2. Observations and reduction

2.1. Observations

The observations were carried out with the ATCA in July 1992 using the 6D configuration including baselines between 77 m and 5878 m. The observed region extends from to in R.A. and to in Dec. (positions are in J2000). This area of 12 square degrees has been divided into 6 8 fields. The centre position of each of the 48 fields is given in Table 1 (click here). Three observations were taken per field with a time seperation of about 4 hours. The total integration time was 12 minutes for each field. Observations at two frequencies, 1.38 GHz and 2.378 GHz, have been carried out simultaneously. The angular resolution at 1.4 GHz is 7'' and at 2.4 GHz it is 4''. The primary beam has a HPW of 32' at 1.4 GHz and 22' at 2.4 GHz. The bandwidth of 128 MHz at each frequency has been separated into 32 spectral channels. The primary flux and bandpass calibrator was 1934-638 (16.2 Jy at 1.4 GHz and 13 Jy at 2.4 GHz), the secondary (phase and gain) calibrators were 0407-658 and 0252-712. For data reduction the Astronomical Image Processing System (AIPS) in the special ATCA version was used.

 

Field No. R.A. [h m s] C-level [mJy] Field No. R.A. [h m s] C-level [mJy]

(J2000) 21 cm 13 cm (J2000) 21 cm 13 cm

1D1 05 20 00.0 -67 15 00 7.5 5.5 1D5 05 20 00.0 -69 15 00 5.1 2.2

2D1 05 25 35.2 -67 15 00 5.5 4.5 2D5 05 25 35.2 -69 15 00 11.0 3.3

3D1 05 31 10.5 -67 15 00 5.5 2.0 3D5 05 31 10.5 -69 15 00 4.0 2.0

4D1 05 36 46.1 -67 15 00 6.0 6.0 4D5 05 36 46.1 -69 15 00 35.0 13.0

5D1 05 42 21.4 -67 15 00 3.0 2.0 5D5 05 42 21.4 -69 15 00 13.0 6.6

6D1 05 47 52.0 -67 15 00 2.0 2.0 6D5 05 47 52.0 -69 15 00 6.0 2.1

1D2 05 20 00.0 -67 45 00 8.0 3.0 1D6 05 20 00.0 -69 45 00 6.0 5.2

2D2 05 25 35.2 -67 45 00 5.5 3.1 2D6 05 25 35.2 -69 45 00 not complete

3D2 05 31 10.5 -67 45 00 9.0 8.7 3D6 05 31 10.5 -69 45 00 6.0 2.5

4D2 05 36 46.1 -67 45 00 5.0 2.1 4D6 05 36 46.1 -69 45 00 7.0 6.6

5D2 05 42 21.4 -67 45 00 5.0 1.9 5D6 05 42 21.4 -69 45 00 12.0 4.1

6D2 05 47 52.0 -67 45 00 7.0 2.2 6D6 05 47 52.0 -69 45 00 4.5 2.4

1D3 05 20 00.0 -68 15 00 5.5 2.0 1D7 05 20 00.0 -70 15 00 3.0 4.6

2D3 05 25 35.2 -68 15 00 6.0 2.4 2D7 05 25 35.2 -70 15 00 4.0 3.0

3D3 05 31 10.5 -68 15 00 3.8 2.3 3D7 05 31 10.5 -70 15 00 4.5 2.4

4D3 05 36 46.1 -68 15 00 3.0 2.3 4D7 05 36 46.1 -70 15 00 4.5 3.1

5D3 05 42 21.4 -68 15 00 8.0 4.0 5D7 05 42 21.4 -70 15 00 3.0 3.0

6D3 05 47 52.0 -68 15 00 4.0 3.5 6D7 05 47 52.0 -70 15 00 5.0 3.4

1D4 05 20 00.0 -68 45 00 3.0 2.3 1D8 05 20 00.0 -70 45 00 7.0 4.8

2D4 05 25 35.2 -68 45 00 3.5 2.5 2D8 05 25 35.2 -70 45 00 14.0 7.4

3D4 05 31 10.5 -68 45 00 4.0 2.3 3D8 05 31 10.5 -70 45 00 4.0 2.7

4D4 05 36 46.1 -68 45 00 8.0 2.6 4D8 05 36 46.1 -70 45 00 4.5 2.9

5D4 05 42 21.4 -68 45 00 10.0 2.3 5D8 05 42 21.4 -70 45 00 6.0 3.9

6D4 05 47 52.0 -68 45 00 11.0 5.2 6D8 05 47 52.0 -70 45 00 7.0 3.4

 

2.2. The maps

The images were made using the AIPS routine MX, which combines the Fourier Transform imaging with the deconvolution of the synthesized beam. The deconvolution process uses the clean algorithm of Clark (1980) with 500 iterations. We chose the inner 18 channels of the frequency band to compute the images using the technique of multi-frequency synthesis.

