ACS was designed to provide a deep, wide-field survey capability from the visible
to near-IR using the Wide Field Camera (WFC), high resolution imaging from the near-UV to near-IR with the now-defunct High Resolution Camera (HRC), and solar-blind far-UV imaging using the Solar Blind Camera (SBC). The discovery efficiency of ACS’s Wide Field Channel (i.e., the product of WFC’s field of view and throughput) is 10 times greater than that of WFPC2.
The failure of ACS’s CCD electronics in January 2007 brought a temporary halt to
CCD imaging until Servicing Mission 4 in May 2009, when WFC functionality was restored. Unfortunately, the high-resolution optical imaging capability of HRC was not recovered.
|Plate scale of ~0.027
arcsec/pixel that provided critical sampling at 6300 Å before the CCD electronics failure in January 2007.
ACS is a versatile instrument that can be applied to a broad range of scientific
programs. For example, WFC’s high sensitivity and large field of view at red and near-infrared wavelengths make it the preferred camera for deep imaging programs in this wavelength region.
Before January 2007, HRC provided full sampling of the HST PSF at
λ > 6000 Å and could be used for high precision photometry in stellar population studies. The HRC coronagraph could be used for detection of circumstellar disks and QSO host galaxies.
ACS CCDs are thinned, backside-illuminated devices cooled by thermo-electric
coolers (TEC) and housed in sealed, evacuated dewars with fused silica windows. The spectral response of WFC CCDs is optimized for imaging at visible to near-IR wavelengths. HRC’s CCD covered wavelengths similar to WFC, but its spectral response was optimized for the near-UV. Both CCD cameras produce a time-integrated image in the ACCUM data-taking mode. HRC also operated in target acquisition (ACQ) mode for coronagraphic observations.
As with all CCD detectors, there is read noise and overhead associated with reading
out the detector following an exposure. The minimum exposure time was 0.1 seconds for HRC and is 0.5 seconds for WFC. Minimum time between successive identical full-frame exposures was 45 seconds for HRC and is ~135 seconds for WFC. These times can be reduced to ~36 seconds for WFC subarray readouts. The dynamic range for a single exposure is ultimately limited by depth of the pixel well (~85,000 e−
for WFC and 155,000 e−
for HRC), which determines the saturation limit of any one pixel. Hot pixels and cosmic rays affect all CCD exposures. CCD observations should be broken into multiple dithered exposures to allow removal of hot pixels and cosmic rays in post-observation data processing.
The SBC MAMA is a photon-counting detector which provides two-dimensional
imaging optimized for far-UV wavelengths. The MAMA can only be operated in ACCUM mode. SBC observations are subject to both scientific and absolute brightness limits. At high local (>= 50 counts sec-1
) and global (> 285,000 counts sec-1
) illumination rates, counting becomes nonlinear in a way that is not correctable. At slightly higher illumination rates, MAMA detectors can be permanently damaged. Lower absolute local and global count rate limits have been imposed that define bright-object screening limits for each SBC configuration. Targets that violate these screening limits cannot be observed in the proposed configuration.
ACS has two main optical channels, one dedicated to WFC and one shared by HRC
and SBC. These channels are shown in Figures 3.2 and 3.3 of the ACS Instrument Handbook
. Each channel has independent corrective optics to compensate for HST’s spherical aberration. WFC has three optical elements coated with silver to optimize visible-light throughput. The silver coatings cut off at wavelengths shortward of 3700 Å. WFC shared two filter wheels with the HRC, which enabled internal WFC/HRC parallel observing for some filter combinations.
The HRC/SBC optical chain comprises of three aluminized mirrors overcoated
. HRC was selected by inserting a plane fold mirror into the optical path so that the beam was imaged on HRC’s detector through the WFC/HRC filter wheels. The SBC channel is selected by moving the fold mirror out of the beam and allowing light to pass through the SBC filter wheel onto the SBC detector. The aberrated beam coronagraph was deployed with a mechanism that inserted a window with two occulting spots at the aberrated telescope focal plane and an apodizer at the re-imaged exit pupil. For detector health and safety reasons, use of the coronagraph with SBC is forbidden.
ACS detectors exhibit significantly more distortion than previous HST instruments.
