Technical Instrument Report NICSMOS 2002-001
SM3B NICMOS Focus Check
A.B. Schultz, G. Schneider (Steward Observatory), E. Roye, S. Malhotra
June 12, 2002
ABSTRACT
We report determinations of the NICMOS foci following cool down of the NICMOS detectors after installation and activation of the NICMOS Cooling System (NCS). A focus
sweep in all three cameras executed on May 3, 2002 (program ID: 8977). Visual inspection of the data indicated that a small amount of coma had been introduced since January
1999. Phase retrieval (STScI) and encircled energy (University of Arizona) measurements
were used to determine the best focus positions. The results of the two methods compared
quite favorably with each other. Adjustments to the current Pupil Alignment Mechanism
(PAM) positions were implemented on May 9, 2002 for Camera 1 & 2 (PAM1 & PAM2).
No focus adjustments were implemented for Camera 3 (PAM3) or for the Camera 2 coronagraphic focus (PAMC). Adjustment of the intermediate focus position between Camera 1
& 2 (PAMI) will be managed starting during week 2002.133 by SMS.
Introduction
The optimal focus for each NICMOS camera was determined on a regular basis during
Cycles 7 and 7N with the last of the pre-NCS focus measurements performed on January
4, 1999 (Suchkov et al. 1998). These results can be found on the NICMOS web page
(URL: http://www.stsci.edu/hst/nicmos/performance/focus). Images of the star cluster
NGC3603 were obtained on May 3, 2002 to determine the locations of the NICMOS
detectors with respect to the f/24 input beam. The images were not dithered, and the data
are a series of images in and out of focus passing through best focus. We note that the
detector temperatures were regulated at ~76.5 K during the focus sweep. The temperatures
had stabilized by the time of the focus sweep observations.
Copyright© 1999 The Association of Universities for Research in Astronomy, Inc. All Rights Reserved.
Technical Instrument Report NICSMOS 2002-001
Overview of NICMOS
The NICMOS has three infrared cameras (NIC1, NIC2, NIC3) with different focal
ratios (f/80, f/45.7, f/17.2). Camera operations are independent but non-confocal. Confocal operation was a goal, but not a design requirement for NICMOS. The three cameras
have adjacent but not spatially contiguous fields-of-view (FOVs). Each camera has a 256
X 256 pixel HgCdTe focal plane array (NICMOS 3 detector architecture). With the dewar
“anomaly,” the detectors were moved forward toward the dewar face plate. This resulted in
non-confocal imaging at the detectors.The focus interface for NIC3 is beyond the adjustable range of the Pupil Alignment Mechanism (PAM) (described in the NICMOS
Instrument Handbook, Schultz et al. 2001).
Axial PAM movement will not only move the focus position, but will also translate the
field-of-view (FOV) of a camera on the sky. A star will appear to move in the camera FOV
between PAM positions. This is due to the PAM mirror sampling a different region on the
sky and relaying it to the detectors. Spacecraft pointing offsets need to be specified to keep
a target star in the same position on a detector during a focus sweep, and an adjustment is
made to the x-y tilt of the PAM for each position to adjust for differentially induced coma.
Five NICMOS PAM states; PAM1, PAM2, PAM3, PAMI (CAMERA-FOCUS=1-2),
and PAMC (coronagraphy), are managed for normal operation. In normal Camera 3 operation, the field offset mirror (FOM) is moved in concert with the PAM to alleviate the
vignetting and thermal emission from the -Y edge of the FOV (FOMYPOS=-16). The
focus values for PAM1, PAM2, and PAM3 are held in the fight software (FSW), while
PAMI and PAMC are uploaded through an SMS. The last adjustment to the PAM1 and
PAM2 values was performed on June 8, 1997, and the PAM3 value was last set on April 5,
1997. The PAMI transition state was implemented on SMS 1997.342 (December 8, 1997),
and the PAMC transition state was implemented on SMS 1997.363 (December 29, 1997).
Data
The southern open star cluster NGC 3603, also observed for the Cycle 7 focus monitor
programs, was observed on May 3, 2002 (program ID: 8977). Focus sweep images were
obtained separately (different visits) in each camera over a range of PAM focus settings,
measured in millimeters of displacement from the mid-range of axial travel. For Camera 1
and 2, the sequence of observations is a series of images, in and out of focus, passing
through best focus. Seventeen MULTIACCUM images were obtained over a range of +/- 8
mm of PAM travel in 1 mm increments. For Camera 3, the focus sweep was conducted at
PAM settings of -0.5 mm to -9.5 mm of motion, and consisted of only ten MULTIACCUM images. The detectors were operated in a non-parallel mode, and no problems
were reported in execution of the observations.
