5.2	Important Considerations for ACS Imaging

5.2.1 Optical Performance

5.2.2 CCD Throughput Comparison

5.2.3 Limiting Magnitudes

5.2.4 Signal-To-Noise Ratios

5.2.5 Saturation

5.2.6 Faint Horizontal Striping in WFC CCDs

The following characteristics of ACS should be considered when planning ACS observations or archival research:

•	The ACS/WFC filter set is more limited than that of WFC3/UVIS. In particular, WFC3/UVIS has an extensive set of ultraviolet, narrow-band, and Strömgren filters unavailable with ACS/WFC. On the other hand, ACS/WFC has polarizers and ramp filters unavailable with WFC3/UVIS.

•	For sources subtending less than 5 arcsec, the WFC1-CTE aperture (10 arcsec from readout amplifier) will ensure minimal CTE losses.

•	ACS ISR 2012-04 contains plots of likely minimum sky backgrounds as a function of exposure time. Observers should determine if their sky backgrounds are likely to be less than 20 electrons. If this is the case, they should consult the ACS Web pages for details of the post-flash capability.

•	SBC observations of red targets should generally be made with two different long pass filters (e.g. F140LP and F165LP), so that the difference in the count rates can be used to isolate the true FUV flux in the presence of filter red-leak contamination.

•

The ACS cameras are intended to be used with a single filter. Unfiltered or two-filter imaging yields significantly degraded PSFs (particularly for the WFC), so these modes are typically used only for polarization observations or HRC coronagraphic acquisitions. The polarizers have zero optical thickness, so they can and should be used with another filter.

•	The geometric distortion of the WFC is significant and causes the projected pixel area to vary by ±9% over the field of view. This distortion affects both the photometric accuracy and the astrometric precision, and must be accounted for when the required accuracy is better than 10%.

•	The cosmic ray fluxes of HRC and WFC are comparable to those of STIS and WFPC2, respectively. Typical observations should be dithered for cosmic ray rejection. See Section 4.3.5 for more information about dithering strategies for removing cosmic rays and hot pixels.

•

At wavelengths longer than ~ 8000 Å, internal scattering in the HRC CCD produced an extended PSF halo. Only a small number of observations were affected because WFC was mostly used at these wavelengths. The WFC CCDs were treated with a front-side metallization that eliminates the large angle, long wavelength halo problem for wavelengths less than 9000 Å. For observations of red targets with the F850LP refer to Section 9.3.2.

•	The WFC and HRC share two filter wheels. Consequently, when the cameras were used simultaneously, only the filter for the primary camera was selectable.

•	The ratios of in- and out-of-band transmission for HRC UV filters are similar to those of WFPC2, after accounting for differences in the camera’s QE. The ACS F220W, F250W, and F330W filters have small red leaks, which are documented in ACS ISR 2007-03.

5.2.1 Optical Performance

Following the fine-alignment and focus activities of the SM3B Orbital Verification period, the optical qualities of all three ACS channels were judged to have met their design specifications. The encircled energy values for the WFC, HRC, and SBC obtained are given in Table 5.4. Refer to ACS ISR 2011-02 (Bohlin, R. 2011) for more details.

Table 5.4: Encircled energy measurements for the ACS channels.

Channel	Encircled energy
	Center of field	Edge of field
WFC at 632.8 nm in 0.25 arcseconds diameter	80.0%	79.4%
HRC at 632.8 nm in 0.25 arcseconds diameter	81.8%	81.6%
SBC at 121.6 nm in 0.10 arcseconds diameter	28%	---

5.2.2 CCD Throughput Comparison

Figure 5.7 compares the wavelength-dependent throughputs of the ACS WFC and HRC with those of WFC3/UVIS, WFC3/IR, NICMOS/NIC3, and WFPC2.

