Proceedings Article | 10 July 2018
KEYWORDS: Sensors, Annealing, Jupiter, Temperature metrology, CMOS sensors, Infrared detectors, Imaging systems, Mercury cadmium telluride, Radiation effects, Ionizing radiation effects
MAJIS (Moons and Jupiter Imaging Spectrometer) is the imaging spectrometer of JUICE (JUpiter and ICy moons Explorer), an ESA large-class mission scheduled for launch in 2022 that will start scientific operations in the Jovian system on 2030. MAJIS will deliver observations of the surfaces and exospheres in a wide spectral range from 0.5 up to 5.5 µm, including several flybys of the icy Galilean moons at spatial sampling down to ~ 75 m, as well as a 9-month Ganymede orbit eventually providing full coverage at 3 km/pixel [Langevin et al., DPS 2016].
To achieve these objectives, two focal plane arrays sharing the same telescope's optic will be used, with a splitter in the 2.25-2.35 µm wavelength range. MAJIS is a push-broom spectral imager: the image over one spatial direction will be projected on several hundred spatial x spectral pixels of both HgCdTe CMOS detectors after wavelengths dispersion.
Those observations will be highly affected by the intense magnetosphere of Jupiter which traps high-energy particles that may decrease the performances of photo-sensitive layer of detectors or even the electronics. The expected total ionizing dose (TID) for the entire mission is 80 krad in Silicium with the appropriate Aluminium shielding and the equivalent ~ 55 MeV proton fluence integrated over the mission will be 10^10/cm² [JUICE environment specifications, ESA 2015].
We performed a TID and a proton radiation test on candidate MAJIS detectors (SOFRADIR NGP MWIR, 1024x1024 pixels, 15 µm pitch). They both took place at UCL (Belgium) facilities to investigate expected degradations to MAJIS detectors. The TID tests have been conducted with ⁶⁰Co radioactive source emitting mostly 1.3 MeV γ photons: the detector was exposed to radiations for 14 days with 1-hour interruptions twice a day to monitor time evolution by taking images. The cumulative dose was measured at 93 krad on the last day. The protons tests have been conducted with a cyclotron, which provided 10¹⁰ protons/cm² over the 4 hours test with energies distributed between 50 and 62 MeV at detector level, behind 2 mm Aluminium shielding.
We characterized detector degradations by monitoring three main parameters: the amount of "bad" (hot, spurious, variable) pixels, the average level of dark current, and the DSNU (Dark Signal Non-Uniformity), for various integration times, especially low values (~ 100 ms) typical of MAJIS integration times. Such short integration times are required by the high photons fluxes from the targeted icy planetary surfaces and the on-board data processing algorithm that will remove electron spikes by selecting data within a set of sub-integrations associated with each acquisition [Langevin and Piccioni, EPSC 2017].
The analysis of data collected before, during and after the TID show no significant increase of the number of bad pixels and no significant variation of the mean dark current level on the candidate detector, for a total dose (93 krad) greater than the one expected during the whole MAJIS mission (80 krad).
After proton irradiation we observe a significant but limited increase in the number of bad pixels. Only 0.2% of detector pixels become bad after irradiation for a typical MAJIS integration time of 100 ms. Such a low value does not endanger the operability of the detector during the mission. An update in the amount of protons in Jovian environment, recently delivered by ESA, may result in an increase by a factor of 10 of the expected fluence at Jupiter. According to SSC/EPFL 2009 ESA Courses, proton-induced damages increase linearly with energy after 20 MeV. Assuming that this remains true at lower energies as we are dealing with (3 MeV), and since on these detectors any pixel has the same probability of being deprecated by protons regardless of their initial state, a linear extrapolation of the number of new bad pixels would give 2% of lost pixels at the end of the mission, which is still adequate to fulfill science objectives.
The overall conclusion of those test campaigns is a satisfying resilience of the candidate SOFRADIR HgCdTe CMOS detectors to typical operation conditions (integration times, dose rates) of MAJIS.
