3 Observations, data reduction and analysis

 

 

Table 2: Antennas participating in the VLBI observations. Note that the three Tidbinbilla antennas are at essentially the same location and only one antenna is used at any one time, depending on availability. The SEFD (system equivalent flux density) is given by the ratio between system temperature ($^\circ$K) and gain $^\circ$K/Jy.

Key	Antenna	Size	SEFD (Jy)
		(m)	2.3 GHz	8.4 GHz
P	Parkes(ATNF)	64	90	90
M	Mopra (ATNF)*	22	400	400
C	Narrabri-CA (ATNF)	22	400	400
3	Tidbinbilla (DSS43)	70	15	20
5	Tidbinbilla (DSS45)	34	165	130
2	Tidbinbilla (DSS42)	34	100	130
H	Hobart (Univ.Tasmania)	26	750	650
G	Perth (ESA)	15	3300	-
E	Harteebeesthoek	26	400	950

$\parbox{14.5cm}{$^*$\space \footnotesize{The Mopra antenna became operational in late 1991.}}$

 

 

Table 3: Observations and images presented in this paper. The dynamic range indicates the ratio of the peak brightness in the image to the peak in the residual map. The $\sigma _{\rm rms}$ entry refers to the residual rms noise in regions devoid of any source emission. Entries in the last column refer to the figure in which each image appears.

Source	Epoch	Freq.	Duration	Stations	$B_{\max}$*	Beam	Dynamic	$\sigma$	Figure
		GHz	(hours)		$M\lambda$	mas   $^\circ$	Range	mJy/beam	Ref.
0023-263	1991.34	2.29	11.5	32PHC	10.5	31 $\times$ 14  -82	60:1	4.5	3
	1991.90	8.42	11	32PHCM	29.9	13 $\times$ 5   -77	40:1	1.7	3
	1992.05	4.99	6	MERLIN	3.6	168 $\times$ 86  -2	130:1	3.5	4
0252-712	1993.14	2.29	10.25	2PHCMGE	8.9	32 $\times$ 19  -73	65:1	8	5
1151-348	1991.34	2.29	11.75	35PHC	10.7	29 $\times$ 13  89	70:1	4.5	6
	1992.24	8.42	11	35PHCM	30.3	11 $\times$ 5   81	35:1	3.6	6
1306-095	1993.14	2.29	10.5	32PHCGE	10.7	48 $\times$ 15  -73	20:1	5.7	7
1814-637	1993.14	2.29	12	32PHCMGE	9.5	32 $\times$ 17  -86	50:1	5.8	8
1934-638	2.29, 4.85 & 8.42 GHz images in King (1994) & Tzioumis et al. (1996)								10
2135-209	1991.55	2.29	12	3H	3.6	Modelfit only
	1992.91	4.99	6.25	MERLIN	3.6	163 $\times$ 34  10	130:1	2.3	9
* Maximum baseline with significant dectection of each source. No detections to Perth (G) or Hartebeesthoek (E).

Figure 3: ( Left) PKS 0023-263 at 2291 MHz from the SHEVE array. The peak level is 1.07 Jy/beam and contours are shown at -2, -1, 1, 2, 4, 8, 16, 35, 65, 80% of the peak. ( Right) PKS 0023-263 at 8419 MHz from the SHEVE array. The peak level is 0.41 mJy/beam and contours are shown at -2, -1, 1, 2, 4, 8, 16, 35, 65, 80% of the peak.

The VLBI observations of all sources were made with the SHEVE network (Preston et al. 1993; Jauncey et al. 1994) between April 1991 and February 1993, using the MK2 recording system (Clark 1973). The characteristics of the antennas and receivers in the SHEVE network are listed in Table 2. The details of each observation are given in Table 3.

All sources were observed at 2.3 GHz, but for 2135-209 only a single baseline (Tidbinbilla-Hobart) was obtained in a 12 hour observation. The sources 0023-263, 1151-348 and 1934-638 were also observed at 8.4 GHz and 1934-638 at 4.9 GHz as well. The GPS source 1934-638 was observed as part of another program and the details of the observations and results are presented elsewhere (King 1994; Tzioumis et al. 1996) while some discussion of the source properties is presented here.

The tapes recorded at each station were processed at the Caltech-JPL Block2 correlator in Pasadena and the NRAO AIPS package was used for global fringe fitting. The amplitude calibration and editing of the visibilities were subsequently performed using the Caltech VLBI package. Self-calibration and imaging of the data were performed using DIFMAP (Shepherd et al. 1994) and the AIPS packages. The single-baseline data for 2135-209 was used to obtain a simple model of the source using the program MODELFIT from the Caltech VLBI package.

