Space Telescope Science Institute
Table of Contents Previous Next Index Print

Near Infrared Camera and Multi-Object Spectrometer Instrument Handbook for Cycle 17 > Chapter 5: Coronagraphy, Polarimetry andGrism Spectroscopy > 5.1 Coronagraphy

5.1 Coronagraphy
NICMOS Camera 2 (NIC2) has a coronagraphic observing mode which is provided by an occulting hole in the Camera 2 FDA mirror, together with an internal cryogenic mask located at the Lyot plane. The internal cold baffling serves to screen out residual radiation from the edges of the HST primary and secondary mirrors as well as the secondary mirror support structures (pads, spider, and mounts).
Since the FDA mirror is in the image plane, a bright star can be placed such that its core falls onto the occulting hole in the FDA mirror. The star is subsequently re-imaged onto the detector at the focal plane, with its central region occulted. At a radius of 0.3 arcsec, in an idealized PSF, a natural break occurs in the encircled energy profile at 1.6 micron with 93% of the energy in the PSF enclosed. Beyond this radius, the encircled energy profile flattens out toward larger radii.
Figure 5.1: Image of the coronagraphic hole in NIC2. The rough edges created by the final drilling process are evident in this figure.
Figure 5.2: Star images. Star imaged outside the hole (left) and within the coronagraphic hole (right). Images were obtained with the F160W filter. Two of the three bright glint regions are marked with arrows. The other is on top of the hole. Images displayed to same stretch.
The light pattern about the coronagraphic hole is not symmetric due in part to the coronagraphic optics and to the Optical Telescope Assembly (OTA) input PSF. The spectral reflections from the roughness about the hole, and imaged in Camera 2, will vary depending upon the location of the target in the hole. There is one azimuth region where the residual light pattern, historically called glint, is brightest. Figure 5.2 presents enlarged images of the same target outside and positioned within the coronagraphic hole displayed to the same stretch. The structure of the scattered light pattern about the hole is different from the NICMOS stellar PSF pattern. The presence of glint brings the useful coronagraphic radius at the detector to ~0.4 arcsec.
The FDA mirror and the Camera 2 ƒ/45 optics image planes are not exactly parfocal. For nominal Camera 2 imaging, the PAM is positioned to achieve optimal image quality at the detector. For coronagraphic imaging, the PAM is adjusted slightly for optimal coronagraphic performance. The PAM is moved to produce a focused star image at the position of the coronagraphic hole. This results in a very slight degradation of the image quality at the detector. The PAM movement is automatic whenever Opmode=ACQ or Aperture=NIC2-CORON are specified on the Phase II exposure line. If a series of exposures need the CORON focus position, only one move is performed.
The tilt of the PAM is changed to compensate for translation from the nominal to coronagraphic setting, and to remove off-axis aberrations.
The NICMOS dewar anomaly caused the coronagraphic hole to migrate to different locations on the detector, during Cycle 7 and 7N. The position of the hole on the detector had been observed to move as much as ~0.25 pixel in three orbits. During the interval April-December 1998, the hole moved about 1 pixel. The movement of the hole is not linear. Rather, the hole “jitters” back and forth along an X-Y diagonal by as much as 0.5 pixel. As an example of post-NCS behavior, Fig 5.3 shows the movement of the hole during Cycle 11.
Figure 5.3: Coronagraphic hole location (detector coordinates) during Cycle 11.
The movement of the hole may cause a problem for coronagraphic observations. Repeat positioning of targets in the coronagraphic hole to a fraction of a pixel is necessary for PSF subtraction. For this reason, the acquisition software is set up to locate the hole position for every acquisition of an astronomical target.
Coronagraphic imaging requires an acquisition sequence at the beginning of the observation to position the target in the coronagraphic hole as the size of the coronagraphic hole is smaller than the typical HST blind-pointing errors. The procedure for a coronagraphic acquisition is to first image the target in the NIC2-ACQ aperture (128128 pixel aperture) using blind-pointing and then use either an onboard acquisition, reuse target offset acquisition, or interactive acquisition to acquire the target. A telescope slew is calculated and commanded to move the image of the target over the position of the hole.
