Radiation-tolerance screening for missions of all sizes

3Decreasing costs for satellite launches have created new opportunities for nontraditional in alternative orbits. The application of expensive radiation-hardened (rad-hard) certified components is often not warranted for these applications. Rather, technology can provide these smaller projects with both substantial savings and an enhanced safety margin. Reliable methods to determine if commercial off-the-shelf () components are radiation tolerant can allow designers to screen components for the target radiation criteria that meet the mission requirements.

The global aerospace and defense industry is estimated to grow by three percent in 2016, according to Deloitte’s 2016 global aerospace and defense sector outlook. Innovations in component miniaturization and ever-decreasing satellite launch costs represent much of this revenue growth. These factors, in turn, have led to the dramatic increase in the number of small-satellite constellations. Companies such as Planet Labs launch and operate constellations, which are groups of many small satellites; the company’s FLOCK 1 solution, to take one example, consists of 28 small Low Earth Orbit (LEO) satellites working together for earth-mapping applications. For communications applications, constellations can be made up of hundreds of small satellites working collectively in order to remain in constant contact with ground stations. Several companies are working to launch even larger constellations in the coming years and are considering how to reduce the affiliated volume costs.

Selecting components certified as rad-hard under the adopted (RHA) procedures of MIL-HDBK-814, MIL-HDBK-816, and MIL-HDBK-817 is the way to go for designers. The rigorous RHA standards certify entire production lines of parts to various specified standards for total ionizing dose (TID), neutron displacement damage, and single-event effects (SEE). While this guaranteed quality affords designers great peace of mind, the added expense of rad-hard components can give designers and financial managers pause. Costs of single dies of certain or integrated circuits can easily run into hundreds of dollars. With these high individual component costs, the bill of materials (BOM) cost of even simple subsystems can balloon into the tens of thousands of dollars.

The most inexpensive alternative to using rad-hard parts is simply to use COTS products. With no guarantee of radiation tolerance, however, the potential for unforeseen and critical failures often outweighs the savings.

The compromise between using rad-hard parts and using completely untested COTS parts is to use parts from selectively screened COTS lots. With this approach, specific lots of components are screened for RHA tolerance. Selective screening of lots of readily available and lower-cost COTS components is much easier than certifying entire production lines. Furthermore, screening can have other advantages, such as uncovering tolerances to mission-specific levels while testing to the same rigorous standards set forth in MIL-STD-750 and MIL-STD-883.

Rad-hard overkill

In order for manufacturers to brand their semiconductor or hybrid products with JAN (Joint Army and Navy)- or RHA (radiation-hardness assured)-level designators from MIL-PRF-38534, MIL-PRF-38535, or MIL-PRF-19500 specifications, a strict, time-consuming, and costly certification process must be followed, which naturally leads to a price tag premium for these rad-hard parts. Military handbooks typically recommend parts designated to at least RHA level R [1.0E05 rad(Si)] as preferred for any application; this ”R” level is fairly high on the designator totem pole, further driving up the price.

The good news for smaller projects, such as those requiring launches of low-weight LEO satellites, is that these high standards are unnecessary. While LEO satellites experience substantially more aerodynamic drag than those in high orbit, they also see less radiation. As an example, with a typical LEO radiation dose rates of 1.0E-04 rad(Si) per second, a 10-year-lifespan satellite would only accumulate 3.0E04 rad(Si), substantially less than the 1.0E05 rad(Si) level required of RHA level R. Figure 1 depicts the gate threshold of two wafers of COTS and one wafer of RHA MOSFETs reacting to TID. Commercial devices will almost always have a more significant reaction in such tests, but it is still possible to find devices that pass parametric limits at mission-specific doses of radiation.

21
Figure 1: MOSFET threshold vs. TID.
(Click graphic to zoom by 1.9x)

Although these LEO-accumulated radiation levels are usually substantially lower than those used in rad-hard certifications, using unscreened parts can be quite risky. Untested components can exhibit unknown behavior after accumulating TID or being exposed to displacement damage or SEE. The failure of a key COTS part can easily lead to overall system failure. Without characterizing how these parts respond to radiation exposure, it is not possible to predict when failure or behavioral changes will occur.

The lower the project budget, the more it makes sense to rely on readily available products, due to both cost efficiency and ease of replacement. Modern COTS components, especially from manufacturers counting on business from the aerospace and defense industry, continue to push the boundaries of both quality and performance; these products are often manufactured alongside their RHA equivalents with very similar processes. Except for their unproven reliability against RHA effects, COTS parts have little reason to be viewed as inferior, as long as their guaranteed specifications are kept in mind.

Screening COTS

Lots or wafers of COTS semiconductors or hybrids can be screened for radiation tolerance to the same strict standards of quality and reliability as those demanded of RHA-certified parts. In fact, Defense Logistics Agency (DLA)-approved radiation-testing facilities conform to all relevant regulations applicable to the devices being tested, commonly using standards MIL-STD-750 or MIL-STD-883.

The higher the confidence of the designers in the required response of the parts to TID, displacement damage, or SEE, the better the chance of finding COTS wafers suitable for use in their projects. Out of the large number of mass-produced devices in a product line, there are always various wafers that happen to respond well to one or more of the types of radiation, even if they shift out of their original specifications. A COTS equivalent to a JAN high-reliability bipolar junction transistor (BJT) would likely fail one or more parameters in a post-radiation MIL-PRF-19500 Group D inspection, but it could still be entirely suitable for use in a project if the parametric shift is predictable, and therefore accounted for in the overall design.

The ultimate goal of a COTS radiation-tolerance-screening program is to make a project more cost-effective by avoiding RHA premiums wherever possible. However, this approach does come with some pitfalls that would defeat this purpose. Since finding wafers of devices that have favorable characteristics is random, deciding how many wafers of each device to screen is a significant challenge. If too many are screened, the associated cost no longer seems favorable, but testing only a single wafer pits the significant cost of radiation screening against the odds of the wafer responding in acceptable fashion. Selecting at least a few different wafers should improve the odds while still maintaining cost-effectiveness.

An additional strategy to assuring cost-effectiveness is to set up the tests that best characterize the conditions the device will experience in the final product. (See Figure 2.) This approach includes test points at well-spaced exposure levels that establish exactly when negative effects can endanger intended use, and – if relevant – considering annealing tests to determine if the part can shift back towards normal specifications.

22
Figure 2: Low-dose-rate radiation testing of components enables designers to screen parts for suitability for use in space. Photo courtesy of .
(Click graphic to zoom)

Victor Brisan is senior test engineer for VPT Rad in Chelmsford, Massachusetts. He holds a Bachelor of Science degree in Electrical and Computer Engineering from the Worcester Polytechnic Institute. Readers may reach Victor at vbrisan@vptrad.com.

VPT Rad www.vptrad.com