Rad-hard for the long haul

Story

June 20, 2016

Doug Patterson

Aitech

Managing the long-term effects of radiation in space environments is essential for designers of components headed for space. Mission length, more than any other factor, contributes most heavily to the sustainability of electronic components in radiation-laden applications.

Next in line is how critical the mission is – is it a question of loss of life in the event of a failure, a catastrophic communication breakdown, or merely a blip in the timetable update of a local bus or railway network?

Different levels of Earth’s orbits contribute separately to radiation exposure as well. High geostationary (GEO) orbits, for example, have harmful galactic cosmic rays and solar flares (protons and electrons) at much higher levels than lower orbits like medium Earth orbit (MEO) and low Earth orbit (LEO), because of the lack of the inherent shielding effects of the Earth’s magnetic shield. However, what is not to be ignored in LEO and MEO is the higher concentrations of trapped particles localized within the Van Allen belts.

There are also anomalies in the Earth’s magnetic field – most notably the South Atlantic Anomaly (SAA) – where those trapped particles wreak havoc with unshielded or non-radiation-tolerant commercial off-the-shelf (COTS) electronics (Figure 1). Orbiting the Earth 16 times in a 24-hour period, the International Space Station and Hubble Telescope both often pass through and must therefore deal with the SAA’s effects by either shutting down sensitive electronics or suspending imaging tasks during the transition.

Figure 1: South Atlantic Anomaly (SAA): Depiction of radiation density of the Van Allen belt over the SAA, with the areas in red being the most intense. Image courtesy of NASA/Goddard Space Flight Center.

(Click graphic to zoom by 1.9x)

Component reliability in space applications comes down to endurance in the face of radiation exposure, both from accumulated levels and the myriad of radiation types that can affect total dose, single-event effects, and other deleterious effects that can include gate burnout, substrate latchup, heavy ion bombardment, neutron and heavy ion-induced lattice atoms displacement (i.e., destruction of the base silicon or sapphire crystal structures at the atomic level), or thermal cycling.

The question of qualification

Mission length and mission criticality, such as manned versus unmanned, will dictate whether or not to use the tried-and-true NASA EEE-INST-002 (Electronic, Electrical, Electro-mechanical) components guidelines to guarantee maximum reliability or think about using different component selection criteria. A major consideration is to either select components not yet flight-proven, but that can be qual-tested and are therefore characterized for space flight, or to select strictly industrial or extended-temperature COTS components.

Table 1: High-level summary of space radiation environments and their effects on LSI/MSI devices.

(Click graphic to zoom by 1.9x)

COTS components without any space heritage or characterization can have one-tenth to one one-thousandth the reliability and longevity of the NASA EEE parts in certain radiation types and exposure times. (Table 1.) Although the COTS parts may cost less, they can fail catastrophically in hours or days versus years – depending on their suborbital and/or orbital application – and the price of failure can be immeasurable.

The EEE to COTS component cost differences between these varied approaches can be several orders of magnitude, but because reliability and mission lengths can be drastically reduced due to unreliable components, component choice needs to be careful and wise.

The long view approach to in-orbit systems

All the supposed advantages regarding size, weight, power, and cost (SWaP-C) and other issues that pertain to a terrestrial military and aerospace program are invalid when it comes to manned or unmanned spacecraft and radiation-induced failures. For university experiments with shoestring budgets, standard COTS components will work, but not for long. Radiation total ionizing dose (TID) will wipe out a command and data handling system, perhaps in only a couple months. Particles and heavy ions can and do accumulate in the component substrates and cause latent damage, powered or not.

Orbital launch platforms are not cheap and launch options are still limited, so it still costs thousands of dollars per pound, per unit mass, whether launch is of one large telecommunications or military-communications satellite or a group of “shoestring” Nanosats built from Arduino processor boards bought one Sunday from the local electronics hobby shop.

Mission length is something that cannot be ignored, and for good reason: In only a few days in orbit, COTS components can fail the mission-success-criticality factors, making the premise of using those “cheaper” COTS components a misnomer.

