Instrument Science Report NICMOS 2006-04
T. Wiklind & T. Wheeler
July 25, 2006
A BSTRACT
NICMOS observations are dependent on a cold and stable thermal environment. In this
ISR we describe the temperature monitoring and how a stable thermal environment for the NICMOS detectors is achieved. A comparison between the thermal properties of the cryogenically cooled dewar during Cycle 7 and the present cooling using using the
NICMOS Cooling System, is presented. A complete thermal history from the start in 1997 up to January 2006 is presented.
The NICMOS detectors show several effects that are dependent on the thermal environment. Most of these effects can be removed through standard calibration, as long as the temperature of the detector at the time of observation is similar to that of the corresponding calibrations. Since calibrations are generally not done for every science observations, this makes it important to maintain the temperature as stable as possible over long time scales.
The temperature of the NICMOS dewar and its detectors is presently monitored using the
NIC1 mounting cup sensor. This sensor is located behind the NIC1 camera array. The temperature is read-out every 30 seconds and is down-linked as part of the normal HST engineering telemetry. Due to the location of the temperature sensor behind the mounting cup of the NIC camera, it does not measure the temperature at the detector. The temperature at the mounting cup sensor may therefore differ from the actual temperature of the different NICMOS detectors. The dewar temperature is maintained at an
1

approximately constant temperature through manual set–point changes in the Neon inlet temperature of the NICMOS Cooling System (NCS; Jedrich et al. 2003). The nominal temperature is 77.15 K and is maintained within +/-0.1 K.
In this Report we briefly describe the thermal history of NIMCOS, the factors that influence the dewar temperature, and how the temperature affects the NICMOS detectors.
Finally, we describe the monitoring and corrective actions done by the NICMOS Team to keep the detector temperatures as stable as possible. We also present the thermal history of the NICMOS dewar over its entire history in graphical form.
A Brief Thermal History of NICMOS
NICMOS thermal history can be divided into two distinct epochs. It was initially cooled to ~61 K by a cryogenic dewar containing solid nitrogen. The cryogenically cooled dewar suffered a shorter lifetime than expected due to a thermal short, leading to a faster sublimation of the nitrogen. Following nitrogen exhaustion in January 1999, the
NICMOS instrument warmed up to a temperature of around 260 K. As this was much too warm for any useful science observations, NICMOS stayed dormant until the third
Servicing Mission in March 2002 when a mechanical cooling machine, using a closedloop reverse-Brayton cycle (Cheng et al. 1998; Jedrich et al. 2003), was installed. The dewar temperature is regulated by the new NICMOS Cooling System (NCS) and is used to maintain the detectors at 77.15 K, about 15 K warmer than with the solid nitrogen cooling.
Operations during the first epoch, with solid nitrogen as the coolant, the detector temperature slowly increased from 61 K to 62.5 K from the start until right before nitrogen exhaustion (see Fig. 1). This variation in operating temperature made it necessary to obtain calibration data, i.e. darks, flats and photometric calibration, quite frequently to keep up with the changing thermal environment. During the NCS era, the dewar temperature is higher but also significantly more stable. Temperature variations over the time period from March 2002 up to the present is typically less than ~0.1 K (see
Fig. 2).
The dewar temperature is presently monitored using the NIC1 mounting cup sensor. This is essentially a diode, with an operating temperature limit of 77.4 K. With the present nominal operating temperature of 77.15 K, we are fairly close to the upper limit of this sensor. Should the temperature increase, alternate means of measuring the detector temperatures becomes necessary. One possibility is to use the bias obtained during dark observations. This method of monitoring the temperature is presently being explored.
2

Figure 1. The dewar temperature at the NIC1 detector mounting cup during the period
February 1997 until the warm-up in January 1999. An almost linear increase in the temperature, from 61 K to 62.5 K can be seen.
The NICMOS detectors show a number of effects that are sensitive to temperature. It is therefore of outmost importance to maintain the thermal environment as stable as possible over long time scales. Any variation of the detector temperature will require a substantial calibration effort.
Apart from the temperature sensor located at the NIC1 mounting cup, we can monitor the NCS Neon inlet temperature. The latter can be changed by adjusting the NCS and is our only way of controlling the dewar temperature. While the inlet temperature remains very stable, not being influenced by parasitic thermal loading, the NIC1 mounting cup shows more temperature variations (see Fig. 3).
The temperature in the dewar depends on several factors, among them the NCS Neon cooling capacity and set–point. The Neon temperature in the closed-loop is maintained at
3

Figure 2. Same as Fig. 1 but starting in March 2002 (top figure) when the NICMOS
Cooling System was installed. The bottom figure shows the dewar temperature between
November 2004 and November 2005. The red horizontal line is the nominal operating temperature of 77.15 K.
4

