Instrument Science Report WFPC2 97-09
Results of the WFPC2 post-ServicingMission-2 Calibration Program
J. Biretta, I. Heyer, S. Baggett, S. Casertano, A. Fruchter, S. Gonzaga, J. Krist, M.
Lallo, M. McMaster, M. Mutchler, C. O’Dea, M. Stiavelli, A. Suchkov, B. Whitmore.
January 14, 1998
ABSTRACT
WFPC2 underwent an extensive series of tests and calibrations following the 1997 HST
servicing mission. Herein we summarize the results of these tests. The servicing mission
appears to have caused essentially no change in the performance and calibration of
WFPC2. The photometric calibration, baseline far-UV response, focus, PSF, flat field,
read noise, A-to-D gain, dark current, and optical alignment all appear identical to those
before the servicing mission. The only discernable effect of the servicing mission was a
small and temporary increase in the rate at which contaminants collected on the cold
CCD windows. This will have a small effect on UV data taken in March and April 1997;
these observations will require non-standard contamination corrections which are discussed herein.
1. Introduction
The Second HST Servicing Mission (SM97 or SM-2) was successfully completed on
February 19, 1997. During the weeks that followed, WFPC2 was run through an extensive
series of calibrations whose purpose was to protect WFPC2 from contamination and to
verify that its performance had not been affected by the HST servicing. Nearly one-thousand calibration images were obtained and analyzed. The results show that the postservicing performance of WFPC2 is essentially identical to that before the servicing mission. The only significant change was a small and temporary increase in the rate at which
contamination grows on the CCD windows, and a corresponding increase in the amplitude
of far-UV throughput variations. All other calibration parameters appear unaffected by the
recent HST servicing mission.
The sections which follow summarize the motivation and results of each WFPC2 Servicing Mission / Orbital Verification (SMOV) calibration program. More detailed
1
information can be found in separate Technical Instrument Reports and Instrument Science Reports listed in the references. A summary of WFPC2 SMOV plans can be found in
Biretta, et al. (1997).
Table 1.1: Table of Contents
Section
Title
PID
2
UV Monitoring and Cool-down Procedure
7016, 7122
3
3
Lyman-α Throughput Check
7018, 7029
9
4
Relative Photometry Check
7020
12
5
SMOV Flat Field Check
7019
20
6
Focus During SMOV
7016, 7017
24
7
WFPC2 PSF Stability
7021
28
8
WFPC2 Internal Monitoring for SM97
7022
30
9
Conclusions
----
34
2
Page
2. UV Monitoring and Cool-down Procedure (PID 7016 and 7122)
by M. Stiavelli, J. Biretta, S. Baggett, S. Gonzaga, and M. Mutchler
2.1 Background
The UV throughput of WFPC2 is known to be affected by the growth of contaminants
on the cold CCD windows. For filters like F170W the standard contamination rate ranges
from 0.5 to 1 per cent per day depending on the chip. Contaminants are evaporated during
monthly decontamination procedures that are able to restore the UV throughput to its original value.
The elevated contamination levels immediately after the First Servicing Mission, and
the fact that the pick-off mirror of WF/PC was completely non-reflective in the UV when
it was brought back to the ground, called for particular care in monitoring the UV throughput of WFPC2 after SM2, and in cooling the CCD down to its operational temperature.
This resulted in the precaution of avoiding the bright Earth (Bright Earth Avoidance, or
BEA) for 10 days after SM2 and for a specific cool-down procedure defined by J. Biretta
and J. MacKenty and described in WFPC2 ISR 97-03. The requirement of avoiding the
bright Earth made it necessary to establish pre-SMOV first epoch observations for a number of new photometric standard stars, since our normal standards would not allow us to
avoid the bright Earth.
2.2 Observations
Six new spectrophotometric standard stars lying in the CVZ were selected from a
spectrophotometric atlas of white dwarfs (Wegner and Swanson, 1991) and observed in
the period from November 1996 to January 1997. The objects were located such that at
least one of them would have been observable in the post-SM2 period, even allowing for
significant launch delays. The list of stars is given in Table 2.1.
Table 2.1: Pre-SMOV standard star obs. The count rates are given for the PC
Object
F555W-cal
GRW+70D5824
F555W-uncal
F170W-cal
F170W-uncal
3793.3
3966.6
158.7
172.1
WD0134+833
2715
2824
101.3
109.8
WD0214+568
1683
1759
81.75
88.20
WD0310-688
13044
13466
335.9
362.1
WD1042-690
3489
3593
156.1
168.4
WD1042-690a
3540
3644
158.3
170.2
WD2126+734
3410
3563
81.2
88.1
a. repeated observation
3
The count rates given in Table 2.1 were derived by averaging the measurements on two
images for each star and each filter. In Table 2.1 for each separate frame we list the count
rate as measured on both the calibrated and the uncalibrated frame. The internal scatter of
these images was not significantly worse for the uncalibrated frames than for the calibrated. As a consequence, for uncalibrated frames photometry seemed to be a viable
means to quickly identify major trends in the data. In order to provide rapid turn-around,
uncalibrated data were used for UV monitoring.