5.2.3 Limiting Magnitudes

Table 5.5 contains the detection limits in Johnson-Cousins V magnitudes for unreddened O5 V, A0 V, and G2 V stars, generated using the ETC. WFC and HRC values used the parameters CR-SPLIT=2, GAIN=2, and a 0.2 arcsecond circular aperture. For the SBC, a 0.5 arcsecond circular aperture was used. An average sky background was used in these examples. However, limiting magnitudes are sensitive to the background levels; for instance, the magnitude of an A0 V in the WFC using the F606W filter changes by ±0.4 magnitudes at the background extremes. Figure 5.8 shows a comparison of the limiting magnitude for point-sources achieved by the different cameras with a signal to noise of 5 in a 10 hour exposure. Figure 5.9 shows a comparison of the time needed for extended sources to attain ABMAG=26.

Figure 5.7: HST total system throughputs as a function of wavelength. The plotted quantities are end-to-end throughputs, including filter transmissions calculated at the pivot wavelength of each broad-band filter.

Figure 5.8: HST Limiting Magnitude for point sources in 10 hours, as a function of wavelength. Point source limiting magnitude achieved with a signal to noise of 5 in a 10 hour long exposure with optimal extraction.

Figure 5.9: HST Limiting Magnitude for extended sources in 10 hours, as a function of wavelength.

Table 5.5: V detection limits for ACS, HRC, and SBC direct imaging.

Camera	Filter	V limit (S/N = 5, exposure time = 1 hour)
		O5 V (Kurucz model)	A0 V (Vega)	G2 V (Sun)
WFC	F606W	27.8	27.8	28.0
WFC	F814W	26.7	27.0	27.7
HRC	F330W	26.8	24.8	24.1
HRC	F606W	27.3	27.3	27.5
SBC	F125LP	27.8	23.2	13.5

5.2.4 Signal-To-Noise Ratios

Chapter 10 contains plots of exposure time versus magnitude for a desired signal-to-noise ratio. These plots are useful for determining the exposure times needed for your scientific objectives. More accurate estimates require the use of the ACS ETC (http://etc.stsci.edu/).

5.2.5 Saturation

Both CCD and SBC imaging observations are subject to saturation at high total accumulated counts per pixel. For the CCDs, this is due either to the depth of the full well or to the 16 bit data format. For the SBC, this is due to the 16 bit format of the buffer memory (see Section 4.3.1 and Section 4.5.2).

5.2.6 Faint Horizontal Striping in WFC CCDs

Subsequent to the replacement of the ACS CCD Electronics Box during SM4, all WFC images show horizontal striping noise that is roughly constant across each row of read-out in all four WFC amplifiers. This striping is the result of a 1/f noise on the bias reference voltage, and has an approximately Gaussian amplitude distribution with standard deviation of 0.9 electrons. The contribution of the stripes to the global read noise statistics is small, but the correlated nature of the noise may affect photometric precision for very faint sources and very low surface brightnesses.

During Cycle 17, STScI developed and tested an algorithm for removing the stripes from WFC science images. The algorithm is effective when the science image is not excessively crowded such that a row-by-row background level becomes difficult to estimate. Because the stripe removal code is not universally effective, it is not currently applied as part of the ACS calibration pipeline. Instead, STScI has released the stripe removal algorithm to the community as a stand-alone task that can be run on ACS data retrieved from the HST archive. This task, acs_destripe, has been written in Python as part of the acstools package in the public release of STScI_Python. As a Python task, it can be run from PyRAF, any Python interpreter or even the operating system command-line, to correct post-SM4 pipeline-calibrated images (_flt.fits). Please see the ACS Web site for details on running this code. Further details regarding the WFC striping and its mitigation are provided in the ACS ISR 2011-05 (Grogin et al. 2011).

Because the WFC bias striping noise is so consistent among the four read-out amplifiers, and because it also manifests within the WFC pre-scan regions, STScI has incorporated a pre-scan based de-striping algorithm into the ACS calibration pipeline CALACS. This permits consistent striping-noise mitigation for all post-SM4 WFC full-frame images, including calibration images as well as arbitrary science images, given the trade-off of slightly less precise striping-noise reduction for “low-complexity” science images. This pre-scan based de-striping algorithm, as well as its implementation in CALACS, will be described in an upcoming ACS Instrument Science Report (Anderson & Grogin, in prep.). For correcting WFC sub-array images, it is still necessary to use the acs_destripe software described above.