Figure 3. An example of the dewar temperature measured at the NIC1 mounting cup (top blue line) and the Neon inlet temperature (green line). The latter is directly associated with the NCS cryocooler and controlled through the compressor. A change in the neon set-point is seen towards the end of the period shown here.
~72 K and can be adjusted to achieve a desirable detector temperature. While the inlet temperature remains very stable, not being influenced by parasitic thermal loading, the
NIC1 mounting cup shows more temperature variations (see Fig. 3). The sources of
'parasitic' heat loading influencing the dewar temperature are:
• Annual cycle . The heat load on HST varies with the Sun distance, with the warmest period in January and the coldest in June. This is seen as a slow change in the average dewar temperature if a constant NCS cooling set-point temperature is maintained.
• Day/Night Cycle . Moving from direct sunlight to night changes the heat loading on HST. The time scale is short enough so that the NICMOS dewar does not respond. The day/night cycle does, however, provide a thermal driving force that can be detected in the detector temperature.
• HST Sun/Roll Angle . The HST sun/roll angle determines which part of the HST aft shroud is in direct sun light. Prolonged pointing can cause more or less heat
5

loading on the NICMOS dewar, with temperature effects of the order 0.1\,K on times scales of several days.
• NICMOS Operation . Use of the NICMOS detectors, whether for science operations, calibrations or autoreset mode affect the detector temperature.
• Other Aft Shroud Instruments.
Instruments other than NICMOS can cause heat loading inside the aft shroud and thereby affecting the NICMOS detector temperature on short time scales. This is, however, a relatively small effect.
• Autoreset Mode . When not used for science exposures, the NICMOS detectors are continuously read out. The reason for this is to remove trapped charges caused by passages through the South Atlantic Anomaly (see below). This cause a heat loading on the NICMOS detectors, with small (<0.1 K) temperature jitter (Fig. 1).
Due to its relatively large thermal mass, the temperature response time of the NICMOS dewar as a whole is ~120 minutes. Local temperature deviations can, however, occur on a much smaller time scale when the source of heating is close to the location in question.
This is the case for the NICMOS detectors, where heating from nearby amplifiers can cause temperature variations on time scales of minutes. Thus, on different time scales, the various thermal loadings listed above induce detectable temperature fluctuations at the
NICMOS detectors. In Fig. 4 we show three power spectral density (PSD) plots of the
NIC1 mounting cup temperature sensor for a period of about two weeks during (a) preautoreset mode in Cycle 7, with autoreset mode in Cycle 7, and with the NCS and autoreset mode during Cycle 14. Four distinct periods are distinguishable in all three cases: ~24 hours, 101 minutes, 49 minutes and 33 minutes. In addition there is a strong peak at ~3.5 minutes (Fig. 5). The latter peak is not related to the temperature of the detector, but is likely to be an artifact caused by the electronics involved in the temperature sensing itself. The 3.5 minutes period is present in both pre-NCS and NCS data and is not affected by the autoreset mode (see Fig. 5).
The 24 hour period is related to passage through the South Atlantic Anomaly (SAA), which occurs a few times every day. To protect the detectors from charge deposits, they are powered off during passage of the SAA. The 101, 49 and 33 minute periods are associated with the orbital period of the HST and, hence, the day/night cycle. Other parasitic thermal loadings are not of periodic nature and hence do not show up in the plot.
Although the sampling is too short to show the annual variation, it would be a very weak signal as the setting of the neon temperature set-point is manually regulated to compensate for the slow heating and cooling associated with the seasonal changes.
6

Figure 4. The Power Spectral Density (PSD) of temperature fluctuations at the NIC1 mounting cup for three different thermal epochs: pre-autoreset mode in Cycle 7 (green), autoreset mode in Cycle 7 (red) and NCS with autoreset in Cycle 10-14 (blue). Peaks corresponding to periodicities of ~24 hours, 101, 49 and 33 minutes are seen in all three epochs. The two first periods correspond to thermal driving from SAA passages and orbital night/day, respectively. The last two periods are harmonics of the orbital periods.
Notice that the use of autoreset (autoflush) mode consistently gives a higher power spectral density. This corresponds to the extra thermal load imposed by the continuous reading of the detectors while they are not in use for scientific exposures.
7

Figure 5. The Power Spectral Density (PSD) of temperature fluctuations at the NIC1 mounting cup for three different thermal epochs: pre-autoreset mode in Cycle 7, post autoreset mode in Cycle 7 and Cycle 14. The flat part of the frequency response corresponds to white noise and the prominent peak at ~4.8 mHz (corresponding to a periodicity of ~3.5 minutes) is not related to thermal fluctuations, but an intrinsic electronic signal. The periods seen in Fig. 3 can be seen at the left. The curves have been offset along the y-axis since they would otherwise overlap each other.
The performance of the NICMOS detectors show several dependencies on temperature.
The four most important ones are (1) linear dark current, (2) detector quantum efficiency
(DQE), (3) the shading residual and (4) the saturation level.
• The linear dark current increases linearly with exposure time (hence the name). It is a residual after amplifier glow and shading have been removed. The dark current levels of all three NICMOS cameras are stable, and does not exceed the values expected for the new operating temperature. The increased operating temperature in the NCS era has increased the number 'hot' pixels, i.e. pixels with higher-than-average linear dark current. This can, however, be accounted for by dithering NICMOS exposures.
• The detector quantum efficiency (DQE) of the NICMOS detectors depends on the temperature in the sense that the DQE has increased with the new higher
8

