Instrument Science Report WFPC2 2002-06
Results of the Observatory
Verification for WFPC2 after
Servicing Mission 3B
Anton M. Koekemoer, Shireen Gonzaga, Inge Heyer, Lori M. Lubin,
Vera Kozhurina-Platais, and Brad Whitmore
October 7, 2002
ABSTRACT
Servicing Mission 3B (SM3B) took place in March 2002, and the Wide Field and Planetary Camera 2 (WFPC2) resumed operations on March 23, 2002 after having been placed
in an inactive mode during the Servicing Mission. We describe here the results from an
extensive series of tests and observations that we carried out with WFPC2 as part of the
subsequent Observatory Verification phase. These tests included UV monitoring of possible contamination, performance checks of the biases, darks and other internal calibrations, as well as PSF measurements and flatfield verification. The results from these tests
show that there are no significant changes in the operational behavior of the camera with
respect to its pre-SM3B performance.
1: Introduction
The Wide Field and Planetary Camera 2 (WFPC2) has been the principal imaging
instrument on board the Hubble Space Telescope (HST) since its installation during the
first Servicing Mission in December 1993. It is located at the center of the HST focal
plane and comprises four cameras (PC, WF2, WF3, WF4), each of which contains an
800x800-pixel Loral CCD that is sensitive to wavelengths in the range 1150 - 11000 Å.
During March 2002, Servicing Mission 3B (SM3B) was carried out, which involved the
addition of several new hardware items including the Nicmos Cryo-cooler System (NCS)
and a new imaging instrument, the Advanced Camera for Surveys (ACS). While it is
Instrument Science Report WFPC2 2002-06
expected that a wide range of new science programs will benefit from ACS, it is also
important to retain WFPC2 as a fully operational instrument, since it has a number of
unique scientific capabilities that are not addressed by ACS. Thus, the primary goals of the
WFPC2 SM3B plan involved firstly protecting the health and safety of WFPC2 during and
immediately after the Servicing Mission, and secondly evaluating and calibrating any possible changes in its performance after completion of the Servicing Mission.
WFPC2 Activities During and Immediately After SM3B
SM3B commenced with the launch of the space shuttle Endeavour on February 28,
2002 and rendezvous grapple of HST on March 2, 2002, followed by 5 days of astronaut
activity from March 3 to March 7, 2002, and release of the telescope from the shuttle on
March 9, 2002, at 10:04 UT. In preparation for the inactivity of the telescope during the
Servicing Mission, the normal operations of WFPC2 were suspended on March 1, 2002
and the instrument was transitioned into Protect Safe mode. In this mode, power was
turned off for the detectors, the low-voltage power supply (LVPS), the mechanical systems
and the thermo-electric coolers (TECs), while power was kept on for the heat pipe heater
and the bay 5 replacement heaters. The shutters also remained closed and the seldom-used
F785LP filter was put in place, in order to minimize the influx of potential contaminants
from the HST hub area into the WFPC2 instrument.
During SM3B the HST Power Control Unit (PCU) was replaced, which necessitated
completely powering down all the science instruments for a period of approximately 8
hours. In preparation for this, WFPC2 was transitioned from Protect Safe to Protect Decon
(LVPS on, CCD and heat pipe heater on, bay 5 replacement heaters off) and a thermal
blanket was installed by the astronauts over Bay 5. This period of warmup ensured that the
temperature of WFPC2 would remain sufficiently above its critical temperature of -24˚ C
during the 8 hour power-off period. After the PCU changeout had been completed,
WFPC2 was successfully powered up once more and transitioned into Protect Safe mode,
and was maintained in this state during the remainder of SM3B. After deployment of HST
from the space shuttle, WFPC2 was transitioned to Protect Decon and remained in this
state during the 12-day Bright-Earth Avoidance (BEA) period.
During the BEA period, HST was maintained at an attitude that ensured no reflected
light from the bright earth would directly enter the primary aperture, since this could
potentially cause permanent polymerization of any contaminants that may have been
deposited on the mirror surfaces. Throughout this period the WFPC2 camera shutter
remained closed and the F785LP filter was kept in place, to continue minimizing the risk
of potential contaminants entering the instrument. Furthermore, in this mode the heatpipe
heater is on and the TECs are off, thereby increasing the temperature of the camera heads
and further minimizing the risk of any deposition of contaminants inside the instrument.
The BEA period was concluded on March 22, 2002, and was immediately followed by
a Vehicle Disturbance Test (VDT) aimed at the anti-sun pointing (-V1), lasting for approx-
2
Instrument Science Report WFPC2 2002-06
imately 10 hours. The purpose of this test was to validate the vibrational stability of the
spacecraft, related to the activation of the NCS. During this time, WFPC2 continued to
remain in Protect Decon mode. Immediately after completion of the -V1 VDT, the
recomissioning activities for WFPC2 were commenced through the execution of a carefully planned sequence of operations, observations and verifications. At 07:29 UT on
March 23, 2002, the recovery of WFPC2 was initiated by transitioning from Protect
Decon mode to Standby, and then to Operate mode. This transition included an engineering matrix verification plan to ensure nominal instrumental functionality. After about 3
hours the camera had cooled down and stabilized at its nominal operating temperature of 88 C, and an extensive program of calibration observations were commenced that would
extend over the next four weeks, to verify that the camera performance and characteristics
remained essentially unchanged since before the Servicing Mission. This series of calibration observations for the Servicing Mission Orbital Verification phase (SMOV3B) for
WFPC2 was formally completed on April 19, 2002. This Report documents the WFPC2
observations associated with SMOV3B, as well as the subsequent analysis carried out by
the WFPC2 group to validate the camera performance.
Recomissioning Plan and Precautions to Mitigate Contamination
The recommissioning of WFPC2 and verification of its performance and calibration
closely followed the plans laid out in the Instrument Science Report WFPC2-2001-11,
with minor deviations and additions which will be identified where appropriate.
A major concern for WFPC2 is UV-obscuring contamination, especially since contaminating substances condense preferentially on the cold (-88˚ C) camera heads and
progressively reduce the UV throughput of the camera. During normal operations, the
total throughput at 170 nm decreases by about 0.6% per day, and the camera heads are
warmed every 4 weeks to evaporate contaminants and restore the UV throughput. The rate
of accumulation of contaminants increases immediately after servicing missions, and the
recommissioning program after Servicing Mission 3B was designed to verify that the total
contamination would never exceed the safe level of 30% at 170 nm.
Another contamination concern is the possibility of depositing a very thin layer of
contaminants on the optical surfaces exposed to the HST hub area, especially the WFPC2
pick-off mirror. This concern arises from the experience with WF/PC1, whose pick-off
mirror eventually lost almost all its reflectivity at the shortest wavelengths. The effect of
thin layer contamination is very wavelength-dependent, and is especially noticeable at
Lyman α (122 nm). Once deposited, such thin layers of contaminants might evaporate
again over time, unless polymerized by strong UV radiation - such as could be caused by
Earth light shining inside the mirror. Therefore, HST was constrained not to point at the
illuminated Earth at any time within the first 12 days after the Servicing Mission (the socalled period of Bright Earth Avoidance, or BEA), and the end of BEA was tied to verification of the absence of significant far-UV contamination.
3
Instrument Science Report WFPC2 2002-06
Other activities carried out during SMOV3B for WFPC2 included measurements of
the HST focus position, as well as verification of the WFPC2 flat field, dark current, bias
level, point-spread function (PSF), and photometric throughput. The programs used to
carry out all these activities are listed in Table 1.1.
Organization of this Report
The remainder of the present report is divided into seven Sections, each corresponding
to a different area of activity and verification. The Sections are listed in Table 1.1. In the
interest of efficiency, each of these areas has been addressed by a different subset of the
WFPC2 group. The respective authors are identified in each Section. Separate reports have
also been released on PSF verification (WFPC2 TIR-2002-01, Kozhurina-Platais et al.),
flat field monitoring (WFPC2 TIR-2002-02, Koekemoer et al.), UV contamination monitoring (WFPC2 TIR 2002-03, Koekemoer et al.), Photometric throughput (WFPC2 TIR2002-04, Whitmore & Heyer), and Lyman α contamination (WFPC2 TIR-2002-05, Lubin
et al.).
Table 1.1: WFPC2 Calibration and Verification Programs for SMOV3B
Section
Title
Program ID
Page
2
UV Contamination Monitoring and Throughput Check
8950
4
3
Lyman α Contamination Monitoring
8951
13
4
Photometric Throughput
8950, 8953
20
5
Point Spread Function
8954
32
6
Camera Internal Monitoring
8950
39
7
Flat Field Verification
8952
54
2: SMOV3B WFPC2 UV Contamination Monitoring
and Throughput Check
by Anton Koekemoer, Shireen Gonzaga, Lori Lubin, Brad Whitmore, and Inge Heyer
Summary: The throughput of the WFPC2 cameras in the UV is potentially susceptible to
significant decreases as a result of contaminants deposited on the cold CCD windows,
once the instrument has been cooled down to -88˚ C. An important part of our SMOV3B
checkout involved intensive monitoring of this contamination using the F170W filter, to
ensure that the throughput never dropped below levels that can safely be removed by the
regular decontamination procedures. We report here the results of these measurements,
4
Instrument Science Report WFPC2 2002-06
showing that the throughput remained at safe levels throughput SMOV3B, and in addition
demonstrating that the decontamination procedures during SMOV3B returned the
throughput to its nominal levels. We also find that the daily contamination rates have now
returned to their pre-SMOV3B levels.
Note: These results have been released independently as WFPC2 Technical Instrument
Report (TIR) 2002-03, SMOV3B WFPC2 UV Contamination Monitoring and Throughput
Check, by Anton Koekemoer, Shireen Gonzaga, Lori Lubin, Brad Whitmore, Inge Heyer.
Introduction
A critical aspect of the initial cool-down and activation of WFPC2 after Servicing
Mission 3B (SM3B) involved a period of intensive monitoring and verification of the
throughput of the instrument at UV wavelengths, in order to ensure that the camera
throughput was not permanently degraded by contamination deposited on the cold CCD
windows (-88˚ C). In particular, the throughput in the F170W filter needed to be monitored to verify that the decrease in flux due to contamination never exceeded the safe limit
of a 30% drop in total throughput, which corresponds to the maximum level that is known
to be removed by the regular decontamination procedures.
Furthermore, we needed to verify whether the decontamination procedures scheduled
during SMOV3B, which are aimed at removing contaminants by warming up the cameras
for a period of several hours, were successful in returning the throughput to its nominal
levels. Finally, we needed to measure the daily contamination rate and carry out a comparison with pre-SMOV3B values.
Observations
The WFPC2 camera was cooled down to -88˚ C on 23 March 2002, within a day after
the end of the Bright Earth Avoidance (BEA) period which had lasted 12 days from the
time that HST had been released from the space shuttle. The UV contamination monitoring plan for WFPC2 was based on observations of the WFPC2 primary standard star
GRW+70d5824 through the F170W filter, in all 4 cameras of the instrument. The program
consisted of two phases: an intensive period, beginning immediately after the cooldown to
-88o C and lasting for 6 days, until the first decontamination procedure took place one
week after cooldown; and a second less intensive phase, lasting throughout the rest of
SMOV3B for WFPC2, consisting of observations immediately before and after each of the
subsequent decontamination procedures that occurred at intervals of 1-2 weeks.
The intensive phase began with a set of observations immediately after cool-down,
consisting of two observations of GRW+70d5824 through the F170W filter in each of the
four chips. This was subsequently repeated at 3, 6, 12, 18, 24, 36 hours, and 2, 3, 4, 5, 6
days after cooldown (see Table 2.1). At each epoch, the two observations for each chip
5
Instrument Science Report WFPC2 2002-06
consisted of a dither-pair where the second exposure was offset by 0.25" along both the x
and y axes in a simple DITHER-LINE pattern. All exposures were 40s in length and used
GAIN=15. The timing of each set of observations varied slightly as a result of scheduling
constraints and the fact that each visit required one to two orbits, but nevertheless all the
observations were executed in accordance with the requirements of the program.
Table 2.1:WFPC2 Observations of GRW+70d5824 in F170W during SMOV3B.
