WFC3 Instrument Handbook for Cycle 21

As described above, the UVIS channel uses two 4096×2051 CCDs, butted together to yield a 4096×4102 array with a ~31 pixel (1.2 arcsec) gap. Because the detector is tilted along its diagonal axis 21° with respect to the incident beam, the field of view projected onto the sky is rhombus-shaped, 162 arcsec on a side, with an angle of 86.1° between the sides at amplifiers B and C (Figure 6.1). The pixels projected onto the sky are also rhomboidal, ~0.04 arcsec on a side.

Distortions due to the WFC3 optics cause the nominally square field of view of the UVIS detector to map onto the sky as a rhombus with small higher order distortion. Geometric distortions in both channels are discussed in more detail in Appendix B:Geometric Distortion.

Distortion must be taken into account when exposures are flat-fielded, photometrically calibrated, used for astrometric measurements, or combined with other dithered exposures. The AstroDrizzle software appropriately carries out those operations; a combination of software packages in DrizzlePac can be used to optimize the combination of dithered exposures. (See the DrizzlePac documentation.)

There are three different coordinate systems defined for use with the CCDs in the UVIS channel, each tailored to specific purposes. They are shown in Figure 6.1 and are as follows:

Figure 6.1 illustrates the fiducial points of the full-detector apertures (UVIS, UVIS1, UVIS2, and UVIS-CENTER), and the outlines of the 2Kx2K, 1Kx1K, and 512x512 subarray apertures. Also indicated are the positions of the four readout amplifiers (A, B, C, and D). The regions imaged by the UVIS detector (represented by blue fill) and by the IR detector (represented by red fill) are also indicated. The POSition TARGet coordinate system for the UVIS-CENTER aperture, with its origin at that aperture's fiducial point, is illustrated. Although the POSition TARGet coordinate systems for the other apertures are not illustrated, they are oriented the same, but have origins at each aperture's fiducial point
(Section 6.4.5). (U2 = –V2 and U3 = –V3).

The image-based coordinate system (Axis1, Axis2, as shown in Figure 6.1) is an orthogonal system of detector pixel readouts. Axis1 is aligned with the detector data rows and Axis2 with the columns. It is used by the calibration pipeline and other data-analysis software and is sometimes also referred to as the user frame. When a detector image is displayed on a computer screen, this system has the X-axis (Axis1) increasing to the right and the Y-axis (Axis2) increasing to the top of the screen, with 1 being the conventional index of the first pixel in both axes. For WFC3/UVIS, each chip has its own origin and Axis1, Axis2 system. The image-based coordinate system is used in most figures in this handbook, as well as in the aperture definitions available in the Science Instrument Aperture File.

The origin of this coordinate system is context specific. Logical or image coordinates have the origin at the lower left corner of the image, even for subarray images or images containing overscan data. These coordinates are shown as the “Image X,Y” coordinates in DS9. Physical or CCD coordinates have their origin at the lower left corner of the chip's light sensitive area. Thus, pixels to the left of the science area within the physical overscan area (see Figure 6.14) actually have negative coordinates in the physical image-based system. These physical coordinates are displayed as “Physical X,Y” coordinates in DS9, which determines the coordinates using the LTV1 and LTV2 FITS header keywords. The lengths of the axes, in pixels, are stored in the FITS header keywords NAXIS1 and NAXIS2.

The POS TARG reference frame, sometimes referred to as the spacecraft system, is used for specifying the placement of an astronomical target relative to the aperture reference point (sometimes called the fiducial point) in the instrument field of view. Its units are arcseconds. For the UVIS channel, the POS TARG system is defined such that the POS TARG Y axis is parallel to Axis2 at the reference point of the aperture in use. The POS TARG X axis is orthogonal to the POS TARG Y axis; it is not parallel to Axis1 due to geometric distortion.

As is the case for other HST instruments, the POS TARG origin is defined to be at the reference point (fiducial point) of the user-selected UVIS aperture (such as the geometric center of a particular chip, or the optimum center of a quadrant, etc.; see Table 6.1 below for the names of the various UVIS channel apertures). Figure 6.1 illustrates the POS TARG reference frame for the “UVIS” aperture, whose center is near the middle of the WFC3 UVIS field of view; the POS TARG directions are indicated by arrows labeled Xpos and Ypos.

The HST-based, or vehicle (V2, V3), system is an orthogonal reference frame tied to the telescope and is used operationally for alignment, pointing, and slewing purposes. The V1 axis lies along the optical axis while V2,V3 run parallel and perpendicular, respectively, to the solar-array rotation axis (see Figure 2.2). Note that the (undistorted) diagonals of the WFC3 CCD detector run along the V2,V3 directions. Because WFC3 is on-axis, the origin of the V2,V3 system lies near the center of the WFC3 field of view. However, the V2,V3 (and U2, U3) coordinate axes have been shifted for clarity in Figure 6.1. HST observers may be more familiar with the U2,U3 coordinate system than V2,V3; for example, the specification of the ORIENT angle Special Requirement in APT uses the position angle of the U3 axis. The U2,U3 coordinates are defined as U2 = –V2 and U3 = –V3, and are marked in Figure 6.1. Observations of an astrometric field are made to locate the detector in the (V2, V3) system (WFC3 ISR 2009-35).