All ACS observations must be corrected for distortion before any photometry or astrometry is derived. For a thorough discussion of ACS Geometric Distortion, we refer the reader to the Multidrizzle Handbook
The principal cause of ACS distortion is that the optics have been designed with a
minimum number of components, consistent with correcting for spherical aberration induced by the Optical Telescope Assembly (OTA), without introducing coma. The result is a high throughput, but focal surfaces far from normal to the principal rays. The WFC detector is tilted at 22°
giving an elongation of 8% along the diagonal. HRC and SBC detectors have a 25°
tilt giving an elongation of 12%. In each case, the scale in arcseconds per pixel is smaller along the radial direction of the OTA field of view than along the tangential direction. When projected on the sky, this causes each detector to appear “rhombus-shaped” rather than square. The angle on the sky between the x
-axes is 84.9°
for WFC1, 86.1°
for WFC2 and 84.2°
Orientations of the ACS detector edges are approximately in line with the V2 and
V3 coordinate axes of the telescope. Consequently, the eigenaxes of the scale transformation are along the diagonals for WFC, and the apertures and pixels appear non-rectangular in the sky projection. For HRC and SBC, the situation is even more irregular because the aperture diagonals do not lie along a radius of the HST field of view. Figure 1.1
shows the ACS apertures in the telescope’s V2/V3 reference frame and illustrates the “rhombus” shape of each detector. A telescope roll angle of zero degrees would correspond to an on-sky view with the V3 axis aligned with North and the V2 with East. The readout amplifiers (A, B, C, and D) are also indicated for each detector.
If these were the only distortions present, their impact on photometry and
mosaicing/dithering could be simply computed. A more problematic effect is the variation of scale and pixel area across each detector. For WFC this amounts to a change of ~10% in scale from corner to corner. For HRC and SBC this variation is only about 1%, since these detectors cover much smaller fields of view. The area on the sky covered by a WFC pixel varies by ~18% from corner to corner, allowance for which must
be made in photometry.
Dithering and mosaicing are complicated by the fact that an integral pixel shift near
the center of the detector will translate into a non-integral displacement for pixels near the edges. This is not a fundamental limitation, but will imply some computational complexity in registering images and will depend on an accurate measurement of distortions.
The readout amplifiers (A,B,C,D) are indicated on the figure. When ACS data are processed through multidrizzle
in the HST pipeline, the resulting drizzled images are oriented with their x,y axes corresponding approximately to the x,y axes shown here. Thus, the WFC data products are oriented so that WFC1 (chip 1, which uses amplifiers A and B) is on top (also see Figure 2.2
), and the HRC data products are also oriented such that amplifiers A, B are on top.
ACS suffered component failures in its Side 1 and Side 2 electronics in June 2006
and January 2007, respectively, that afterwards prevented operations of its WFC and HRC cameras. SBC was unaffected by these failures and remained available for scientific use throughout this problematic period. WFC was restored to operation after the successful installation of a replacement CCD electronics box (CEB-R) and power supply during Servicing Mission 4 (SM4) in May 2009. Unfortunately, additional damage to the HRC power harness during the original failure prevented recovery of the HRC during SM4. Consequently, HRC remains unavailable for scientific use.
Tests conducted shortly after SM4 showed that read noise, linearity, pixel full-well
depth, and amplifier cross-talk of the restored WFC are as good or better than the pre-failure levels in January 2007 (Table 1.1
). The WFC’s dark current, hot-pixel fraction, and charge-transfer efficiency (CTE) have degraded to the levels expected after extended exposure to HST’s trapped radiation environment. Bias frames obtained under default CEB-R operation show improved read noise and a 5
DN gradient spanning the rows and columns of each image quadrant. This bias gradient is stable over the time in-between consecutive calibration reference files, so it is precisely removed during normal image reduction and processing.
All WFC images exhibit faint horizontal stripes along the rows of pixels and across
the quadrant boundaries. The stripes are constant along each row, but they are not stable from frame to frame. They are caused by low-frequency (1
mHz to 1
noise on the bias reference voltage generated by the CEB-R. The contribution of the stripes to global read noise statistics is small (peak-to-peak deviation is approximately 2
DN), but the correlated nature of the noise may affect photometric precision for very faint sources. STScI has developed algorithms for removing the stripes from most science images, in the form of a task called acs_destripe
, that’s available in the stsdas.hst_calib.acs
package. Because it requires manual user intervention, this task is independent of calacs
. As the effect of the stripes on most science programs will be insignificant, STScI does not expect to add stripe removal to standard calacs
processing in the OPUS1
The default CEB-R mode induces a signal-dependent bias shift whose cause is
related to that of the bias gradient. The DC level of the dual-slope integrator is sensitive to changes in the CCD output voltage in such a way that the pixel bias level is shifted positively by 0.02%
0.30% (depending on the amplifier) of the signal from the previously integrated pixel. This phenomenon is well understood and can be cleanly removed. It is anticipated that a correction for this effect will be incorporated into the OPUS pipeline in 2011.