Time since
Cooldown
PC
WF2
WF3
WF4
0h
u8cz1301
u8cz1302
u8cz1303
u8cz1304
u8cz1305
u8cz1306
u8cz1307
u8cz1308
3h
u8cz1501
u8cz1502
u8cz1503
u8cz1504
u8cz1505
u8cz1506
u8cz1507
u8cz1508
6h
u8cz1601
u8cz1602
u8cz1603
u8cz1604
u8cz1605
u8cz1606
u8cz1607
u8cz1608
12h
u8cz1701
u8cz1702
u8cz1703
u8cz1704
u8cz1705
u8cz1706
u8cz1707
u8cz1708
18h
u8cz1801
u8cz1802
u8cz1803
u8cz1804
u8cz1805
u8cz1806
u8cz1807
u8cz1808
24h
u8cz2001
u8cz2002
u8cz2003
u8cz2004
u8cz2005
u8cz2006
u8cz2007
u8cz2008
36h
u8cz2301
u8cz2302
u8cz2303
u8cz2304
u8cz2305
u8cz2306
u8cz2307
u8cz2308
48h (2d)
u8cz2401
u8cz2402
u8cz2403
u8cz2404
u8cz2405
u8cz2406
u8cz2407
u8cz2408
3d
u8cz2601
u8cz2602
u8cz2603
u8cz2604
u8cz2605
u8cz2606
u8cz2607
u8cz2608
4d
u8cz2801
u8cz2802
u8cz2803
u8cz2804
u8cz2805
u8cz2806
u8cz2807
u8cz2808
5d
u8cz2901
u8cz2902
u8cz2903
u8cz2904
u8cz2905
u8cz2906
u8cz2907
u8cz2908
6d
u8cz3001
u8cz3002
u8cz3003
u8cz3004
u8cz3005
u8cz3006
u8cz3007
u8cz3008
7d pre-decon
u8cz4107
u8cz4108
u8cz4109
u8cz410a
7d post-decon
u8cz4807
u8cz4808
u8cz4809
u8cz480a
14d pre-decon
u8cz5107
u8cz5108
u8cz5109
u8cz510a
14d post-decon
u8cz5407
u8cz5408
u8cz5409
u8cz540a
26d pre-decona
u8cz7107
u8cz7108
u8cz7109
u8cz710a
26d post-decon
u8cz7407
u8cz7408
u8cz7409
u8cz740a
a. Due to guide star problems, the 26-day pre-decon data were not obtained.
6
Instrument Science Report WFPC2 2002-06
The data during the first 24 hours were all transferred using expedited delivery and
were available for analysis within 2-3 hours of the observations being taken; the remainder
of the data for the first week were generally available 12-24 hours within the time of the
observations. In all cases the data were analyzed immediately upon receipt, typically
within 1-2 hours, thereby maintaining a total turn-around time well within the 12 hours
requirement. In all cases the resulting flux measurements showed that the contamination
rate remained within the safety margin.
The remainder of the UV observations during the rest of SMOV3B, obtained before
and after each subsequent decontamination procedure, were primarily intended to track
the expected decrease in contamination rate as a function of time after HST release and to
afford a comparison with pre-SM3B values. These differed slightly from the intensive
observations during the first week, in that only a single exposure was obtained for each
chip, and the observations were obtained as part of a photometric sweep that included several other filters in the UV and optical. The star was also placed on a different part of the
chip, to provide consistency with other long-term photometric programs. However, all the
F170W exposures were still 40s in length with GAIN=15, thus they afford a direct comparison with the more intensive observations from the first week after cooldown.
Analysis
The data were first calibrated using the normal WFPC2 pipeline, which included
applying standard flatfields, A-to-D conversion, and bias subtraction. For each camera, the
pair of CR-split images were not combined using cosmic-ray rejection algorithms but
were instead treated as separate images, and the flux of the star measured independently
on each image. This was considered to provide more robust identification of outlier flux
measurements that may be affected by cosmic rays, warm pixels or other effects. In addition, this allows a direct comparison with the later photometric points from the remainder
of SMOV3B, where only a single exposure in F170W was obtained for each chip.
In each image, cosmic rays were identified and removed by hand, and standard aperture photometry was carried out on the resulting image. In general cosmic rays were not a
serious problem, due to the relatively short exposure times that were used (40s); only
about 10% of the images had cosmic rays sufficiently close to the star, and of sufficient
intensity, that editing was necessary. During the first 24 hours, which was the most critical
part of the monitoring, the editing and photometry was carried out independently by all
five members of the team, in order to verify consistency in terms of cosmic ray removal as
well as photometry. In general, extremely good consistency was achieved (typically to
well within the formal measurement errors associated with the photometry, which were
generally in the range 0.3 - 0.5%). The measurements from all the team members were
collected into a single database and the points were then input into a regression routine to
obtain a fit to the contamination rate as a function of time; the database and the regression
fit were updated immediately upon receipt of each new set of measurements.
7
Instrument Science Report WFPC2 2002-06
Results
In Figure 2.1 we show the rate of throughput decrease as a function of time, resulting
from the growth of contaminants on the CCD windows of the four detectors. The dashed
line in each case shows the nominal count-rate based on measurements of the star before
SMOV3B; here we describe several features apparent in the plots.
Firstly, the initial throughput in all four cameras measured during the first few hours
after cooldown was ~ 1 - 2% below the nominal expected throughput, even though the
instrument had just come out of a decontamination period that had lasted almost two
weeks. Analysis of the PSF revealed that the telescope was substantially out of focus during the first day or so after the end of BEA, therefore some flux was lost from the aperture
surrounding the star, thereby reducing the apparent count rate. This effect is verified by the
fact that the throughput only began to show a clear downward trend after the first 2-3 days,
when the focus had recovered to its nominal values and the measured count rates were no
longer being substantially affected by the quality of the PSF. The effect was also verified
by showing that for smaller apertures, even more light was lost during the first few measurements; larger apertures retained more light but were not used due to the increased
effects of contamination from low-level cosmic rays and other defects.
It can be seen that for all four cameras, the overall decrease in flux by the time of the
first decontamination at 7 days after cooldown was no more than 10-15% below the nominal 100% throughput value, thereby remaining safely above the 30% decrease limit.
The rate of change of throughput was also measured separately for each camera and
are tabulated in Table 2.2 where we compare them to the normal (pre-SMOV) rates of
change, as well as those from the last servicing mission.
Table 2.2:Measured rate of throughput decrease in F170W (expressed in %/day).
Normal
(2001-2002)
SMOV3A
SMOV3B
PC
0.330
1.4
0.65
WF2
0.486
2.0
0.79
WF3
0.516
2.0
0.93
WF4
0.373
1.5
0.38
It can be seen that the SMOV3B rates are somewhat higher than normal, up to a factor
of 2. This is likely due to an increased presence of contaminants related to servicing mission activities and new instruments aboard the telescope. However, the rates are still
significantly below those that were seen in SMOV3A, and at all times the cameras
remained well within their safe limits for contamination rates. This is likely due to the fact
that the WFPC2 cameras were not cooled down until 12 days after HST release, thereby
allowing substantially more time for contaminants to escape than had been the case during
8
Instrument Science Report WFPC2 2002-06
WFPC2 F170W GRW+70d5824 Measured count-rates
PC
WF4
counts/s
200
180
160
140
counts/s
200
WF2
WF3
180
160
140
0
Normalized counts/s
2
3
4
5
Time since cooldown (days)
6
0
1
2
3
4
5
Time since cooldown (days)
6
WFPC2 F170W GRW+70d5824 Normalized count-rates
-0.65%/day +/- 0.16%
PC -0.38%/day +/- 0.17%
WF4
1.05
1.00
0.95
0.90
-0.79%/day +/- 0.14%
1.05
Normalized counts/s
1
WF2
-0.93%/day +/- 0.18%
WF3
PC1
1.00
WF4
WF2 WF3
0.95
0.90
0
1
2
3
4
5
Time since cooldown (days)
6
0
1
2
3
4
5
Time since cooldown (days)
6
Figure 2.1: Measured decrease in throughput of GRW+70d5824 during the first week after
cooldown, plotted for each of the cameras separately. The top panels show the measured count-rates,
while the bottom panels show count-rates normalized to the pre-SMOV3B values. The bottom panels
also show the rate of change of percentage throughput per day, fitted separately for each camera.
9
Instrument Science Report WFPC2 2002-06
SM3A (where cooldown occurred within the first week after HST release). Moreover, the
new instruments installed this time had been extensively treated before flight to ensure a
minimum of outgassing material, and this has likely contributed to the reduced contamination rates that we observe for SM3B.
In addition to measuring the decrease in throughput relative to the nominal count-rate,
we also measure the rate of change of throughput relative to the first measurements taken
after cooldown. Although this is subject to some slight uncertainties in the measurements
as a result of the somewhat degraded PSF after the end of the BEA, we nevertheless obtain
results that are consistent with those measured for the cameras independently. In
Figure 2.2 we show the average contamination rate measured for all four cameras, normalized to the flux of the first observation after cooldown. In this figure we differentiate
between the first and second observation of each split pair, and we explain here the reasons
why.
WFPC2 F170W Contamination: PC,WF2,WF3,WF4
104
Zeropoint =
Slope =
Zeropoint =
Slope =
Normalized counts/s (%)
102
99.98 +/- 0.23 %
-0.70 +/- 0.08 % /day
99.60 +/- 0.32 %
-0.71 +/- 0.12 % /day
PC
WF2
WF3
WF4
100
98
96 PC1 WF4
WF2 WF3
94
92
0
1
2
3
4
5
Time since cooldown (days)
6
Figure 2.2: Measured decrease in throughput of GRW+70d5824 during the first week after
cooldown, for each of the chips, plotted on the same set of axes and normalized to the count-rate of
the first data-points after cooldown. At each epoch there are two split observations per chip; we use
solid symbols to denote the first exposure in each pair and open symbols to denote the second exposure. The second exposure is often systematically brighter than the first by about a percent, which is
likely attributable to mild “pre-flashing” of the pixels by the first exposure. The reason that the effect
is so noticeable is probably a result of the negligible background at F170W, combined with the short
exposure times. The two lines correspond to fits with and without the second exposure in each pair,
showing consistent slopes but a constant offset.
10
7
Instrument Science Report WFPC2 2002-06
We found that the second observation of each pair of observations typically gave a
somewhat higher flux measurement than the first, particularly for measurements taken
after the first day. One possible cause of this is that the first observation results in a mild
“pre-flashing” of the pixels; although the second observation was offset from the first, it
was by a dither of only a few pixels in each direction, thus many of the pixels sampling the
star in the second observation would also have sampled it during the first observation. This
mild pre-flashing would result in less CTE loss in the second observation, thereby increasing it by up to a few percent relative to the first. The reason why this effect is so strong is
likely related to the fact that in the F170W the background is negligible, combined with
the fact that the exposures are only 40s each and thus relatively short.
Therefore, in Figure 2.2 we differentiate between the two points of each pair by plotting the first point with a solid symbol and the second with an open symbol. We also
carried out two sets of fits to the data, one involving all the points and the second fit
involving only the first point of each pair. It can be seen that the average rate of decline for
both fits is almost identical, at around 0.7% / day, but that the two lines are offset from one
another by around 0.4%, corresponding to the contribution from the increased flux in the
second measurement of each split pair.
Finally, we show in Figure 2.3 the results of the subsequent UV monitoring observations, carried out during the rest of SMOV3B after the first week. Unfortunately, a guide
star acquisition failure prevented us from obtaining a measurement just before the 26-day
decon, but we were nonetheless successful in obtaining measurements before and after
each of the other decons. It can be seen that the measurements just before the 14-day
decon generally have a much smaller decrease in throughput compared to those immediately before the 7-day decon, and this is attributable to the successful removal of
contaminants by the decontamination procedure, as well as the fact that more time has
elapsed for material to escape from the system. Moreover, the points immediately after the
14-day and 26-day decontamination procedures show that the F170W throughput is essentially recovered each time, to within the 1-2% errors related to the actual measurement as
well as systematic uncertainties, and the contamination rate is also back to nominal.
