|Annu. Rev. Astron. Astrophys. 1991. 29:
Copyright © 1991 by . All rights reserved
2.6 Dark Current
Our data in Figure 4 show a sharp break in the darkest background at a wavelength around 1600 Å near the middle of the figure. The break arises because Figure 4 has been made from the conjoined data from two independent Johns Hopkins spectrometers. The dark current was different in the two spectrometers. The central question is, what is left once this dark current is subtracted? That dark current differences are so readily apparent in the original of Figure 4 tells us at once that the accuracy of dark current subtraction is a serious matter. How serious depends quite fundamentally on the field of view of the instrument. This extremely important point must be kept in mind when considering the reliability of claimed measurements of the diffuse ultraviolet background discussed below. It can be appreciated by a numerical example. Suppose the dark current is accurately known to be 1 count per second. There is often not adequate opportunity to measure this with precision in flight, so suppose only 0.9 counts per second are subtracted from the observed signal. The artificial excess will then erroneously be attributed to diffuse ultraviolet background radiation. How large is the resulting error? Suppose, reasonably, that the efficiency is 0.01 counts per photon, that the area is 10 cm2, and that the passband is 100 Å. Then the deduced spurious diffuse background, before taking into account the instrumental field of view, will appear as a flux of 0.01 photons cm-2 s-1 Å-1. Now for comparison, typically reported values of the ``true'' cosmic background, at high galactic latitudes, are 300 units. If our instrument field of view is 4.4 x 10-2 steradians, as with William G. Fastie's Apollo 17 experiment, our spurious background will translate into 0.2 units, which is negligible. If, on the other hand, the field of view is as small as 10-5 steradians - and this is not uncommon in past experiments (62, 75, 76, 101) - the spurious background will be 1000 units, which of course is very serious. This problem is even more serious if the experiment is in Earth orbit rather than in interplanetary space, because the dark count rate can be highly time-variable in earth orbit due to variable impact particle flux from the Earth's radiation regions such as the South Atlantic Anomaly.