operating temperature. Pixels with a lower than average response (i.e. 'hot' pixels) show a higher than average DQE increase with temperature, resulting in a flatter overall response. The average response at 77.15 K increased by about 60% at J,
40% at H, and 20\% at K.
• The NICMOS detectors exhibit a noiseless signal gradient orthogonal to the direction of primary clocking, which is commonly referred to as shading. It is caused by changes of the pixel bias levels as a function of temperature as well as time since the last readout ( Δ -time). The shading can be completely removed during pipeline processing once it has been calibrated with Δ -time and accurate detector temperature.
• The saturation level, or dynamic range, of a given detector pixel is defined by the well depth, i.e. the amount of charge loaded into the pixel during the detector reset. The reset voltages of the NICMOS detectors vary with temperature. The pixel saturation level will therefore also vary with temperature. For NIC1 and
NIC2, the saturation level is ~15% lower when operating at 77.15 K than at 61 K, for NIC3, the difference is only ~7%.
It has recently been found that the expected behavior of the dewar temperature can be modeled in a qualitative way by convolving the sun-time, i.e. the time per orbit that HST is exposed to direct sunlight, and the annual phase (see Fig. 6). While this does not constitute an absolute modeling of the dewar temperature, it does describe the general trend of the dewar temperature fluctuations over time scales of days to weeks. This model is presently used on order to adjust the neon set-point temperature and, hence, the dewar temperature. Such changes are done manually when the dewar temperature deviates more than 0.1 K from the nominal 77.15 K, unless the model predicts a return to the nominal temperature within a reasonably short period of time.
The dewar temperature, as measured at the NIC1 mounting cup, for the entire life span of
NICMOS, can be found at http://www.stsci.edu/hst/nicmos/performance/temperature .
Here the dewar temperature is plotted for each month of operation. If an observer suspects that image artifacts are not removed by the usual procedures, it is advisable to check if the data was obtained during a period when the dewar temperature deviated from the nominal calibration temperature of 77.15 K.
9

Figure 6. The HST Sun-time, i.e. the time per orbit that HST is exposed to direct sun light (blue line) and the solar intensity due to Earth’s orbital parameters (red line). A combination of the Sun-time and solar intensity serves as a qualitative model of the dewar temperature on time scales of several days to weeks.
The NICMOS detectors show several effects which are temperature dependent. Although most of these can be handled in the pipeline, a successful data reduction require accurate knowledge of the detector temperature. The easiest way of assuring good data is to keep a constant detector temperature, corresponding to the environment when calibration data was obtained. In the NCS era this is achieved through monitoring of the NIC1 mounting cup temperature sensor and manually changing the NCS Neon temperature set-point.
Since the re-start of NICMOS in March 2002, we have achieved a temperature stability of ~0.1 K. Despite this, there are small scale temperature fluctuations on short time scales. These are mainly caused by passages through the SAA, the orbital night/day cycle and intermittent parasitic thermal loads. The autoreset mode, started in Cycle 7 to flush out trapped charges, induces a small scale temperature jitter in the NICMOS detectors.
Observers are advised to check the temperature logg at the NICMOS web-site whenever they suspect that artifacts are present after applying proper calibration procedures.
10

Cheng, E.S., Smith, R.C., Jedrich, N.M., Gibbon, J.A., Cottingham, D.A., Swift, W.L. \&
Dame, R.E. 1998, SPIE Proc.,3356, 1149
Jedrich, N.M., et al. 2003, SPIE Proc. 4850, 1058
Schultz, A., Noll, K., Barker, E., Arribas, S., Bergeron L., de Jong, R., Malhotra, S.,
Mobasher, B., Wiklind, T. \& Xu, C., {\it NICMOS Instrument Handbook, version 8},
Space Telescope Science Institute, Baltimore, 2005
11

Figure 7. Dewar temperature data for Cycle 7. During this period the NICMOS detectors were cooled by a cryogenic dewar containing solid nitrogen. The dewar temperature increased slowly as the nitrogen was consumed. Each panel covers a period of 6 months.
12

13

14

Figure 8. Dewar temperature with the NCS, starting March 2002 and running until
January 2006. During this period the NICMOS detectors were cooled through a mechanical cryocooler, the NICMOS Cooling System (NCS). The temperature is more stable than during Cycle 7, but the nominal temperature is higher. Each panel shows the temperature over a period of 6 months, starting in March 2002 and running to January
2006.
15