Since the first SMOV data were obtained at a CCD temperature of -43 C for which no
past photometric data were available, there was a concern that our measurements might
have been affected by a large Charge Transfer Efficiency problem. Indeed, it is known that
the effect of CTE is reduced by a large factor when the temperature is reduced from -75 C
to -88 C. Some data were found in the archive taken at -22 C and they seemed to show that
the counts were 5% to 10% lower at that temperature than at -88 C. Given that the main
goal of the monitoring was to measure contamination rates rather than the absolute
throughput it was decided that CTE did not present a significant concern for the -43 C
measurements.
2.3 Results
On February 23rd the first SMOV WFPC2 data of CVZ target WD0310-688 at a temperature of -43 C were obtained and analyzed. They showed only a modest (less than
10%) decrease in countrate with respect to the -88 C pre-SMOV measurements. Much of
the decrease could be attributed to CTE effects at -43 C, and in any case it was apparent
that the pick-off mirror had not been significantly contaminated. The second set of measurements on February 25th indicated a preliminary contamination rate of about 1% per
day (see Table 2.2 for the -43 C data). On their basis the decision was made to proceed
with a decontamination and then the cool-down to -88 C.
On March 3rd, at the end of our BEA constraint, we changed the UV monitoring standard from WD0310-688 back to GRW+70D5824 which is normally used for WFPC2
monitoring.
The results of the standard monitoring at both -43 C and -88 C are shown in Figures
2.1 and 2.2 for the PC and the WF chips, respectively. From the figures it is clear that
while the contamination rate was initially higher than the normal one experienced by
WFPC2, the tendency was for it to flatten down to the pre-SMOV value. Note how the
decontaminations at JD 50506.27 (27 Feb 1997 and coincident with the cool-down to -88
C), 50511.43 (4 Mar 1997), 50528.15 (21 Mar 1997), and 50543.37 (5 Apr 1997) were
able to restore the throughput to its previous values. The observed contamination rates
4
between our decontamination cycles and the pre-SMOV (WFPC2 ISR 96-04) values are
given for each WFPC2 camera in Table 2.2.
Table 2.2: Daily Contamination Rates (in percent per day) for F170W.
Chip
pre-SMOV
4.54-20.92 Mar 97
21.17 Mar - 5.37 Apr 97
PC
0.53 +/- 0.04
0.82 +/- 0.04
0.55 +/- 0.06
WF2
0.95 +/- 0.02
1.28 +/- 0.04
0.86 +/- 0.06
WF3
0.95 +/- 0.02
1.24 +/- 0.04
1.02 +/- 0.05
WF4
0.77 +/- 0.02
1.04 +/- 0.04
0.90 +/- 0.06
In an effort to keep the total contamination accumulation within levels previously
experienced, decontamination procedures were run at an accelerated schedule for the first
few months after the servicing mission. Instead of the usual four-week interval between
decons, the decons were scheduled at increasing intervals ranging from a few days to three
weeks. Table 2.3 gives the decontamination dates following the servicing mission.
Table 2.3: Decontamination Dates Bracketing SM97
Date
Time
Day of Year
MJD
Comments
23 Feb. 1997
19:08
054
50502.7978
CCD’s to -43C, TEC on 19:40:46
27 Feb. 1997
06:31
058
50506.2721
CCD’s to -88C, TEC on 07:10:48
04 Mar. 1997
10:16
063
50511.4278
CCD’s to -88C, TEC on 10:55:46
21 Mar. 1997
03:35
080
50528.1494
6 hour decon
05 Apr. 1997
08:50
095
50543.3681
6 hour decon
25 Apr. 1997
23:00
115
50563.9583
6 hour decon
15 May 1997
20:18
135
50583.8460
6 hour decon
07 Jun. 1997
13:06
158
50606.5461
6 hour decon
24 Jun. 1997
11:04
175
50623.4612
6 hour decon
24 Jul. 1997
18:42
205
50653.7795
6 hour decon
In summary, the post-SM WFPC2 cool-down proceeded successfully and no permanent contamination occurred during SM2. Even through the contamination rate was
initially higher, our standard decontamination procedure restored the uncontaminated
throughput.
5
2.4 Contamination Corrections
A consequence of the increased contamination growth rate is that WFPC2 UV observers will need to make larger contamination corrections when performing photometric
calibration of data taken within a month or two after the servicing mission. The calculation
is the same as for pre-SM97 data, but the constants are larger, since the growth is more
rapid. (See WFPC2 Instrument Handbook, v. 4, Section 6.9.) Table 2.2 shows the
throughput loss in percent per day for the F170W filter in the different CCDs. The third
and fourth columns give the contamination rates following the 4 March 13h UT and 21
March 04h UT decontaminations, respectively. Each decon effectively resets the throughput loss to zero, with the constants giving the decline per day following the decon.
The results after the 21 March decon (except for WF4) are generally consistent with
pre-SM97 rates, and show that contamination levels have returned to pre-SM97 levels.