A fourth coordinate system (the detector-based reference frame in pixel units) is described here for completeness, but observers are unlikely to encounter this system other than in technical documents created during the development and ground-testing of WFC3. The detector-based system (Xdet, Ydet) is used by the flight software for commanding the detectors. It is a right-handed system based upon the orientation of the CCD serial registers, with its origin at Amplifier A (the four amplifiers are in the outer corners of the detectors, as shown in Figure 6.1). The +Xdet and +Ydet axes map to the –Axis2 and +Axis1 axes, respectively. Unlike the image-based Axis1, Axis2 system, the detector system is 0-indexed. Parallel shifting is performed along constant Xdet, and serial shifting is done along the constant Ydet direction (Section 6.7.2).

While the default WFC3 UVIS readout mode is full-frame (i.e., both CCD chips), subarrays may be employed to read out and store only a portion of the full field of view. Subarrays may be used, for example, in cases where the full-field data-volume and/or instrument-overhead requirements would constrain the observations unacceptably. There are also circumstances in which on-chip binning may be desirable. Both modes have implications for the inclusion of physical overscan information in the readout, as listed in Table 6.1 and discussed in Section 6.7.2.

Beginning in Cycle 18, a wider range of subarray sizes was provided: 512×512, 1k×1k, and 2k×2k. User-defined subarrays, and subarrays that span quadrant boundaries, are no longer supported. Subarrays are invoked via the appropriate Aperture parameter in the Phase II observing proposal; these apertures contain the string “SUB” in their names. Figure 6.2 shows the supported subarrays. See Section 6.4.5 and Table 6.1 for the reference points (default target positions) of these apertures.

For the special case of quad filters (which are optical elements that include four different bandpasses simultaneously, one bandpass per WFC3 UVIS quadrant), the observer must select one of the “QUAD” Aperture values in the Phase II proposal, in conjunction with the desired quad filter via the filter parameter. This combination of quad aperture and quad filter ensures that the target is automatically placed in the correct quadrant for imaging with the requested quad bandpass. Furthermore, specification of the subarray quad aperture (UVIS-QUAD-SUB) instructs the ground system to read out only the 2k×2k quadrant of interest. If “-SUB” is omitted from the quad aperture name (i.e., UVIS-QUAD, UVIS-QUAD-FIX), the target is positioned in the proper quadrant for the bandpass requested in the filter parameter, but the entire frame, both CCDs, is still read out. Table 6.1 indicates which apertures place the target at the geometric center of the subarray, and which apertures place it at a substantial offset from the center. See the Phase II Proposal Instructions for updates in aperture definitions at the beginning of a new cycle.

Use of any quad filter aperture, full-frame or subarray, is fully supported. However, if data volume and overhead time are not an issue, the observer is encouraged to allow full-frame readouts of quad-filter exposures, enabling serendipitous discoveries in the other quadrants of the image as well as enrichment of the HST data archive.

For greater sensitivity when observing very faint targets, or for acquiring full-frame images as rapidly as possible, the BIN optional parameter in the observing proposal may be used to specify that the CCDs be read out in binned mode. Legal values for the parameter are NONE (the default), or 2 or 3, for 2×2 and 3×3 on-chip binning, respectively. On-chip binning is only allowed with full-frame readouts; it may not be used in conjunction with any subarray mode. To perform on-chip binning, multiple rows are parallel-shifted into the serial register, which is shifted multiple pixels at a time into a single output pixel, which is then read once. This type of binning improves the signal-to-noise in an image by minimizing readout noise: a 2×2 or 3×3 binned pixel contains approximately the same amount of read noise as a single pixel in an unbinned readout (see Section 5.4.7), but 4 times or 9 times the signal.

In addition to the S/N improvement, on-chip binning results in a significant reduction of readout overhead time, not only because of the shorter time needed to read out 1/4 or 1/9 of the usual full-frame number of pixels (~50 and ~23 sec, for 2×2 and 3×3 binning respectively, compared to 93 sec for unbinned), but the image sizes are also smaller, i.e., more images will fit into the instrument Science Data Buffer before it must be dumped to the SSR. However, binning results in a loss of spatial resolution and increases the area affected by bad pixels and cosmic rays; the observer must weigh the benefit of binning against the cost of degraded resolution and greater contamination by bad pixels and cosmic rays in the science images.

Binning affects the overscan data associated with an exposure. At the frame edges, some science and overscan pixels are binned together. Section 6.7.2 gives details concerning the rows and columns affected by binning mixed science and overscan data.

The APERTURE parameter in the Phase II observing proposal defines two quantities: the active region of the detector to be read out (full frame or subarray), as well as the positioning of the target within the region (reference point). The default is to place the target at the reference point (also called the fiducial point) of the chosen aperture, but a POS TARG Special Requirement may be added to offset the target from this position.