The science exposures are then specified using any of the NICMOS observing modes and any of the NIC2 filters. The science observations following the ACQ need to specify the Aperture = NIC2-CORON.
Onboard Acquisition (Mode-2 Acquisition)
The Mode-2 Acquisition for coronagraphy includes two steps: first, the position of the coronagraphic hole is located; second, the target is acquired and placed in the hole.
The location of the coronagraphic hole is determined from pointed flat field observations. Two short F160W filter exposures (7.514 seconds each) with calibration Lamp 1 on (flat field) and two identical exposures with the lamp off (background) are obtained before the acquisition images. The background images are needed because NICMOS does not have a shutter and the flat field images are also imaging the sky. The flight software (FSW) combines the two background and two flat field images by performing a pixel-by-pixel minimum to eliminate cosmic rays (using the lower valued pixel of the two frames). The processed background is subtracted from the processed lamp flat and the image is inverted by subtracting the image from a constant. A small 32  32 pixel subarray containing the hole is extracted and a small checkbox (15  15 pixels) is used to find the centroid and a weighted moment algorithm is applied to determine the flux-weighted centroid within the checkbox. The location of the hole is temporarily stored onboard, but it is not saved in the engineering telemetry sent to the ground.
The target needs to be positioned within the NIC2-ACQ aperture, a square 128  128 pixel area on the detector (center at 157, 128) of size 9.6  9.6 arcseconds. Two images of equal exposure are obtained. (The Phase II exposure time is not split.) The two images are pixel-by-pixel minimized to eliminate cosmic ray hits and a constant value (data negative limit) is added to the processed image. The brightest point source in the acquisition aperture is determined by summing the counts in a checkbox of size 3  3 detector pixels. The algorithm passes the checkbox over the entire acquisition aperture. The brightest checkbox is selected and the location of the target is determined by centroiding the X,Y center of the 3  3 checkbox.
The observer needs only to specify a NICMOS onboard Acquisition (ACQ) to acquire the target. The software schedules the background and flat field observations first, followed by the observations of the target. The exposure times for the pointed background and flat field observations are 7.514 seconds. As an aid to coronagraphic observers, Figure 5.4 presents a plot of counts in the peak pixel for a centered point source obtained with the F160W filter as a function of integration time. The right-hand axis indicates the percentage of full-well for that peak pixel. The true responses of the pixels where the target falls within the FOV will vary. Thus 70% full well should be a reasonably conservative goal for the peak counts needed for a successful acquisition. Over plotted on the figure are diagonal lines which indicate the counts in the peak pixel of a PSF for H-band magnitudes from 8 to 18 (labeled). The shaded region in the lower-right indicates a domain where relatively hot pixels’ dark current will result in more counts than faint point-sources, which will cause the acquisition to fail.
Figure 5.4: Coronagraphic point source acquisition exposure times. A goal of 70% full well is recommended for the peak pixel. Over plotted are diagonal lines indicating the estimated counts for different H-band magnitudes. The shaded region on the lower right shows the regime where hot pixels might confuse the acquisition.
Note that the telescope is not slewed to position the target out of the FOV for the background and flat field observations. If the target saturates the NIC2 detector in 7.514 seconds with the F160W filter, a residual image will be created that will contaminate the onboard target ACQ observation.
Very bright targets will cause saturation, leading to poor results in the centroid solution, and in the subsequent placement behind the occulting hole. To avoid this, a narrow band filter may have to be used to reduce the target flux. Targets brighter than H ~4.0 will saturate the central pixel of the PSF when observed with the F187N filter (narrowest Camera 2 filter) using the shortest ACQ integration time of 0.228 seconds. Since the NICMOS filters are essentially at a pupil plane, there will not be an image shift introduced by using a different filter for the acquisition than for the science observations. Shading will be a problem for centroiding when the target lands near the shading break, as no dark subtraction is performed.
The location of the target and the slew are saved and sent to the ground together with the science observation.