Throughput corrections for other filters can be roughly estimated by comparing the
pre-SMOV corrections for them against F170W (see WFPC2 Instrument Handbook v.4
Table 6.10). Below we give ratios of (filter / F170W) throughput corrections for various
filters. For example, using these two tables, we would predict the F336W throughput loss
in WF3 on 15 March to be (11 days)(1.24% +/- 0.04% per day)(0.15) = 2.0% +/- 0.1%.
Table 2.3: Ratios of Throughput Corrections
Filter
Ratio of Throughput Corrections
F218W
F218W / F170W = 0.88
F255W
F255W / F170W = 0.47
F336W
F336W / F170W = 0.15
F555W
F555W / F170W = 0.07
6
Figure 2.1: Relative throughput normalized to its maximum pre-SMOV value for the
PC (upper panel) and WF2 (lower panel) in the F555W filter (open triangles) and the
F170W filter (filled squares). The dates of decontamination procedures are indicated
with vertical dotted lines. Note the much higher contamination rate for F170W and
how decontaminations restore the throughput to its original value. Data before JD
50506 refer to -43 C. CVZ targtet WD0310-688 was used before 3 Mar 97 and
WFPC2 standard GRW+70D5824 was used thereafter.
7
Figure 2.2: Same as Figure 1 for WF3 and WF4.
8
3. Lyman-α Throughput Check (PID 7018 and 7029)
by C. O’Dea, S. Baggett, and S. Gonzaga
3.1 Observations and Analyses
Observations were made of the UV-bright star WD0310-688 at several epochs in order
to measure any far-UV thoughput decrease due to any contamination of the WFPC2 pickoff mirror. The properties of the star are summarized in Table 3.1. The post-SMOV data
was compared with pre-SMOV data taken on 18-Nov-96. Observations are made in farUV filters F122M and F160BW. However, both of these filters have a significant read leak.
Thus, images were also taken with these filters crossed with F130LP, so as to isolate
counts contributed by the red leak in F122M and F160BW. The observations were taken in
both the PC1 and WF3 chips. Two observations were taken in each filter/chip combination. In addition, a F555W image was taken in PC1 as an overall sanity check.
The observations in PID 7018 consisted of three epochs of observation, each within 12 days after a DECON. Because of a delay in the commissioning of the STIS MAMAs,
and the need to continue monitoring far-UV throughput, the program was extended (PID
7029) for another six epochs at roughly weekly intervals up to 05-May-97.
A rapid turnaround was not required for these results, so the data were calibrated via
the normal pipeline procedures. Because a flat field reference file is not available for the
filters crossed with F130LP, a crude flat field was applied in three ways and the results
compared, (1) these data were flatfielded with a dummy file of unity value, (2) they were
corrected by hand using a local median filtered flat field value from the uncrossed filter flat
field, and (3) recalibrated using the flat field from the uncrossed filter. The data were
reduced by Gonzaga using scripts initially created by Ritchie and modified by Gonzaga.
Selected data sets were analysed independently by O’Dea and Baggett, obtaining consistent results. The two observations in each filter/chip combination were reduced separately
and the results compared afterwards. Cosmic rays were removed by hand and the centers
of the stars were determined by eye due to asymmetries in the PSF. O’Dea also reduced
selected data implementing CR removal with “crrej” and obtained similar results. The
stars were checked for saturation with none found. Aperture photometry was obtained
with the task “phot” using an aperture radius of 0.5 arcsec (5 pixels in WF3 and 11 pixels
in PC1) and a sky subtraction annulus between 1.5 and 2.0 arcsec.
The count rates in the F130LP crossed filters (the red counts) were subtracted from the
uncrossed filters (the total counts) to produce the “approximate” red leak corrected (blue)
count rates.
9
3.2 Results
A comparison of the uncrossed and crossed filter count rates shows that the red leak
contribution is high for this star, roughly 70%. Thus, in Figure 3.1 we show the count rates
with an approximate correction for the red leak. These results are consistent with no general trend in the throughput, but with a large scatter of ~15-25%. While it is clear that the
data have much scatter, it is also clear that the pre-SM-2 (MJD ~50404) and post-SM-2
(MJD 50500-50600) results show good agreement on the PC1 in Figure 3.1. Thus, we
conclude that the far-UV throughput of WFPC2 was not significantly affected by contamination due to the second servicing mission. Any change is less than the ~20% uncertainty.
Table 3.1: Properties of WD0310-688 (LB3303) in March 1997
Property
Value
RA (J2000.0)
3:10:30.970
DEC (J2000.0)
-68:36:03.05
Type
DA3
F555W (STMAG) in 0.5” aperture
12.3
B-V
0.02
U-B
-0.62
Avg count rate F555W
13140 DN/s
Avg count rate F122M
39 DN/s
Red Leak Fraction in F122M
72%
Avg count rate F160BW
142 DN/s
Red Leak Fraction in F160BW
78%
10
Figure 3.1: Red Leak corrected normalized count rates through F160BW and F122M
in the PC1 (left) and WF3 (right). Dotted lines represent decontamination dates.
