Radiation in space: Expert Q&A on engineering, software, and testing

This post compiles all the questions asked by attendees and the expert answers provided during the webinar “Addressing Radiation Environment on Spacecraft in Orbit,” presented by Katrina Elsey. Covering topics such as radiation engineering, simulation software, testing methods, and the impact on COTS components, this Q&A offers valuable insights into how to evaluate and mitigate radiation effects in space missions.

Whether you’re working in New Space or traditional programs, you’ll find practical answers to common technical challenges.

Q. How about ELDR effect? is there any plan for ELDR effect estimation?

A. A way of determining the average ELDR is by dividing the TID on the device by its lifetime. Concerning tests, according to the ESCC 22 900 document, there are two dose rate windows:

• “Standard rate”: 0.36 to 180 krad(Si)/hour (3.6 to 1800 Gy(Si)/hour)
• “Low rate”: 36 to 360 rad(Si)/hour (0.36 to 3.6 Gy(Si)/hour)

ELDR is lower than 36 rad(Si)/hour and occurs in bipolar technologies. It is therefore not linked to simulation but is to be taken into account during tests.

Q. Is it possible to consider any minimum TID radiation hardness for COTS if no test data are available?

A. Unfortunately, this is not possible, the TID sensitivity varies too much from component to component or lot to lot for the same component reference. It can even vary from wafer to wafer depending on whether the ovens are horizontal or vertical, for example.

Q. Would you recommend any open-source database that gathers admissible TID / TNID values for different components/materials?

A. We can recommend the open-source database by the ESA, ESCIES. You can also find this kind of information in IEEE XPLORE (there is a small admission fee).

Q. COTS parts also suffer from build quality. Do you do screening on them for recommendation, apart from a recommendation on radiation tolerances?

A. We apply the RHA for the NewSpace procedure (using mainly COTS). Regarding the COTS screening, various approaches are available. The first question to answer is: What is the mission profile?

Depending on this, we use either one of two approaches:

• A standard CLASS 1, 2 or 3 approach based on the ECSS Q QT 60 13c document
• Depending on the device classification and also its technology maturity, we can adapt the screening flow to a subset. This can be done even though nothing has yet been done except for the radiation to be studied. If necessary, a radiation test will be performed.

We can support you in this process according to your specific requirements. This can be provided in both forms: technical support for defining the tests and performing the tests.

Q. For the NewSpace activity, is it possible to request (purchase) the SEE/SEL data available at Alter without the need to subcontract the radiation analysis study? (The analysis is performed by the developer)

A. We can look into all available databases to find as much data as we can. Concerning the data available at ALTER FR, the test data generally belongs to the company that ordered the test, and therefore, we may not be allowed to disclose the results. However, we do have the possibility to do crowd testing.

Q. What options are there to make components radiation hard, besides changing the part to a radiation-hardened one?

A. Hardening can be achieved both by software (ECC) and hardware (triplication for example). In the case of ASIC development, there are various rad-hard technologies that exist, and also specific design libraries.

Q. What is the use of the NewSpace software you have developed compared to the other ones?

A. The software we have developed is useful for companies that do not have a radiation engineering team, and therefore do not have a complete understanding of the wide range of models and parameters to choose from. We would provide you with a license to this software for which your only inputs are the altitude, inclination, lifetime, and shielding thickness of your device, and the software returns an approximation of the TID/TNID/SEE rate on orbit. The models are implemented and chosen by us, therefore you do not have to choose which one to use. In this way, the software is usable by any user.

Q. Has the new space cube modeling (SEU4NS and TID4NS) been compared with OMERE or SPENVIS for different types of orbits? What is the worst-case deviation between both models?

A. In both cases, feel free to contact us if you wish to see detailed information comparing both pieces of software.

• In the case of TID4NS, the model is based on a solid cube geometry, whereas OMERE and SPENVIS use solid spheres. The maximum deviation between the two is about 2%.
• In the case of SEU4NS, the solid sphere method is used. The average deviation with OMERE is 6%.

Q. Could you add electrical calculations simultaneously with your software to know the effect on the circuit’s functionality?

A. This is not something we had planned to work on so far, but we are open to all suggestions.

Q. What is the limitation of these simulation tools in terms of altitude & inclination angle?

A. For our NewSpace tools, you can input an altitude between 100 km and 36 000 km, and the inclination angle ranges from 0° to 98°.