Conclusions
We have presented the results from our program of intensive monitoring of the UV
throughput at F170W due to growth of contaminants on the cold (-88˚ C) CCD windows.
We have verified the following:
•
At no point did the contamination on any of the CCD windows exceed a 15% drop
in throughput, which is well above the safety limit of a 30% drop in throughput.
11
Instrument Science Report WFPC2 2002-06
WFPC2 F170W GRW+70d5824 Measured count-rates
PC
WF4
counts/s
200
180
160
140
counts/s
200
WF2
WF3
180
160
140
0
0
5
10
15
20
25
Time since cooldown (days)
1.00
0.95
PC1 0.90 WF4
WF2 WF3
1.05
Normalized counts/s
10
15
20
25
Time since cooldown (days)
WFPC2 F170W GRW+70d5824 Normalized count-rates
PC
WF4
1.05
Normalized counts/s
5
WF2
WF3
1.00
0.95
0.90
0
5
10
15
20
25
Time since cooldown (days)
0
5
10
15
20
25
Time since cooldown (days)
Figure 2.3: Measured decrease in throughput of GRW+70d5824 during the first month after
cooldown, for each of the chips, plotted on the same set of axes and normalized to the pre-SMOV3B
count-rates. The vertical dashed lines indicate the decontamination visits that were carried out 7, 14
and 26 days after cooldown. It can be seen that, within our measurement uncertainties, the sensitivity
of the instrument is effectively recovered after each decontamination.
12
Instrument Science Report WFPC2 2002-06
•
While the contamination rate was measured to be slightly higher in each of the
cameras (up to a factor of 2 above normal), the rates were still well below those
observed during the previous servicing mission. This is likely due to a combination
of factors, including not cooling down the camera until 12 days after release
(allowing extra time for contaminants to escape), as well as possibly a reduced
level of contaminants from the new instruments compared to previous servicing
missions.
•
The decontamination procedures carried out during SMOV3B demonstrated that
the UV throughput in the F170W filter was successfully recovered to its nominal
value (within our measurement errors and systematic uncertainties of 1-2%).
Moreover, the contamination rates for each camera now appear to have returned to
their nominal values.
Thus, we conclude that our program of delayed cooldown, pro-active UV monitoring,
and frequent decontaminations during SMOV3B were successful in fully retaining the UV
throughput capabilities of WFPC2 as measured in the F170W filter.
References
Casertano, S. Gonzaga, S., Baggett, S., Balleza, J., Biretta, J., Heyer, I., Koekemoer,
A., M., O’Dea, C. P., Riess, A., Schultz, A. B., and Wiggs, M. S. 2000, “Results of
WFPC2 Observatory Verification after SM3A”, ISR WFPC2-2000-02
Koekemoer, A. M., Gonzaga, S., Heyer, I., Lubin, L., Casertano, S. and KozhurinaPlatais, V. 2001, “Summary of WFPC2 SM3B Plans”, ISR WFPC2-2001-11
3: Lyman α Contamination Monitoring
by Lori Lubin, Brad Whitmore, Anton Koekemoer and Inge Heyer
Summary: The far-UV throughput of WFPC2 was monitored on March 23, 2002 for any
signs of unexpected throughput degradation following Servicing Mission 3b. No significant changes were detected.
Note: These results have been released independently as WFPC2 Technical Instrument
Report (TIR) 2002-05, SMOV3B WFPC2 Lyman-α Throughput Check, by Lori Lubin,
Brad Whitmore, Anton Koekemoer and Inge Heyer.
Introduction
Servicing Mission 3B (SM3B) occurred in March 2002. A major concern for WFPC2
during any servicing mission is the possibility of contamination due to the servicing activities. The CCD windows can be heated to remove any contaminant buildup; this is
13
Instrument Science Report WFPC2 2002-06
routinely and successfully accomplished with decon procedures about once a month during normal science operations. However, during a servicing mission, contaminants could
potentially settle on the pick-off mirror (POM) which is exposed in HST’s hub area and
can not be heated. The Lyman-α throughput tests, based upon observations taken with the
F122M and F160BW filters, are designed to monitor the far-UV throughput for any signs
of degradation which may be due to a layer of contaminants on the POM.
The concern over the WFPC2 POM stems from the near total loss of Lyman-α reflectivity of the WF/PC-1 pick-off mirror during its stay in orbit. This behavior has not to date
been observed in orbit with WFPC2 (MacKenty & Baggett 1996). In addition, Lyman-α
monitoring after the previous two Servicing Missions (SM2 and SM3A) also detected no
significant change in throughput to within an uncertainty of 20% (O’Dea et al. 1997; Baggett & Heyer 2000). We present here the results from the Lyman-α monitoring (proposal
8951) during Servicing Mission Orbital Verification 3b (SMOV3B).
Data
New Observations
We have used the standard photometric monitoring target GRW+70D5824 to perform
the Lyman-α test. The photometric properties of this star are summarized in Table 3.1.
The Lyman-α contamination check was performed on March 23, 2002. These observations were taken 12 hours after the WFPC2 was cooled down to -88˚ C. To assess the
Lyman-α contamination, the target was observed in the far-UV using the filters F160BW
and F122W, by themselves, and crossed with F130LP. The crossed filters are used to estimate the contribution of the red leak to the total signal in the F122M filter and to split the
F160BW filter in order to calculate the signal in a pseudo “F120N” filter, thereby providing an additional measure of the Lyman-α throughput. Two images of the standard star in
each of the four filter combinations (F122M, F122M+F130LP, F160BW, and
F160BW+F130LP) were taken on both the PC and WF3 chip.
Table 3.1:Target used for Lyman Alpha Throughput Monitoringa
Property
GRW+70D5824
RA (J2000.0)
13:38:51.77
DEC (J2000.0)
70:17:08.5
spectral class
DA3
V
12.77
B-V
-0.09
U-B
-0.84
a. Positions, spectral class, UBV data from Turnshek et al. (1990).
14
Instrument Science Report WFPC2 2002-06
Individual exposure times were 100 sec and 40 sec for the F122M and F160BW filter
combinations, respectively. Images were calibrated using the standard STSDAS calwp2
task. After calibration, we used a photometric monitor script maintained by Shireen
Gonzaga to measure the flux from the standard star. This script includes the removal by
hand of cosmic rays and hot pixels within the region of interest using the IRAF imedit
task, centering of the star with the imcntr task, and photometry from the phot task in
IRAF package noao.digiphot.apphot, where we use an aperture radius of 0.5" (5 pixels in
WF3, 11 pixels in PC) and a sky subtraction annulus from 1.46"-1.96". The pairs of observations were reduced and analyzed individually. To check repeatability, all datasets were
reduced and analyzed by the authors individually. A comparison of these results indicate
that the values were consistent within the measurement uncertainty; therefore, for the final
analysis, we only present results from one person.
Pre-Servicing Mission Observations
In order to make a comparison to data taken before the Servicing Mission 3B, we have
collected the historical monitoring data of the standard star GRW+70D5824 in the single
and crossed filters. Some of these data are publicly available on the WFPC2 website.
These data were taken over many years and at various times after the normal monthly
decon procedures. Therefore, we must first correct accurately the F122M, F160BW,
F122M+F130LP, and F160BW+F130LP data for the throughput loss due to contaminates.
We have used the historical data to measure the rate of contamination by plotting measured flux versus Days since Decon and performing an unweighted least-square fit to all
available data. Each point has been normalized to the best-fit y-intercept. The results are
shown in Figure 3.1. The final contamination rates for the two single filters and the two
crossed filters, which are used to correct the historical data in the following analysis, are
listed in each panel. In the case of the crossed filter F122M+F130LP, there is limited historical data; however, there is no obvious trend in the measured count rate with Days since
Decon indicating that the contamination rate in this crossed filter is negligible. This result
is expected since the crossed filter samples only the red tail of the F122M filter. In the case
of the crossed filter F160BW+F130LP, we observe a similar decline in count rate with
Days since Decon as we do for the single F160BW filter. The close similarity between the
results for the single and crossed filter is expected since the F130LP filter cuts off only the
small fraction of the F160BW filter which extends blueward of 1250 Å.
In the panels on the top two rows of Figure 3.1, we also show how the data for the redleak-corrected F122M filter and the pseudo “F120N” filter created from the single and
crossed F160BW filters declines with the number of Days since Decon. We label these
panels as F122M-F130LP and F160BW-F130LP, respectively. We calculate the red-leakcorrected F122M data by subtracting the count rate measured in the crossed filter
F122M+F130LP from that measured in the single F122M filter. Since we have shown in
Figure 3.1 that the contamination rate in the F122M+F130LP filter is negligible, we have
15
Instrument Science Report WFPC2 2002-06
simply calculated the average of all measured count rates in this crossed filter and subtracted this value from all measurements made in the single F122M filter. The resulting
decline in the red-leak-corrected F122M data is approximately 2-3% per day. To create the
pseudo “F120N” filter, we subtract the count rate measured in the crossed filter
F160BW+F130LP from that measured in the single F160BW filter. Since we have shown
that both the single and the crossed filter are affected by contaminates, we only use data
from days where there are measured count rates in both the single and crossed filter. The
results from these data indicate a decline of approximately 1.5-2.0% per day. We note that,
due to the small number of data points in the F160BW-F130LP panels, the errors on the
measured contamination rates are large, approximately +/- 0.6%.
Figure 3.1: Normalized count rate vs. Days since Decon from the historical data in the F122M, F160BW,
F122M+F130LP, and F160BW+F130LP filters for the PC and WF3, respectively. The panels on the top two
rows indicate the red-leak-corrected data for the F122M filter and the data from the pseudo “F120N” filter
created from the F160BW filter combinations (see text for details). Solid line indicates the best-fit, leastsquare line. The measured contamination rates are listed in the upper right of each panel.
16
Instrument Science Report WFPC2 2002-06
Results
The normalized counts for the new observations and the historical data are shown in
Figure 3.2 and Figure 3.3 for the F122M and F160BW filter, respectively. The bottom and
middle panels indicate data taken in the single filter and the crossed filter, respectively.
The red-leak-corrected F122M data and the pseudo “F120N” data created from the single
and crossed F160BW filters are shown in the top panels. Left and right panels represent
the data on the PC and WF3 chip, respectively. The historical data are indicated by solid
dots. The new observations taken during SMOV3B are indicated by open stars. As mentioned earlier, these data were taken in pairs, and the results from each exposure are shown
separately. The two measurements agree to 3% or better. The second observation of each
pair typically gives a slightly higher flux measurement than the first, possibly as a result of
a mild “pre-flashing” of the pixels by the first exposure (for a discussion, see Whitmore,
Heyer & Casertano 1999; Koekemoer et al. 2002).
Historical count rate measurements made in the single filters, F122M and F160BW,
and the crossed filters, F122M+F130LP and F160BW+F130LP, have been corrected for
the appropriate contamination rate as determined in Figure 3.1. As discussed above, we
calculate the red-leak-corrected F122M data by measuring the average count rate of all
data available in the crossed filter F122M+F130LP and subtracting this value from the
count rates measured in the single F122M filter. To calculate the throughput from the
pseudo “F120N” filter, we subtract the count rate measured in the crossed filter
F160BW+F130LP from that measured in the single F160BW filter on days where there
are measured count rates in both the single and crossed filter. Each data point has been
normalized by the average value of all fluxes measured at MJD > 51200. The solid and
dotted lines in each panel indicate the mean and one-sigma standard deviation, respectively, of the historical data.
As we can see from these figures, the data taken during SMOV3B are quite consistent
with the historical data. All data taken in the single filter F160BW and the crossed filter
F160BW+F130LP are within 2 sigma of the mean value. The data in the pseudo “F120N”
filter (F160BW-F130LP) is lower than the mean by, on average, 0% on the PC and 8% on
the WF3 chip. The biggest deviations are observed in the F122M filter where the count
rates measured during SMOV3B are lower than the average value by 2-3 sigma. This corresponds to a reduction in the red-leak-corrected count rates by an average of 14% in the
PC and 18% in the WF3 chip. We note, however, that there is a trend in these data of an
overall decrease in count rate in the F122M filter with time since MJD ~ 51000, most
likely due to the effect of CTE (see Whitmore & Heyer 2002). The new observations are
reasonably consistent with this decline. Specifically, if we perform a linear fit to this trend,
we would predict that the count rates measured during SMOV3B to be lower than the
mean by 8% and 6% in the PC and WF3 chip, respectively. We also note that the level of
the observed deviations in the red-leak-corrected F122M data is consistent with that found
after the previous two Servicing Missions (O’Dea et al. 1997; Baggett & Heyer 2000).
