After we overcame the two problems discussed above, we obtained some good exposures of spectra. It was immediately apparent that the response of the detector was very nonuniform. We suspected that the photocathode had been damaged, rather than the CCD. This hypothesis was confirmed after the flight, see §7.5 and §7.6. The entire picture format was filled with spots where the response inside each spot was about half as large as the interspot regions. It was also clear that the overall sensitivity of the system was much less than expected. Post-flight analysis of spectral intensities showed that the interspot regions had a photoefficiency about 1/3 what it should have been (as judged by the IMAPS sensitivity on earlier sounding rocket flights). The overall impression one had when viewing a spectrum was that it was being viewed through a mud splattered window: see Fig. 4 in §7.2.
To cope with the photocathode damage, we invoked two strategies in our observing. First, we limited our observations to very bright, unreddened stars. To some degree, this changed the character of our scientific program, but there were still plenty of worthwhile targets to choose from. Second, for each star we took exposures that had slightly different pointings, so that certain features would not consistently fall inside or, worse yet, on the edge of a spot. An explanation of how we coped with the modulation of the spectrum in the data reduction is given in §7.5.
We feel that the most likely explanation for the damage was our inability to control the humidity inside the payload for a period of about 3 days before the Shuttle launch. Up to the time that the payload bay doors were closed, we could maintain a purge of dry nitrogen. However at the time of door closing, we had to remove this purge since access to IMAPS would no longer be possible. During the 3 days that the payload was not under purge, air of normal humidity in the bay could seep inside through a decompression vent and small gaps in the front aperture doors. As explained in §3.4, the photocathode is open to the instrument's interior environment, and the active material, KBr, is strongly hygroscopic. We feel that the damage occurred either during the last day or so before launch, or when adiabatic decompression of the air inside the payload produced a fog that lasted for about 15 seconds during the ascent of the Shuttle.