The raw data were independently calibrated and reduced by the NICMOS group at
STScI and the NICMOS Project at the University of Arizona (UA), Tucson, AZ. The
STScI group used calnica with a standard set of calibration reference files including an
2
Technical Instrument Report NICSMOS 2002-001
appropriate temperature-specific flat field and a temperature dependent dark. For Camera
1, the temperature dependent dark left an imprint of the flat field upon the images, and a
new dark was created from dark data obtained in the SMOV 8974 DQE program. The UA
Group used software and calibration reference files developed by the NICMOS IDT for
calibration and data reduction.
Data Characteristics
The observations were offset from the star cluster core to alleviate field crowding
thereby permitting star selection for analysis with no spoilers or overlapping of the point
spread function (PSF). There is a small amount of induced coma which was essentially
absent in the Cycle 7 and 7N observations. This implies that a small amount of induced
differential shear has occurred since January 1999. The presence of a small amount of
coma did not affect the determination of the “best focus” by either of the two employed
analysis methods. Coma corrections were determined separately, and subsequently, to the
8977 focus sweep observations. Images closest to best focus showing the different FOVs
for each camera are presented in Figure 1.
NIC1
NIC2
NIC3
Figure 1: 8977 focus sweep camera field-of-views (FOVs). NIC3 was obtained not at the
nominal FOM position of (0,-16), but at FOM position 0,0.
The FOM offset during the 8977 Camera 3 focus sweep was NOT at its nominal offset
position of +16" (FOMYPOS= -16"). This was done to reduce the thermal constraints on
frequent PAM operations. The FOM is moved to alleviate vignetting at the -Y edge after
moving the PAM to the “best focus” (-9.5 mm, the extreme range of PAM travel). Without
moving the FOM, Camera 3 images exhibit at the -Y edge reduced optical throughput and
at long wavelengths, brightening of the background from thermal emission from the aperture bulkhead, apparent in the 8977 Camera 3 F222M filter images. This signature was
removed during post-processing before the encircled energy (EE) analysis.
3
Technical Instrument Report NICSMOS 2002-001
Analysis
The processed data sets were independently analyzed using recovery of Zernike polynomial coefficients by phase retrieval (STScI) and minimization of PSF core flux density
dispersal by encircled energy (UA).
Phase Retrieval (STScI)
Phase retrieval is the process of deriving the aberrations in an optical system from an
observed PSF image. The IDL phase retrieval software used at STScI (Krist and Burrows
1995) compares model PSFs to the observed one, optimizing the assumed Zernike-type
aberrations (defocus, coma, astigmatism, etc.) using iterative, nonlinear least-squares fitting. This program is used to measure defocus (Z4) from images taken at each PAM
position in a NICMOS focus sweep (Krist and Burrows 1997, Galas 1998, Roye and
Schultz 2002). Corrections for defocus caused by telescope breathing are then applied.
These Z4 values are transformed to equivalent (“measured”) PAM positions using conversion factors derived from optical ray tracing analyses. The measured PAM position is
subtracted from the actual one, providing an estimate of the offset from best focus for each
measurement. The mean offset is computed from all of these values, except in Cameras 1
and 2, where the few points near best focus are omitted because of focus sign ambiguities.
All measurements are used for Camera 3, which historically has been sufficiently out of
focus at all PAM positions to prevent such ambiguities. Because focus varies over the field
of view in NICMOS, the best focus position is computed for the detector center using relations described in Suchkov et al. 1998 and the observed PSF position.
NIC1
NIC2
NIC3
Figure 2: Stars selected for the phase retrieval analysis are identified in the respective
frame. Images processed by STScI.
The 8977 phase retrieval focus values are the mean focus values of 3 stars from NIC1,
4 stars from NIC2, and 7 stars from NIC3. Five of the PAM values close to the best focus
positions for Camera 1 and 2 were not used in the phase retrieval analysis due to focus
sign ambiguities. The positions of the stars used in the phase retrieval analysis are marked
4
Technical Instrument Report NICSMOS 2002-001
in Figure 2. An example of the phase retrieval output for one star in each camera is shown
in Figure 4.
Encircled Energy (UA)
For Cameras 1 and 2, encircled energy measurements were made from six stars in the
field at each focus position. Stars chosen (shown in Figure 2) were well-isolated (i.e., no
overlaps in the wings of the PSFs), and well-exposed but not saturated in the PSF cores.
NIC1
NIC2
NIC3
Figure 3: Stars selected for the encircled energy analysis are identified in the respective
frame. Images processed by the NICMOS IDT.