17
Instrument Science Report WFPC2 2002-06
Figure 3.2: Normalized count rate versus Modified Julian Date (MJD) in the F122M filter, the F122M filter
crossed with the F130LP filter (F122M+F130LP), and the red-leak-corrected F122M filter (F122MF130LP) for the PC (left panels) and the WF3 (right panels). Data points are normalized by the mean value
of all points with MJD > 51200. The solid points indicate the historical data (which have been corrected for
contamination; see text). The open stars indicate measurements made during SMOV3B. The solid and dotted
lines indicate the mean and one-sigma standard deviation, respectively, of the historical data.
18
Instrument Science Report WFPC2 2002-06
Figure 3.3: Same as Figure 3.2 but for the F160BW filter, the F160BW filter crossed with the F130LP filter
(F160BW+F130LP), and the pseudo “F120N” filter (F160BW-F130LP).
Conclusions
We find that the Lyman-α throughput measurements made after Servicing Mission 3B in
the PC and WF3 chip did not show any significant changes compared to pre-SMOV3B
levels, although there is an indication of an overall trend of decreasing throughput in the
F122M filter over the past four years. We, therefore, conclude that the WFPC2 pick-off
mirror appears to have been protected from any significant contamination related to SM3B
activities.
19
Instrument Science Report WFPC2 2002-06
References
Baggett, S., and Heyer, I. 2000, “Results of the WFPC2 SM3a Lyman-Alpha Throughput Check,” Technical Instrument Report WFPC2-2000-02
Koekemoer, A.M., Gonzaga, S., Lubin, L.M., Whitmore, B., & Heyer, I. 2002,
“SMOV3b WFPC2 UV Contamination Monitoring and Throughput Check,” Technical Instrument Report WFPC2-20 02-03
MacKenty, J., and Baggett, S. 1996, “WFPC2 Throughput Stability in the Extreme
Ultraviolet,” Technical Instrument Report WFPC2-1996-07
O’Dea, C., Baggett, S., and Gonzaga, S. 1997, “Results of the WFPC2 SM2 LymanAlpha Throughput Check,” Technical Instrument Report WFPC2-1997-05
Turnshek, D.A., Bohlin, R.C., Williamson, R.L., Lupie, O.L., and Koornneef, J. 1990,
AJ 99, 1243
Whitmore, B., Heyer, I., and Casertano, S. 1999, PASP, 111, 1559
Whitmore, B., and Heyer, I. 2002, Technical Instrument Report WFPC2-2002-04.
4: Photometric Throughput
by Brad Whitmore and Inge Heyer
Summary: A check of the photometric throughput of the WFPC2 was performed March
31, 2002 (program ID: 8953). The standard star GRW+70d5824 was observed with a
selection of filters and the standard star was centered in each of the four CCDs. The data
indicate that any changes in the photometric throughputs due to SM3B are less than 1% in
most of the visible wavelength filters, and less than a few percent in the UV filters. The distribution shows a mean value of 0.34 +/- 0.26 sigma (where sigma is defined separately
for each filter-chip combination) for all the observations, essentially consistent with an
unchanged photometric throughput. This corresponds to about 0.4% on an absolute scale.
Hence, the response of the WFPC2 was essentially unchanged by the servicing mission.
We also find that the long-term throughput decline is consistent with the expected CTE
loss.
Note: These results have been reported independently in WFPC2 Technical Instrument
Report (TIR) 2002-04, SMOV3B WFPC2 Photometry Check, by Brad Whitmore and Inge
Heyer.
20
Instrument Science Report WFPC2 2002-06
Introduction
The goal of the relative photometry check (proposal 8953) was to verify that the photometric accuracy remained unchanged at the 1-2% level after the latest servicing mission
SM3B. Our regular photometric standard star, GRW+70D5824, was observed in a variety
of filters (F160BW, F170W, F185W, F218W, F255W, F300W, F336W, F439W, F555W,
F675W, and F814W), with the star centered in each of the 4 CCDs (one orbit per CCD).
This is the same photometry check performed after the previous servicing missions, which
showed no decline in the photometric performance. The observations were obtained on
March 31, 2002, (MJD 52364) about 0.3 days after WFPC2 was in a 9-hour special Decon
on March 30, 2002.
Observations
Contaminants collect on the cold CCD windows and reduce the UV throughput of the
WFPC2. A warm up decontamination (Decon) procedure is performed monthly to evaporate contaminants from the CCD windows which is followed by a cool down. Following a
Decon on March 30, 2002 (MJD=52363.7516), observations were taken for the purpose of
determining if any chance in the photometric performance had occurred as a result of
SM3B.
Observations of standard star GRW+70d5824 (mv = 12.7, B-V = -0.09) for the photometry check were obtained on March 31, 2002 (program ID: 8953) which was
approximately 0.3 days following the Decon on March 30, 2002. The star was positioned
in the center of a camera during four 1-orbit visits, one visit per camera. The respective
single camera images were read out and sent to the ground for analysis. Observations were
obtained with filters F160BW, F170W, F185W, F218W, F255W, F300W, F336W, F439W,
F555W, F675W, and F814W. Table 4.1 lists the observations.
21
Instrument Science Report WFPC2 2002-06
Table 4.1:8953 Photometry Monitor Observations.
filter
PC1
PC1
exptime
(sec.)
WF2
WF3
WF4
WF
exptime
(sec.)
F160BW
u8bd0101r
200.0
u8bd0201r
u8bd0301r
u8bd0401r
100.0
F170W
u8bd0102r
40.0
u8bd0202r
u8bd0302r
u8bd0402r
40.0
F185W
u8bd0103r
100.0
u8bd0203r
u8bd0303r
u8bd0403r
100.0
F218W
u8bd0104r
40.0
u8bd0204r
u8bd0304r
u8bd0404r
40.0
F255W
u8bd0105r
80.0
u8bd0205r
u8bd0305r
u8bd0405r
40.0
F300W
u8bd0106r
10.0
u8bd0206r
u8bd0306r
u8bd0406r
10.0
F336W
u8bd0107r
14.0
u8bd0207r
u8bd0307r
u8bd0407r
12.0
F439W
u8bd0108r
14.0
u8bd0208r
u8bd0308r
u8bd0408r
8.0
F555W
u8bd0109r
3.5
u8bd0209r
u8bd0309r
u8bd0409r
2.3
F675W
u8bd010ar
8.0
u8bd020ar
u8bd030ar
u8bd040ar
4.0
F814W
u8bd010br
14.0
u8bd020br
u8bd030br
u8bd040br
7.0
Calibration and Data Reduction
The OPUS pipeline calibrated data were used for the analysis. No other calibration
steps were performed. Photometry was performed using the APPHOT task phot with the
star positions automatically identified for each camera as input. For the PC1 frames, a
photometry aperture radius of 11 pixels was used with the sky fitting region parameters set
to annulus=32 pixels and dannulus=11 pixels. For the WF frames, a photometry aperture
radius of 5 pixels was used with the sky fitting region parameters set to annulus=15 pixels
and dannulus=5 pixels. The sky fitting algorithm was set to “ofilter.” The centering algorithm was set to “centroid”. The photometry values are listed in Tables 4.2-4.5, with table
headers:
•
filter - WFPC2 filter used.
•
mjd - modified Julian Date (Julian Date - 2400000.5) for the observation.
•
ct_rate - countrate (DN/s) for the respective aperture.
22
Instrument Science Report WFPC2 2002-06
Table 4.2:PC1 photometry.
filter
mjd
ct_rate
F160BW
52364.22
88.301
F170W
52364.23
170.478
F185W
52364.23
92.641
F218W
52364.24
128.388
F255W
52364.24
150.319
F300W
52364.24
941.872
F336W
52364.24
740.793
F439W
52364.24
886.434
F555W
52364.25
3724.889
F675W
52364.25
2063.758
F814W
52364.25
1355.537
Table 4.3:WF2 photometry.
filter
mjd
ct_rate
F160BW
52364.29
76.366
F170W
52364.29
192.428
F185W
52364.30
102.416
F218W
52364.30
139.858
F255W
52364.30
159.448
F300W
52364.30
955.157
F336W
52364.31
756.947
F439W
52364.31
883.425
F555W
52364.31
3726.059
F675W
52364.31
2098.310
F814W
52364.31
1377.432
23
Instrument Science Report WFPC2 2002-06
Table 4.4:WF3 photometry.
filter
mjd
ct_rate
F160BW
52364.36
66.767
F170W
52364.36
152.701
F185W
52364.36
82.069
F218W
52364.37
127.220
F255W
52364.37
157.696
F300W
52364.37
959.739
F336W
52364.37
753.756
F439W
52364.38
859.329
F555W
52364.38
3681.950
F675W
52364.38
2036.548
F814W
52364.38
1338.025
Table 4.5:WF4 photometry.
filter
mjd
ct_rate
F160BW
52364.09
67.101
F170W
52364.09
167.927
F185W
52364.10
94.994
F218W
52364.10
138.287
F255W
52364.10
157.344
F300W
52364.10
984.035
F336W
52364.11
750.429
F439W
52364.11
865.116
F555W
52364.11
3735.275
F675W
52364.11
2075.017
F814W
52364.11
1363.641
Results
The 8953 calibration program executed approximately 0.34 - 0.63 days following the
Decon on March 30, 2002.
Figure 4.1 through Figure 4.4 show the evolution of the photometric measurements
from 1995 until the present, for nine filters and all four chips. The star represents the post
-SM3B data point. The data were first normalized to a mean of 1.0 based on the historical
dataset. The residual trend was then fitted from MJD 51000 to the present, with the excep-
24
Instrument Science Report WFPC2 2002-06
tion of F675W for PC1 and WF2, which were fitted from MJD 50200 and 50700,
respectively, due to the paucity of recent data. The 1-sigma scatter (based on the empirical
scatter) of the baseline measurements are shown by dashed lines.
The observed long-term trend is a decline of the throughput, to varying degrees in the
different filters. The post-SM3B data point seems to follow the most recent trend. The
overall data for several UV filters (most notably F255W) seem to show a curious sharp
downward trend around MJD 51000, and we are currently investigating possible causes.
Figure 4.5 shows the statistical distribution of the post-SM3B photometry data points
around the predicted values based on the fits shown in Figure 4.1 through Figure 4.4, and
normalized by the 1-sigma error bars shown by the dotted lines. We find that the distribution has a mean value of 0.34 +/- 0.26 sigma for all the observations, hence there is no
obvious change in the throughput before and after SM3B. This compares to a value of 0.17 +/- 0.14 for SM2 (Biretta et al. 1997). The mean value in the PC appears to be
slightly higher, with a value of 1.31 +/- 0.56 sigma. However, even this is only about a 2sigma result, and is probably due to low number statistics. We note that in SM2 a similar
2-sigma result was found for the PC, but in the opposite sense (i.e. a decline instead of an
increase). The fact that the width of the distributions in Figure 4.5 are larger than 1 sigma,
suggests that there are other sources of uncertainty besides the historical random uncertainties, such as large fluctuations in focus or breathing.
Figure 4.6 shows the observed throughput decline compared with that predicted from
Dolphin’s CTE equations (2001) for F336W, F555W, and F814W. For all three filters the
throughput decline is consistent with the expected CTE loss, within the uncertainties.
25
Instrument Science Report WFPC2 2002-06
Figure 4.1: Photometric Throughput for PC1 in 9 filters. The star represents the post-SM3B data point. The
data were first normalized to a mean of 1.0 based on the baseline observations, and then the residual trend
was fitted from MJD 51000 to the present, with the exception of F675W which was fitted from MJD 50200
due to the paucity of recent data. The 1-sigma scatter (based on the empirical scatter) of the baseline measurements are shown by dashed lines.
