The average full well depths for the ACS CCDs are given in Table 3.1 as
for the WFC and 155,000 e−
for the HRC, but the pixel-to-pixel values vary by ~10% and ~18% across the fields of view of the WFC and the HRC, respectively. When the CCD is over-exposed, pixels will saturate and excess charge will flow into adjacent pixels along the column. This condition is known as “blooming” or “bleeding.” Extreme overexposure does not cause any long-term damage to the CCDs, so there are no bright object limits for the ACS CCDs. When using ADC gains of 2 for the WFC and 4 for the HRC, the linearity of the detectors deviates by less than 1% up to the full well depths. On-orbit tests using aperture photometry show that this linearity is maintained when summing over pixels that surround an area affected by charge bleeding, even when the central pixel is saturated by 10 times its full well depth (see ACS ISR 2004-01
The ACS camera has a very fast shutter; even the shortest exposure
times are not significantly affected by the finite travel time of the shutter blades. On-orbit testing reported in ACS ISR 2003-03
verified that shutter shading corrections are not necessary to support 1% relative photometry for either the HRC or WFC.
Four exposure times are known to have errors of up to 4.1%; e.g., the
nominal 0.1 seconds HRC exposure was really 0.1041 seconds. These errant exposure times are accommodated by updates to the reference files used in image pipeline processing. No significant differences exist between exposure times controlled by the two shutters (A and B), with the possible exception of non-repeatability up to ~1% for WFC exposures in the 0.7 to 2.0 second range. The HRC provided excellent shutter time repeatability.
The read noise levels in the physical overscan and imaging regions of
the four WFC amplifiers were measured for all gain settings during the orbital verification period following the installation of the CEB-R in SM4. The average WFC read noise is about 25% lower than before SM4. This reduction is due to the use in the CEB-R of a dual-slope integrator (DSI) in the pixel signal processing chain instead of the clamp-and-sample method of pixel sampling used in the original CEB. The DSI method offers lower signal processing noise at the expense of a spatially variable, but temporally stable, bias level.
Prior to failure of the original CEB, the WFC read noise had been
generally constant with time. After a transit through the South Atlantic Anomaly (SAA) on 29 June 2003, the read noise of amplifier A changed from ~4.9 to ~5.9 e−
rms. Although the telemetry did not show any anomaly in any WFC components, it is likely that the sudden increase was caused by radiation damage. Amplifier A was the only amplifier that showed this anomaly. The amplitude of the variation (~1 e−
) was the same for ADC gains 1 and 2. After the following anneal (see Section 4.3.5
), the read noise dropped to ~5.5 e−
and remained constant for 27 days. After the following two anneal cycles, the read noise stabilized at ~5.6 e−
. Figure 4.4
shows the read noise in the image area for amplifier A during the instability period. The read noise of the other amplifiers was very stable between the installation of ACS in March 2002 and the failure of ACS in January 2007.
Although amplifier A has higher read noise than the other amplifiers,
most WFC broadband science observations are sky limited. Narrowband observations are primarily read noise limited.
Prior to January 2007, the read noise of the HRC was monitored using
only the default amplifier C and the default gain of 2 e−
/DN. No variations were observed with time. The read noise measured in the physical overscan and image areas were consistent with the pre-flight values of 4.74 e−
(see Table 4.2
and Table 4.3
All ACS CCDs are buried channel devices which have a shallow n-type
layer implanted below the surface to store and transfer the collected signal charge away from the traps associated with the Si-SiO2
interface. Moreover, ACS CCDs are operated in Multi-pinned Phases (MPP) mode so that the silicon surface is inverted and the surface dark current is suppressed. ACS CCDs therefore have very low dark current. The WFC CCDs are operated in MPP mode only during integration, so the total dark current figure for WFC includes a small component of surface dark current accumulated during the readout time.
Like all CCDs operated in a low earth orbit radiation environment, the
ACS CCDs are subject to radiation damage by energetic particles trapped in the radiation belts. Ionization damage and displacement damage are two types of damage caused by protons in silicon. The MPP mode is very effective in mitigating the damage due to ionization such as the generation of surface dark current due to the creation of trapping states in the Si-SiO2
interface. Although only a minor fraction of the total energy is lost by a proton via non-ionizing energy loss, the displacement damage can cause significant performance degradation in CCDs by decreasing the charge transfer efficiency (CTE), increasing the average dark current, and introducing pixels with very high dark current (hot pixels). Displacement damage to the silicon lattice occurs mostly due to the interaction between low energy protons and silicon atoms. The generation of phosphorous-vacancy centers introduces an extra level of energy between the conduction band and the valence band of the silicon. New energetic levels in the silicon bandgap have the direct effect of increasing the dark current as a result of carrier generation in the bulk depletion region of the pixel. As a consequence, the dark current of CCDs operated in a radiative environment is predicted to increase with time.
