Charge Transfer Efficiency and Charge Injection in the WFC3 UVIS Detector
SUMMARY: Like all CCD detectors in low-Earth orbit, the WFC3/UVIS CCDs experience a degradation of their Charge Transfer Efficiency (CTE) over time, introduced by their exposure to energetic radiation. The WFC3 team has been monitoring this degradation. Beginning in Cycle 19, Charge Injection will be offered to observers in LINE17 mode, as an option that can significantly mitigate the photometric and astrometric effects of CTE degradation. A summary of the UVIS CTE monitoring programs, the CTE evolution until September 2010, and an introduction to Charge Injection and its noise penalties is given in the WFC3 UVIS CTE Whitepaper.
Updated results on the UVIS CTE, and on tests of the performance of Charge Injection, are presented in the WFC3 UVIS CTE and Charge Injection Update for Cycle 19 Observer.
Further updates on the UVIS CTE and Charge Injection will be published on this webpage and announced in future STANs as the WFC3 team’s calibration efforts progress. Questions should be directed to your program’s contact scientist, or the STScI help desk.
I. Advice for Observers Who Consider Charge Injection:
- Charge Injection carries some penalties in residual image noise and a fraction of image rows may need to be discarded or assigned lower weights during the data analysis. It is therefore recommended that observers weigh the benefits and disadvantages of CI for their program. If the science target is relatively small, CTE can be strongly mitigated by placing the target on rows close to the read-out amplifiers. CTE effects will be signifi- cant mainly in very low background data (very few electrons per pixel) where point sources of 500 to 2000 electrons, far from the amplifiers, can suffer CTE losses of roughly 10% (March 2011) in 3 pixel radius apertures. CTE Effects are still likely non- zero (~2±2%) for stars with 8000 to 16000 electrons. However, in the presence of a sig- nificant background, 20-30 electrons per pixel, stars with 500 to 2000 electrons show significantly lower CTE losses of only ~2% that may be correctable through empirical relations to a precision that is better than the sources’ Poisson flux errors. The penalties of Charge Injection, enhanced noise and likely loss of Charge-Injected and adjacent rows, are detailed in the UVIS CTE Whitepaper.
- If you choose Charge Injection, the images should be taken in several exposures, with a dither pattern that includes several small steps in detector Y direction, subdividing the 17-row-interval of the Charge Injected rows. Of course, this dither pattern must also include meaningful dithers in X direction. Since the latter rows, and likely some adjacent ones, may need to be discarded or assigned lower weights when combining the images, each of the resulting gaps in single images must be covered by several other images that do not have a Charge Injected row at that sky position. We recommend at least 4 dithered images.
- The online Exposure Time Calculators do NOT account for the enhanced noise in part of the image rows in Charge Injected data, and neither for the loss of the Charge Injected and some adjacent rows. Observers should consult the UVIS CTE Whitepaper for details and modify their exposure time estimates accordingly.
- Observers do not need to obtain special calibrations. The WFC3 team will provide the necessary calibration frames (Charge Injected Bias frames), which will be applied automatically to Charge Injected data by the CALWF3 pipeline.