The NICMOS ACQ exposure times, T_acq, are quantized with a minimum exposure time of 0.228 seconds. For an ACQ exposure with T_acq=0.356 sec, the overhead to complete the hole finding and location of the target is about 3 minutes which includes the telescope slew to move the hole over the target. A full description of the overheads for Mode-2 Acquisitions is given in Chapter 10.
For T_acq exposures longer than ~5 minutes the probability of cosmic ray hits occurring in the same pixel in each of the two acquisition images is sufficiently high that observers must instead use an early acquisition to avoid their observations failing due to a false center determination. Early acquisitions are described in the next section. In practice, this should not be a severe restriction as in the F160W filter one will reach a signal-to-noise of 50 at H=17 in only 2–3 minutes.
The flight software processed images are not saved, but the two background, two flat field, and two acquisition ACCUM images are sent to the ground. These images, which are executed in a single target acquisition observation, will be packaged into one data set with the same root name but with different extensions. A full description of the extensions is given in the HST Data Handbook.
Starting from Fall 2003, calnica calibrates the ACQ Accum and Accum background images. The Accum F160W flat field is calibrated up through dark subtraction, using a temperature dependent dark constructed on the fly given the exposure time and detector temperature.
Reuse Target Offset (RTO) and Interactive Acquisitions
Bright targets will saturate the NICMOS Camera 2 detector, resulting in possible failure of the onboard software (Mode-2 Acquisition) to successfully acquire and position the target into the coronagraphic hole. Any target that will saturate the detector in the shortest possible Mode-2 ACQ exposure time, 0.228 seconds, should be considered a bright target. A variation of the Reuse Target Offset (RTO) capability can be used to acquire and position a bright target into the coronagraphic hole. In addition, the onboard acquisition software may not successfully acquire the desired target in a crowded field. For this case, an interactive acquisition (INT-ACQ) may be required to successfully acquire the target.
The following discussion describes the necessary steps for a Reuse Target Offset (RTO) acquisition to acquire a bright target and position the target into the coronagraphic hole. These steps can also be used for an interactive acquisition of a target in a crowded field. It is recommended for RTO acquisition that two orbits be used when observing a bright target except possibly for an INT-ACQ. The first orbit is used for the acquisition and the second orbit for the coronagraphic observations.
Images of the target and coronagraphic hole are obtained a few orbits in advance of the coronagraphic observations, and sent to the ground for analysis (RT ANALYSIS). The target exposures should be offset from the NIC2-CORON aperture fiducial point to avoid having the target fall in the hole.
An RTO slew is limited to ~10", taking ~26.5 seconds of time to complete. A target offset of ~9" or less from the hole position is highly recommended to avoid exceeding the 10" limit due to the uncertainty in the target coordinates and proper motion.
The observer needs to specify at least two background, two flat field, and two on-target exposures in the Phase II template. The background and flat field observations should be offset by 18-25 arcseconds from the target position to avoid the diffraction spike from the image of an overexposed target crossing the coronagraphic hole and introducing errors in the measured position of the coronagraphic hole. The recommended pairs of images are needed to remove cosmic ray hits (Schultz, 1998, NICMOS-ISR-031).
OPUS staff will assist the PI in identifying the target, centroiding, and determining offsets. OPUS staff will then provide the offsets to the Flight Operations Team (FOT) at the Space Telescope Science Institute for uplink to the spacecraft in advance of the coronagraphic observations. The ultimate responsibility for determining the offsets will be the PI (or the PI’s representative), who must be present at STScI at the time of the target/hole location observations.
Both the total encircled energy rejection (from the occulted core of the PSF) and the local contrast ratio obtainable in a coronagraphic image depend on the accuracy of the target centering on the coronagraphic hole. The goal is to center the PSF of the occulted source to a precision of a 1/6 pixel at a position x=–0.75, y=–0.05 pixels from the center of the hole, “the low scatter point” (during Cycle7 and 7N the low scatter point was at x=-0.75, y=-0.25). The decrease in the fractional encircled energy due to imprecise centering of the core of an idealized PSF in the coronagraphic hole is 0.3 percent for a 1/4 pixel offset, and 4.4 percent for a 1 pixel (75 milliarcseconds) offset at 1.6 microns.
