Comparison of Methods to Calculate the Dark Current Nonuniformity

Abstract

Analytical and Monte Carlo DCNU prediction methods are compared. The difference is studied according to the different assumptions used for the calculations. The analysis is performed as a function of the incident proton fluence. The convergence of the DCNU toward a Gaussian distribution predicted by the central limit theorem is also investigated. Index terms − Space environment, Displacement damages, DCNU, Image sensors, Monte Carlo. I.

Statistical spread on the displacement damage degradation of irradiated semiconductors

“…These damage distributions evolve with the incident applied fluence, and depend on the type and the energy of the incident particles. They can be calculated according to a Monte Carlo algorithm [54][55][56][57]. As can be seen on the Fig.…”

“…The damage distribution is an interesting function commonly evaluated for image sensors radiation hardening analysis. The dark current increase, whose dispersion is called theDark Current Non Uniformity (DCNU) has been extensively studied, and over the years different numerical methods have been developed for its evaluation [54][55][56][57][58][59][60][61][62][63][64]. It finds a particular interest for radiation evaluation of star trackers, which functioning can be strongly disturbed by hot pixels, presenting high induced dark current [57].…”

“…We have developed a Monte Carlo based model [54][55][56] well suited for the statistical analysis proposed here. The only difference with DCNU prediction, is that here, the calculation is limited to the evaluation of distribution of damage.…”

“…We thus obtain a damage distribution which average is equal to NIEL. The principle of calculation of damage distribution by this Monte Carlo toolkit can be found in references [54][55][56]. It can be summarized with the algorithm of Fig.…”

“…For a given pixel, damage induced by each interaction is summed whatever their nature. The number of atomic displacements, produced by nuclear interactions (elastic, inelastic), are modelled thanks to GEANT4 Monte Carlo toolkit [51][52][53][54][55][56]. GEANT4 is a C++ library dedicated to transport of particle through matter.…”