WF3, F160BW
1.4
1.2
1.2
normalized countrate
normalized countrate
PC, F160BW
1.4
1
.8
1
.8
.6
.6
50400
50450
50500
MJD
50550
50600
50400
50450
50550
50600
50550
50600
WF3, F122M
1.4
1.4
1.2
1.2
normalized countrate
normalized countrate
PC, F122M
50500
MJD
1
.8
1
.8
.6
.6
50400
50450
50500
MJD
50550
50600
50400
11
50450
50500
MJD
4. Relative Photometry Check (PID 7020)
by B. Whitmore, S. Gonzaga, and I. Heyer
4.1 Observations and Analyses
The goal of the relative photometry check (proposal 7020) was to verify that the photometric accuracy remained unchanged at the 1-2% level. 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) on
March 5, 1997 (MJD 50512), with the star centered in each of the 4 CCDs (one orbit per
CCD). Each orbit started with an observation with the F555W filter in the PC1, in order to
monitor any drifts in the throughput. This program was run one day after the “protect
decon” which executed 13 days after HST release from the shuttle. WFPC2 was at its normal operating temperature of -88C. The results were compared to observations of the
same object taken prior to the servicing mission.
Photometric count rates were measured independently by using two methods, using
pipeline-calibrated data. Method #1 used IRAF scripts developed by Ritchie for routine
photometric monitoring of WFPC2 over the past several years. Method #2 used more
automated IRAF scripts developed by Whitmore and Heyer. Both cases used an aperture
radius of 0.5 arcsecs (5 pixels in the WF chips; 11 pixels in the PC1 chip) for the targets,
and an annulus from radius 30 to 35 pixels for determining the sky values.
The agreement between both methods was very good (method #1 / method #2 =
0.9993 +/- 0.0002). The values from method #1 were used for the comparative analysis,
since this method was also used to establish the baseline measurements. The baseline measurements have been corrected for contamination using the contamination rates in ISR
WFPC2 96-04.
Figures 4.1a - i show the baseline observations as filled circles for the past 2 years
(since MJD = 49750, i.e. Feb. 2, 1995) and the March 5, 1997, observations from proposal
7020 for the full set of filters and chips. The measurements have been normalized to a
mean of 1.000 based on the baseline observations. The 1-sigma scatter (based on the
empirical scatter) of the baseline measurements are shown by dashed lines. In cases with
less than 10 baseline measurements (i.e., WF2 and WF4 for all but F170W), the scatter
from WF3 has been adopted. In cases where no baseline measurements were available the
value on WF3 was used.
The SMOV observations were taken on March 5, 1997 (MJD 50512.22 to 50512.44),
roughly one day after a decontamination (MJD 50511.43). One observation with the
F255W filter in the WF3 chip was excluded from the analysis because the star’s central
pixel was saturated. A contamination rate of 1.7 times the pre-SMOV values has been
adopted, based on measurements from proposal 7016 (see Chapter 2). The contamination
12
correction is very small in all cases (maximum of 1.3% for F160BW on WF4), since the
observations were taken within 1 day of the decontamination. The observed scatter in the
F555W observations on the PC1 taken at the start of each orbit is less than the predicted
scatter, hence no correction for a “drift” from orbit to orbit was made.
4.2 Results
Figures 4.1a-i show the SMOV data (March 5, 1997) as open stars. In all cases the
observed SMOV count rates are in good agreement with pre-SMOV data, showing that
WFPC2 throughput was essentially unaffected by the 1997 servicing mission. A flat mean
has been adopted in all cases, except for F160BW and F170W on the PC1 chip, where a
slow increase in the throughput as a function of time has been observed previously (see
ISR WFPC2 96-4). For these two cases a linear fit was adopted, using baseline data, to
predict the value for March 5, 1997.
Figures 4.2a and 4.2b show the observed versus predicted values for all cases with
more than 10 baseline measurements (N=22 in the resulting sample) both in terms of the
ratio of observed to predicted values, and as the 1-sigma scatter (shown as dashed lines in
Figures 4.1a - i). Overall, the agreement is excellent, with only a very small (1 sigma) hint
that the throughput has decreased by about 0.16% (i.e. observed/predicted = 0.9984 +/0.0023). However, there is some evidence for a slightly larger effect on the PC1 alone, at
about the 2-sigma level (i.e. N=12, observed/predicted = 0.9958 +/- 0.019 for the PC1,
N=8, observed/predicted = 1.0010 +/- 0.015 for the WF3). Including all the data (i.e. chips
WF2 with two baseline observations and WF4 with five baseline observations) yields a
value of observed/predicted = 0.9968 +/- 0.0026.
It is interesting to note that the observed value of the scatter is less than the predicted
scatter (i.e. the width in Figure 4.2a is 0.64, less than 1.0). This indicates that there is more
structure in the long-term baseline data than just random Gaussian noise. Examples can be
seen in the F170W WF3 and F218W WF3 baseline observations (Figures 4.1b and 4.1c),
where the throughput during days 49750-50000 appears to be about 3% higher than the
more recent observations.
In summary, the throughput of WFPC2 was essentially unaffected by the 1997 servicing mission. The mean change in throughput is observed/predicted = 0.9984 +/- 0.0023.
There is evidence for a slight (2-sigma) tendency for the PC1 chip to be 0.4% lower following the servicing mission; this is under further study.
