The nonlinear photon transfer curve of CCDs and its effects on photometry

TL;DR

This paper investigates the nonlinear photon transfer curve in CCDs caused by charge sharing, which affects photometry and star shape measurements, especially at high signal levels, and presents a model to quantify this effect.

Contribution

It introduces a model for charge-sharing PSF to explain the nonlinear PTC and its impact on photometry and star shape measurements in CCDs.

Findings

Nonlinear PTC begins at about 1/3 full well in the CCD.

Charge sharing causes the variance to drop at high signal levels.

The model allows quantification of effects on photometry and star shape.

Abstract

The photon transfer curve (PTC, variance vs. signal level) is a commonly used and effective tool in characterizing CCD performance. It is theoretically linear in the range where photon shot noise dominates, and its slope is utilized to derive the gain of the CCD. However, recent researches on different CCDs have revealed that the variance progressively drops at high signal levels, while the linearity shown by signal versus exposure time is still excellent and unaffected. On the other hand, bright stars are found to exhibit fatter point spread function (PSF). Both nonlinear PTC and the brighter-fatter effect are regarded as the result of spreading of charges between pixels, an interaction progress increasing with signal level. In this work we investigate the nonlinear PTC based on the images with a STA1600FT CCD camera, whose PTC starts to become nonlinear at about 1/3 full well. To explain the phenomenon, we present a model to characterize the charge-sharing PSF. This signal-dependent PSF can be derived from flat-field frames, and allow us to quantify the effects on photometry and measured shape of stars. This effect is essentially critical for projects requiring accurate photometry and shape parameters.