26
Instrument Science Report WFPC2 2002-06
Figure 4.2: Same as Figure 4.1 for WF2. In this case, F675W was fitted from MJD 50700 due to the paucity
of recent data.
27
Instrument Science Report WFPC2 2002-06
Figure 4.3: Same as Figure 4.1 for WF3.
28
Instrument Science Report WFPC2 2002-06
Figure 4.4: Same as Figure 4.1 for WF4.
29
Instrument Science Report WFPC2 2002-06
Figure 4.5: Statistical distribution of the post-SM3B photometry data points (number of data points vs.
sigma), relative to the predicted values based on the fitted decline rates in Figure 4.1 through Figure 4.4.
30
Instrument Science Report WFPC2 2002-06
Figure 4.6: Actual throughput (solid lines with non-zero slope) vs. expected decline from Dolphin’s CTE
equations (dashed lines) for F336W, F555W, and F814W for the PC1 chip. This shows that the long-term
trends are due to CTE loss, within the measurement uncertainty. The discontinuity for the early data points
for F814W is due to a change in the exposure times.
31
Instrument Science Report WFPC2 2002-06
Conclusions and Recommendations
The WFPC2 calibration program 8953 was executed post-SM3B to check that the relative photometric throughput had not changed. Observations of the standard star
GRW+70D582 were obtained with each camera and with the normal selection of monitor
filters (11 filters). Each camera performed as expected, and WFPC2 was unaffected by the
SM3B Servicing Mission. The distribution shows a mean value of 0.34 +/- 0.26 sigma for
all the observations, consistent with an unchanged photometric throughput. Hence, the
response of the WFPC2 was essentially unchanged by the servicing mission. We also find
that the long-term throughput decline is consistent with the expected CTE loss.
References
Baggett, S. and Gonzaga, S. 1998, WFPC2 Long-Term Photometric Stability, WFPC2ISR-1998-03
Biretta J., Heyer I., Baggett S., et al. 1997, Results of the WFPC2 Post-Servicing Mission-2 Calibration Program, WFPC2-ISR-1997-09
Biretta J., McMaster M., Baggett S., and Gonzaga S. 1997, Summary of WFPC2
SM97 Plans, WFPC2-ISR-1997-03
Casertano, S., et al., editors, The 1997 HST Calibration Workshop Proceedings,
STScI, 1998, p.318
Dolphin, A. E. 2001 (April 2), private communication, (http://www.noao.edu/staff/dolphin/wfpc2_calib/)
Gonzaga, S., Ritchie, C., Baggett, S., Casertano, S., Whitmore, B., and Mutchler, M.
2000, “Standard Star Monitoring Memo”, (http://www.stsci.edu/instruments/
wfpc2/Wfpc2_memos/wfpc2_stdstar_phot3.html)
5: Point Spread Function
by Vera Kozhurina-Platais, Lori Lubin, and Anton Koekemoer
Summary: We examine the Point-Spread-Function (PSF) of the WFPC2 after the Servicing Mission 3b using dithered observations of the rich globular cluster, Omega Cen.
Comparisons to previous measurements of the PSF which were made after the last two
servicing missions (SM2 and SM3A) indicate that no significant variations in the PSF
have occurred. We also present an analysis of the shape of the PSF as a function of five
positions on the Planetary Camera (PC) chip.
32
Instrument Science Report WFPC2 2002-06
Note: These results have been reported independently in WFPC2 Technical Instrument
Report (TIR) 2002-01, SMOV3B Check of the WFPC2 Point-Spread-Function, by Vera
Kozhurina-Platais, Lori Lubin, and Anton Koekemoer.
Observations and Data Reduction
Following the procedures used after the previous two servicing missions (SM2 and
SM3A) (see Biretta et al. 1997; Casertano et al. 2000), images of the globular cluster
Omega Cen were obtained to characterize the Point-Spread-Function (PSF) of the WFPC2
after the Servicing Mission 3B (Proposal 8954). Four dithered images of Omega Cen, each
with a duration of 100 seconds, were obtained in the broad-band F555W and F814W filters on March 25, 2002. The dithered images were sub-stepped by one-third of a pixel in
each coordinate to provide a critically sampled PSF. We note that the focus had stabilized
by the time these observations were taken.
In order to make a comparison to previous measurements of the PSF made with Omega
Cen, we analyze and present the results in the F555W filter. The images were combined
using the dither routine in IRAF, and in particular the IRAF scripts of A. Fruchter (private
communication), according to the procedures described in the HST Dither Handbook
(Koekemoer et al. 2002). To produce our final drizzled images, we have used a scale that is
one-half of the original pixel size and a pixel footprint of 0.7.
In order to compare our measure of the PSF with previous measurements made after the
last two Servicing Missions, we have performed a similar analysis to that presented in
Biretta et al. (1997) and Casertano et al. (2000). In these studies, a composite PSF was
made using a number of bright stars which were not saturated, were well isolated, and
were located near the center of the PC chip. The PC was used because its pixel scale is
more than a factor of two smaller than in the Wide Field (WF) cameras. As a result, the
observed PC PSF is a better indicator of the telescope and instrumental optics than the
observed WF PSF which is dominated by the pixel response function.
For our study, we use the task psf in the IRAF package noao.digiphot.daophot. This task
implements the empirical PSF fitting as a sum of an analytical function and look-up
table(s) of residuals between the actual PSF and the fitting function (Stetson 1987; Stetson, Davis & Crabtree 1990). This lookup table is used as additive corrections from the
integrated analytic function to the actual observed empirical PSF. Five additional lookup
tables containing the first derivatives of the PSF in x and y and the second order derivatives of the image with respect to x2, x*y, and y2 are also written. This model permits the
PSF fitting process to take into account smooth linear or quadratic changes in the PSF
across the frame. It is well known that a Lorentzian function provides the best representation of the PSF in WFPC2 data because it has an undersampled core and extended wings.
However, we have instead used a Gaussian analytic function with the six lookup tables
described above. To construct a composite PSF, we select about 30 bright, unsaturated,
33
Instrument Science Report WFPC2 2002-06
and isolated stars at the center and the four corners of the PC chip and use the task psf to
create a two-dimensional, composite PSF from the stars in each of the five regions. This
technique allows us to make a comparison between previous measurements of the PC PSF
made in the center of the chip, as well as to characterize its variation across the chip.
Results
We use the task radprof in the IRAF package noao.digiphot.apphot to measure the radial
profile of the composite stellar image made from the stars in the central region of the PC
chip. The resulting radial profile is shown in Figure 5.1. The counts have been normalized
to 1 at the profile center and have been plotted as a function of the pixel distance from the
profile center. The best-fit Gaussian, with a full-width-half-maximum (FWHM) of 0.066",
is shown for comparison. The error on the FWHM is approximately +/- 0.002". The shape
of the PSF is unchanged with respect to the previous measurements. In Biretta et al.
(1997) and Casertano et al. (2000), the PC PSF is well described by a Gaussian curve
within the half-maximum point. They find that the FWHM of the best-fit Gaussian is
0.064", consistent with our measurement. In addition, we find an excess of up to 10% of
the peak counts with respect to the best-fit Gaussian outside that radius. That is, at distances greater than about 1 PC pixel from the profile center, we see that the stellar counts
have intensities relative to peak which are approximately 10% greater than the best-fit
Gaussian curve. This behavior was also observed in the measurements made in 1997 and
2000.
Figure 5.1: Radial profile of the PSF in the central region of the WFPC2 PC image of Omega Cen taken
after SM3B. A composite stellar image was created using about 30 isolated, unsaturated stars near the center
of chip with the IRAF task psf. The radial profile was measured using the IRAF task radprof. The best-fit
Gaussian model with FWHM = 0.066" is shown by the solid line.
34
Instrument Science Report WFPC2 2002-06
We have also examined how the PSF varies with chip position. To examine this behavior,
we have calculated the radial profile of the composite stellar images by making cuts along
both the x and y direction. In Figure 5.2, we show the x and y radial profiles as a function
of five chip positions. In the five panels, we plot the PSF, from top to bottom, for the center, lower left, lower right, upper right, and upper left, respectively. The observed PSF is
shown as the dotted line, while the solid line represents the analytic Gaussian function.
The width of the Gaussian model (in PC pixel units) is listed in each panel. To enhance the
visibility of the fine details in the observed PSFs, we show in Figure 5.3 the difference
between the observed profile and the fitted model. We can easily see from the residuals
that there is a difference of about 10% in the profile wings and about 20% in the profile
core which is due to the poor representation of the observed PSF by the analytic Gaussian
function.
In Figure 5.4., we show the contour plots of the composite stellar image as a function
of chip position. As previously noted (Krist & Burrows 1995), the observed PSF varies as
a function of the position on the PC chip, with the off-center PSFs being noticeably more
asymmetric due to coma and astigmatic aberrations. This behavior is consistent with the
level of residuals observed in Figure 5.3.
35
Instrument Science Report WFPC2 2002-06
Figure 5.2: Composite stellar profile as a function of position on the PC chip. Radial profiles were made by
making cuts along the x (left) and y (right) axis. The dotted and solid lines indicate the observed profile and
the fitted Gaussian model, respectively. The width of the Gaussian model (in PC pixels) is listed in each
panel. The panels, from top to bottom, represent the observed PSF in the center, lower left, lower right,
upper right, and upper left of the chip, respectively.
36
Instrument Science Report WFPC2 2002-06
Figure 5.3: Residuals (in percent) between the observed PSF (dotted lines in Figure 5.2) and the best-fit
Gaussian model (solid lines in Figure 5.2) as a function of chip position. The panels are the same as shown
in Figure 5.2.
37
Instrument Science Report WFPC2 2002-06
Figure 5.4: Contour plots of the composite PSF as a function of position on the PC chip. The panels, from
top to bottom, represent the observed PSF in the center, lower left, lower right, upper right, and upper left of
the chip, respectively.
38
Instrument Science Report WFPC2 2002-06
Conclusions
We have measured the PSF on the PC chip of WFPC2 after the Servicing Mission
SM3B. Comparisons made with previous measurements of the width and shape of the PSF
indicate that no substantial changes in the PSF have occurred. The PSF can be defined by
a Gaussian curve within the half-maximum point. The FWHM is approximately 0.066",
fully consistent with measurements made after the previous two Servicing Missions. We
also characterize the change in the PSF as a function of chip position by measuring the
PSF at the center and the four corners of the PC chip. As previously noted, there are clear
differences in the shape of the PSF, with the off-center PSF being noticeably more
asymmetric.
References
Biretta, J. et al. 1997, “Results of the WFPC2 Post-Servicing Mission-2 Calibration
Program,” Instrument Science Report WFPC2-1997-09.
Casertano, S. et al. 2000, “Results of the WFPC2 Observatory Verification after Servicing Mission 3a,” Instrument Science Report WFPC2-2000-02.
Koekemoer, A.M. et al. 2002, “HST Dither Handbook,” Version 2.0 (Baltimore:
STScI).
Stetson, P.B. 1987, PASP, 99, 191.
Stetson, P.B., Davis, L.E., and Crabtree, D.R. 1990, in ASP Conf. Ser. 8, “CCDs in
Astronomy,” ed. G.H. Jacoby, 289
Krist, J., & Burrows, C. 1995, Applied Optics 34, 4952.
6: Camera Internal Monitoring
by Anton Koekemoer, Vera Kozhurina-Platais, and Shireen Gonzaga
Summary: We report on the stability of WFPC2 with special attention to any potential
changes due to SM3B as determined from internal calibration observations, including
read noise, dark current, internal flats, gain ratios, and position of the K-spots. We find no
significant changes in the read noise at either gain 7 or 15. Dark current increases by a
few percent (2 to 8% depending on chip), consistent with the steady increase found by
Baggett et al (1999). Small changes (0.1% to 1%) in the pattern and relative intensity of
the INTFLAT and VISFLAT images are compatible with differential variations in the light
output of the lamps. Overall, the WFPC2 appears to be very stable, exhibiting only the
minor changes expected due to known low level variability.
