due to some form of expansion since all
reasonable estimations place the synchrotron emitting lifetime of the material
to about 1000 years, has been observed to continue on MERLIN-scale images.
Observations by Spencer (1984) have been unable to differentiate between
power-law (indicating adiabatic or sub-adiabatic expansion) or exponential
decay.
This paper presents preliminary results of recent 5 GHz observations of SS433
using MERLIN,
including a measurement of its distance and an attempt to determine the
behaviour of the knot brightness decay.
Phase-referencing observations of SS433 at 5 GHz were performed with the
MERLIN synthesis array
on five occasions in December, 1991 and January 1992, as detailed in
Table 1. The naturally-weighted data were imaged using a
combination of proprietorial
MERLIN software, the
NRAO AIPS package and the
Caltech Difmap program. The r.m.s. noises obtained in the images were between
100 and 
, comparable with that expected for
twelve-hour observations, except for epoch (b) when a cryogenics failure at the
Cambridge telescope increased the map noise by a factor 
over the
other maps.
The observation on epoch (d) was affected by the loss of the telescopes
comprising the shortest baselines in the MERLIN array. This has resulted in the map being
insensitive to extended emission in the jets. This observation was repeated a
day later (epoch (e)).
Table 1: Dates of the five observations and map quality information.
Figure 1: Contour images of the maps obtained for the naturally weighted data on
7th, 12th, 22nd December, and 3rd, 4th January (a to
e). The contours are -0.5, 0.5, 1, 2, 4, 8, 16, 32, 64, 128 mJy/beam for
each image, the first contour being about 
, 
, 
,
and 
for maps a to e. Each image has been
restored with a 70 mas beam. The expected Kinematic model for SS433's measured
distance of 4.7 kpc is shown. The labelled knots are discussed in the text.
Each of the images shown in Figure 1 exhibit the general shape of the
characteristic curve that would be expected given the Kinematic Model. The
Eastern jet is the predominantly approaching one and so it is visibly longer
than the Western jet which is foreshortened by light travel-time effects. K
A false colour version of this sequence of maps will be shown if you click
\
of bright emission are visible in the jets, and some of these can be identified
from one image to the next. Although 18 knots could be traced for more than
one epoch, only five (termed A, C, M, N, Z) were present for all images whilst
displaying well-behaved brightnesses. Some knots, although present on several
maps, behaved in an erratic manner, and it is possible that these were
artifacts caused by under-resolution of knots smaller than the beam.
The apparent transverse velocity of each knot was measured by fitting a
straight line to its radial-position vs time behaviour. As can be seen from
Figure 2 the knots were well-fitted by such lines, indicating that
their motion was ballistic with no appreciable deceleration during the
observations. Given that the velocity of each knot was constant, the dates that
the knots were ejected were extrapolated from the fits (see
Table 2).
Figure 2: Graph showing the radial position of the knots during each
observation and the best-fit straight lines.
Table 2: Ejection times and angular velocities of the radio
knots. The results of the brightness evolution fits are also
displayed. The power law decay index and exponential decay constants
marked 
for knots M and N are those obtained by ignoring the
7th December observation. Note that the exponential decay
constants shown here have had the effects of light travel-times taken
into account.
The distance to SS433 was measured by comparing the observed velocity of the
knots (after taking light travel-time effects into account) with their
predicted velocities transverse to the line-of-sight, assuming that the
optically determined velocity of 
also applies to the radio-emitting
material. This calculation leads to a distance to SS433 of 
,
which is consistent with previous measurements made using this method, but
covering different scales and/or frequencies.
As the knots in the jets of SS433 age and move away from the core, their peak
brightness decreases, but previous studies have been unable to differentiate
between power law and exponential decay of brightness with time since ejection
(Spencer (1984)). In an attempt to resolve this uncertainty, straight lines
were fitted to log(brightness) vs log(age) plots (for power law) and to
ln(brightness) vs age plots (for exponential decay). Unfortunately, direct
comparisons of the 
parameters of the fits (see
Table 2) show that knot A was better fitted by a power law, and
knots C and Z by exponential decay, although the 
s are not
drastically different in these cases.
