WFC3: Understanding and mitigating
UVIS charge transfer efficiency losses
and IR persistence effects
149.03
SPACE
TELESCOPE
SCIENCE
INSTITUTE
Operated for NASA by AURA
S. Baggett, J. Anderson, K. S. Long, J. W. MacKenty, K. Noeske, J. Biretta, and the WFC3 team (STScI)
CTE in the UVIS
 HST’s low-earth orbit environment damages CCDs, generating
new hot pixels, increasing dark current, & decreasing CTE.
 The decline in CTE reduces detected source flux as defects trap
and release charge during readout. Flux losses depend on:
1) Distance from amplifier: more transfers=more traps
2) Signal level: higher fractional loss in faint sources
3) Image background: higher background fills traps
4) Observing scene: sources preceding target source can
fill traps and improve charge transfer
 Flux losses range up to ~50% for faint sources on faint
backgrounds, ~5-15% for more typical sources/backgrounds.
Ensure sufficient background
WFC3/UVIS CTE losses especially for faint targets can be
significantly reduced by ensuring images contain a minimum
of 12 e- total background (dark + sky + post-flash if needed).
Summary of mitigation strategies
 Place small targets close to readout amp.
 Ensure sufficient image background (e.g. broader filters,
longer exposure times and/or post-flash).
 Apply formulaic corrections to aperture photometry results.
 Correct images using pixel-based image correction.
 Keep exposure levels significantly below saturation.
 Dither whenever possible.
 Avoid dithering bright objects across regions on the detector
where targets of interest will fall.
 Group exposures with saturated sources at the end of the visit.
 Check if persistence is an issue using MAST.
 Evaluate archival persistence products: use them to mask
areas of persistence or (with caveats) analyze the persistencecorrected products directly.
Persistence in the IR
Fig 5. Traps in the IR detector
cause after-images (left panel)
due to bright objects in prior
data. Exposure to light changes
voltage levels, trapping charge
which is released during later
image readouts (Smith et al. 2008). Right panel shows the same
image after a persistence correction (see discussion below).
Fig 6. Amplitude and duration
depend on exposure levels of,
and time elapsed since, the
preceding images. The effect is
low for signals < ~half saturation
but reaches 0.3-0.5 e-/s ~1000s
after 2x–20x saturation.
Full-well
(~75ke-)
‘truth’ image
No background
16e- background
Fig 1. Traps capture and
release charge during readout,
generating trails above the
sources (middle panel). A
modest amount of background
can reduce CTE losses.
Perform pixel-based correction
Hot pixels in dark frames are used to empirically model CTE
losses on a pixel-by-pixel basis (Massey 2010; Anderson &
Bedin, 2010). The model is applied to science images to
restore charge to its original location within the image.
Fig 2. Measured flux losses
as a function of background
level, faint and bright
sources (top, bottom
panels), and distance from
the amplifier (X axis). A
higher image background
reduces CTE losses.
Fig 4. WFC3 image
subsections farthest from the
amplifier before (top) and
after (bottom) CTE
correction. Further
refinements of the model are
in progress. Plans are to
incorporate the code into the
calibration pipeline in 2014.
Apply formulaic corrections
A standalone version of the pixel-based CTE correction
software is available for download from the WFC3 CTE page.
Fit coefficients can be used
to estimate CTE losses or
correct photometric results.
The CTE correction works well but the nature of the
algorithm is such that it can not completely recover
what was lost, particularly at the faintest levels. To
avoid amplification of read noise, the algorithm is
conservative in its reconstruction at the low
background levels where losses are non-linear.
Fig 3. CTE losses as a function
of observation date and source flux
fit with 2-parameter polynomials
(lines). Image backgrounds are
worst-case: 1-2 e-/pixel.
Abstract
A panchromatic instrument, Wide Field Camera 3 (WFC3) contains a UVIS channel with a 4096 x 4096 pixel
e2v CCD array as well as an IR channel with a 1014x1014 Rockwell Scientific HgCdTe focal plane array
(FPA). Both detectors have been performing well on-orbit since the installation of the instrument in the
Hubble Space Telescope (HST) in May 2009. However, as expected, the harsh low-earth orbit environment
has been damaging the UVIS CCDs, resulting in a progressive loss of charge transfer efficiency (CTE) over
time. We summarize the magnitude of the CTE losses, the effect on science data, and the pre- and postobservation mitigation options available. The IR FPA does not suffer from accumulating radiation damage but
it does exhibit persistence i.e. an after-glow from sources in previous exposures, an anomaly commonly seen
in these types of IR arrays. We summarize the characteristics of persistence in WFC3, suggest methods for
reducing the effects during observation planning, and describe the calibration products which are available
via the Mikulski Archive for Space Telescopes (MAST) for addressing persistence in IR science data.
The decay time follows a power
law (slope ~ -0.9). Repeat
exposures (no dithering) show
stronger and longer-lived effects.
Check persistence products
Fig 7a. The MAST History page
can be used to determine whether
persistence is a problem.
Make use of the
persistence output files.
Fig 7b. An estimate of persistence
for every science image is in
MAST . Results are based on a
model of persistence as a function
of time since, and exposure level of, the brightest preceding
exposure (Long et al., 2012). Products include a list of prior
exposures, the number of saturated pixels they contain, a map of
the persistence, and a persistence-corrected flt file.
The model may not work well in all situations.
Rescaling the persistence map and subtracting it
from the pipeline flt file can improve the results.
For assistance please email help@stsci.edu.
References and further information
\
 WFC3: www.stsci.edu/hst/wfc3
STScI general help desk: help@stsci.edu
 UVIS CTE: www.stsci.edu/hst/wfc3/ins_performance/CTE
 IR persistence: http://www.stsci.edu/hst/wfc3/ins_performance/persistence/
 Anderson, J., & Bedin, L., An Empirical Pixel-Based Correction for HST/ACS, PASP 122, 1035, 2010.
 Baggett, S., & Anderson, “WFC3/UVIS Sky Backgrounds”, WFC3 ISR 2012-12.
 Long et al., “Characterizing persistence in the IR detector”, Proc.SPIE, 8442, 84421W-9, 2012.
 Noeske, K., et al., “WFC3/UVIS CTE Monitor Results”, WFC3 ISR 2012-09.
 Massey, R., et al., “Pixel-based correction for Charge Transfer Inefficiency in HST/ACS”, MNRAS 401, 2010.
 Smith, R., et al., Proc. SPIE Vol 7021, 7021J, 2008.