SPEQS-1: Radiation Tests

Publications about radiation test results:

How do we quantify the particle radiation environment in LEO?
The radiation flux in Low Earth Orbit (LEO) is expected to be "modest" because of the protection afforded by the Earth's magnetic fields. Satellites in LEO (below 1000 km) are generally below the inner radiation belt. However, as physicists who are thinking of sending instruments into LEO (see SPEQS), we really want some numbers on the radiation flux for protons as they have the potential to be highly damaging to our electronics.  We also expect the Earth to be a magnet so the magnetic field lines must pinch together near the poles - how does this affect the flux at different regions of space? Are there regions of space we should avoid?

We decided to use a piece of software called the Space Environment System (SPENVIS) that builds on years of data accumulated by NASA and ESA. SPENVIS enables us to look at radiation flux at different altitudes, and we generated a graph for the electron/proton flux shown below. The picture shows increased flux radiation near the poles (as expected), and another region of high intensity called the South Atlantic Anomaly (SAA) which is a little bit more surprising. The SAA is due to the Earth's magnetic axis not being completely aligned with the rotational axis.

This graphs from SPENVIS tell us that if we want to avoid regions of high flux intensity we should try and keep our satellites to a low equatorial orbit. We did some additional analysis and summarized our results for a combination of two inclination angles and two altitudes in the figures below. The figures relate only to proton flux because we expect that electrons are not capable of penetrating much beyond the walls of the satellite.  In the figure on the left (a), we present the proton flux spectra - note that altitude has a dramatic effect for equatorial orbits.  These spectra can be combined with the graph on the right (b) displaying the Non-Ionising Energy Loss (NIEL) deposited in a unit sample of Si for protons of  specific energy.  NIEL plots can obtained from the literature for a number of materials, but as our single photon detectors are Silicon based, we concentrate on the NIEL value for Si. Together, the spectra and NIEL values enable us to estimate the displacement damage dose from protons for different orbits.

Simulating in-orbit Radiation

Armed with this knowledge provided by SPENVIS we had a number of our critical devices (single photon detectors, memory modules and micro-controllers) tested to different dose levels of proton and gamma-ray irradiation. Gamma-ray irradiation was used to simulate ionising dosage, and this test was performed at the Centre for Ion Beam Applications at NUS.  For proton irradiation, we travelled to the University of California at Davis.  


 Cliff Cheng and Dr Carlos Castaneda from UC Davis
with the SPEQS electronics in the proton beam path.
SPEQS electronics undergoing preparation for proton irradiation. 

Some conclusions

After irradiation, we measured the electrical characteristics of the components.  We were pleasantly surprised to find that for most of the components, the radiation tolerance was very good - even after a dose equivalent to 1 year in orbit the components were still performing within specifications.  As expected, the commercial single photon detectors suffer the most effects but we found that even in the worst-case scenario of a high altitude polar orbit, the SPEQS unit should still perform for 2-3 months during which time sufficient data would have been accumulated.

In the future, we would like to continue studying proton irradiation effects on different kinds of commercial single photon detectors.  We have a hypothesis that single photon detectors with smaller active areas (for absorbing photons) would have a higher tolerance to proton irradiation because of the reduced cross-section. We will need to understand if the trade-off between a large active area (and ease of collection and detection efficiency) and radiation tolerance is acceptable.