Ground testing of the WFC CCDs, radiated with a cumulative fluence
equivalent to 2.5 and 5 years of on-orbit exposure, predicted a linear growth of ~1.5 e−
The dark current in ACS CCDs is monitored three days per week with
the acquisition of four 1000 seconds dark frames (totaling 12 images per week). Dark frames are used to create reference files for the calibration of scientific images, and to track and catalog hot pixels as they evolve. The three daily frames are combined together to remove cosmic rays and to extract hot pixel information for any specific day. The dark reference files are generated by combining two weeks of daily darks in order to reduce the statistical noise. The hot pixel information for a specific day is then added to the combined bi-weekly dark. In order to study the evolution of the dark current with time, the modal dark current value in the cosmic-ray free daily darks is calculated. As expected, the dark current increases with time. The observed linear growth rates of dark current are 2.1 and 1.6 e−
/pixel/hour/year for WFC1 and WFC2 respectively, and 2.1 e−
/pixel/hour/year for the HRC CCD. These rates are in general agreement with the ground test predictions.
At the beginning of the side-2 operation in July 2006 the temperature set
point of the WFC was lowered from -77 °
C to -81 °
C (documented in ACS TIR 2006-021
). Following SM4, we measure a dark current of 20 - 25 e−
/pixel/hour among the four amplifiers. We have reverted to 12-hour anneals. We expect to see the earlier dark current rates of 2.1 and 1.6 e−
/pixel/hour/year for WFC1 and WFC2, respectively, but it is too early to confirm this.
Dark current and hot pixels depend strongly on the operating
temperature. The reduction of the operating temperature of the WFC CCDs reduced the number of hot pixels by almost 50% (See Figure 4.5
and Figure 4.5
). The dark rate shows a clear drop on July 4, 2006, when the temperature was changed. Figure 4.6
illustrates the before-and-after effect directly. The new operating temperature brought the dark current of the WFC CCDs back to the level eighteen months after the launch.
The figures above show hot pixel growth rates (DQ flag 16) in the WFC, and in the HRC during
its entire mission. The sawtooth patterns correspond to ACS anneal cycles. In the HRC, the growth rate increased slightly when the anneal duration was switched from 12 hours to 6 hours. The slight drop in hot pixels coincided with the switch to Side 2 electronics. In the lower figure for WFC, the sawtooth patterns also correspond to ACS anneal cycles. The growth rate slightly increased when the anneal duration was switched from 12 hours to 6 hours. Post-SM4 anneals are once again 12 hours long. The temperature change from -77 C to -81 C resulted in a significant drop in hot pixels.
In the presence of a high electric field, the dark current of a single pixel
can be greatly enhanced. Such pixels are called hot pixels. Although the increase in the mean dark current with proton irradiation is important, of greater consequence is the large increase in dark current nonuniformity.
We have chosen to classify the field-enhanced pixels into two
categories: warm and hot pixels. The definition of “warm” and “hot” pixel is somewhat arbitrary. We have chosen a limit of 0.08 e–
/pixel/seconds as a threshold above which we consider a pixel to be “hot”. We identify “warm” pixels as those which exceed by about 5 σ (was 0.02 e–
/pixel/second but after SM4 it is 0.04 e–
/pixel/second) the normal distribution of the pixels in a dark frame up to the threshold of the hot pixels (See Figure 4.6
) for a typical dark rate pixel distribution.
Warm and hot pixels accumulate as a function of time on orbit. Defects
responsible for elevated dark rate are created continuously as a result of the ongoing displacement damage on orbit. The number of new pixels with a dark current higher than the mean dark date increases every day by few to several hundreds depending on the threshold. The reduction of the operating temperature of the WFC CCDs has dramatically reduced the dark current of the hot pixels and therefore many pixels previously classified as hot are now warm or normal pixels.
Most of these new hot pixels are transient. Like others CCDs on HST
, the ACS devices undergo a monthly annealing process. (The CCDs and the thermal electric coolers are turned off and the heaters are turned on to warm the CCDs to ~19 °
C.) Although the annealing mechanism at such low temperatures is not yet understood, after this “thermal cycle” the population of hot pixels is greatly reduced (see Figure 4.5
). The anneal rate depends on the dark current rate; very hot pixels are annealed more easily than warmer pixels. For pixels classified as “hot” (those with dark rate > 0.08 e–
/pix/sec) the anneal rate is ~82% for WFC and ~86% for HRC.
