Gaia aims to create the most accurate stereoscopic map ever created of our galaxy, the Milky Way, by constantly scanning the sky with two telescopes that will collect the light from more than a billion of celestial objects. Together with ESA's scientists and the industry, the DPAC (Gaia Data Processing and Analysing Consortium), which is composed by 300 researchers spread other 15 countries, is currently preparing the algorithms that will transform the collected photons into extremely valuable information: source position, proper motion, radial velocity, and composition. Gaia will help us in the understanding of the formation and the evolution of our Galaxy but not only, it will also strongly affect every fields of today's astrophysics.

To do so the accuracy that needs to be achieved is extreme e.g., the position on the focal plane of a 10th magnitude star photo-centre has to be determined with an error of 7 micro-arcseconds: a beetle crawling on the surface of the moon could be seen from Earth with "Gaia eyes". Technically speaking the focal plane of Gaia is formed by a set of 100 CCDs (Charge Coupled Device, a light detector) and the allowed error in the position measurement has to be as small as 300 CCD's silicon atoms sitting next to each other. Flare protons are very energetic, when they collide with one of these atoms, they knock it out of the very well ordered silicon lattice. The disorder thus created changes the electrical property of the device itself.

Indeed displaced atoms let vacancies that can gather with the silicon's impurities or with other vacancies to create a charge trap. A CCD collects photons thanks to the photoelectric effect: each time a photon collides with the CCD, an electron is created inside the CCD. These "photo-electrons" are transferred from the pixel where they were produced to the electronic system that can ultimately send the information to earth, by multiple transfers from one pixel to its neighbour. Astronomical CCDs are very efficient devices; they transfer 99.999999% of the photoelectrons from one pixel to its neighbour. The radiation-induced traps capture the photoelectrons and release them after a time that varies accordingly to their species and increase the charge transfer inefficiency (CTI).

The CTI effects contribute to a direct charge loss added to a distortion of the photoelectron count profile. This leads to more uncertainties in the determination of the star photo-centre position (i.e. centroid bias). According to recent experiments carried on irradiated Gaia CCDs, after five years the centroid bias (or photo-centre shift) could reach the order of a milli-arsecond. Unfortunately the photoelectron capture and release processes are stochastic; their probabilities of occurrence vary with the signal intensity, the CCD state, the CCD illumination history, the sky background and the trap species involved. The radiation induced centroid bias cannot be systematically corrected. The only way to mitigate the CTI effects is to model them.

DPAC plans for radiation damage recovery involve experimental tests on irradiated CCDs and several level of modelling from sub-pixel physical Monte-Carlo charge transfer simulation to fast analytical photo-electron count profile distortion. The detailed physical models are designed to help us to increase our understanding about the semiconductor traps physics, to resolve the shape of the electron cloud inside a pixel, to investigate the field of parameters as a pathfinder for faster models and provide a validated set of physical equations for analytical models. While fast analytical models will ultimately be implemented in the Gaia data processing pipeline in order to reach the required accuracy on measurements and initial Gaia's science requirements.

CEMGA constitutes a platform for any kind of CTI effects models developed in the Gaia mission framework. It provides a detailed description of physical entities such as CCD, pixel or trap, and method to interact with them. CEMGA offers a rigorous environment for testing and validating implemented models as well as comparing them. It comes with a set of preconfigured simulated experiments such as First Pixel Response and various tools to analyse the results (io management methods, centroiding algorithm, visualizing routines) and compare them to experimental data. So far several complementary models have been implemented (two pixel-level Monte-Carlo models and one analytical model), they are currently in validation. CEMGA is developed in JAVA as every Gaia's data processing work packages; a detailed documentation about CEMGA's code is available here:

CEMGA is the main tool of my PhD researches entitled: "Theoretical and empirical modelling of charge transfer inefficiency in astronomical CCDs" and detailed here:

Loading more stuff…

Hmm…it looks like things are taking a while to load. Try again?

Loading videos…