On the very long baselines from the antennas at eastern Australia (ATNF, DSN, Hobart) to Perth (3000 km) and Hartebeesthoek (9500 km), only 1934-638 and VLBI calibrator sources were detected, i.e. all other sources are completely resolved out at these spacings and sensitivities. This implies that there are no very compact components ( $\theta \ll 3$ mas) with flux density larger than about 40 mJy, the detection limit from the intercontinental baselines.

A "typical'' uv-coverage from the SHEVE array is presented in Fig. 2, for the source 0252-712 at 2.3 GHz. Note that all sources except 1934-638, were not detected on the long baselines to Perth or Hartebeesthoek and uv-tracks are not shown.

The limited number of antennas in the network and the lack of short baselines limit the dynamic range achievable from these observations. In particular, for many of the sources the total flux density in the VLBI images is significantly lower than that measured with lower resolution observations. This implies that low surface brightness extended structures with angular sizes between 0.1 and 1 arcsec are often present but invisible by the VLBI interferometers, as discussed in the next section on each source.

Figure 4: PKS 0023-263 at 4996 MHz from the MERLIN array. The peak level is 1.41 Jy/beam and contours are shown at -6, -3, -1.5, -0.75, 0.75, 1.5, 3, 6, 12, 24, 48, 96% of the peak.

For 0023-263 and 2135-209 we also present MERLIN observations at 5.0 GHz, which provide information on more extended structures, not detected by the VLBI observations.

The new images of the six sources observed in this program are presented in Figs. 3 through 9. Restoring beams together with the noise measure on the image plane are given in Table 3.

The overall size of most sources in this sample was not previously well determined at radio wavelengths. The largest angular/linear size for each source was estimated from the images, adopting a "low resolution'' approach (cf. Dallacasa et al. 1995). Most of the sources show distinct and well separated "lobes'', and extended structure can be seen in most of them. Their sizes and flux densities were estimated from the images and are summarised in Table 5. The component angular dimensions were generally estimated directly from the images (except were otherwise stated) by considering the lowest reliable contour; they are given as full-width axis sizes. This approach was adopted because almost all components show extended structures and it is then difficult to fit single Gaussian components. The measurements provided values well in excess of the beam size, and they were not deconvolve given they would not change significantly; therefore, all sizes in Table 5 should be treated as upper limits.

The separation between the strongest component in each of the lobes was also measured from the images. This separation is a very sensitive parameter since it is well determined by the "beating'' in the visibilities and can be used to search for any source expansion in time, even with sparse uv coverage. The results are presented in Table 4.

For the two sources observed at 2.3 and 8.4 GHz, a two-point spectral index for each lobe has also been determined.

 

 

Table 4: Overall angular and linear sizes of the sample sources and separation of strongest components in the two lobes.

			Component separation
Name	$\theta_{\max}$	Lin. size	$\theta$	l	Orientation
	mas	pc	mas	pc	$^\circ$
0023-263	680	1970	654	1900	-34
0252-712	240	900	145	540	7
1151-348	170	425	91	228	72
1306-095	460	1600	373	1290	-41
1814-637	410	328	239	191	-20
1934-638*	70	140	42	84	89
2135-209	$\sim$250	$\sim$1000	167	650	52
* Tzioumis et al. (1996).

 

 

Table 5: Size and flux density of the components in the images. Component angular sizes are estimated from the images (full width), except were otherwise noted.

Source	Comp.	$\nu$	$\theta_1$	$\theta_2$	Lsize1	Lsize2	$S_{\rm peak}$	$S_{\rm tot.}$	$\alpha^{8.4}_{2.3}$
		GHz	mas	mas	pc	pc	Jy/beam	Jy
0023-263	NW	2.3	80	75	230	220	1.07	1.6	-0.7
		8.4	40	23	110	65	0.41	0.65
	SE	2.3	100	90	290	260	0.33	1.3	-1.3
		8.4	60	30	175	85	0.047	0.25
0252-712	North	2.3	110	90	400	330	1.14	2.1
	South		120	90	440	330	0.47	1.4
1151-348	NE	2.3	85	45	210	110	1.57	2.1	-0.6
		8.4	40	30	100	75	0.437	0.94
	SW	2.3	85	45	210	110	0.78	1.24	-1.0
		8.4	35	20	90	50	0.12	0.36
1306-095	SE	2.3	110	70	380	240	0.304	0.76
	NW		80	50	275	170	0.107	0.15
1814-637	North	2.3	90	80	70	65	1.58	2.8
	South		130	70	100	55	0.380	1.8
1934-638a
2135-209b	NE	5.0	25	9	100	350	1.02	1.1
	SW		90	40	350	155	0.2	0.36

^a Components and spectra for 1934-638are given in Tzioumis et al. (1996).
^b Component sizes from fitting Gaussians (FWHM) since the beam is comparable to the source size and components cannot be estimated graphically.