WFC3 Space Telescope Analysis Newsletter - Issue 21, July 2015
- 1 Cycle 23 Phase II Deadline Reminder and Updates to APT
- 2 Source-Dependent Master Sky Images for the WFC3/IR Grisms
- 3 UVIS Read Noise Monitoring
- 4 Charge Transfer Efficiency (CTE) in WFC3/UVIS
- 5 Impact of "Blobs" on WFC3/IR Stellar Photometry
- 6 New Documentation
Cycle 23 Phase II Deadline Reminder and Updates to APT
L. Dressel, K. Peterson
Observers who have been awarded time on HST for Cycle 23 are reminded that the submission deadline for Phase II proposals is July 23, 2014. For more information regarding Phase II development and submission, please visit the Resources for HST Phase II Proposal Development page. Note that phase II proposals must be submitted through the latest version of the Astronomers Proposal Tool (APT 23.2). This version of APT includes the following updates specific to WFC3, as listed in What's New for Cycle 23 Phase II:
- • New WFC3/UVIS apertures - The apertures UVIS2-C512C-CTE and UVIS2-C1K1C-CTE enable observers to automatically place a small target near a readout amplifier to reduce CTE losses. See the article in WFC3 STAN Issue 20.
- • New diagnostic message about FLASH for WFC3/UVIS - A message is generated for exposures that may use too low or high a value of FLASH to mitigate CTE losses for a point source without needlessly increasing the noise. (Discussed below.)
• New sample sequence for WFC3/IR - The sample sequence SPARS5, intermediate between RAPID and SPARS10, is intended to allow the efficiency and uninterrupted time series of SPARS10 and yet be short enough in cadence to better isolate an exoplanet-host star from a nearby stellar companion in spatially-scanned observations using an IR grism. See the
article in WFC3 STAN Issue 20.
The diagnostic message about FLASH has been added to APT to remind observers when it may be needed and to provide a guide to the recommended level. The phenomenon of charge trailing and loss due to declining charge transfer efficiency (CTE) is discussed and illustrated in
Section 6.9 of the WFC3 Instrument Handbook. CTE losses are worse when the background level in the image is low, when the target is faint, and when the target is far from a readout amplifier. The background level can be increased above the level provided by the sky and dark current by briefly flashing the detector with light using the exposure level optional parameter FLASH in APT (see
Section 6.2 in the WFC3 Instrument Handbook). The recommended background level for the efficient transfer of point sources over the full height of one of the two UVIS CCD chips is 12 electrons per pixel. Higher background levels increase the noise while providing diminishing returns in CTE loss mitigation. APT will provide a diagnostic message if a background level including FLASH (if any) is <11 electrons/pixel or if a background level which includes FLASH is >15 electrons/pixel, based on ETC calculations. (Extended emission in the target or a very high density of point sources provide CTE loss mitigation that is not taken into account.) Since FLASH is applied at the exposure level in APT, each subexposure (member of a pattern, CR-SPLIT, or set of multiple iterations) gets the same specified level of FLASH. If exposure times are adjusted to vary within such a set, different members of the set may have too little FLASH, adequate FLASH, or excess FLASH. Only the most deficient and most excessive FLASH are reported for the set, as a guide to where to start iterating by changing exposure times or the level of FLASH.
Users needing assistance are encouraged to contact the STScI Helpdesk by sending an email to email@example.com.
Source-Dependent Master Sky Images for the WFC3/IR Grisms
G. Brammer, N. Pirzkal, R. Ryan
Slitless spectroscopy observations with the WFC3/IR grisms are almost always background- limited as the exposures are often long (e.g., two 1500 s exposures in an orbit) and the grism is effectively a very broad passband. In IR grism exposures, each pixel on the detector sees background light from nearby areas of sky at different effective wavelengths and vignetting of spectral orders results in a vertical "curtain" structure across the detector. Modeling and removing this background structure is critical for enabling robust spectral extraction (continuum, in particular) for objects across the entire WFC3/IR field of view.
Representative empirical background-only images can be constructed from the body of archival science exposures. For example, Kummel et al. (2011) prepared such "master sky" images for the WFC3/IR G102 and G141 grisms from the first ~100 science images obtained after the instrument was commissioned. The aXe slitless spectroscopy analysis software scales and subtracts these master background images from the science exposures.