Q. Is it also possible to change the sensitive volume depth when using SEU4NS? And what interpolation is used for the proton cross-section fitting?

A. This is currently not an option as the goal of this software is to keep the parameters as simple as possible. In the program, we have set the sensitive volume depth to 2 microns.

The proton cross-section is fitted with a step function because this approximation is very close to the test results (doc available upon request).

Q. Can you repeat the part when you don’t have information of a component, how do you perform the SEE analysis? From where do you take the cross-section for example?

A. In SEU4NS, if you do not yet have any information on your component, you can simply input the orbit parameters and shielding thickness, and the software will return an SEU rate given in n° of SEU per Mbit per day. This value is calculated using the Weibull parameters of a highly sensitive device, and thus gives an approximation of the worst-case SEU rate on orbit.

Q. What type of hardware do you use to perform the simulations? Do you use multi-core workstations, a cluster of computers, or a desktop PC?

A. The software we use mostly runs on desktop computers. However, depending on the complexity of the model, the use of a cluster might be needed.

Q. Are SEE rate calculation software different for classical space and new space?

A. For the classic space community, the software used is OMERE and SPENVIS. For the New Space community, we have developed our own SEE tools. In order to obtain the results, we calculated the SEE rate at a geostationary level and created an algorithm to attenuate the rate with the variation of altitude and inclination. The accuracy of this method is about 6% when compared to OMERE/SPENVIS. (doc available upon request).

And of course, both types of communities are free to use whichever software they wish, our software is just a much simpler version of those currently available, to make radiation simulation more accessible to all.

Q. What about SET in analogue devices?

A. Design rules mainly drive the way to avoid a SET, thanks to a decoupling capacitor: you would need to perform heavy ion/proton/laser testing.

Q. Are there ways to avoid testing (TID, TNID, DD) with analysis of COTS electronics?

A. No, this is not possible; the TID & TNID levels even vary from lot to lot.

Q. Is there any support granted from ALTER on radiation effects on non-EEE parts, such as exposed materials on external surfaces?

A. We have yet to perform tests on external surface materials, but we have performed TID tests on phototransistor lenses and cables. We are able to go up to an ionizing dose of 750 krad/h and can go up to 300 Mrad in total.

Q. I understand that customers may handle WCA (Worst Case Analysis) based on radiation data extracted through analysis. But do you offer any support to customer teams in circuit analysis, for example, in Single Event Transients propagations in designs? or SEFI impacts on circuits?

A. This is not something that we do on a regular basis. If you have a precise idea of the analysis you need for your project, feel free to contact us, and we can discuss the possibilities.

Q. Did you refer to a component database? Which kind of database is it? Moreover, what kind of facility/instrumentation do you use for testing particle (electron, proton, ..) effects?

A. Databases: there are many to choose from but the main ones we use are IEEE Xplore, ESCIES by the ESA, and our in-house data.

Test facilities:

• TID: We use several sources, for example, TRAD’s facility in Toulouse and/or RADLAB’s facility in Sevilla (Spain). This provides us with enough flexibility to handle all the TID requests.
• TNID: Generally speaking, we use UCL in Belgium and PSI in Switzerland.
• SEE: we work with many particle accelerators, but the ones we use most frequently are RADEF at the University of Jyvaskyla in Finland, HIF at UCL in Belgium & PIF at PSI in Switzerland.

Q. Do you have a service for testing the radiation shielding capacity of material?

A. Our test engineers frequently test the effects on devices, but it would be possible to perform such a test. Feel free to contact us if you have a particular project in mind.

Q. Possibility of SEE at board level? Size of the active area?

A. These tests are possible but they depend heavily on the beam characteristics (penetration, beam size). We usually work with beams roughly 4cm wide (RADEF). For high penetration irradiation, we can perform our tests at TAMU in the USA. Feel free to contact us and let us know the test you have in mind, we would be happy to discuss the different options available.

Q. How do you calculate the atmospheric or ground radiation environment?

A. We use the atmospheric environment model described in IEC 62 396 ed.2.

Q. Are TID and TNID always negligible at atmospheric/ground heights?

A. The proton and electron fluxes in the atmosphere are much lower than those in space; therefore, over the lifetime of a device, the cumulative effects, TID & TNID, are negligible.