Sensor Anomalies


  • Andrei Nomerotski (BNL), [anomerotski(AT)]
  • Pierre Astier (LPNHE) [pierre.astier(AT)]

The sensor anomalies working group deals with non-ideal behaviors of CCD sensors used for astronomy at visible wavelengths, in particular by LSST. CCD sensors are affected by distortions of the image induced by alterations of the drift field. These alterations are either permanent (e.g. at the side of the sensors, or due to spatial variations of doping in the silicon bulk) or induced by the charges accumulating during image integration. These two classes of anomalies are commonly labeled static and dynamic. The dynamic effects cause a non-linear response of sensors to incoming light, and in particular the fact that bright point sources appear bigger than their fainter counterparts, causing the so-called brighter-fatter effect. 

These anomalies have always been at play in CCD sensors, but the advent of thick, deep-depleted CCDs has made them more striking. Together with the stringent performance requirements on sensors for an LSST-like survey, these effects should be mitigated. For example, the brighter-fatter effect on LSST CCDs will amount to a few percent increase of point-source size from zero flux to saturation, while exploiting the weak lensing measurements at the LSST survey scale require an estimation of the PSF size for faint objects accurate to about one part in a thousand. 

The current approach of the SAWG working group is to devise methods to correct the images for both sets of effects. The static effects are thought to be easy to handle because they exhibit slowspatial variations and can be accurately mapped (either in the laboratory, on the sky). The dynamic effects are more difficult to handle because undoing them requires an estimation of how stored charges alter the electric field inside the sensor. This electric field pattern can be constrained from statistical correlations of pixels in flat-field exposures, but the inversion is not unique. One can then try to constrain a full electrostatic model of the sensor using the measured brighter-fatter effect, and the measured flat-field correlations. 

Many SAWG members also belong to the LSST camera sensor team, or to the LSST software team, but SAWG is independent from their managements. SAWG acts as a forum to present measurements, calculations and methods related to sensor anomalies. The working group aims at producing and validating the algorithms that will mitigate the impact of these anomalies on LSST measurements. 

 Top: Charge distribution in a single pixel filled with 80,000 electrons. 

Bottom: position of effective pixel boundaries (and effective pixel areas), 
when the central pixel contains 80,000 electrons. The electrostatic
setup is tuned to reproduce laboratory measurements of a real LSST
candidate sensor. Electrostatic computations and figures by Craig Lage (UC Davis).