13
4.1b: F170W
4.1a: F160BW
Figure 4.1: The normalized counts rates vs. MJD for all filters.
14
4.1d: F255W
4.1c: F218W
Figure 4.1 continued
15
4.1f: F439W
4.1e: F336W
Figure 4.1 continued
16
4.1h: F675W
4.1g: F555W
Figure 4.1 continued
17
4.1i: F814W
Figure 4.1 continued
18
Figure 4.2a: Observed vs. Predicted Values
Figure 4.2b: Distribution in Terms of Sigma
19
5. SMOV Flat Field Check (PID 7019)
by J. Biretta and M. Wiggs
5.1 Introduction
We compare WFPC2 Earthflats taken before and after the 1997 servicing mission.
Most of the field-of-view shows no change (<<1%) in flat field calibration. The largest
changes occur within a few pixels of the CCD corners and reach ~1.5%; these are likely
attributable to long-term changes in the camera / OTA geometry, rather than SM97.
As part of our post-servicing check-out of WFPC2, we have observed a series of
F502N Earthflats to test the flat field stability. The goal of these observations is to test for
any unexpected obscuration or contamination in the OTA or WFPC2 due to HST servicing. 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 1995).
5.2 Observations and Analyses
The Earthflat observations made for SMOV were identical to those obtained for the
routine Cycle 4 to Cycle 6 calibration programs. They are 1.2 second exposures of the
bright Earth made with gain 15 in filter F502N. Since we are interested primarily in
changes to the flat field, we have also selected a set of Cycle 6 Eathflats to compare with
those taken during SMOV.
For the pre-SMOV observations we examined 48 Earthflats in F502N taken between
May 1996 and Feb. 1997 as part of proposal 6909. From these we discarded ones with
mean counts on WF3 below 1000 DN, ones with less than 500000 good pixels on any
CCD (usually due to saturation), and ones with streaks exceeding about 1% peak-to-peak
amplitude. In order to evaluate this last criteria, the raw Earthflats were multiplied by the
current F502N flat field reference file, and then examined with IMSTAT and displayed.
The remaining seven images, u3ek1b03t, u3ek1l04t, u3ek1v03t, u3ek2603t, u3ek2604t,
u3ek6603t, and u3ek9604t, were combined with the task STREAKFLAT to produce an
averaged, de-streaked pre-SMOV flat.
Similar criteria were applied to the 108 Earthflats taken after the servicing mission as
part of program 7019. This resulted in five images being selected, u3sc0502r, u3sc0703r,
u3sc0803r, u3sc1102r, and u3sc1503r, which were combined with STREAKFLAT to produce an SMOV flat.
20
We then divided the SMOV flat by the pre-SMOV image, and normalized so that the
center 400 x 400 pixels of WF3 had a mean of unity. The resulting SMOV / pre-SMOV
ratio image shown in Figures 5.1 and 5.2.
5.3 Results
As Figures 5.1 and 5.2 show, the pre- and post-servicing flats are essentially identical
The most significant changes are seen in the corners of WF2 and WF4 near the pyramid
apex, where departures from unity reach about 1.5% within about 10 pixels of the pyramid
apex. Similar effects are seen in the corners of WF3. These effects in the CCD corners are
most likely due to small changes in the camera vignetting, which in turn result from small
changes in the geometry of WFPC2 and the OTA. Other evidence of small on-going geometric changes from K-spot images is described by Mutchler and Stiavelli (1997).
There are also faint traces of the “worm” features on WF2 at the 0.2% peak-to-peak
level, which are related to contamination on the CCD windows. This is to be expected,
since we have not restricted our selection of Earthflats with regards to time delay after
decontaminations.
The pixel-to-pixel fluctuations (over 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.
No change in the chip-to-chip sensitivity is seen. The average ratio of post-SMOV /
pre-SMOV counts over the central 400 x 400 pixels of each CCD is 0.9995 for PC1,
0.9998 for WF2, 1.0000 for WF3 (by definition), and 1.0003 for WF4.
21
Figure 5.1: Ratio of SMOV / pre-SMOV flats taken in F502N. The display greyscale
ranges from 0.98 (black) to 1.02 (white). The CCDs are indicated by the key at the
upper left. For display purposes the original 1600 x 1600 pixel image has been
smoothed using a 5x5 pixel box average.
PC1
WF2
WF4
WF3
22
Figure 5.2: Same as Figure 5.1, except that the display greyscale ranges from 0.995
(black) to 1.005 (white).
PC1
WF2
WF4
WF3
23
6. Focus During SMOV (PID 7016 and 7017)
by S. Casertano, M. Lallo, A. Suchkov, and J. Krist
6.1 Data and Analyses
The measurements described in this report derive from WFPC2 observations of a single, isolated bright star placed near the center of the PC, through broad-band, non-UV
filters such as F555W (primarily), F439W, F675W and F814W. The primary source of
data are the two SMOV proposals 7017, especially designed for focus check and including
a series of eight exposures in each of three orbits, and 7016, primarily designed for contamination monitoring but which also includes two F555W exposures in the PC for each
execution. Additional data have been taken from photometric monitoring proposals and
the filter sweep, and a few observations of rich star fields in Omega Centauri. Since there
appears to be no systematic dependence of the results on target, filter or proposal, we will
not distinguish between different types of measurements in the following.
