Loral 5 is a Lesser-thinned, two-side-buttable, Loral 2kx2k CCD. The backside passivation is acheived through a platinum flash gate. Cosmetically, it is not a very good device, since there are several nasty bad columns and a few deep traps. There are some peculiarities of the CTE at high signal levels in the upper area of the device, but the few hundred rows runnning parallel to the horizontal register are relatively unaffected. This, combined with the excellent UV response of the device means that it is best used on the Coude' spectrograph Gecko.
During optimization, we discovered that Loral 5 is susceptible to contamination in a manner similar to the ill-fated Orbit device, now scrapped. This takes the form of a non-uniformly reduced sensitivity, particularly in the blue, of an F-shaped region in the middle of the CCD. This is counter to our expectations because a PPt surface layer is supposed to be relatively unaffected by surface contamination. We suspect that the contamination is caused by water vapour. Dan McKenna has discovered that it is possible to reverse the effect by repeated soaks in oxygen. Nevertheless, we are very cautious about this problem and encourage observers to obtain frequent calibration data and, upon reducing their data, to cross-check repeatedly between different periods in their observations.
The device is a little noisy, at about 8 electrons/pixel, but this is more than compensated by the excellent QE. CTE is also somewhat poor, being no better than 0.99995 in both directions, but the reduced region parallel to the serial register as used in Gecko means that this is unlikely to affect observations.