Annealing has no effect on the normal pixels that are responsible for the
increase in the mean dark current rate. Such behavior was also seen with STIS and WFC3 CCDs during ground radiation testing. Since the anneals cycle do not repair 100% of the hot pixels, there is a growing population of permanent hot pixels (see Figure 4.7
and Figure 4.6
In principle, warm and hot pixels could be eliminated by the superdark
subtraction. However, some pixels show a dark current that is not stable with time but switches between well defined levels. These fluctuations may have timescales of a few minutes and have the characteristics of random telegraph signal (RTS) noise. The dark current in field-enhanced hot pixels can be dependent on the signal level, so the noise is much higher than the normal shot noise. As a consequence, since the locations of warm and hot pixels are known from dark frames, they are flagged in the data quality array. The hot pixels can be discarded during image combination if multiple exposures have been dithered.
The standard CR-SPLIT approach allows rejection of cosmic rays, but
hot pixels cannot be eliminated in post-observation processing without dithering between exposures.
For example, a simple ACS-WFC-DITHER-LINE
pattern has been developed that shifts the image by 2 pixels in X and 2 pixels in Y along the direction that minimizes the effects of scale variation across the detector. The specific parameter values for this pattern are given on the ACS dithering Web page at:
Given the transient nature of hot pixels, users are reminded that a few
hot pixels may not be properly flagged in the data quality array (because they spontaneously “healed” or because their status changed in the period spanning the reference file and science frame acquisition), and therefore could create false positive detections in some science programs.
Studies have been made of the characteristics of cosmic ray impacts on
the HRC and WFC. The fraction of pixels affected by cosmic rays varies from 1.5% to 3% during a 1000 second exposure for both cameras, similar to what was seen on WFPC2 and STIS. This number provides the basis for assessing the risk that the target(s) in any set of exposures will be compromised. The affected fraction is the same for the WFC and HRC despite their factor of two difference in pixel areas because the census of affected pixels is dominated by charge diffusion, not direct impacts. Observers seeking rare or serendipitous objects, as well as transients, may require that every single WFC pixel in at least one exposure among a set of exposures is free from cosmic ray impacts. For the cosmic ray fractions of 1.5% to 3% in 1000 seconds, a single ~2400 second orbit must be broken into 4 exposures of 500 to 600 seconds each to reduce the number of uncleanable pixels to 1 or less. Users seeking higher S/N (lower read noise) may prefer the trade-off of doing 3 exposures of ~800 seconds, where CR-rejection should still be very good for most purposes. But we do NOT recommend 2 long exposures (i.e. 1200 seconds), where residual CR contamination would be unacceptably high in most cases. We recommend that users dither these exposures to remove hot pixels as well as cosmic rays (see Section 7.4
The flux deposited on the CCD from an individual cosmic ray does not
depend on the energy of the cosmic ray but rather the distance it travels in the silicon substrate. The electron deposition due to individual cosmic rays has a well defined cut-off below 500 e–
and a median of ~1000 e–
(see Figure 4.8
and Figure 4.9
The distribution of the number of pixels affected by a single cosmic ray
is strongly peaked at 4 to 5 pixels. Although a few events are seen which encompass only one pixel, examination of these events indicate that at least some, and maybe all of these sources are actually transient hot pixels or unstable pixels which can appear hot in one exposure (with no charge diffusion) and normal in the next. Such pixels are very rare but do exist. There is a long tail in the direction towards increasing numbers of attached pixels.
Distributions of sizes and anisotropies can be useful for distinguishing
cosmic rays from astrophysical sources in a single image. The size distribution for both chips peaks near 0.4 pixels as a standard deviation (or 0.9 pixels as a FWHM). This is much narrower than for a PSF and is thus a useful discriminant between unresolved sources and cosmic rays.
Charge transfer efficiency (CTE) is a measure of how effectively the
CCD moves charge between adjacent pixels during read out. A perfect CCD would be able to transfer 100% of the charge as the charge is shunted across the CCD and out through the serial register and its CTE would be unity. In practice, small traps in the silicon lattice compromise this process by holding on to electrons, releasing them at a later time. Depending on the trap type, the release time ranges from a few microseconds to several seconds. For large charge packets (several thousands of electrons), losing a few electrons along the way is not a serious problem, but for smaller (~100e–
or less) signals, it can have a substantial effect.