The key focus measurements were obtained using the phase retrieval code developed
by John Krist and Chris Burrows (see Krist and Burrows, 1995). This code determines the
focus position from an observed stellar image by optimizing the match with a model PSF,
computed using the known optical characteristics of the OTA and of WFPC2 itself; for
details on its use see also Casertano (1995). The optimal focus position is defined as the
position that minimizes the rms wavefront error at the image plane. The typical errors due
to the phase retrieval alone are known to be less than 1 micron. The error in the derived
position of the secondary mirror is dominated by OTA terms, such as the well-known
breathing.
We have previously shown (Suchkov and Casertano, 1997) that the aperture correction
for very small apertures (1-2 pixels) correlates strongly with focus position. Therefore we
determined the aperture correction for all SMOV focus images, to serve as a sanity check
on the focus position determined from the phase retrieval.
The focus position measured at a specific moment in time differs from the average
focus position during that orbit because of the OTA “breathing”. The observer is affected
by the actual, instantaneous focus position, but the orbit-averaged position should be used
to determine the OTA status and in any trending of the long-term position of the secondary
mirror. Hasan and Bely (1994) give a simple formula to determine the breathing correction
for any observation, using the temperatures measured at the four light-shield temperature
sensors. Typical breathing corrections found during SMOV range from -2 to +2 micron,
although larger corrections were found in a few cases. Breathing corrections have been
used in all trending plots.
24
6.2 Results
All focus measurements since the last pre-SMOV mirror move (October 30, 1996) are
plotted in Figure 6.1. Shortly after the resumption of WFPC2 activities, it became apparent that the focus position was systematically positive by about 3-4 micron (see Fig. 6.1),
in agreement with the rather sparse pre-SMOV measurements since the previous focus
move (October 30, 1996). Obviously, SM activities had no significant impact on the position of the secondary mirror. The focus position remained mostly around 3-4 micron for
the following two weeks, and a secondary mirror move of -2.4 micron was executed on
March 18, 1997, bringing the average focus position to about 1 micron. The focus measurements after that date reflect this new position, and it is expected that there will be no
need for additional moves of the secondary mirror over the next 6-8 months, or longer
depending on the desorption rate.
The results of the aperture correction measurements largely confirm those of the phase
retrieval. Aperture corrections are closely related to the magnitude of the focus positions
(see Suchkov and Casertano, 1997). For PC observations in filters such as F555W, a focus
displacement of 4 micron corresponds to a 9% change in the fraction of the flux within 1
pixel radius, or a 0.10 mag change in the aperture correction for that aperture. This is easily measurable in individual images. Figure 6.2 shows how the 1-pixel aperture corrections
correlate with the focus position measured with the phase retrieval method for a subset of
the SMOV data, with the solid line indicating the quadratic relationship determined from
pre-SM data (see Suchkov and Casertano, 1997). The pre- and post-SM data are in excellent agreement with one another, indicating that the encircled energy in small apertures is
essentially unchanged compared to pre-SM measurements.
25
Figure 6.1: Measured focus position since October 30, 1996, including all SMOV
data.
26
Figure 6.2: Small-radius aperture corrections vs. focus position. The solid line represents the pre-SM relationship. Dots and crosses represent data for different stars: dots
for GRWd70+5824, and crosses for S121-E. The outlying point at (3.4, -1.1) is from
an image heavily affected by a cosmic ray.
27
7. WFPC2 PSF Stability (PID 7021)
by A. Fruchter and M. McMaster
7.1 Observation and Analyses
The WFPC2 PSF is a function both of the WFPC2 optics and the WFPC2 electronics.
In the Wide Field Cameras, in particular, the core of the PSF is dominated by the pixel
response function1 (PRF), and therefore the shape of the PSF does not provide a sensitive
test of the optics. In the PC the size of the pixel relative to the intrinsic PSF of the optics is
reduced by more than a factor of two. Although the PRF remains a significant component
of the final PSF shape, PC observations allow far more sensitive test of the telescopic and
instrumental optics than do WF images. Therefore, in the discussion that follows we will
emphasize PC observations.
WFPC2 F439W images of ω Cen were obtained both in July 1996 (pre-SMOV) and
March 1997 (post-SMOV). Due to guide star constraints, the fields imaged at these two
epochs were not identical. Sixteen dithered positions, each with a single 40s integration,
were obtained on each field. The dithers were placed on a 4x4 grid with a gridsize of 0.125
arcseconds. Thus the planned offsets were, to an accuracy of better than one percent, multiples of 1.25 and 2.75 pixels in the WF and PC, respectively.
The data were combined using the “drizzle” and “blot” routines which have been
developed to image, and to remove cosmic rays from, dithered WFPC2 images (Fruchter
and Hook, PASP submitted). The final images were drizzled onto output images with a linear scale four times finer than that of the original images.