We have recently discovered ( Brammer et al. 2014) that the diffuse WFC3/IR background arises from two dominant sources: zodiacal continuum from sunlight reflected by dust in the Solar System and an emission line at 1.083 μm from metastable Helium in the upper sunlit Earth atmosphere. The dominant background component determines the structure seen in the 2D grism background, which can be different for multiple exposures in a visit and even for multiple reads of the detector within a single exposure. Here we present master sky background images for the G102 and G141 grisms generated from an ensemble of individual reads of the detector where the two background components can be effectively isolated.
The figure below shows the master zodi continuum and He 1.083μm line background images for both the G102 and G141 grisms. The differential background structure for the two components can be instantly appreciated: the shapes and contrast of the 100–200 pixel wide bands on both the left and right sides of the detector are very different. We note that we first divided the grism exposures by the corresponding F105W and F140W imaging flat-field images to separate the additive background and multiplicative flat-field effects in the analysis.
The background of a given IR grism exposure will be a linear combination of these two spectral components, again with the relative scaling between the two likely varying for exposures within a single visit (the zodi component should be roughly constant and with a variable contribution from the He line component). The aXe software is not currently set up to allow fitting the combination, but users can fit and subtract the backgrounds manually and then run aXe with the background-subtraction steps disabled (axeprep backgr="NO"). Typically this would involve 1) masking object spectra and 2) fitting for the normalizations of the background images using, e.g., least-squares techniques ( Fig.2 ).
The construction of the multi-component master sky images will be described in detail in a forthcoming ISR (Brammer in prep), which will also provide a more detailed prescription and example code for how the images can be applied to science exposures. The sky background images can be downloaded from http://www.stsci.edu/~brammer/grism_sky/.
Fig. 1: Master background images for the IR grisms, separated by the sky spectral component.
Fig. 2: Examples of sky-subtracted grism exposures. The left panel shows the science exposure with the bright object spectra masked. The middle panel shows the best-fit sky component model as a combination of the two images in Fig.1. The right panel shows the residuals in the sky-subtracted images.
UVIS Read Noise Monitoring
With 6 years of on-orbit WFC3 data now in MAST [Mikulski Archive for Space Telescopes], we took this opportunity to analyze the images for changes and trends in the read noise over time and evaluate the health and stability of the instrument. The read noise was measured using 2400 full frame UVIS bias files taken during this 2088 day period, employing the serial physical overscan and serial virtual overscan regions, as well as the CCD science pixel area using full-frame UVIS bias image pairs. The Serial Virtual Overscan shows very little trend over time (only 0.05 to 0.24% over the 6 years) with average read noise values of 2.9-3.03 electrons for the various amps, and errors of about 0.01 electrons. From the Serial Physical Overscan, which is more indicative of hardware characteristics, we find read noise levels of 2.95-3.1 (electrons) for the 4 amps with typical errors of about 0.01 electrons. The Serial Physical Overscan results show a slight increase in the read noise over time: 0.46% to 1.6%, depending on the amp, over about 6 years. The science pixel area results show an upward trend of read noise with time. Linear fits show that the initial read noise values were between 2.91-3.01 (electrons) and reveal an upward trend with a change of 0.2-0.3%/year, i.e. the instrument has been relatively stable over time. The small increase in the read noise is attributed primarily to radiation damage and general aging of the WFC3 instrument. The results of this study can be found in a forthcoming ISR entitled WFC3 UVIS Read Noise.
Charge Transfer Efficiency (CTE) in WFC3/UVIS
S. Baggett, C. Gosmeyer
The long-term behavior of CTE in WFC3/UVIS is routinely monitored using observations of external star clusters. Results from the analysis of data taken 2009-2015 are presented in ISR 2015-03.
Flux loss due to CTE degradation is a function of the source’s distance from the detector amplifier, the source signal level, the background within the image, and the epoch of the observations. The worst-case flux losses occur in images with extremely low backgrounds. In such data, based on photometry within a 3-pixel radius aperture and losses measured across 2048 pixels, the flux losses in early 2015 for faint sources (500-2000 e-) can be as high as ~50+/-2%. Losses for brighter sources (8000-32000 e-) are considerably less: ~5+/-1%.
