Princeton University Observatory

Peyton Hall
Princeton, New Jersey 08544

September 8, 1982

TO: J. B. Rogerson
FROM: L. Spitzer
SUBJECT: Sensitivity Loss in Copernicus

Summary. Comparison of Copernicus signals from a sealed photomultiplier tube (with a LiF window) with those from the open-face U1 photomultiplier shows that at least 80 percent of the large sensitivity loss observed at 1340 Å with U1 during the final few years of Copernicus operation must be attributed to degradation of the open-face tube, rather than to decrease in reflectance from the mirrors and grating.

In my memorandum of January 5, 1982, I summarized evidence that bombardment by atmospheric atoms produced the high voltage glitches and, to some extent, the sensitivity degradation experienced by Copernicus. The detailed mechanism of sensitivity degradation was not discussed.

Loss of sensitivity by far-UV open-face photoelectric detectors is potentially a very serious problem for the projected far-UV satellite, now designated as the FUSE program. Don York's request that I forward to Al Boggess any memoranda summarizing our Copernicus experience in this area has reawakened my interest in this problem.

Several years ago Walter Upson pointed out that the sensitivity loss of the closed U3 phototube through orbit 30,000 (March 1, 1978) was "significantly less than a factor two"; in contrast, the sensitivity of the U1 open-face photomultiplier, measured at 1340 Å, was about a factor two. As you know, the nominal wavelength of the U3 tube is about 1340 Å, and while the mirror directing light to this tube is covered with a metal disc (so that this tube can be used to measure the background produced by energetic particles), a weak additional signal is received by this tube when the telescope is pointed at a very bright star; the variation of this signal with stellar apparent magnitude and spectral type is consistent with the assumption that U3 is measuring stellar photons at about 1340 Å.

This effect pointed out by Walter is of great importance, since it suggests that the loss of sensitivity was not produced by the optical system --primary mirror, secondary mirror and concave grating --which feeds these two phototubes. Instead, the loss of sensitivity of U1 may be attributed in large part to degradation of this open-face phototube, presumably of the KBr coating on the photocathode; degradation of the BeCu dynodes in the multiplier section would not account for the increased sensitivity loss at the shorter UV wavelengths. Since the U3 photomultiplier is a conventional sealed unit, with a CsI photocathode and a LiF window, the presence of contamination vapours in the spectrometer would not be expected to degrade the performance of this tube, though other factors could produce degradation.

Since the U1 degradation by 1978 was not very large, this argument did not seem very decisive. To obtain more definitive data, I have now examined the U3 signal from the bright star Beta Cen as a function of time until orbit 41,000 (June 1980), and compared this with the normalized U1 sensitivity at 3,400 Å plotted in the Final OAO-C Copernicus Operations Report, 1982 (Fig. 2-1). The results are shown in the accompanying figure. Each U3 point represents the average 14-second count in a group of adjacent orbits (usually 2 to 5 in number), with 2 to 4 scans averaged in each orbit, and 16 individual counts averaged in each scan. The U1 points are based on smoothed representations of the observations.

Evidently the sensitivity at 1340 Å, near the end of the satellite's operational life at about orbit 44,000 (February 1981), amounted to about 15% of its initial value for U1, as compared to about 60% for U3, some four times as much. The data in the figure seem to show conclusively that degradation of the optical system cannot be responsible for most of the sensitivity loss observed with U1 at 1340 Å. Since the observed small degradation of U3 could be produced by increasing opacity of the LiF window on this tube, the sensitivity loss attributable to reduced reflectance of mirrors and grating cannot be determined, but at 1340 Å cannot exceed one fourth of the total degradation observed with U1.

At shorter wavelengths there is no firm evidence on the relative importance of mirror reflectance and photocathode sensitivity in the overall degradation noted. The degradation appears somewhat greater on U2 than on U1, suggesting that reflectance losses at the diagonal mirror in the U2 optical train are contributing significantly to the degradation. However, it is also possible that the difference between these two phototubes is a result of differences between the two photocathodes, possibly resulting from their different location and orientation in the spectrometer.

cc: A. Boggess
D. C. Morton
C. R. O'Dell
Copernicus astronomers