39
Instrument Science Report WFPC2 2002-06
Introduction
We describe the results of the internal WFPC2 calibration observations (program 8950)
that were obtained during March and April 2002 as part of SMOV3B for WFPC2. The
goal of these observations was to ensure that there was no significant change in the basic
WFPC2 instrument health and internal calibrations.
The internal monitoring observations for program 8950 were carried out during the
four weeks after WFPC2 cooldown, following the Bright Earth Avoidance period (BEA),
and included bias frames, dark frames, INTFLATs, VISFLATS, and Kellsall-spot (K-spot)
images. These observations were repeatedly obtained at a number of time periods, in the
following visits:
•
Visit 14: starting 3 hours after cooldown
•
Visit 40: 7 days after cooldown, pre-decon
•
Visit 47: 7 days after cooldown, post-decon
•
Visit 50: 14 days after cooldown, pre-decon
•
Visit 53: 14 days after cooldown, post-decon
•
Visit 60: 21 days after cooldown
•
Visit 70: 28 days after cooldown, pre-decon
•
Visit 73: 28 days after cooldown, post-decon
In Table 6.1 we give a breakdown of the structure of each of these visits for which biases,
darks, INTFLATS and K-spots were obtained. Each visit had the same structure. We note
that the VISFLATS were obtained only once, in a separate visit (number 49) about 26
hours after cooldown, and they are presented separately in Table 6.12 on page 50, in the
section dealing specifically with VISFLATS.
Table 6.1:Structure of each visit for the internal biases, darks, INTFLATs and K-spots.
Type of
Observation
Number of
Exposures
Exposure
Time
Gain
Filter
(if applicable)
BIAS
2
0s
7
-
BIAS
2
0s
15
-
DARK
5
1800s
7
-
INTFLAT
2
10s
7
F555W
INTFLAT
2
20s
15
F555W
KSPOTS
1s
1
15
-
KSPOTS
2.6s
1
15
-
40
Instrument Science Report WFPC2 2002-06
Bias Frames and Read-Out Noise
The goal was to measure any changes in the biases to a precision of about 0.1 DN for
gain 15 and 0.2 DN for gain 7 (thus about 1.4 electrons/pixel). We obtained a total of sixteen biases in each of these two gain settings during the SMOV3B period. We retrieved the
same number of pre-SM3B biases from proposals 8933 and 8939 during January - February 2002 (starting just before SM3B and working backwards). There were some small bias
jumps in the data but no other anomalies. For each chip, we then measured the read-noise
(standard deviation) of each frame, as well as the standard deviation of this set of measurements. The results are given in Table 6.2 and Table 6.3, for the two gain settings of 7 and
15. We converted from DN to electrons using the current gain values in Table 4.4 of the
WFPC2 Instrument Handbook for Cycle 11, V 6.0 (Biretta et al., 2001).
Table 6.2:Post/Pre-SM3B WFPC2 read-noise measured with GAIN=7.
Pre-SM3B
Chip
Number
Post-SM3B
e-/pixel
DN/pixel
e-/pixel
DN/pixel
Post-SM3B Pre-SM3B
change
(e-/pixel)
mean
stddev
mean
stddev
mean
stddev
mean
stddev
PC
0.7677
0.0157
5.466
0.112
0.7660
0.0040
5.454
0.029
-0.012 0.116
WF2
0.7668
0.0108
5.459
0.076
0.7684
0.0088
5.471
0.062
+0.012 0.098
WF3
0.7619
0.0100
5.257
0.069
0.7517
0.0112
5.187
0.077
-0.070 0.103
WF4
0.7986
0.0195
5.670
0.138
0.7610
0.0182
5.403
0.129
-0.267 0.189
Table 6.3:Post/Pre-SM3B WFPC2 read-noise measured with GAIN=15.
Pre-SM3B
Chip
Number
DN/pixel
Post-SM3B
e-/pixel
DN/pixel
e-/pixel
Post-SM3B Pre-SM3B
change
(e-/pixel)
mean
stddev
mean
stddev
mean
stddev
mean
stddev
PC
0.5541
0.0210
7.751
0.294
0.5566
0.0166
7.786
0.234
+0.035 0.376
WF2
0.5508
0.0301
7.987
0.436
0.5334
0.0162
7.734
0.235
-0.253 0.495
WF3
0.6033
0.0131
8.416
0.183
0.5909
0.0112
8.243
0.156
-0.173 0.240
WF4
0.6300
0.0167
8.789
0.233
0.6088
0.0161
8.493
0.224
-0.296 0.323
We also combined the biases to make a “master” image for each gain, pre- and postSM3B. We found no obvious differences, either in the images themselves or in the row
and column averages, above very low levels of 0.01 - 0.02 DN.
In Figure 6.1and Figure 6.2 we present plots showing the measured standard deviation
of the bias frames, for each of the four WFPC2 chips (PC, WF2, WF3, WF4), for gain 7
and 15, comparing the pre-SM3B and post-SM3B values.
41
Instrument Science Report WFPC2 2002-06
Figure 6.1: Comparison of the standard deviation of bias frames taken
before and after SM3B, for the four WFPC2 chips (PC, WF2, WF3, WF4),
at gain=7.
Figure 6.2: Comparison of the standard deviation of bias frames taken
before and after SM3B, for the four WFPC2 chips (PC, WF2, WF3, WF4),
at gain=15.
It can be seen from these tables and figures that there are no statistically significant differences across SM3B for the read noise in any of the chips, for both gain settings, to
levels well below our formal measurement errors. Therefore, we conclude that he overall
result from these measurements is that there is no significant change in the read-noise
from post-SM3B as compared with pre-SM3B.
42
Instrument Science Report WFPC2 2002-06
Dark Current
The goal was to measure any changes in the darks to ~1.4 electrons/pixel (0.2 DN for
gain 7). We obtained a total of forty 1800s darks during the SMOV3B period (March April 2002). For the pre-SM3B data we extracted from the archive the same number of
1800s darks taken under programs 8932, 8933 and 8935 in gain 7 from Jan - Feb 2002,
starting just before SM3B and working backwards in time. We combined the darks using
imcombine (with crrej) to remove cosmic rays and determined the median value in several regions, corresponding to (1) most of the chip, (2) the center 400x400 pixels and (3)
the edge columns following Baggett et al. (1998) and Casertano et al. (2000).
Results are given in Table 6.4 for the pre-SM3B darks and in Table 6.5 for the postSM3B darks. The total exposure time for each set of darks is 72,000s, and the median
counts/s in 1800s of exposure time is given for comparison with the measurements of Baggett et al. (1998). The dark currents were converted to electrons/sec using the gains in
Table 4.4 of the Cycle 11 WFPC2 instrument Handbook (V.6.0, June 2001).
The data show no strong changes at all in the dark current between January/February
2002 and March/April 2002, above values of a few percent (which are well below our
measurement errors). Figure 6.3 shows the dark current in DN/1800s as a function of time
from December 2001 to April 2002. Differences between pre- and post-SM3B darks also
showed no significant structure in the central 400x400 pixels. A very small change (~ 0.2 0.3%) is found in the row averages, but not in the column averages. Direct inspection
shows that the variations are due to spurious features in the pre-SM3B data which are not
present in the post-SM3B data. However, the mean values of the dark current for each chip
across SM3B remain effectively the same, to well within a few percent, and there appears
to be no significant effect from SM3B on the dark current.
Figure 6.3: Mean value of the dark current (in DN) as a function of time from December 2001 to
April 2002. The dark current is for the total mean exposure time of 1800s, in order to facilitate comparison with Baggett et al. (1998).
43
Instrument Science Report WFPC2 2002-06
Table 6.4:Dark Current, Pre-SM3B.
a.
Chip
Number
Pixel Area
median DN/pixa
(in 1800s)
Dark Current
(e-/sec/pixel)
PC1
[200:600,200:600]
1.708
0.00676
WF2
[200:600,200:600]
1.156
0.00457
WF3
[200:600,200:600]
1.680
0.00644
WF4
[200:600,200:600]
1.534
0.00606
PC1
[50:750,50:750]
1.665
0.00659
WF2
[50:750,50:750]
1.118
0.00442
WF3
[50:750,50:750]
1.598
0.00613
WF4
[50:750,50:750]
1.477
0.00583
PC1
[10:100,50:750]
1.528
0.00604
WF2
[10:100,50:750]
0.991
0.00392
WF3
[10:100,50:750]
1.318
0.00505
WF4
[10:100,50:750]
1.311
0.00517
The median DN/pixel in 1800s is given for comparison with the
measurements of Baggett et al. (1998).
Table 6.5:Dark Current, Post-SM3B.
a.
Chip
Number
Pixel Area
median DN/pixa
(in 1800s)
Dark Current
(e-/sec/pixel)
∆DN (Post-SM3B Pre-SM3B)
PC1
[200:600,200:600]
1.780
0.00704
+0.00028 (+4.1%)
WF2
[200:600,200:600]
1.064
0.00421
-0.00036 (-7.9%)
WF3
[200:600,200:600]
1.661
0.00637
-0.00007 (-1.1%)
WF4
[200:600,200:600]
1.581
0.00624
+0.00018 (+2.9%)
PC1
[50:750,50:750]
1.715
0.00678
-
WF2
[50:750,50:750]
0.994
0.00393
-
WF3
[50:750,50:750]
1.622
0.00622
-
WF4
[50:750,50:750]
1.532
0.00604
-
PC1
[10:100,50:750]
1.531
0.00606
-
WF2
[10:100,50:750]
0.917
0.00363
-
WF3
[10:100,50:750]
1.367
0.00524
-
WF4
[10:100,50:750]
1.284
0.00507
-
The median DN/pixel in 1800s is given for comparison with the measurements of Baggett et al. (1998).
44
Instrument Science Report WFPC2 2002-06
INTFLATs
Our goal was to measure changes in the INTFLATs of about 1% accuracy, for each of
the four chips, in both gain settings, between post-SM3B and pre-SM3B. In addition, we
used this program to verify the stability of the gain ratios for each of the four chips, by
obtaining a measurement of the ratios of counts in the two gains before SM3B, and comparing this with a measurement of the gain ratios after SM3B. Finally, we carried out these
comparisons separately for INTFLATS obtained using Shutter Blade A versus those
obtained using Shutter Blade B. This is because the INTFLATS are obtained by reflecting
light from the INTFLAT lamp off the interior surface of each shutter, hence this provides a
sensitive tracer of whether the shutter behavior has changed at all across SM3B.
We used the IRAF task imcombine (rejecting cosmic rays using the crrej option) to
create four POST-SM3B master flats in the F555W filter: two at GAIN=7 (one with Shutter Blade A and the other with Blade B), and the other two at GAIN=15, again using
Shutter Blades A and B separately. We created another four pre-SM3B master flats using
the same number of images taken before SM3B, from the January - February 2002 timeframe, starting just before the Servicing Mission and working backwards in time. The
images used are summarized in Table 6.6.
We generated ratios of the post-SM3B master images with each of the earlier sets, and
compared their statistics with the ratio of early to mid-2001 master images. The results are
presented in Table 6.7 and Table 6.8, for Shutter Blade A and B respectively. It can be seen
that, overall, the counts with Shutter Blade A have remained very stable, in both gain settings, with only very small changes less than about 0.1-0.3%, well below our typical
measurement uncertainties of 0.5-1%. Such changes are consistent with the long-term
behavior of the Carley bulbs in the INTFLAT lamps, which are well-known to exhibit
time-dependent changes of order a few tenths of a percent (O’Dea et al. 1999). With Shutter Blade B the changes are slightly higher, up to 0.5% for WF4 and 0.6% for the PC.
However, these changes are still below our formal measurement uncertainties of 0.7% and
1% for these two chips respectively. For the other two chips the changes are less than
0.25% in both gains.
We next considered the information on gain ratios from the INTFLATs. Statistics of
images were taken over the whole chip [50:750,50:750] of the ratio of INTFLAT F555W
masters with gain 7 divided by gain 15. The ratio was normalized by the total exposure
times for the master images (10s and 20s for gain=7 and 15 respectively). The statistics of
the ratio images (gain 7/gain 15) are given in Table 6.9 and Table 6.10. The last column is
the ratio of the post-SM3B ratio divided by the pre-SM3B ratio. In general the ratios of
INTFLAT count rates in the two gains are extremely steady, for both shutters, to levels
well below 0.05%.