7.2 Results
The accompanying Figure 7.1 shows radial profile plots of single stars taken near the
center of PC1 in the pre- and post-SMOV ω Cen visits. The post-SMOV image has a
somewhat broader PSF than that of the pre-SMOV image; however, this is consistent with
the change in focus found between the two epochs (see WFPC2 TIR 97-03). Both images,
however, show a full-width at half-maximum of approximately 0.064 arcseconds.
Except for the small change in focus, the pre- and post-SMOV PSFs appear identical.
1. The instrumental PSF is made up of the combined effects of the optics on the image, and any scattering or diffusion of charge and photons that occurs in the detector. This latter component is frequently
referred to as the pixel response function.
28
Figure 7.1: Radial profile plots from PC images taken of ω Cen taken both before
and after SMOV. The stars have been normalized to equal flux within a 5-PC-pixel
circular aperture. A gaussian profile with a full-width at half-maximum of 0.064 arcseconds is plotted for comparison.
29
8. WFPC2 Internal Monitoring for SM97 (PID 7022)
by M. Mutchler and M. Stiavelli
8.1 Observation and Analyses
The goal of WFPC2/SMOV proposal 7022 was to verify that there was no change in
basic instrument health and internal calibrations. A variety of internal exposures were
obtained in order to provide a monitor of the integrity of the CCD camera electronics in
both bays (gain 7 and gain 15), a test for quantum efficiency in the CCDs, and an internal
check on the alignment of the WFPC2 optical chain.
Internal WFPC2 observations were made at a gain of both 7 and 15. Each visit consisted of two bias frames at both gains (to verify the electronic integrity of WFPC2 and to
estimate noise) and two F555W flat fields at both gains (to verify optical integrity and gain
ratios). Two Kelsall spot (“KSPOT”) images (one optimized for the PC and one for the
WFs), and five 1800s dark frames were also taken at gain=15. This entire set ran weekly,
on a non-interference basis. The first visit started shortly after WFPC2 became operational
on February 23, 1997, with the CCDs at a temperature of -43C. The remaining visits
began after WFPC2 was cooled -88C degrees on February 27.
A goal of this program was to measure any changes in bias and dark frames to ~1.4
electrons per pixel, and to measure any changes in the flat fields to ~1% accuracy. The
read noise was estimated by running imstat on the bias frames. Bright cosmic rays were
ignored by setting the upper pixel value cutoff to be ten sigmas above the mean. The table
below shows that the noise level was high by a factor of 2 or 3 for the first visit on February 23 (when the CCDs were at a temperature of -43C), and that the noise returned to
nominal levels after WFPC2 was cooled to -88C.
Table 8.1: Read Noise
CCD
Gain
Noise on Feb 23 at -43C
Noise on Feb 27 at -88C
Nominal pre-SMOV
noise at -88C
PC1
7
15.12e
5.31e
5.24 +/- 0.30
PC1
15
17.00e
7.70e
7.02 +/- 0.41
WF2
7
16.82e
5.30e
5.51 +/- 0.37
WF2
15
18.71e
8.11e
7.84 +/- 0.46
WF3
7
20.09e
5.21e
5.22 +/- 0.28
WF3
15
21.14e
7.94e
6.99 +/- 0.38
WF4
7
17.72e
5.21e
5.19 +/- 0.36
WF4
15
20.15e
7.27e
8.32 +/- 0.46
30
The gain ratio was verified by comparing the flux measured in flat field exposures
through the same shutter with different gains. Shutter B was used for all the flat field exposures. The table below shows that the gain ratios for each chip were nominal after WFPC2
was cooled down to -88C on February 27.
Table 8.2: Gain Ratio
CCD
counts/sec at gain=7
counts/sec at gain=15
gain ratio at -88C
Nominal pre-SMOV
gain ratio
PC1
58.7
28.8
2.04
1.99 +/- 0.02
WF2
197.2
98.0
2.01
2.00 +/- 0.02
WF3
207.5
105.1
1.97
2.01 +/- 0.02
WF4
144.1
73.9
1.95
1.96 +/- 0.02
To check the internal alignment of the WFPC2 optical chain, we compared the positions of the Kelsall spots in pre-SM97 and post-SM97 exposures. We found a small shift
on the order of 5-10 milliarcseconds (see Figure 8.1) which is consistent with the general
trend over WFPC2’s on-orbit lifetime (see Figure 8.2).
Figure 8.1: Shifts of WFPC2 Kelsall spots after the 1997 servicing mission. The
pre-SM97 coordinates (from February 9) were all normalized to 0,0. The relative
location of each KSPOT on February 27 is plotted.
31
The 2nd servicing mission occurred on February 11-25, 1997 (days 1064-1075). Figure 8.3 shows that the small shifts measured between February 9 (day 1062) and February
27 (day 1080) -- i.e. before and after the servicing mission -- are within the expected
range.
Figure 8.2: Average shifts of WFPC2 Kelsall spots from March 15, 1994 (day 0), to April
5, 1997 (day 1117).
32
Figure 8.3: Average WFPC2 Kelsall spot shifts around the 2nd HST servicing mission.
