Intra-pixel response characterization of a HgCdTe near infrared detector with a pronounced crosshatch pattern

TL;DR

This paper investigates the intra-pixel quantum efficiency variations caused by the crosshatch pattern in HgCdTe near-infrared detectors, revealing that these variations impact photometric accuracy and are not corrected by flat-field calibration.

Contribution

It demonstrates that the crosshatch pattern induces sub-pixel QE variations in HgCdTe detectors, which can affect high-precision astronomical measurements and are characterized using a specialized testbed.

Findings

Crosshatch pattern causes intra-pixel QE variations.

Flat-field calibration does not remove these variations.

Potential impact on photometry, astrometry, and spectroscopy.

Abstract

The "crosshatch" pattern is a recurring "feature" of HgCdTe arrays, specifically the Teledyne HAWAII family of near infrared detectors. It is a fixed pattern of high frequency QE variations along 3 directions generally thought to be related to the crystal structure of HgCdTe. The pattern is evident in detectors used in Hubble WFC3, WISE, JWST, and in candidate detectors for Euclid and WFIRST. Using undersampled point sources projected onto a HAWAII-2RG detector, we show that the pattern induces photometric variations that are not removed by a flat-field calibration, thus demonstrating that the QE variations occur on scales smaller than the 18 micron pixels. Our testbed is the Precision Projector Laboratory's astronomical scene generator, which can rapidly characterize the full detector by scanning thousands of undersampled spots. If not properly calibrated, detectors showing strong crosshatch may induce correlated errors in photometry, astrometry, spectroscopy, and shape measurements.