Knots M and N were not fit well by either method. This was caused by a low
value for the measured brightness for each knot during the first epoch of
observation on 7th December. Both of these knots were within 20
milliarcseconds from the core, 
at 4.7 kpc
distance - well
within the Brightening Zone. It seems likely, therefore, that these knots
increase in brightness as they move into the Brightening Zone between
7th and 12th December. By ignoring the first observation, a much
improved fit was obtained in both cases although the fit for power-law decay
was better.
Figure 3: (a) Log(Brightness) plotted against Log(Age) for each knot, and
best-fit straight lines. The lines shown for knots M and N are those derived by
ignoring the earliest data points. Note that the measured brightnesses for the
observations on 3rd and 4th January have been averaged and
plotted as a single point. (b) Ln(Brightness) plotted against Age for
each knot, and best-fit straight lines. The lines shown for knots M and N are
those derived by ignoring the earliest data points.
The time constants obtained from the fits to exponential decay ranged from
days to 21, whereas the power law decay index showed a higher
range of variation from 
to 
. Moreover, the power law decay index
had a larger magnitude for those knots seen at a later stage of their
evolution. It is possible that this represents an intrinsic difference between
the knots' physical conditions, but this behaviour is broadly consistent with
an attempt to model the decay as power law, when it is intrinsically
exponential. Figure 4 shows a plot of the best-fit power law decay
index (
) against the mean age of each knot during the observations. The
straight line represents the evolution of 
that would be expected for a
knot whose brightness decays exponentially with a time constant of 16 days.
Although this line is not a particularly good fit to the data, differences
could be explained by postulating variations in the physical parameters of each
knot.
Figure 4: The power law decay index obtained from the slopes of the lines in
Figure 3, against the mean observed age of each
knot. The line represents the behaviour that would be expected from a
knot which decayed exponentially with a time constant of 16 days.
The 7th and 12th December maps (Figures 1a and 1b) show two
knots (AA and BB) which lie at a position off to the side of the main jet.
Since these knots are not exactly symmetrical about the core, and are also
present for more than one image, they are not caused by spurious symmetrization
during the self-calibration cycles and are likely to be real. Knot AA is also
present on Figures 1c and 1e (but may have been too weak to
be detected on the 3rd January image), and has an apparent velocity on the sky,
with no apparent deceleration, of 
milliarcseconds per day, which
is about half the minimum that can be obtained from the Kinematic Model. The
projected ejection date for knot AA is MJD 8516, which, given a somewhat large
error of 20 days, is consistent with its observed position angle of
. Upon the assumption that the knot was ejected in the direction
predicted by the Model, this gives its ejection velocity at 
(for 4.7
kpc distance). Variations of the ejection velocity of the jet material of the
order of a few thousand 
have previously been posed as explanations
for residuals in the fit of the Kinematic Model to the optical line doppler
shifts (termed ``jitter''(Margon & Anderson (1989)), but the deviation measured for knot
AA is several times larger than this. Interestingly, Iijima (1993) found
large Doppler Shift residuals on MJD 
, but no measurements were
made for a period of 20 days before this, during which knot AA was ejected.
This paper has presented initial results from studying the naturally-weighted
images obtained from the observations. The images have shown that the central
engine of SS433 ejects ``blobs'' of matter, usually in antiparallel pairs, at
velocities of 
, and that this behaviour is well described by the
Kinematic Model. The blobs, which are seen as knots in the images, move away
ballistically from the core with no deceleration even out to 3300 AU. The blobs
decay in brightness exponentially, which may indicate that they are expanding
exponentially. Finally, the distance to SS433 was measured at 
.
It is hoped that analysis of the uniformly-weighted images (in preparation)
which are showing unprecedented detail, will eliminate some of the
uncertainties and help to tie down the brightness evolution behaviour of the
knots.
MERLIN is a national
facility operated by the University of
Manchester on behalf of
PPARC
- Abell G. & Margon B., 1979,
Nature, 279, 701.
- Iijima T., 1993,
ApJ, 410, 295.
- Margon B. & Anderson S.F., 1989,
ApJ, 347, 448.
- Shepherd M. et al., 1993,
Difmap: An interactive program for synthesis imaging,
preprint.
- Spencer R.E., 1984,
MNRAS, 209, 869.
- Thomasson P. et al.1994
MERLIN user guide, (University of Manchester).
- Vermeulen R.C. et al., 1993,
A&A, 270, 177.
YERAC 94 Account
Wed Feb 22 18:51:28 GMT 1995
