Modelling electron distributions within ESA's Gaia satellite CCD pixels to mitigate radiation damage

TL;DR

This paper develops a detailed 3D microscopic model of Gaia satellite CCD pixels using the ATLAS simulation framework to understand electron trapping effects caused by radiation damage, aiding in calibration.

Contribution

It presents the first application of the ATLAS device simulation software to model Gaia CCD pixels, calibrating its parameters and benchmarking against measurements.

Findings

ATLAS successfully models Gaia CCD pixels and matches experimental data.

Calibration of the fixed oxide charge parameter is essential for accurate simulations.

Different doping approximations significantly affect the simulation results.

Abstract

The Gaia satellite is a high-precision astrometry, photometry and spectroscopic ESA cornerstone mission, currently scheduled for launch in 2012. Its primary science drivers are the composition, formation and evolution of the Galaxy. Gaia will achieve its unprecedented positional accuracy requirements with detailed calibration and correction for radiation damage. At L2, protons cause displacement damage in the silicon of CCDs. The resulting traps capture and emit electrons from passing charge packets in the CCD pixel, distorting the image PSF and biasing its centroid. Microscopic models of Gaia's CCDs are being developed to simulate this effect. The key to calculating the probability of an electron being captured by a trap is the 3D electron density within each CCD pixel. However, this has not been physically modelled for the Gaia CCD pixels. In Seabroke, Holland & Cropper (2008), the first paper of this series, we motivated the need for such specialised 3D device modelling and outlined how its future results will fit into Gaia's overall radiation calibration strategy. In this paper, the second of the series, we present our first results using Silvaco's physics-based, engineering software: the ATLAS device simulation framework. Inputting a doping profile, pixel geometry and materials into ATLAS and comparing the results to other simulations reveals that ATLAS has a free parameter, fixed oxide charge, that needs to be calibrated. ATLAS is successfully benchmarked against other simulations and measurements of a test device, identifying how to use it to model Gaia pixels and highlighting the effect of different doping approximations.