Mitigation strategies are available. Ensuring a modest amount of background can reduce the losses substantially: ~12e-/pix added via the WFC3 post-flash reduced the flux losses to ~ 15+/- 1% and ~4+/-1% for faint and bright sources, respectively. Application of the empirical pixel-based CTE correction algorithm can also reduce flux losses: to ~10+/-1% and ~0+/-1% (unflashed images, no background) and to 3+/-1% and 0.5+/-1% (post-flashed), for the faint and bright sources, respectively.
For details of the CTE monitoring analysis, including tabulated empirical aperture photometry corrections, please see ISR 2015-03.
The standalone CTE correction code is available at the WFC3 CTE tools page.
Impact of "Blobs" on WFC3/IR Stellar Photometry
The presence of “blobs” in WFC3 IR channel images was observed shortly after WFC3’s installation in 2009, and was first described in Pirzkal, Viana, & Rajan (2010). Blobs are small, roughly circular regions of moderate attenuation caused by particulate matter on the CSM mirror. McCullough et al. (2014) have provided a complete census of all known blobs, including position, radius, and date of first appearance. They have also created a flat field designed to correct for blob attenuation, which we have tested for the case of point-source photometry.
By comparing photometry of stars that fell within blobs to photometry of the same stars after the application of the blob flat field, we have found that:
- • Blobs introduce up to 0.1 mag of attenuation for stellar sources that fall within them.
- • The blob flat field effectively corrects photometry for stars that fall in strongly attenuated blob regions (i.e. the attenuation is removed and the quality of the photometry on flat-fielded stars is comparable to that of stars that are unaffected by blobs).
- • There is no apparent color dependence of blob attenuation or flat field correction.
We note that blobs do not significantly affect most observations, as blobs occupy only 1% of the detector area and their effects can be mitigated by dithering and drizzling. The use of the blob flat to improve stellar photometry is most effective in crowded fields where there has been no dithering or where stellar photometry is performed on individual exposures.
For more information, please refer to ISR 2015-06.
ISR 2015-01: IR "Snowballs": Long-Term Characterization - M. Durbin, M. Bourque, S. Baggett
ISR 2015-02: Standard Astrometric Catalog and Stability of WFC3/UVIS Geometric Distortion - V. Kozhurina-Platais, J. Anderson
ISR 2015-03: WFC3/UVIS Charge Transfer Efficiency 2009 - 2015 - S. Baggett, C. Gosmeyer, K. Noeske
ISR 2015-04: Optimizing pixfrac in Astrodrizzle: An example from the Hubble Frontier Fields - R. Avila, A. Koekemoer, J. Mack, A. Fruchter
ISR 2015-05: WFC3 Cycle 21 Proposal 13561: UVIS Gain - H. Gunning
ISR 2015-06: The Impact of Blobs on WFC3/IR Stellar Photometry - M. Durbin, P. McCullough
ISR 2015-07: WFC3 Cycle 22 Calibration Program - E. Sabbi & The WFC3 Team
ISR 2015-08: A Study of the Time Variability of the PSF in F606W Images taken with WFC3/UVIS - J. Anderson, M. Bourque, K. Sahu, E. Sabbi, A. Viana
ISR 2015-09: Combining WFC3 Mosaics of M16 with DrizzlePac - J. Mack
ISR 2015-10: IR Grism Wavelength Solutions Using the Zero Order Image as the Reference Point - R. C. Bohlin, S. E. Deustua, N. Pirzkal
ISR 2015-11: The Internal Flat Fields for WFC3/IR - R. E. Ryan, S. M. Baggett
ISR 2015-12: WFC3/UVIS Shutter Characterization - K. Sahu, C. M. Gosmeyer, S. Baggett
To unsubscribe from the WFC3 STAN, Please send a blank email to wfc3_news-signoff-request@MAILLIST.STSCI.EDU
The Space Telescope Science Institute is operated by the Association of Universities for Research in Astronomy, Inc., under NASA contract NAS5-26555.