Résumé
We used an SRAM memory based monitoring device as a Single Event Upset (SEU) monitor to mixed-field radiation environments. The device was tested and evaluated in the H4IRRAD experimental test area that has been designed to reproduce the LHC tunnel and shielded areas at CERN [1]. In this area, a 400CeV/c proton beam impinges on a copper target. There are two zones in which we placed our monitors: an internal zone where the target is located, and an external zone, located behind a 20cm concrete shielding. A mixed radiation field is produced in these two zones composed of neutrons, protons, pions, electrons, photons and muons. The devices received a cumulated fluence of a few 1010 hadrons cm2/week for a time period of two weeks. The monitor embeds three commercial 90nm SRAM memories, an FPGA and an anti-latchup circuit. A Finite State Machine (FSM) is implemented inside the FPGA, which is responsible for the application of the tests while the three memories are in static mode, and also the communication with an external computer for the recording of the occurring upsets. Testing in static mode, requires the memory to be written with a known sequence (all bits set to ‘1’ for example), and read back after a certain time window for the detection of SEUs. The first factor we considered would be the effect of Total Ionizing Dose (TID) on our device’s functionality and more specifically on the bit cross section of the embedded SRAMs. The second factor would be the ability of the SRAMs to sense the fluctuations of the beam intensity through the increase or decrease of the recorded SEUs over the time. From the occurring results, we observed that the bit cross sections of the SRAMs in different positions remain constant for a maximum total dose of 76Gy at the internal zone. We also observed that the bit cross sections were the same with the ones recorded in our previous experiments where the same SRAM memories were irradiated with a mono-energetic neutron beam for different energies [2]. As a result, in mixed-field radiation environments with particle spectra similar to the LHC zones, the major particle contributing to the SER of an SRAM is neutrons. Finally considering the SEU count for each hour of exposure, while removing the blind periods in which the beam was not functional, or the device was under a latchup, we found that there is a constant ratio between the beam intensity and the occurring SEUs.The constant ratio between the occurring upsets and the beam intensity proves that such a device can be used for monitoring particle accelerators. Also, since the monitor can withstand a certain TID its long term utilization inside particle accelerators can be guaranteed. Further testing is to be performed in the future for the calibration of our results.