45
Instrument Science Report WFPC2 2002-06
Table 6.6:INTFLAT Observations from before and after SM3B.
INTFLAT Observations Obtained Before SM3B
GAIN=7
GAIN=15
Filename
Observation Date
ExpTime Shutter
Filename
Observation Date
u6hie205m
2002-01-22 00:10
10s
u6hie206m
2002-01-22 00:12
u6hief05m
u6hief06m
ExpTime Shutter
A
u6hie207m
2002-01-22 00:14
20s
A
10s
A
u6hie208m
2002-01-22 00:16
20s
A
2002-01-23 17:27
10s
B
u6hief07m
2002-01-23 17:27
20s
B
2002-01-23 17:29
10s
B
u6hief08m
2002-01-23 17:29
20s
B
u6hg3005m
2002-01-28 10:53
10s
B
u6hg300bm
2002-01-28 11:05
20s
B
u6hg3006m
2002-01-28 10:55
10s
B
u6hg300cm
2002-01-28 11:07
20s
B
u6hg3205m
2002-02-04 04:18
10s
B
u6hg320br
2002-02-04 04:30
20s
B
u6hg3206m
2002-02-04 04:20
10s
B
u6hg320cr
2002-02-04 04:32
20s
B
u6hg3305m
2002-02-11 14:51
10s
A
u6hg330bm
2002-02-11 15:03
20s
A
u6hg3306m
2002-02-11 14:53
10s
A
u6hg330cm
2002-02-11 15:05
20s
A
u6hjf205m
2002-02-18 01:27
10s
B
u6hjf207m
2002-02-18 01:31
20s
B
u6hjf206m
2002-02-18 01:29
10s
B
u6hjf208m
2002-02-18 01:33
20s
B
u6hjff05m
2002-02-19 03:35
10s
A
u6hjff07m
2002-02-19 03:39
20s
A
u6hjff06m
2002-02-19 03:37
10s
A
u6hjff08m
2002-02-19 03:41
20s
A
u6hg3405m
2002-02-25 11:40
10s
A
u6hg340bm
2002-02-25 11:52
20s
A
u6hg3406m
2002-02-25 11:42
10s
A
u6hg340cm
2002-02-25 11:54
20s
A
INTFLAT Observations Obtained After SM3B
GAIN=7
GAIN=15
Filename
Observation Date
ExpTime Shutter
Filename
Observation Date
u8cz1405r
2002-03-23 12:13
10s
u8cz1406r
2002-03-23 12:15
u8cz4005m
u8cz4006m
A
u8cz1407r
2002-03-23 12:17
20s
A
10s
A
u8cz1408r
2002-03-23 12:19
20s
A
2002-03-29 10:41
10s
B
u8cz4007m
2002-03-29 10:45
20s
B
2002-03-29 10:43
10s
B
u8cz4008m
2002-03-29 10:47
20s
B
u8cz4705r
2002-03-30 18:43
10s
A
u8cz4707r
2002-03-30 18:47
20s
A
u8cz4706r
2002-03-30 18:45
10s
A
u8cz4708r
2002-03-30 18:49
20s
A
u8cz5005m
2002-04-05 06:09
10s
B
u8cz5007m
2002-04-05 06:13
20s
B
u8cz5006m
2002-04-05 06:11
10s
B
u8cz5008m
2002-04-05 06:15
20s
B
u8cz5305m
2002-04-06 05:45
10s
B
u8cz5307m
2002-04-06 05:49
20s
B
u8cz5306m
2002-04-06 05:47
10s
B
u8cz5308m
2002-04-06 05:51
20s
B
u8cz6005m
2002-04-13 00:49
10s
B
u8cz6007m
2002-04-13 00:53
20s
B
u8cz6006m
2002-04-12 00:51
10s
B
u8cz6008m
2002-04-13 00:55
20s
B
u8cz7005m
2002-04-16 23:06
10s
A
u8cz7007m
2002-04-16 23:10
20s
A
u8cz7006m
2002-04-16 23:08
10s
A
u8cz7008m
2002-04-16 23:12
20s
A
u8cz7305m
2002-04-18 06:00
10s
A
u8cz7307m
2002-04-18 06:04
20s
A
u8cz7306m
2002-04-18 06:02
10s
A
u8cz7308m
2002-04-18 06:06
20s
A
46
ExpTime Shutter
Instrument Science Report WFPC2 2002-06
Table 6.7:Post/Pre-SM3B INTFLAT Count Ratios with Shutter Blade A.
Chip
Number
Gain=7
Gain=15
median
stddev
median
stddev
PC
0.999571
0.0124542
0.999588
0.0087281
WF2
0.999470
0.0051768
0.999535
0.0037075
WF3
0.997728
0.0058004
0.997905
0.0044924
WF4
0.999125
0.0063868
0.999101
0.0045931
Table 6.8:Post/Pre-SM3B INTFLAT Count Ratios with Shutter Blade B.
Chip
Number
Gain=7
Gain=15
median
stddev
median
stddev
PC
0.994063
0.0136805
0.993786
0.0099220
WF2
1.002250
0.0084388
1.002420
0.0077554
WF3
0.999820
0.0071063
1.000060
0.0055974
WF4
0.994562
0.0080512
0.995050
0.0065167
Table 6.9:Post/Pre-SM3B INTFLAT Gain Ratios with Shutter Blade A.
Chip
Number
Pre-SM3B
Post-SM3B
Post/PreSM3B
median
stddev
median
stddev
PC
1.94109
0.02239
1.94111
0.02245
1.00001
WF2
1.96854
0.00902
1.96845
0.00911
0.99995
WF3
1.97375
0.00911
1.97348
0.00918
0.99986
WF4
1.92595
0.01135
1.92598
0.01162
1.00002
Table 6.10:Post/Pre-SM3B INTFLAT Gain Ratios with Shutter Blade B.
Chip
Number
Pre-SM3B
Post-SM3B
Post/PreSM3B
median
stddev
median
stddev
PC
1.94105
0.02383
1.94152
0.02533
1.00024
WF2
1.96905
0.00941
1.96905
0.00939
1.00000
WF3
1.97408
0.01034
1.97352
0.00950
0.99972
WF4
1.92695
0.01322
1.92604
0.01241
0.99952
47
Instrument Science Report WFPC2 2002-06
Finally, we examined the stability of the illumination pattern of the INTFLATs, by
comparing the images of the ratio of the pre- and post-SM3B INTFLATs for F555W, as
shown in Figure 6.4. The large-scale variations are summarized in Table 6.11. We find that
in most of the chips there are large-scale features with amplitudes of only about 0.1-0.3%,
although in a couple of cases the features range up to one percent. The largest variations
occur in the corners and edges of the images, while the inner 400x400 regions are relatively more stable.
These small changes in the illumination pattern are likely caused by the fact that the
four Carley bulbs do not vary in the same manner. As documented previously by O’Dea et
al. (1999), these changes are wavelength dependent which suggests that the changes in
brightness are associated with changes in temperature. These results have implications for
the use of INTFLATs for pre-flashing to reduce CTE. The master INTFLAT which is used
to subtract the signature of the INTFLAT from the background should be constructed of
INTFLATs taken close to the observation - determination of the exact time period requires
further study.
In conclusion, we have determined that the INTFLATS have remained stable across
SM3B, with the small observed changes (a few tenths of a percent) being most likely due
to well-documented changes in the lamps, along with some slight differences between the
shutters which is to be expected. However, we see no evidence of any significant changes
in the sensitivity of the chips themselves.
48
Instrument Science Report WFPC2 2002-06
intflat_gain15_A_ratio_post_pre_sm3b_mosaic.hhh - INTFLAT_GAIN15_A_RATIO_POST_PRE_SM3B_MOSAIC[1/1]
1500
1000
500
100
200
300
400
500
z1=0.98 z2=1.02 ztrans=linear Con=1.00 Brt=0.50 cmap=Grayscale ncolors=138
0
0
NOAO/IRAF koekemoe@wallaby.stsci.edu Thu Sep 26 13:10:18 2002
0
1
1
1
Figure 6.4: Ratio of post-SM3B/pre-SM3B INTFLATs, covering a +/- 2% scale centered on a value
of 100%. Overall the changes are less than 1%, and are all characteristic of well-documented longterm changes in the internal flat lamps (O’Dea et al. 1999), and are not the result of SM3B.
Table 6.11: Summary of Large Scale Variations in INTFLAT Ratios.
Region
Post/Pre-SM3B
GAIN=7
GAIN=15
PC1 Rows
0.1%
0.25%
PC1 Columns
<0.1%
0.1%
PC1 Center Columns
<0.1%
0.1%
WF2 Rows
0.1%
0.2-0.3%
WF2 Columns
0.1%
0.3-0.7%
WF2 Center Columns
0.1%
0.l2-0.4%
WF3 Rows
0.2-0.4%
0.1%
WF3 Columns
0.2-0.4%
0.2%
WF3 Center Columns
0.1-0.2%
0.1%
0.1%
0.1%
WF4 Columns
0.3-0.7%
0.5-1.0%
WF4 Center Columns
0.3-0.5%
0.5-1.0%
WF4 Rows
49
Instrument Science Report WFPC2 2002-06
VISFLATs
The goal was to measure changes in the VISFLATs to ~ 1%. Due to the decline in the
VISFLAT lamps with usage (Stiavelli & Baggett 1996), VISFLATs are now taken very
rarely. The most recent sweep before SM3B was obtained in April 2001, through 4 photometric filters plus the FR533N filter This filter sweep was repeated in the post-SM3B data
set taken on March 31 2002. In Table 6.12 we summarize the VISFLAT datasets from
before and after SM3B that were used. All data were obtained at GAIN=7.
Table 6.12:VISFLAT datasets from before and after SM3B.
VISFLAT Observations Obtained Before SM3B
Filename
Observation Date
ExpTime
Filter
Gain
u69q2901r
2001-04-18 02:41
40
F439W
7
u69q2902r
2001-04-18 02:43
40
F439W
7
u69q2903r
2001-04-18 02:46
1.6
F555W
7
u69q2904r
2001-04-18 02:48
1.6
F555W
7
u69q2905r
2001-04-18 02:51
0.6
F675W
7
u69q2906r
2001-04-18 02:53
0.6
F675W
7
u69q2907r
2001-04-18 02:56
0.6
F814W
u69q2908r
2001-04-18 02:58
0.6
F814W
7
u69q2909r
2001-04-18 03:01
60
FR533N
7
u69q290ar
2001-04-18 03:04
120
FR533N
15
VISFLAT Observations Obtained After SM3B
Filename
Observation Date
ExpTime
Filter
Gain
u8cz4905n
2002-03-31 00:05
600
F336W
7
u8cz4906n
2002-03-31 00:17
600
F336W
7
u8cz4903n
2002-03-31 00:00
40
F439W
7
u8cz4904n
2002-03-31 00:02
40
F439W
7
u8cz4907n
2002-03-31 00:30
1.6
F555W
7
u8cz4908n
2002-03-31 00:32
1.6
F555W
7
u8cz4909n
2002-03-31 00:35
0.6
F675W
7
u8cz490an
2002-03-31 00:37
0.6
F675W
7
u8cz490bn
2002-03-31 00:40
0.6
F814W
7
u8cz490cn
2002-03-31 00:42
0.6
F814W
7
u8cz4901n
2002-03-30 23:52
60
FR533N
7
u8cz4902n
2002-03-30 23:55
120
FR533N
15
50
Instrument Science Report WFPC2 2002-06
In order to check the stability of the flat fields and the lamp illumination, we combined
the pairs of images at each filter using crrej and divided the pre-SM3B images by the
post-SM3B images. The statistics of the ratio images. for the section [50:750,50:750], are
given in Table 6.13. Due to previously document changes in the VISFLAT lamps (O’Dea
et al. 1999), the ratios averaged over the chips are not quite unity. The lamps have
decreased in brightness by about 0.6-0.9% at F439W, 1.2-1.5% at F555W, 1.1-1.5% at
F675W, and 0.8-1.1% at F814W. The mean values of the images have a standard deviation
of 1.6% and 0.8% for the PC and WF chips, respectively.