For Cameras 1 and 2, in each of the 17 calibrated images (at each PAM focus setting)
the selected PSFs were coregistered, resampled, and combined, to form a suite of “composite” PSFs. PSF re-registration was performed based upon iterative weighted-moment
PSF-core centration. Extracted stars were then “shifted” (and combined) via bi-cubic
interpolation apodized by a sinc function kernel (Park and Schowengerdt 1983) with a
width appropriate for the FWHM of the PSF cores. For Camera 3, afocal at even the maximum range of compensatory PAM travel, only a single very well exposed and isolated
star at each PAM focus setting was used to determine the focus.
At each PAM focus setting the flux densities encircled within pre-defined radii of the
center of the (NIC1 & NIC2 composite or NIC3 selected) PSFs were measured from the
calibrated images (in units of instrumental count rates). For Cameras 1 and 2, which used
the F110M and F165M filters, respectively, the energy encirclement fully enclosed the
spatially-resolved PSF cores at 2.6 and 2.3 pixels (at the first Airy minimum of the stellar
diffraction pattern). In the case for Camera 3, because of the significant undersampling
near best-focus, the first Airy maximum was encircled to a radius of 3 pixels (as was easily
permitted due to the brightness of the star used).
For each camera: (1) the measured encircled energies were normalized to the maximum measured count rate, (2) the square-root of the normalized count rates as a function
5
Technical Instrument Report NICSMOS 2002-001
of PAM setting were fit to a 2nd order polynomial, (3) the "best focus” was determined as
the point of inflection for each fit (see Figure 5).
Coma
As part of the phase retrieval analysis of each stellar image, the x- and y-coma are determined and written to the output file. The coma value change with the PAM position due to
the change in the FOM position necessary to recenter the field on the detector during a
focus sweep. Therefore, coma values from images near best focus can be used to determine the actual values. A full analysis of coma measurements and a discussion of the
necessary PAM tilt correction will be presented in a separate TIR.
Results
Two methods (phase retrieval and encircled energy) were used to determine the PAM
focus positions for each camera. In addition, two independent analysis were performed
(and reported here) using phase retrieval. The focus sweep results are presented in Table 1.
The difference between the two phase retrieval results for the PAM2 position is most
likely due to manual selection of the center of the PSF, the PSF extraction box size used
for fitting, and the selection of the areas used to determine the amount of background to
subtract from the images. The difference is larger than expected, but not significant.
Table 1. The phase retrieval and encircled energy focus positions (not corrected for
breathing). The phase retrieval results are the mean of the individual star measurements
per camera translated to the center of the arrays. The dispersion in the phase retrieval
results for the individual star measurements is ~0.05 mm per camera.
PAM1
PAM2
PAM3
Phase Retrieval
+1.80
+0.05
-11.97
Phase Retrieval
+1.76
+0.25
-11.98
Encircled Energy
+1.79
+0.29
-10.88
The PAM positions, to affect best foci at the detectors, appear to have moved marginally in the negative direction for Camera 1 and 2 and in the positive direction for Camera
3. The recommended nominal PAM positions in mm per camera are presented in Table 2.
For the phase retrieval results, there were no correlations between PAM focus positions
and the thermal breathing models. Thereby, no breathing scaling factor was applied to the
phase retrieval results.
The intermediate focus setting between Camera 1 & 2 (PAMI) is positioned at a location which evenly splits the afocal wavefront error in both cameras. Camera 1 PSF is
critically sampled at 1.0 µm, and Camera 2 PSF is critically sampled at 1.75 µm. To
6
Technical Instrument Report NICSMOS 2002-001
evenly split the error in the wavefront error (at 1 µm for Camera 1 and 1.75 µm at Camera
2) the PAM needs to be positioned 36.36% of the distance of travel between the best focus
for Camera 1 and Camera 2. Specifically, this is at a distance in PAM space of +1.218 mm
for the intermediate focus setting (PAMI). At the recommended intermediate focus position, we achieve a wavefront error of λ/35 in both cameras at their critically sampled
wavelength.
The NICMOS optics were designed to form confocal (conjugate) image planes at the
field-divider assembly (FDA) mirror (where the coronagraphic hole is located) and at the
detector. Because the detector bench moved forward as a result of the N2 ice expansion
and deformation of the dewar, the PAM position for “normal” imaging was moved forward to refocus the re-imaged beam at the OTA f/45 detector focus position. This resulted
in an afocal image at the FDA mirror which has no consequence for direct imaging. However, the f/24 image plane falls beyond the FDA mirror (closer to the f/45 relay optics). For
coronagraphic imaging, the f/24 input beam is focused at the FDA mirror (i.e., at the hole)
to keep the PSF core size as designed in the coronagraphic hole. At this time, there is no a
priori reason to change the coronagraphic focus.