With regard to pointing HST, there are two types of apertures: “fixed” and “optimum.” The fixed positions have reference points near the geometric center of the aperture in question, and, as the name implies, their locations will remain fixed in image-based coordinates for the life of the instrument. The “optimum” apertures have reference points that are offset from the geometric center so as to avoid any known CCD features (e.g., bad column, quad filter edge effects) that might compromise the observation of a target at that position. The locations of the “optimum” aperture reference points—in both the image-based coordinate system and the spacecraft V2,V3 coordinate system—may change over time as the characteristics and the characterization of the CCD evolve.

	For updates on subarray and aperture definitions, check the Science Instrument Aperture File (SIAF) page: http://www.stsci.edu/hst/observatory/apertures/wfc3.html

The choice of optimum or fixed aperture depends on the program’s objectives and target. For a very small target, or one which easily fits into the aperture’s field of view, the optimum aperture’s reference point is defined to ensure that the target center does not fall on known problematic areas of the detector. The WFC3 Web site and STAN will announce changes in optimum aperture definitions. On the other hand, when the observer needs precise control of the position of the target on the detector, fixed apertures should be used. Fixed apertures are therefore appropriate when target position relative to detector edges is important, or when mosaics are being designed with edges requiring positional accuracies of better than 10 arcsec or so. The available WFC3 UVIS apertures (Table 6.1) include both fixed and optimum versions for the primary locations: the center of the full two-chip detector (UVIS and UVIS-FIX), and the center of each chip (UVIS1 and UVIS2, and UVIS1-FIX and UVIS2-FIX).

There are also fixed and optimum apertures for use with the quad filters. Because the filter wheel assembly is necessarily offset from the focal plane, the edges between quad filters are blurred at the focal plane, producing regions with contributions from multiple quad filter passbands (Figure 6.8). The UVIS-QUAD-SUB apertures (redefined for cycle 20) and UVIS-QUAD apertures have reference points centered within the useful single-passband regions, while the UVIS-QUAD-FIX apertures have reference points at the geometric centers of the quadrants (closer to the filter edge effects). In programs where targets are placed in different quadrants, the choice of quad aperture will affect the size of offsets and may require new guide star acquisition, as described in Section 10.2.

Subarray apertures pictured in Figure 6.2 are all “optimum”; no “fixed” apertures are available for these subapertures. The reference points for these apertures have initially been defined near the geometric center of the subarray, except for the 2K2 apertures, where the reference points match those of the UVIS-QUAD apertures.

Aperture	Over- scan¹	Region	Reference (fiducial) point
UVIS	P, V	Full detector	Optimum point near center (displaced from interchip gap) of full two-CCD field of view
UVIS-CENTER	P, V	Full detector	Geometric center of full two-CCD field of view
UVIS-FIX	P, V	Full detector	Near geometric center (displaced from interchip gap) of full two-CCD field of view
UVIS1	P, V	Full detector	Optimum point near center of CCD Chip 1
UVIS1-FIX	P, V	Full detector	Geometric center of CCD Chip 1
UVIS2	P, V	Full detector	Optimum point near center of CCD Chip 2
UVIS2-FIX	P, V	Full detector	Geometric center of CCD Chip 2
UVIS-IR-FIX	P, V	Full detector	Pointing matched to IR-FIX aperture in Table 7.1
G280-REF2	P, V	Full detector	Reference point for undispersed exposures (coincides with reference point of exposures made with the G280 grism)
UVIS-QUAD	P, V	Full detector	Optimum point in the quadrant corresponding to the selected quadrant filter (offset from the center of the quadrant toward the nearest corner of the detector by about 8 to 10 arcsec in X and in Y; see Figure 6.8)
UVIS-QUAD-FIX	P, V	Full detector	Geometric center of quadrant corresponding to selected quadrant filter
UVIS-QUAD-SUB	P	2048×2050, quadrant corresponding to selected quadrant filter	Same as UVIS-QUAD (optimum point in the relevant quadrant) starting in Cycle 20. (Previously geometric center of the quadrant.)
UVIS1-2K2A-SUB UVIS1-2K2B-SUB UVIS2-2K2C-SUB UVIS2-2K2D-SUB	P	2048×2050	Optimum point for the corresponding quadrant filter (for matching N,M,W filter exposures to UVIS-QUAD aperture exposures), about 14 arcsec from the center of the 2k × 2k subarray
UVIS2-M1K1C-SUB		1024×1024, quadrant C near detector center	Optimum point near center of 1k × 1k subarray
UVIS2-C1K1C-SUB	P	1024×1024, near Amplifier C	Optimum point near center of 1k × 1k subarray
UVIS2-M512C-SUB		512×512, quadrant C near detector center	Optimum point near center of 512×512 subarray
UVIS2-C512C-SUB	P	512×512, near amplifier C	Optimum point near center of 512×512 subarray;