Astron. Astrophys. 343, L35-L39 (1999)

2. Observations and data reduction

2.1. VLT imaging

Broad-band B and Ly narrow-band images of 0047-2808 were obtained on the night of 1998 August 30, with the VLT UT1 test camera as part of the Science Verification programme (Leibundgut et al. 1998). The CCD pixel size is . However the CCD was binned , so the pixel size in all the frames was . The narrow-band filter has a central wavelength 5589Å and width 20Å FWHM. Integration times were sec (B) and sec (Ly). The seeing was . Procedures followed for bias subtraction and flatfielding were mostly standard. However the flatfielded narrow-band images required a correction for large-scale gradients. This was achieved by firstly combining the deregistered frames, clipping out objects. The resulting frame was smoothed, and normalised, and the flatfielded frames were divided by this correction frame.

2.2. UKIRT imaging

Broad-band K images of 0047-2808 were obtained with the UKIRT IRCAM3 instrument on the nights of 1997 September 12 and 13. The pixel size was . The final image is a mosaic from two positions, one centred on 0047-2808, total integration time sec, and another at a position to the SSW centred on a second distant elliptical, total integration time sec. At each position several sequences of 9-point dithers were summed. The seeing averaged . The data were flat-fielded using a sequence of twilight sky exposures, and then an appropriate sky frame, formed from a running median filter through the stack of images, was subtracted from each data frame, and the resulting frames registered and summed.

2.3. Results

Fig. 1 shows the rgb colour image resulting from combining the B (=b ), Ly (=g ), and K (=r ) images. The ring stands out strongly in green because of the strong Ly line in the narrow-band filter, while the lensing galaxy is very red and is visible inside the ring. A minimum fit of a de Vaucouleurs model for the light profile of the lensing galaxy in the K-band image was computed by convolving two-dimensional profiles with the psf, measured from a star in the frame. (The K-band image is best for fitting the galaxy profile because the ring is not detected at this wavelength, and the contrast between the galaxy and the ring is maximised.) The model was then convolved with the Ly-band psf, scaled to the central counts in the Ly image, and subtracted. The resulting image of the ring, rebinned to a pixel size of , is shown in the top left-hand panel of Fig. 2.

Fig. 1. True-colour image of the field of the Einstein-ring gravitational lens 0047-2808, showing a region . North is up and East to the left. The image is a rgb combination of the VLT B and narrow-band images and the UKIRT K image. Because of the strong Ly emission line in the narrow-band the ring () appears as green. The lensing galaxy () is red because the galaxy is bright in the K band but faint in the B and narrow bands. There is a second massive elliptical galaxy in the field, visible as the brightest red image in the SW quadrant, which has a redshift .

Fig. 2. False-colour images displaying the results of modelling the mass distribution in the lens. The size of each square image is on a side. North is up and East to the left. The top LH image shows the VLT Ly image after subtraction of the image of the lensing galaxy, and rebinned to a pixel size of . The lower LH image shows the model of the source, and the 3-image and 5-image caustics. The lower RH image shows the image that would be observed at infinite spatial resolution, and the upper RH image is the lower RH image convolved with the psf to emulate the observing conditions. The upper RH image therefore models the upper LH image with zero noise. The good correspondence of the two upper images gives a measure of the accuracy of the mass model.

© European Southern Observatory (ESO) 1999

Online publication: March 1, 1999
helpdesk.link@springer.de