The dark current was estimated by combining calibrated (i.e. bias subtracted) dark
exposures with crrej to remove cosmic rays, and then measuring the median pixel value
for each CCD for each shutter setting. All the dark exposures were 1800s at gain=7 at 88C. Here we use the three DARKs with shutter B.
Table 8.3: Dark Current
CCD
Shutter
Total exp time
Median counts
per pixel
Dark current
e/sec/pix
Nominal pre-SMOV
dark current e/sec/pix
PC1
B
5400s
3.515 DN
0.0046
0.0058
WF2
B
5400s
2.304 DN
0.0030
0.0025
WF3
B
5400s
3.187 DN
0.0041
0.0038
WF4
B
5400s
2.734 DN
0.0035
0.0038
The pre-SM97 and SMOV dark currents are in reasonable agreement. The small difference for PC1 is likely due to the strong dependence of its dark current on the cosmic ray
rate (see WFPC2 Handbook version 4, page 74).
In general, analysis of the data from program 7022 shows that the 2nd Hubble Space
Telescope servicing mission of February 1997 caused no significant changes to the basic
health and internal calibration of WFPC2.
33
9. Conclusions
WFPC2 was placed through an extensive series of tests following the 1997 HST servicing mission during which NICMOS and STIS were installed. Herein we have
summarized the results of these tests.
The servicing mission appears to have caused essentially no change in the performance and calibration of WFPC2. The photometric calibration, baseline far-UV response,
focus, PSF, flat field, read noise, A-to-D gain, dark current, and optical alignment all
appear identical to those before the servicing mission. The only discernable effect of servicing has been a small and temporary increase in the rate at which contaminants collect
on the cold CCD windows. This will have a small affect on UV data taken in March and
April 1997; these observations will require non-standard contamination corrections which
are discussed in Section 2.4.
Further details of the WFPC2 SMOV tests will be found in the individual Technical
Instrument Reports listed in the References.
We acknowledge important contributions to the WFPC2 SMOV program by Merle
Reinhart, George Chapman, John Trauger, Carl Biagetti, John MacKenty, Chris Burrows,
and the OPUS, PRESTO, and CNS groups.
References
•
Biretta, J., 1995, “WFPC2 Flat Field Calibration,” in Calibrating HST: Post Servicing
Mission, eds. A. Koratkar and C. Leitherer, STScI, p. 257
•
Biretta, J., 1996, ‘WFPC2 Instrument Handbook, “ version 4.0
•
Biretta, J., McMaster, M., Baggett, S., and Gonzaga, S., 1997, “Summary of WFPC2
SM97 Plans, “ Instrument Science Report WFPC2 97-03
•
Biretta, J., and Wiggs, M., 1997, “SMOV Flat Field Check,” Technical Instrument
Report WFPC2 97-08
•
Casertano, S., 1995, “Focus Monitoring and Recommendation for Secondary Mirror
Move”, Instrument Science Report OTA-18
•
Casertano, S., Lallo, M., Suchkov, A., and Krist, J., 1997, “OTA Focus during SMOV,”
Technical Instrument Report WFPC2 97-03
•
Fruchter, A., and Hook, R., 1997, PASP, submitted
•
Fruchter, A., and McMaster, M., 1997, “SMOV Check of WFPC2 PSF Stability,”
Technical Instrument Report WFPC2 97-06
•
Hasan, H., and Bely, P. Y., 1994, in “The Restoration of HST Images and Spectra II”,
eds. R. J. Hanisch and R. L. White (Baltimore, STScI), p. 157
•
Krist, J., and Burrows, C. J., 1995, Appl. Opt. 34, 4951
34
•
O’Dea, C., Baggett, S., and Gonzaga, S., 1997, “Results of the WFPC2 SM-2 Lymanα Throughput Check,” Technical Instrument Report WFPC2 97-05
•
MacKenty, J.W., Baggett, S., 1996, “WFPC2 Throughput Stability in the Extreme
Ultraviolet,” Instrument Science Report WFPC2 96-07
•
Mutchler, M., and Stiavelli, M., 1997, “WFPC2 Internal Monitoring for SM97,” Technical Instrument Report WFPC2-97-07
•
Stiavelli, M., Biretta, J., Baggett, S., Gonzaga, S., and Mutchler, M., 1997, “SM-2 UV
Monitoring and Cool-down Procedure,” Technical Instrument Report WFPC2 97-02
•
Suchkov, A., and Casertano, S.,1997, “Impact of Focus Drift on Aperture Photometry,” Instrument Science Report WFPC2 97-01
•
Wegner, G., Swanson, S.R., 1991, ApJS, 75, 507
•
Whitmore, B., Gonzaga, S., Heyer, I., 1997, “Results of the WFPC2 SMOV Relative
Photometry Check,” Technical Instrument Report WFPC2 97-01
•
Whitmore, B., Heyer, I., and Baggett, S., 1996, “Effects of Contamination on WFPC2
Photometry,” Instrument Science Report WFPC2 96-04
•
Whitmore, B., Wiggs, M., 1995, “Charge Transfer Traps in the WFPC2,” Instrument
Science Report WFPC2 95-03
35