Table 2. The recommended 8977 nominal focus positions in mm of PAM motion. No
breathing correction has been applied.
PAM1
PAM2
PAM3
PAMIa
PAMCb
Date
2.36
0.69
-9.50
1.75
2.69
pre-NCS
1.80
0.20
-9.50
1.22
2.69
May 8, 2002
a. PAMI is an intermediate focus position between Camera 1
& 2 (CAMERA-FOCUS=1-2).
b. PAMC is the best focus position for Camera 2 coronagraphy.
Conclusions
We have determined the PAM settings for best focus in each NICMOS camera after the
Servicing Mission 3B and following the NICMOS cool down. The detector temperatures
were regulated at ~76.5 K during the focus sweep. Comparisons made between phase
retrieval and encircled energy measurements are in good agreement. The agreed upon
focus positions for all three cameras have been up linked to the telescope to be written into
EDAC and EEPROM (OPS Request 16762-0, May 9, 2002). No focus adjustments were
implemented for NIC3. The intermediate Camera 1 & 2 focus position (PAMI) will be
managed by SMS. No change to the coronagraphic focus is recommended at this time. In
addition to adjusting the PAM to achieve the best possible focus in all three cameras, a
preliminary assessment was made for coma, and preliminary adjustments were up linked
to the spacecraft to null out the comal aberrations in Camera 1 and 2.
7
Technical Instrument Report NICSMOS 2002-001
The coma question for NIC1 was addressed following execution of the 8977 visit 5
observations (May 10, 2002), and a final recommendation for the NIC1 x-y tilt for science
observations was determined and up linked (May 16, 2002) to the telescope. NIC2 and
NIC3 x-y tilt observations are forth coming. A full analysis of coma measurements and a
discussion of the necessary PAM tilt correction will be presented in a separate TIR.
Acknowledgements
References
Burrows, C.J. and Krist, J.E. 1997, Memorandum, March 19, 1997, “NICMOS Focus”
Galas, G. 1998, “NICMOS Focus Data Reduction and Analysis using Phase
Retrieval,” NICMOS Instrument Science Report, NICMOS-98-016, (Baltimore:STScI)
Krist, J.E. and Burrows, C.J. 1995, Applied Optics, 34, 4951
Krist, J.E. and Burrows, C.J., “STScI Phase Retrieval Software (FITPSF) User’s
Guide”, STScI, Baltimore, MD (September 1997)
Park, S. and Schowengerdt, R. 1983, Comput. Vision, Graphics, Image Processing, 23,
256.
Roye, E. and Schultz, A. 2002, “NICMOS Focus Data Analysis: Using the Updated
Phase Retrieval Software,” NICMOS Instrument Science Report, NICMOS-02-xxx,
(Baltimore:STScI)
Schultz, A., et al. 2001, "NICMOS Instrument Handbook", Version 4.1, (Baltimore:
STScI).
Suchkov, A., Bergeron, L., and Galas, G. 1998, “NICMOS Focus Monitoring,” NICMOS Instrument Science Report, NICMOS-98-004, (Baltimore:STScI)
8
Technical Instrument Report NICSMOS 2002-001
NICMOS focus (not centered) May 3, 2002
NIC1(n): focus (PAM space) = 1.934 +/- 0.50 mm
NIC2(n): focus (PAM space) = 0.141 +/- 0.55 mm
NIC3(n): focus (PAM space) = -12.53 +/- 0.43 mm
Figure 4: Phase retrieval results for one star in each camera. Zernike coefficient Z4, focus,
and thermal breathing correction as a function of PAM position (no breathing correction
applied and thus set to zero) plots are presented. All parameters are measured in mm of
PAM space. The dashed line in the middle column of plots indicates mean focus. No spatial correction was applied (“not centered”) to translate the focus to the middle of the
detectors. The spatial correction is applied following the phase retrieval analysis.
9
Technical Instrument Report NICSMOS 2002-001
NIC3
NIC2
NIC1
Figure 5: Encircled energy (EE) measurements for the NICMOS focus sweep of May 3,
2002 and resulting polynomial fits to find “best focus” as discussed in the body of this
TIR. The commandable PAM setting, which changes the location of the optically relayed
focus in each camera, is plotted in mechanical steps of motion of the mechanism from the
mid-range of travel where STEPS_FROM_ZERO = -1.4531 + (908.96 *
MM_OF_PAM_TRAVEL) as determined by the NICMOS IDT. The “absolute” setting of
the axial translation of the PAM in mechanical steps to affect “best focus” is at the maxima
of the EE fits (but is beyond the range of travel for Camera 3).
10