Properly employing COTS in space

Radiation-tolerant electronics are typically characterized in the range of 15 to 50 krad (Si), with rad-hard over 100 krad (Si). Most standard COTS components rarely make it past 1k, which can be easily hit within the first month in flight, depending on the orbital altitude and angle of inclination. Testing at heavy ion cyclotrons have shown that most of today’s COTS complementary metal-oxide semiconductor (CMOS)-based processors suffer from miserable TID stats: Most, if not all, are well under 300-400 rad (Si) TID, which is accumulated in less than two or three weeks in a typical LEO.

Device physics cannot be ignored – as the number of processor cores increases and the line geometries decrease in the chip, the worse it gets.

For reference, the average launch cost for a Space-X Falcon 9 to LEO is around $5,000 per pound ($2,300 per kg); an Atlas 5 is about $13,000 per kilogram. Moreover, finding an open launch platform can literally take from two to more than five years to secure a slot on a booster, even a Russian one.

Recently, 11 SmallSats were launched, designed to last in orbit between two and three years: How does one measure success and ROI if all of them – based on COTS components – died off in the first couple of months?

If a program truly has a limited mission length of only one or two months of failure-free operation, then the cheaper approach using COTS components may suffice. The designer can take that chance, as long as failure is an option. In manned spacecraft, such as the International Space Station in LEO, or in lunar and interplanetary (Mars) missions, which are measured in years, not months, early failure due to component fatigue is never an option.

Evaluating and selecting radiation-hardened components

Due to a rapid decline in the availability of MIL-STD-883B Class S – or for that matter, 883B Rev. C components – the appropriate parts selection and accurate qualification needed to accommodate the demanding physics encountered in space applications is more critical than ever. Their qualification for space usage now falls on the systems integrators.

For example, the potential for outgassing in a high vacuum can significantly impact system performance. Components selected must minimize outgassing that can create a corrosive or a deleterious atmosphere in a confined capsule wherever possible, regardless of whether it is a manned or unmanned environment. Specifying conformal coatings that will minimize the potential for outgassing problems caused by the high vacuum levels in space (>10-4 Torr) can also help mitigate these effects.

With regard to screened components, insisting on 100 percent environmental and radiation testing – with manufacturer lot and date-code traceability of all components – is mandatory to ensure the radiation hardness and long-term reliability of component performance in the space environment. Neither the probability batch testing of single devices, nor the characterization testing of just one board, is a satisfactory process for certifying space-qualified designs.

Characterizing components

Every batch of semiconductors exhibits some level of process variation. Such variations can encompass overall transistor gains, etch boundaries, transistor-well depths, and even package epoxies which, in turn, affect speed, performance, radiation tolerance, and outgassing properties.

For these reasons, each component used in each design should be individually radiation tested, certified, and tracked against component lot and date codes, with complete and full documentation and traceability. (Figure 2.) Just because a semiconductor design passed testing in an earlier application does not mean that a subsequent production run of the identical design will yield the same results, especially since these components can come from physically different fabs.

Figure 2: Boards qualified for space applications should include individually-characterized components.

(Click graphic to zoom by 1.9x)

By definition, components to be characterized must be from the exact same lot and date code. If they do not, that completely nullifies the testing.

The importance of safety margins and the extreme consequences of potential component failures far outweigh the expenditures associated with proper testing. Performance and certification of appropriately conducted tests and characterizations can increase radiation tolerance levels to nearly five times that of unscreened devices.

Considerations for any space mission

While the need to accommodate the degrees of radiation may vary based on mission longevity and the specific Earth orbit, the effect of radiation on system components in space need to be evaluated. Rad-tolerance in space components defines the success of a mission, so accounting for the proper levels before launch is critical.

There may be less expensive, less qualified alternatives, but the cost of a failure in system operation while in flight needs to be assessed as well. For a small satellite cluster designed to last only a few years, it may not be as important for all satellites to remain operational across the entire mission. In contrast, for a system exploring the deeper regions of space or destined to stay in orbit for several years, reliability far outweighs saving a few dollars in the development phase.

The bottom line? Designers must make sure to account for all design aspects that will affect the success – or failure – of the mission.

Doug Patterson is vice president of the military and aerospace business sector for Aitech Defense Systems, Inc. He has more than 25 years of experience in marketing and business development as well as product management in telecommunications and harsh environment electronics. He served on VITA’s board of directors and holds three patents in advanced metered mailing systems and nonvolatile memory redundancy mapping. Doug holds a BSEE from BEI/Sacred Heart University. He can be contacted at [email protected].

Aitech www.rugged.com