We consider in this section the photometric properties of the deconvolved images, except in the case of 3C 66B for which a reliable calibration of the data is lacking.
The improvement in the quality of the deconvolved image prompts us to reconsider some results presented earlier (Jakobsen et al. 1991; Panagia et al. 1991). Interesting comparison can also be made with the work of Plait et al. (1995). They perform a RLM deconvolution, but on a F501N summed image, thus of higher signal-to-noise ratio than our single exposure image. Jakobsen et al. (1994) present also images obtained with the COSTAR corrected HST. Comparison will be made even though the structure and photometry of the ring has changed between the two observations.
The ratio of the intensity of the ring to the one inside could tell us whether it is really a ring or a shell (see Dwek & Felten 1992). Since we have shown that the RLM deconvolution artificially increases the intensity of the peaks in the ring, the ratio (between 10 and 100) found by Panagia et al. (1991) is thus probably overestimated. The IDEA solution indicates that this ratio might be in the range 30-40, relaxing a little bit the constraint in favor of a ring. Moreover, Plait et al. (1995) argue that such a ratio results from a ring, and not from a shell.
The fluxes of the central star and of the ring has been
measured within the support used in IDEA. For the raw image we
used the values estimated
by Jakobsen et al. (1991) for comparison.
From Table 1 (click here) it is clear that IDEA recovers a
large fraction of the energy from the wings of the
PSF. Note that it is coherent with ground based
observations obtained at the same epoch (see R.
Cumming cited in Plait et al. 1995)
giving a ring intensity of 
erg s
cm
. In addition, the flux of the central star
obtained with IDEA is compatible with a unique decay
rate between days 1275 and 2522 (see Jakobsen et al. 1994).
Table 1: Absolute fluxes in 
erg s
cm
. The value for the raw image has been taken
from Jakobsen et al. (1991)
The profile of the central star of the supernova presented in
Sect. 3.1 shows the presence of an envelope around the star.
This has already been found by Jakobsen et al. (1991).
We measure similar values of the total radius for the
envelope profile: 
, that is (at a distance of 50 kpc)
about 35 light-days, corresponding to an expansion speed of
km s
.
We also note that the central star on the IDEA image is elongated
(PA 
). The elongation is absent on the other
deconvolution images. This value is not very precise
because of the anisotropy of the PSF, but is corroborated
by the results with speckle
interferometry providing an elongation with PA = 
(Papaliolios et al. 1989) and with COSTAR-corrected FOC images
by Jakobsen et al. (1994): PA 
.
Panagia et al. (1991) noted the clumpiness of the light
distribution in the ring. This might have three
origins (Dwek & Felten 1992):
1) anisotropy in the ionising photon propagation from the
star to the ring;
2) clumpiness of the ambiant medium acting on the ring propagation;
3) heterogeneities within the ring (density or emissivity).
On the RLM solution, the ring is really knotty,
probably due to over-resolution. Further evidence
is provided by the faint superposed star, located at PA = 
on the ring, described in Plait et al. (1995).
Its intensity with respect to that of the ring is artificially
increased on the RLM image.
On the opposite, the ring in the IDEA solution is rather
continuous with a few regions (one especially) of increased
brightness defining the ``clumps''. This shows that the ionising
radiation is isotropic and that the ambiant medium has not
disturbed the ring too much during its expansion. However, the
south part of the ring is flatter than an ellipse, indicating
probably a resistance of
the ambiant medium. Some filaments in the ring (note the one to
the East) suggest a certain level of heterogeneities.
Note also how the northern and southern parts of the ring
are brighter than the eastern and western parts. This is probably
explained by the inclination of
the ring with respect to the line of sight (Jakobsen et al. 1991).
The relative smoothness of the ring found with IDEA is
confirmed by the modelisation of the UV emission
(Plait et al. 1995) which slightly
favors the smooth ring model. The IDEA solution is close
to the RLM deconvolution of Plait et al. (1995) (applied to a higher
signal-to-noise ratio image).
The IDEA image is also similar to the COSTAR-corrected images
presented by Jakobsen et al. (1994) even though the ring structure might have
evolved between the two epochs.
From the deconvolutions it can
be seen that the center of the galaxy is
elongated indicating the presence of a nuclear jet, confirming
the statement of Boksenberg et al. (1992). However, the
IDEA deconvolution suggests that the knot at 
from
the nucleus might be an artifact:
it is less prominent on the IDEA deconvolution than on the RLM
deconvolution and is exactly at the position of a ring created
by the PSF.
On the RLM deconvolution, knot D appears as 2 knots, whereas on the IDEA deconvolution it has the appearance of a filament. The filamentary structure of knot E is also more evident on the IDEA deconvolution. This is consistent with the general filamentation found in extragalactic jets (see Sect. 4.3 for another example). The continuity of the jet seems less obvious on the IDEA deconvolution, but there is a slight hint for a double filament in knot F. This is also true for the outer part of the jet (visible on the jet image). The IDEA features are essentially confirmed on WFPC2 images from HST (Ford et al. 1994).
Boksenberg et al. (1992) performed photometry on the raw image because of the lack of confidence in the deconvolution result. We use the IDEA solution after calibration of our raw data in the same manner. We present (Fig. 7 (click here)) the profile of the jet obtained within perpendicular slits. Its width is defined by the support used for the deconvolution.
Figure 7: Profile of the jet integrated over
its whole width in a one-pixel-wide slit
This can be compared with Fig. 4 (click here) of Biretta
et al. (1991) compiled
using data at different wavelengths and a lower resolution.
If we integrate this profile along the jet
to have the total intensity of the knots, for the whole jet
(1
to 19
from the nucleus) we find 1.2 mJy.
From knots F to C, our value
(1.1 mJy) is significantly higher than that found by
Perola & Tarenghi (1980) from IUE data (
mJy).
Perola's values are however compatible
with ours knots A and B (0.81 mJy).
That would indicate that knot C is not entirely in the
IUE 10
20
window.
The image presented in this paper is the first image of this jet
at such a short wavelength. The jet is very faint, but the galaxy
background is nearly absent making galaxy substraction unnecessary.
As can be expected, the structure of the jet at this wavelength
is identical to that at 320 nm (Macchetto et al. 1991),
except may be for ``the fatter structure'' in which
the double-stranded filament seems to be embedded. This could be
naturally explained by the lower signal-to-noise ratio of the
220 nm image. We note that the B knot (Fraix-Burnet et al. 1989)
appears to be filamentary on the IDEA
deconvolution. This could indicate that the double-stranded
filament is continuous from 
from the nucleus outward.
There is no trace of the ``blue knots'' found by Fraix-Burnet
et al. (1989). But because of the
artifacts left by the non-perfectly adapted PSF within
about 
from the nucleus, the detection of such faint
features is certainly hopeless.