The sparse uv-coverage of snapshot observations makes imaging difficult due to poor dynamic range in the maps. Extended emission is poorly sampled and cannot be adequately cleaned. We have tested several ways of selecting the data for mapping. Best results were obtained by restricting the data to baselines longer than 3 to exclude the poorly sampled extended emission. In the region around 30 Doradus only baselines above 6 are useful. We taper the uv data with a Gaussian weighting function with half width to 30% level of at 1.4 GHz and at 2.4 GHz. This gives synthesized beamwidths of about 10'' at 1.4 GHz and 6'' at 2.4 GHz (see Table 2 (click here)). The shape of the synthesized beam depends on the declination. The beam has a circular shape only for fields with declination between and . For higher and lower declination, the shape of the synthesized beam is more and more elongated, in the extreme case (fields at ) it is about at 1.4 GHz. The maps are 49' 49' at 1.4 GHz and overlap slightly. At 2.4 GHz the map size is 31' 31'. The maps have an rms noise level between 0.4 mJy and 1 mJy at 1.4 GHz and between 0.2 mJy and 0.6 mJy at 2.4 GHz. The theoretical rms values are 0.17 mJy at 1.4 GHz and 0.23 mJy at 2.4 GHz based on receiver noise alone. The excess is due to sidelobes from confusing sources aggravated by our limited dynamic range.

 

21 cm 13 cm image size

cellsize 2.9'' 1.8'' uvmin (minimum spacing)

source size <54'' <34'' <73'' <34''

number of CLEAN iterations 500 500

CLEAN loop gain 0.1 0.1 - HPW of synthesized beam

HPW of primary beam 32' 22'

 

To correct the images for primary beam attenuation the program PBCOR was used with the standard AT primary beam shape. We tested this by comparing the peak flux densities of sources detected separately in different overlapping fields. We find that the ratio of the flux densities of the same source in two fields is independent of the difference of its distances from the pointing centres. Only for some very weak sources with a large distance d (above 20') from the field centre does this ratio differ much from one, indicating a higher uncertainty of the primary beam correction for ()² > 780.

A mosaic of the 48 final images is shown in the first figure. 3 4 fields of 49' 49' each have been combined with the AIPS routine LTESS which takes the primary beam attenuation into account. The resultant four mosaicing fields have then been combined with the task COMB, averaging the overlapping areas. The image shown is smoothed to about 1' resolution.

  
Figure 1: Mosaic of all 48 images at 1.4 GHz. The cross feature is the overlapping area of the four combined LTESS-fields. Unreal structures with high intensity can be seen in the area of 30 Dor (05h38m40s, ). At the top right corner strong sidelobes are produced by a high intensity source outside the snapshot field

2.3. Source finding

Very brief observations in the snapshot mode are ideal to study sources which are bright and compact, but the high sidelobe levels of beams synthesized from snapshots exacerbate the problems created by confusing sources. Some sidelobes cannot be completely eliminated in our cleaning process. The fields near 30 Dor are particularly compromised. Due to such sidelobe structure, the minimum detectable source flux is raised. We determined three criteria to decide which objects are real:

• The source must have signal exceeding 5 times the rms noise of the field, and
• the source must be stronger than any remaining sidelobe structure in the field, and
• the source has to be stronger than the absolute flux density value of the largest negative peak in the field.

We used the highest of these values to determine the completeness level, C, above which the sources in the field are accepted as real. These completeness levels are listed in Table 1 (click here) uncorrected for primary beam attenuation. Most fields are complete to about 6 mJy peak flux density at 1.4 GHz and to about 3 mJy at 2.4 GHz. In field 2D6 confusion by the bright extended emission of SNR 0525-696 makes source finding for this area impossible.