In addition, a small error in target centering will create an asymmetric displacement of the PSF zonal structures both in and out of the coronagraphic hole, leading to position dependent changes in the local image contrast ratios.
Temporal variations of the NICMOS PSF due to HST breathing and “wiggling” of the misaligned cold mask in NIC2 are discussed in Chapter 4. Of relevance to coronagraphic observers is that the effects of temporal variations for PSF subtraction can be minimized by obtaining observations of the same PSF in back-to-back orbits or twice in the same orbit, with a roll of the spacecraft between the two observations. The success of this technique is due to the orbital timescale of the PSF temporal variations.
During Cycles 7 and 7N, the NICMOS IDT reported very good results for PSF subtraction when the same target was observed twice in the same orbit with a roll of the spacecraft between observations. Between Cycle 11 and the start of Two Gyro Mode operations during Cycle 14, this observational strategy was available to General Observers (GOs). This strategy was generally not feasible under Two-Gyro Mode. With the return to Three-Gyro Mode in Cycle 17, this strategy will again be available.
Back-to-back coronagraphic observations of the same target with a roll of the spacecraft between observations are scheduled as two separate visits. The two visits are linked close in time by using the Phase II visit-level requirement “AFTER”, such as AFTER 01 BY 15 MINUTES TO 30 MINUTES as a requirement on the second visit. The timing link will be used to schedule both visits. Each coronagraphic visit, including guide star acquisition, ACQ, exposure time, and overhead must not exceed 22 minutes in duration to allow time to roll the telescope between visits.
The roll between two visits must be expressed as a relative orient, such as ORIENT 25D TO 30D FROM 01 as a requirement on the second visit. The permitted amount of spacecraft roll varies throughout the Cycle as the target position changes relative to the Sun. A roll of 6 degrees between visits will take about one minute to complete which does not include the overhead to ramp up and ramp down the motion. A 30 degree roll will take about nine minutes to complete. These roll overheads need to be allowed for in Phase I planning, but they are automatically handled by the scheduling software when the visits are actually scheduled.
Two guide star (GS) guiding is strongly recommended when performing coronagraphic observations. For best coronagraphic results, the target should be centered to better than 1/6 pixel.
Coronagraphic observations executed in back-to-back orbits should be scheduled with the same guide star pair, except possibly if a roll of the HST is performed between orbits. This is also critical for Reuse Target Offset (RTO) Acquisitions, which require the same guide stars be used for all observations. Switching guide stars between the acquisition and science observations will force the respective target to either be positioned away from the coronagraphic hole or on the edge of the hole.
The use of a single guide star is discouraged for coronagraphic observations. The drift about a single guide star is small, but will yield intense residuals for PSF subtraction. If we represent the linear motion due to gyro drift around a star as , where X equals the linear motion, D the distance from the guide star to the aperture, a the angular gyro drift rate, and t the time since the last FHST (fixed-head star tracker) update, then for D = 20 arcmin (worst case) = 1200 arcsec and a = 0.001 arcsec/sec, for one visibility period t = 50 min = 3000 sec we get X = 0.0175 arcsec or less than 1/4 pixel in Camera 2. For two orbits t = 146 min = 8760 sec, X = 0.051 arcsec or a little over 2/3 pixel in Camera 2.
During Cycle 7 and 7N, NICMOS SNAP coronagraphic observations were scheduled with single guide star guiding. Starting with Cycle 11, all NICMOS SNAPSs including coronagraphic observations are scheduled with two-guide star guiding.
