Документ взят из кэша поисковой машины. Адрес оригинального документа : http://www.stsci.edu/hst/wfc3/documents/ISRs/2002/WFC3-2002-05.pdf Дата изменения: Fri Sep 26 03:24:55 2008 Дата индексирования: Thu Jan 15 04:55:44 2009 Кодировка: Поисковые слова: universe

Instrument Science Report WFC3 2002-05

Internal background and dark current of the IR channel of WFC3 in ambient conditions _______________________
Massimo Stiavelli June 4, 2002
____________________________________________________________________________

Abstract
During the ground testing of WFC3 it may be desirable to operate the IR channel enclosure at ambient temperature. This implies that the temperatures of the optics and of the detector would be warmer than the design values. The purpose of this ISR is to estimate the internal thermal background and dark current for a range of temperatures of the various components. 1. Introduction The WFC3 IR channel is actively cooled during normal operations. The optical bench is cooled to 0 C, the refractive corrector plate (RCP), filter wheel and dewar window are cooled to ­30 C and the detector housing is cooled to about ­70 C. The detector itself is operated at 150 K. Warming up these components has different effects depending on their location in the optical train and on the filter used. I will consider for simplicity the J and H filters. The clear filter used for alignment purposes, F093W, and all filters that do not reach the long wavelength cutoff are essentially insensitive to the thermal background and are equivalent to the J filter for the purposes of this ISR. The basic assumptions are briefly described in Section 2. In Section 3 I will discuss the thermal emission by the various optical components, while Section 4 is devoted to the detector dark current.

2. Assumptions
I am relying on the same instrument model used for setting the temperature of the various components. A number of parameters have been updated to better reflect the performance of the instrument as presently understood. In particular, the IR detector QE has

1


been increased from the value of the CEI Specs to that of detector FPA50. The filter throughput has also been increased from 0.8 to 0.9. For the detector dark current we use the measured value for detector FPA50 extrapolating to higher temperature values according to the laboratory measurements and theoretical curves. The background count rates at nominal operating temperature are 0.012 and 0.175 e s-1 pixel-1, respectively, in the J and H bands. In the calculations described below each component will be "warmed up" and resulting effect on the count rate will be determined for both the J and the H band. The shortest integration for the IR detector is 4 seconds. This implies that the total background for a pair of readouts corresponds to an effective integration from the reset point of up to 8 seconds. For measuring the detector noise we assume that one should have a background plus dark current contributing less than the readout noise square ~1600 e pixel-1, i.e., less than ~400 e s-1 pixel-1 for the shortest read. Measurements of basic functionality can instead be carried out as long as the detector does not saturate, i.e., up to ~8,000 e s-1 pixel-1 for the shortest read.

3. Thermal emission from the optical train
Optical elements mounted before the filter wheel contribute a negligible amount of background over the entire range of temperatures considered. The transmission curve of the filter determines what fraction of the thermal background is seen by the detector and, indirectly, what is the emissivity profile of the filter itself. The small solid angle of those components preceding the filter wheel contributes to further reducing their contribution to the background count rate. For measurements with the J filter temperatures up to +20C for the optical bench and the pick-off mirror lead to no measurable increase of the dark counts. For the H filter a temperature of +20C for the bench and the pick-off mirror leads to an increase of the count rate by 0.08 e s-1 pixel-1 and 0.04 e s-1 pixel-1, respectively. The refractive corrector plate, the filter and the dewar window are nominally at ­30C. Increasing their temperature to +20C increases the count rate in the J filter by 2.6 e s-1 pixel-1 and that in the H filter by 3.9 e s-1 pixel-1. This modest increase is due to the relatively small solid angle of these components. A larger effect is observed when the whole detector enclosure, including the baffles, is warmed up. In the table below we give count rates that apply to both the J and H band since the thermal background is dominated by components within the filter wheel.

Temperature Count rate (e s-1 pixel-1 )

-30C -20C -10C 6 22 79

0C 258

+10C 775

+15C 1,305

+20C 2,159

+25C 3,516

Table 1: count rate as a function of the cold enclosure temperature.

Summarizing, in order to carry out measurements of the detector noise one should maintain the temperature of the inner baffle of the cold enclosure to no more than -5C (under the assumption that half of the maximum acceptable count rate is due to the thermal background and the other half is contributed by the dark current). For basic functionality tests

2


even temperatures of up to +25C are acceptable and will limit the integration time to less than about 20 seconds. When operating at 150K, FPA50 has a dark current of 0.17 e s-1 pixel-1. Although the variation of dark current with temperature of this particular device has not been measured in detail, for other, similar, devices the DCL typically measures an increase of about a factor 8 in dark current for each 10 degrees in temperature. In addition to this estimate we have considered also the classic "1/T" model for the variation of the dark current and a twocomponent physically based (but pessimistic) model. The count rates as a function of temperatures for these three models are given in Table 2. As it can be seen from the table, the requirement of being able to integrate for at least 4 second (8 seconds of charge accumulation) without reaching full well would force us to temperatures lower than 190K according to the most pessimistic model. A contribution from the dark current similar to that of a cold enclosure at +25C would be obtained ­ according to the pessimistic model ­ for temperatures lower than 185K. Temperature (K) 150 160 170 180 190 200 210 220 230 Factor 8 per dex (e s-1 pixel-1) 0.17 1.36 10.9 87 696 5,570 44,564 1/T (e s-1 pixel-1) 0.17 1.4 9.0 47 208 789 2,636 7,891 21,420 2-component (e s-1 pixel-1) 0.17 2.1 38 549 6,084 53,032 -

4. Detector dark current

Table 2 : dark current as a function of temperature according to three different models.

5. Conclusions
The thermal background of the IR channel grows relatively slowly with temperature thanks to the relative short wavelength cutoff of the WFC3 detectors. Thus, the main driver is the detector dark current that is increasing very rapidly with temperature. It appears that the IR channel of WFC3 can be operated at ambient temperature if the optical components do not exceed +25C and if the FPA can be cooled to at least 190K. It is possible that even warmer temperatures could be acceptable but they would require additional testing since they are supported only by some of the existing models. Measuring the detector noise would require the optical components not to exceed -5C and the detector temperature to be below 175K.

Acknowledgements
I thank J. MacKenty and M. Robberto for useful comments on the manuscript.

3