The ACS detectors exhibit more distortion than previous HST
instruments. The principal reason is that the optics have been designed with a minimum number of components, consistent with correcting for the spherical aberration induced by the OTA, without introducing coma. The result is a high throughput, but with focal surfaces far from normal to the principal rays. The WFC detector is tilted at 22°
giving an elongation of 8% while the HRC and SBC have a 25°
tilt leading to an elongation of 12%. In each case, the scales in arcseconds per pixel are smaller along the radial direction of the OTA (Optical Telescope Assembly) field of view than along the tangential direction.
There is not only a strong geometric distortion of ACS detectors but a significant variation of the scale across each detector. For the WFC, the scale changes by approximately 10% from corner to corner. For the HRC and SBC this variation is only about 1% as they cover much smaller fields of view. The Science Instrument Aperture File (SIAF)
provides the most accurate values of the scale for all instruments on board HST
. The area on the sky covered by a WFC pixel varies by about 18% from corner to corner, corrections for which must be made in photometry of extended objects. This variation of scale creates a problematic effect in combining ACS images 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. Therefore, image alignment and combining require an accurate geometric distortion model.
The HRC geometric distortion calibration is fully described in ACS ISR 2004-15
. The SBC geometric distortion is described in ACS ISR 2008-02
. The WFC geometric distortion is described in ACS ISR 2015-06
. The geometric distortion model is expressed by high-order polynomials for each of the ACS detectors in the form of reference files called as Instrument Distortion Coefficients Tables (IDCTAB) in the FITS format. On the top of the geometric distortion described by the polynomial equations, there are additional filter-dependent components of the distortion that are presented by a 2-D array and are applied to each row/columns of the WFC or HRC images. This filter dependent component of the distortion is applied via a calibration file called NPOLFILE. In the case of the WFC, there is an additional component of the distortion called pixel-grid irregularities. These are caused by irregularities in the manufacturing process and are also presented by 2-D array to correct each row/column on WFC images. This correction is applied via the FITS formatted D2IMFILE. All distortions reference files are installed in the ACS on-the-fly recalibration (OTFR) pipeline and can be used with the Drizzlepac software distributed by STScI.
The rhombus shape of the WFC is evident in Figure 7.9
. The angle between the X and Y axes is 84.9°
for WFC1 and 86.1°
for WFC2. The geometric distortion map for WFC1 and WFC2 is illustrated in Figure 10.127
. A vector diagram shows the contribution of the non-linear part of a quadratic fit only. At the center of chip WFC1, the scale in the X direction is 0.0493 arcseconds per pixel, and 0.0486 arcseconds per pixel in the Y direction. In the case of WFC2, the scale is 0.0498 arcseconds per pixel in the X direction, and 0.0503 arcseconds per pixel in the Y direction. Between the corner of WFC nearest to the V1 axis and the diagonally opposite corner, the scale increases by 10%. Therefore, WFC1 forms a slightly distorted rectangle 201 by 100 arcseconds in size, while WFC2 is 203 by 103 arcseconds. There is a 2.5 arcsecond gap between the two chips. See the SIAF website
for more details.
ACS ISR 2007-08
shows that the linear terms of the WFC geometric distortion are changing over time and distortion corrected positions could be off by up to 0.3 pixels when compared to the beginning of the instrument's lifetime. The scale of the x-axis changed after SM4 (ACS ISR 2015-06
), but the reason for the change is not yet understood. The time dependent distortion has been calibrated for pre and post-SM4 (ACS ISR 2015-06
) and its parameters are implemented via the IDCTAB, which is used by both the OTFR and the Drizzlepac software.
The High Resolution Channel has its edges aligned approximately along the V2 and V3 axes. In this case, the center of the aperture lies on a line passing through the V2-V3 origin and making an angle of 22°
with the V3 axis. The diagonal of the aperture does not correspond to a radius of the HST
field of view, so the distortion has no particular symmetry with respect to the detector axes. The focal plane of HRC is also 25°
away from the plane normal to the light path, hence the scales along the axes differ by 14%. The full field of view of the HRC is less than 30", therefore the scale variation over the field is much less than for the WFC, about 1%. At the center, the X and Y scales are 0".0284 and 0".0248 per pixel respectively. The average scales across the middle of the detector are 0".02842 and 0".02485 per pixel, making the X and Y widths 29".1 and 25".4 arcseconds. The slightly non-square projected aperture shape is evident in Figure 7.8
. The angle between the X and Y axes on the sky is 84.2°
. The geometric distortion map for HRC is given in Figure 10.129
, where the residuals from the non-linearity are scaled by a factor of 10, with residual as large as 0".14 or 4.9 HRC pixels.
The map of the effective pixel areas of the SBC is shown in Figure 10.132
. Because the pixel area map is normalized to square pixels which are 0.025 arcseconds on a side rather than an area equal to that of the central SBC pixel, the pixel area map ranges from about 1.58 to 1.65. (Meurer et al. 2002
). The maximum deviation from the central value is about 5% (ACS ISR 2007-09
and ACS ISR 2008-02