The same flat field technique, as utilized for the WL flat fields, was used for the lowest contrast ultraviolet (UV) wavelengths at 1700 and 1216A. However, even when the images are defocused by most of the range of the mechanism, quiet sun bright points still show contrast of about a factor of 3 above the average level. As a result, there is always "solar noise" in the derived flats at the ~5 percent level. Since the fall of 1998, the UV flat fields have shown a decrease of sensitivity in the center of the CCD, which has grown steadily over the first 6 years in orbit. Figure 1 shows a 1700A smoothed flat field for 7-Jul-99 with a decrease of about 83% in the center. Based on preflight experiments, the loss was expected to be caused by extreme-ultraviolet (EUV) photons damaging the thin layer of lumogen, which coats the face of the CCD and which makes the CCD sensitivity to the 1700 and 1216A photons. It has been assumed that the centering of active region targets on the CCD is responsible for the localized damage because the pattern looks very much like that of the total EUV dose images derived by adding up all of the exposures. Smoothed flat fields, such as that shown in Figure 1, interpolated in time, will be utilized in correcting the gain in the TRACE UV images.

The Kuhn-Lin flat field results have been well confirmed by analyzing the synoptic disk center images taken (nearly) every other day of the mission at all of the TRACE wavelengths. For the first five years of the mission the images were binned into 16-arcsecond "super-pixels" for convenience and noise reduction. For each super-pixel linear, quadratic, and eventually cubic fits of intensity (DN/sec) versus time were made. These fits demonstrate the decreasing sensitivity in the center of the CCD with time. In fact the synoptic images are utilized in creating the gain corrections for the sensitivity loss. The flat field images are normalized and then smoothed, using an optimized Wiener filter in Fourier space. Such an image in 1700A for 1-May-02 is displayed in Figure 2. The decrease in the image center has reached about 69%. It occurred almost 3 years after the smoothed flat field shown in Figure 1. The smoothed flat field images throughout the mission have then been analyzed together with temporally nearby synoptic disk center images to provide a set of gain correction parameters over time for the 4 UV channels at 1700, 1600, 1550, and 1216A. The 1700A flat field images serve as the basis for the lumogen loss corrections in all but the WL channel. The analysis determines a set of time- and wavelength-dependent power and multiplier parameters interpolated along with a corresponding smoothed flat field image to correct for the sensitivity loss of the TRACE image at a given time and for a given wavelength during the mission.

The lumogen on TRACE CCD continued to demonstrate an exponentially decreasing sensitivity over time until 2004 when it was found that annealing the CCD, warming it to about +40 deg C for about 3 days, restores a significant fraction of the lost sensitivity. In fact the annealing of March 2004 and February 2005 have restored the quantum efficiencies of the various wavelengths back to approximently their 2001 levels. Figure 3 shows a 1700A smoothed flat field for 6-Jun-05 with a center decrease of about 67%, about the same level as that for the May 2002 flat field in Figure 2. The relative response near the center of the CCD for all of the TRACE passbands, with the exception of WL, throughout the mission to date can be seen in Figure 4 to be quantitatively similar to that for 1700A. The annealings have increased the central sensitivity by more than 40% in some of the wavelengths. It is planned to continue the annealing periods twice a year.