Astrometric Errors Introduced by Interpixel Capacitive Coupling in Hybridized Arrays

TL;DR

This paper investigates how interpixel capacitance causes image blurring and astrometric errors in hybridized arrays, especially as pixel sizes decrease, and proposes a numerical correction method based on detailed characterization.

Contribution

It introduces a numerical successive approximation method for correcting IPC effects, improving accuracy over traditional deconvolution techniques.

Findings

IPC causes significant astrometric errors in hybrid arrays.

Correction via successive approximation improves measurement accuracy.

Characterization of IPC's functional relationship enhances correction quality.

Abstract

Interpixel capacitance (IPC) between adjacent pixels in hybridized arrays gives rise to an electrostatic cross talk. This cross talk causes MTF degradation and blurring of images or spectra collected using these devices. As pixel size is driven down from the 18 micron pixel pitch of the H2RG read out circuits to the 10 or 15 micron H4RGs IPC is driven up resulting in greater cross talk, all else being equal. Mounting evidence indicates that IPC varies as a function of depletion state of the photo-active diodes. For single pixel events, increasing the event intensity corresponds to a decreasing fractional coupling. If left uncorrected, IPC can give rise to systematic errors in precision astrometric and photometric measurements, in particular when dealing with confused point sources or spatially extended structures for shape measurements as demonstrated through comparison of registered sources from ESO HAWK-I and HST ACS WFC datasets. Furthermore these errors will be the most significant when operating near the sensitivity limit of these devices. Deconvolution based correction methods are invalidated by this same signal dependence. Instead a numerical method of successive approximation can be used to correct coupling due to a well characterized IPC. Examination of single pixel reset data above flat fields could be used to characterize IPC's functional relationship for neighboring pixels. This higher quality characterization can result in more accurate correction.