For RTO acquisitions, the maximum default slew is 10 arcseconds. This is set by the coordinate uncertainties as specified in the Phase II template. If a slew larger than the default 10 arcseconds is scheduled, it has to be approved by the STScI Commanding Group and the FOT notified that a slew of this size or larger will not force the guide stars out of the FGS field of view (a.k.a. pickle). Increasing the target coordinate uncertainties will increase the slew limit. STScI Commanding will use the coordinate uncertainties to determine the size of the slew request timing. Guide star selection is also affected. If the requested amount of guide star movement will force the guide star out of the pickle, the guide star selection software will not select that star. This may result in single star guiding. One solution to this problem is to decrease the distance between the target star and the hole and correspondingly decrease the target coordinate uncertainties. Note that the NIC2 field-of-view (FOV) is ~19 arcsec on a side.
Coronagraphic observations scheduled over more than one visibility period will most probably be impacted by an SAA passage and possibly be affected by charged particle induced persistence (see Chapter 4 for a discussion on the cosmic ray persistence). To avoid breaking exposures across visibility periods, coronagraphic observations should be scheduled using the exposure level Special Requirement “SEQ <exp. list> NON-INT”, which forces all observations to be within the same visibility window, i.e., without interruptions such as Earth occultations or SAA passages.
One of the coronagraphic calibration problems is “proper” calibration of images near the edge of the hole due to motion of the hole itself. The problem arises from the fact that the OPUS flat field reference files are not contemporary with the coronagraphic images. During Cycle 7 and 7N, the coronagraphic hole moved about 0.1 to 0.2 pixels per month. In addition, there is a second, short term component to the movement along a pixel diagonal (back-and-forth) and, imposed upon this motion, a third component of random jitter composed of a few tenths of a pixel.
The light pattern about the coronagraphic hole is not symmetric due to glint (see Section 5.1 and Figure 5.2), and will vary depending upon the location of the target in the hole. Calibrating with a contemporary flat, which has the coronagraphic hole pattern at the correct location, restores the flux level and re-establishes the light pattern about the hole at the time of the observation. For distances greater than ~0.7 arcseconds from the hole (diameter ~17 pixels), the standard, high S/N flat is the best reference file to use for calibration.
“Proper” calibration of coronagraphic images can be achieved with contemporaneous lamp and background observations. These calibration observations can be scheduled within the time allowed and will increase the scientific return of the science data. Calibration observations are normally obtained as part of the STScI calibration program and GOs are not usually allowed to request calibration data. However, the coronagraphic programs are allowed to obtain lamp and background observations to be used to locate the coronagraphic hole. For RTO Acquisitions, if there are no pressing scientific reasons to fill the remaining acquisition orbit with science observations, then it is recommended that lamp and background observations be obtained to support the coronagraphic science observations.
A Cycle 12 commissioning program showed that the NICMOS Camera 2 polarizing filters can be used successfully in combination with the coronagraph. This significantly enhances the imaging polarimetry mode by enabling polarization measurements of regions near a bright object, such as a star or active galactic nucleus.
Image artifacts within about 2 arcseconds of the coronagraph, caused by scattering off the edge of the hole, are repeatable and induce only a small instrumental polarization (IP ~2%). The calibratable polarization measurements near the hole are limited by this IP component and the current calibration data to about 8–10% (4 sigma in percentage polarization) per 22 pixels (approximate resolution element). However, the observed stability of the IP component implies that future refinement to the calibration and further characterization of the scattering about the coronagraphic hole may improve this limit.
Observers considering the use of the coronagraph combined with the polarizing filters should follow the standard recommendations for two-roll coronagraphic imaging, and remember that images through all three polarizers must be obtained at each roll. We also recommend that observers include observations of an unpolarized standard star in addition to their primary target object, and that the standard star be observed at sufficient depth to obtain similar S/N to the primary target in each polarizer. Thus, a minimum of two orbits per target is typically needed; i.e., target star and unpolarized standard star. A single, well-exposed unpolarized standard star should be sufficient for a multi-target science program.
The decision chart presented in Figure 5.5 helps guide the proposer through the selection process to construct coronagraphic observations when using an onboard acquisition or an early acquisition image. The process for specifying RTO acquisitions of bright target is presented in NICMOS-ISR-031 (13-Jan-1998). The observer is advised to contact the STScI help desk,, for additional information.
Figure 5.5: Coronagraphic Decision Chart

Table of Contents Previous Next Index Print