The CTE numbers for the ACS CCDs at the time of installation are
given in Table 4.6
. While the numbers look impressive, remember that reading out the WFC CCD requires 2048 parallel and 2048 serial transfers, so that almost 2% of the charge from a pixel in the corner opposite the readout amplifier was lost.
Like other CCD cameras aboard HST
, ACS has suffered degraded CTE due to radiation damage since its installation in 2002. Since 2003, specific calibration programs aimed at characterizing the effects of the CTE on stellar photometry were performed (Riess, ACS ISR 2003-09
). First characterizations of the effects of a decreasing CTE for WFC were made by Riess & Mack (ACS ISR 2004-06
). Meanwhile, results from the internal CTE calibration programs showed that CTE appears to decline linearly with time, and that CTE losses are stronger for lower signal levels (Mutchler & Sirianni, ACS ISR 2005-03
). However, in order to derive an accurate correction formula for photometry, a number of observations at different epochs and for different levels of sky background and stellar fluxes have to be accumulated. The results of the observations taken through March 2006 are described by Chiaberge et al. (ACS ISR 2009-01
) and are summarized here. For WFC, significant photometric losses are apparent for stars undergoing numerous parallel transfers (y-direction). Extrapolated to 2011, the losses should be 5-10% for typical observing parameters, rising to ~50% in worst cases (faint stars, low background). The size of the photometric loss appears to have a strong power-law dependence on the stellar flux, as seen for other CCDs flown on HST
The dependence on background is significant, but for faint targets there
is little advantage to increasing the background intentionally (e.g., by post-flashing) due to the added shot noise. However, in some specific cases, it may instead be useful to choose the filter in order to obtain a higher background (e.g. F606W). CTE degradation also has an impact on astrometry (see ACS ISR 2007-04
). Therefore, for astrometric programs of relatively bright objects, the use of post-flash may be considered. No losses are apparent for WFC due to serial transfers (x-direction). Correction formulae are presented in ACS ISR 2009-01
to correct photometric losses as a function of a source’s position, flux, background, and time. Using data obtained both as part of SM4-SMOV and Cycle 17 External CTE Calibration Program (Sept. 2009), we have verified that the predictions of the photometric correction formula are still accurate within the errors. Users are encouraged to check the ACS Web
page for updates.
shows the predicted photometric losses for the WFC due to imperfect parallel CTE as a function of time. These curves are based on the formulae published in ACS ISR 2009-01
. We consider three “typical” science applications:
|A “worst case scenario” of a 20th
mag (Vega mag) star observed with the F502N narrow band filter for a short exposure time (30s), thus giving rise to a very low sky background level;
|a PSF with a flux of 1
/s observed for 1000s with a filter that gives rise to a sky background level of 80 e–
with such an exposure time (e.g. F775W in our specific case).
Predicted impact of CTE on science images for WFC (the three scenarios described above). The
magnitude loss is estimated for a star located at the chip middle point along Y=1024 and refers to counts measured in a 3-pixel aperture. The light gray lines indicate 1-sigma error on the value obtained with the correction formula. The two vertical dashed lines refer to the epoch of the ACS failure in January 2007 and to a time shortly after the SM4 repair in May 2009.
Anderson and Bedin (2010 PASP 122, 1035) have developed an
empirical approach based on the profiles of warm pixels to characterize the effects of CTE losses for WFC. Their algorithm first develops a model that reproduces the observed trails and then inverts the model to convert the observed pixel values in an image into an estimate of the original pixel values. This algorithm will be available as a stand-alone tool by late 2010. Please check the ACS Web
site for the latest information. The ACS-WFPC2 team is also working on implementing the CTE de-trailing software in the WFC calibration pipeline in the near future.
It is important to note that the results of the Anderson & Bedin
correction on stellar fields are found to be in agreement with the photometric correction formula of Chiaberge et al. (ACS ISR 2009-01
). Statistically significant deviations are observed only at low stellar fluxes (~300e–
) and for background levels close to 0e–
. Further testing and improvement of the pixel-based CTE correction code is currently under way at STScI.
In the optical, each photon generates a single electron. However, in the
near-UV, shortward of ~3200 Å there is a finite probability of creating more than one electron per UV photon (see Christensen, O., J. App. Phys
,689, 1976). At room temperature the theoretical quantum yield (i.e., the number of electrons generated for a photon of energy E > 3.5eV (λ
~3500 Å)), is Ne = E(eV)/3.65. The HRC CCDs quantum efficiency curve has not been corrected for this effect. The interested reader may wish to see the STIS Instrument Handbook
for details on the signal-to-noise treatment for the STIS CCDs.