Table 6.13:Statistics of Ratios of Post-/Pre-SM3B VISFLAT Images.
chip
filter
median
mean
stddev
PC1
F439W
0.99311
0.99321
0.01581
WF2
F439W
0.99344
0.99346
0.00747
WF3
F439W
0.99355
0.99356
0.00752
WF4
F439W
0.99181
0.99182
0.00794
PC1
F555W
0.98611
0.98621
0.01585
WF2
F555W
0.98730
0.98733
0.00754
WF3
F555W
0.98704
0.98706
0.00754
WF4
F555W
0.98490
0.98489
0.00789
PC1
F675W
0.98757
0.98776
0.01627
WF2
F675W
0.98832
0.98836
0.00776
WF3
F675W
0.98805
0.98807
0.00779
WF4
F675W
0.98586
0.98586
0.00809
PC1
F814W
0.99005
0.99030
0.01653
WF2
F814W
0.99162
0.99166
0.00794
WF3
F814W
0.99118
0.99120
0.00789
WF4
F814W
0.98918
0.98916
0.00816
The ratio images themselves are generally smooth and featureless, with the only
apparent changes corresponding to those previously documented in the study of long-term
flatfield evolution based on Earthflats (Koekemoer et al. 2002). These are likely due to
slow, long-term changes in the camera geometry.Thus, we conclude that there are no significant changes in either the flat fields or the VISFLAT illumination pattern as a result of
SM3B.
51
Instrument Science Report WFPC2 2002-06
K-SPOTS
The goal was to measure relative shifts of the K-spots to an accuracy of 0.1 pixel. We
determined the initial location of the K-spots by displaying the images and using imexamine. The positions were then used as input to the IRAF task center, which was run on the
K-spot images using the centroid algorithm with a box of 3 pixels. We used exposures of
the 2.6s and 1s for PC1 and the WF chips, respectively. We obtained measurements for
Feb. 19, 2002 (just before SM3B) and two sets after SM3B (March 23 and April 18,
2002). The average measured shifts (in pixels) are tabulated in Table 6.14 and displayed in
Figure 6.5, after having converted to milliarcseconds using the known pixel scales.
Table 6.14:Average K-spot Shifts (in pixels), given in the sense of later date - earlier date.
Chip
Number
Number
of
K-spots
PC1
February 19 - March 23, 2002
March 23 - April 18, 2002
avg x
shift
rms
avg y
shift
rms
avg x
shift
rms
avg y
shift
rms
10
0.0436
0.0208
0.0318
0.0382
-0.1378
0.1623
0.0115
0.0441
WF2
18
0.0069
0.0450
-0.0148
0.0651
-0.0001
0.0415
0.0208
0.0513
WF3
20
-0.0172
0.0617
0.0307
0.0612
0.0276
0.0579
-0.0159
0.0467
WF4
20
-0.0247
0.0399
-0.0076
0.0292
0.0120
0.0507
0.0328
0.0405
February 19 -> March 23, 2002
PC
WF2
WF3
WF4
20
Y-shift (milliarcsec)
March 23 -> April 18, 2002
10
0
-10
-20
-20
-10
0
10
X-shift (milliarcsec)
20
-20
-10
0
10
X-shift (milliarcsec)
20
Figure 6.5: Measured X and Y shifts in the cameras, in milliarcseconds, relative to pre-SM3B.
The shifts for each chip are displaced from zero by a few milliarcseconds, with a comparable scatter. The values for each chip tend to clump together, but there are no large
systematic variations larger than a few milliarcseconds in any of the chips. The average
value of the shifts are all less than 0.14 pixels for the PC, and 0.06 pixels for the WF chips.
These small shifts are generally consistent with the measured time-evolution of the Kspots found by Casertano and Wiggs (2001). Therefore, we conclude that there are no significant changes in the relative camera positions as a result o fSM3B.
52
Instrument Science Report WFPC2 2002-06
Conclusions
We have studied the stability of WFPC2 with special attention to any potential changes
due to SM3B as determined from internal calibration observations. Over the course of the
period January/February 2002 to March/April 2002, there were no significant changes in
the gain 7 or 15 read noise. The data from the dark images are also consistent with no significant increase in the dark current from before and after SM3B.
We see only very small changes (few tenths of a percent) in the INTFLAT intensity, in
both gains, as well as low-level changes in the illumination pattern of the INTFLATs associated with the changes in intensity. Ratios of INTFLATs taken from before and after
SM3B show small amplitude (0.1 - 1%) large-scale variations which are chip and wavelength dependent. These are very likely due to the previously known time-dependent
changes in the Carley bulbs, and are not the result of SM3B.
The gain ratios in all the chips, as measured from the INTFLAT count ratios taken
before and after SM3B, remain extremely constant to levels below 0.02 - 0.05%, indicating that there have been no measurable changes in these signal chains as a result of SM3B.
The ratios of VISFLATs taken in April 2001 and March 2002 show that there are no
substantial changes in either the flat fields or the VISFLAT illumination pattern due to
SM3B. The only changes are on very low levels below 1%, and are entirely consistent
with long-term evolution of the cameras as documented in previous flatfield studies.
Over the period February to April 2002, the Kelsall spots shifted by average values of
5-10 milliarcseconds. These shifts are entirely consistent with the known time evolution of
the K-spot positions, and do not indicate any unusual behavior in the camera geometry.
In summary, WFPC2 appears to be very stable overall, exhibiting only the minor
changes expected due to known low level variability, with no significant changes
apparently attributable to SM3B.
References
Baggett, S., Casertano, S., and Wiggs, M. S., 1998, TIR WFPC2-1998-03
Casertano, S. and Wiggs, M. 2001, “An Improved Geometric Solution for WFPC2”,
Instrument Science Report WFPC2-2001-10
Koekemoer, A. M., Biretta, J. and Mack, J., 2002, “Updated WFPC2 Flatfield Reference Files for 1995 - 2001”, Instrument Science Report WFPC2-2002-02
Mutchler, M. and Stiavelli, M., 1997, TIR WFPC2-1997-07
O’Dea, C. P., Gonzaga, S., McMaster, M., Heyer, I., Hsu, J. C., Baggett, S., and Rudloff, K., 1997, Instrument Science Report WFPC2-1997-04
O’Dea, C., Mutchler, M., and Wiggs, M. 1999, ISR WFPC2-1999-01
53
Instrument Science Report WFPC2 2002-06
7: SMOV3a Flat Field Stability Check
by Anton M. Koekemoer and Inge Heyer
Summary: We compare WFPC2 Earthflats taken before and after the 2002 Servicing Mission 3B. Most of the field-of-view shows no change (<0.3%) in flat field calibration. The
only changes on large scales are at very low levels (<0.1-0.2%) and are likely attributable
to long-term changes in the camera geometry, rather than SM3B.
Note: these results have been reported independently as WFPC2 Technical Instrument
Report (TIR) 2002-02, SMOV3B Flat Field Verification, by Anton Koekemoer and Inge
Heyer.
Introduction
As part of our post-servicing check-out of WFPC2 after SM3B, we have examined a
series of Earthflats to test the flat field stability, over a wide range of narrow-band filter
including F375N, F502N, F656N, and F953N. The goal of these observations is to test for
any unexpected OTA obscuration or contamination in WFPC2 that may have occurred as a
result of the servicing mission. The flats are also capable of revealing changes in the OTA/
WFPC2 geometry, as well as any QE changes localized to one CCD camera or to a small
region of the field-of-view. While internal flats can provide some of this information, the
Earthflats are unique in providing an end-to-end test of the OTA+WFPC2 system. Detailed
discussion of Earthflats and WFPC2 flat fields can be found elsewhere (Biretta et al. 1995;
Koekemoer, Biretta & Mack 2002).
Observations and Analyses
Earthflats were observed as part of the routine calibration proposals 8815 and 8940 for
late Cycle 9 and early Cycle 10, as well as proposal 8495 during SMOV3B. All the F502N
flats are 1.2 second exposures of the bright Earth made with gain 15. Since we are interested primarily in changes to the flat field, we select subsets of these to construct a preSMOV and an SMOV flat.
For the pre-SMOV observations we started with a total of 124 Earthflats in F502N
taken between June 2001 and February 2002 as part of proposals 8815 and 8940. We discarded images with mean counts in the PC1 below 500 DN and mean counts in the three
WFC chips above 3200 DN (to avoid saturation), and furthermore selected only those that
had been obtained within 7 days after a decontamination (to minimize effects from contamination). These images were then examined for streaks (produced by features on the
Earth moving across the detector), after multiplying with the current F502N flat field reference file. The streaks increase the overall RMS of the image; displaying the images and
54
Instrument Science Report WFPC2 2002-06
examining them with IMSTAT allowed the rejection of images with prominent streaks
(more than around 0.8% overall normalized RMS, and exceeding about 1% peak-to-peak
amplitude). Images with prominent “worms” in WF2 were also rejected. The remaining
10 images were combined with the task STREAKFLAT to produce an averaged, destreaked pre-SMOV flat.
Similar rejection criteria were applied to the Earthflats taken after the servicing mission, as part of program 8952. Again the images were displayed and examined with
IMSTAT, rejecting those with too high an RMS and large peak-to-peak streak variations.
The remaining five acceptable images were also combined with STREAKFLAT, to produce the post-SMOV flat.
We then divided the SMOV flat by the pre-SMOV image, and normalized so that the
central 400x400 pixels of WF3 had a mean of unity. The resulting post-SMOV/pre-SMOV
ratio image is shown in Figure 7.1.
Results
Figure 7.1 shows that the mean ratio between the pre- and post-servicing flats is essentially unity. The most significant changes are seen at low levels, on large scales across the
chips, where departures from unity reach about 0.1-0.2%. These effects are well characterized based on pre-SMOV data over the past few years (Koekemoer et al. 2002) and are
most likely due to small changes in the camera vignetting, which in turn result from small
changes in the geometry of WFPC2. Other evidence of such small on-going geometric
changes is also seen in K-spot images and is described by Mutchler and Stiavelli (1997),
and Casertano and Wiggs (2001).
The pixel-to-pixel fluctuations (over the central 400 x 400 pixels) in the ratio image
are typically 0.4% RMS for the WFC CCDs, and 0.8% RMS for PC1, which is entirely
consistent with photon statistical noise. After smoothing with a 10-pixel FWHM Gaussian
function, the fluctuations decrease to <0.1% RMS, as would be expected for photon noise.
No change in chip-to-chip sensitivity is seen in on any levels above ~0.3% in the average ratio of post-SMOV / pre-SMOV counts over the central 400x400 pixels of each CCD.
There is also no significant evidence of obscuration or other changes in the OTA.
55
Instrument Science Report WFPC2 2002-06
streakflat_f502n_ratio_mos - STREAKFLAT_F502N_RATIO_MOS[1/1]
1600
1400
1200
1000
0
500
1000
1500
z1=0.98 z2=1.02 ztrans=linear Con=1.00 Brt=0.50 cmap=Grayscale ncolors=166
0
0
0
1
1
1
Figure 7.1: Ratio of SMOV / pre-SMOV flats taken in F502N. The display greyscale ranges from
0.98 (black) to 1.02 (white). The large-scale changes are of the order of 0.1-0.2%, and are entirely
consistent with well-known long-term changes in the camera geometry.
References
Biretta, J., 1995, “WFPC2 Flat Field Calibration,” in Calibrating HST: Post Servicing
Mission, eds. A. Koratkar and C. Leitherer, STScI, p. 257, 1995
Casertano, S. and Wiggs, M. 2001, “An Improved Geometric Solution for WFPC2”,
Instrument Science Report WFPC2-2001-10
56
Instrument Science Report WFPC2 2002-06
Koekemoer, A. M., Biretta, J. and Mack, J., 2002, “Updated WFPC2 Flatfield Reference Files for 1995 - 2001”, Instrument Science Report WFPC2-2002-02
Mutchler, M. and Stiavelli, M., 1997, “WFPC2 Internal Monitoring for SM97,” Technical Instrument Report WFPC2-1997-07
57