In order to test the accuracy of the photometric calibrations over the whole mosaic, we observed a few selected standard fields in each of the eight CCDs.
For the broad-band, we used the Landolt field PG1525-071, which contains only five stars with different (B-V) colours that range from -0.198 to 1.109. In this way it has been possible to check the photometric homogeneity over the eight CCDs and its dependence from the star colour. The results for the B V R I bands are displayed in Figures 6 and 7 and clearly show that there is no colour dependence.
Moreover, no significant differences are found among the three observing runs. On the other hand, it is clear from Figures 6 and 7 that there is an offset of the instrumental magnitude among the eight CCDs. This effect is summarised in Figure 8, where the flux of the standard stars measured in each CCD relative to the mean value over the 8 CCDs is shown. Within the errors, the trend is the same in the different bands (B, V, R, I) and different times (from April 1999 to April 2000).
A similar test was also done with the intermediate-band filters using the spectro-photometric standards Eg 274, Hill 600 and LTT 6248. These stars were observed in all the three runs in each one of the eight CCDs of the mosaic, in order to check whether there are differences in the factor from one CCD to another. From Figure 9, where the relative flux of the standards measured in each CCD relative to the mean value is shown, we note that the trend does not change significantly from one band to another and in the different runs. Note also that the photometric offset appears to be slightly stronger in the intermediate-band filters with respect to the broad-band ones.
In conclusion, we find an offset in the flux of the stars which introduces an average uncertainty of 3% in the broad-band B V R I filters and and uncertainty of 5% in the intermediate-band filters. This effect is quite constant in time.
This photometric offset, may be attributed to an additional-light pattern caused by internal reflections off the telescope corrector. Such pattern, with an amplitude that depends also on the exposure time, is present in every image, including flat-fields. As such, it should be subtracted from every image before applying the flat-fielding correction, using an adequate scaling factor.
In order to be able to subtract the additional light, one must know the additional-light pattern. Unfortunately, for the WFI at the ESO 2.2m telescope, such pattern is not well determined yet. The ESO 2.2m team is currently working on this problem.