Summary

Spacecraft depend on electronic components that must perform reliably over missions measured in years and decades. Space radiation is a primary source of degradation, reliability issues, and potentially failure for these electronic components. This report of the Committee on Space Radiation Effects Testing Infrastructure for the U.S. Space Program evaluates the nation’s current capabilities and future needs for testing the effects of space radiation on microelectronics to ensure mission success and makes recommendations on how to provide effective stewardship of the necessary radiation test infrastructure for the foreseeable future. Although simulation and modeling is valuable for understanding the radiation risk to microelectronics, there is no substitute for testing, and an increased use of commercial-off-the-shelf (COTS) parts in spacecraft may actually increase requirements for testing, as opposed to simulation and modeling. Although the effort of testing may be difficult and expensive, it is small compared to the cost of a radiation-induced failed mission, which can be hundreds of millions of dollars, not to mention the inability to conduct the mission.

This study was conducted at the request of the U.S. Department of Energy (DOE), the U.S. Air Force, and NASA, all of which have an interest in the testing infrastructure for space electronics. It was initiated because these organizations are concerned about the adequacy of the current testing infrastructure and its ability to meet the growing demands for national security, civil, and commercial space systems.

Radiation that threatens space missions derives from three main natural sources: galactic cosmic rays (GCRs), consisting of protons, electrons, and ionized heavy nuclei; charged particles (mainly protons and electrons) trapped by planetary magnetic fields (e.g., Earth’s Van Allen belts); and solar particle events (SPEs) that occasionally flood regions of space with large fluxes of energetic protons and heavier nuclei. Together, these sources create a radiation environment of high-energy electrons with energies from a few electron volts (eV) to as high as tens of millions of electron volts (megaelectron volts, MeV) (as in the Jovian environment), protons with appreciable fluxes at energies from a few electron volts to hundreds of megaelectron volts, and heavy ions with energies from ~1 MeV per nucleon to several thousand megaelectron volts per nucleon.

Spacecraft electronics are susceptible to radiation effects that emerge from interactions with these energetic particles—both degradation and eventual failure—due to total ionizing dose (TID) and displacement damage dose (DDD) and the instantaneous response of the electronics to single ionizing particles, called single-event effects (SEEs). If electronics are not “hardened” to both cumulative and single-event radiation effects, they will likely experience these effects in space, resulting in performance anomalies and the potential compromise of space missions. SEE testing fulfills several roles, including the following: hardening of electronics technology; measuring

susceptibility of off-the-shelf parts (be they commercial or military); and, assessing system-level hardening in a realistic environment.

After reviewing the facilities, methodologies, and expertise for bounding the threats due to TID, DDD, and electrostatic discharge, the committee believes that the test infrastructure for these effects does not experience the same level of strain and fragility as the infrastructure required for testing SEEs. Most TID testing is done with gamma-ray sources, so the infrastructure requirements are not nearly as intensive as those for SEEs. A gamma irradiator can fit in a small room and does not require all the power, cooling, and other resources that a particle accelerator does. There are safety requirements, but these are manageable. SEEs are also less well understood. Depending on the technology of the affected device, SEEs can result in consequences ranging from self-recovering disturbance of device outputs to catastrophic failure of the device. The infrastructure is not under as much strain with the other threats as it is for SEEs and can meet the threats for the foreseeable future. As such, this report focuses on one specific area of radiation damage: single-event effects.¹

The frequent changes in the microelectronics market also require frequent retesting and requalification of the same parts due to changes made in the design and production of those parts as time progresses. Testing for radiation hardening is difficult and costly. Revealing the susceptibilities of a part requires exposing it to high fluences of high-energy ions. Depending on the device being tested and the goals of the test, the effort may cost from $25,000 to more than $600,000 and require access to an ultra-high-energy heavy-ion accelerator. Moreover, even if one can make do with a more limited test requiring fewer hours at the accelerator, much (~70 percent) of the cost of testing is expended even before the test team reaches the accelerator, limiting the savings that can be realized by running a less ambitious test. Another reason radiation testing of electronics is costly, time-consuming, and challenging is because the parts being tested and the mission radiation environments in which they are to be used are complex and because testing involves the collaborative efforts of many highly trained engineers, technicians, and other professionals.

CURRENT RADIATION TESTING INFRASTRUCTURE

The current infrastructure for radiation testing has the following five elements: facilities for testing, standards and guidelines that detail conditions of testing, modeling and simulation tools that predict the response of parts and systems to radiation, databases that can help minimize the duplication of test efforts, and workforce and training to ensure the ability to conduct the needed tests.

Today, 15 U.S.-based facilities provide the radiation testing infrastructure required. At most facilities, radiation testing is a secondary function, after high-energy physics research (e.g., at National Laboratories such as Brookhaven National Laboratory) and cancer treatment (e.g., Massachusetts General Hospital). These facilities, by necessity, are located throughout the United States (with the exception of the Canadian TRIUMF facility) (see Figure S.1). Overall, as demonstrated by the decommissioning of the proton cyclotron at Indiana University, the bankruptcy of the Scripps proton accelerator, and ongoing financial instability at the critical Lawrence Berkeley National Laboratory (LBNL) heavy-ion accelerator, these facilities are facing financial and other strains. For example, many of the accelerators in use for SEE testing are 30 to 55 years old, past their design life, and a failure of critical systems is becoming more likely—especially if the financial health of the facilities does not permit increasing maintenance and replacement of vulnerable components prior to failure. Radiation testing is usually a secondary feature of these facilities and, therefore, dependent on the primary functions, although income from SEE testing has also been used to modernize and replace antiquated equipment to reduce the risks of critical system failures, benefiting both science and SEE testing programs. While under the financial and demand stresses discussed above, these facilities meet current needs.

Modeling and simulation (M&S) come into play in many different aspects of the radiation hardness assurance process. Analysts use models of radiation environments to predict the response of parts and systems to both the

___________________

¹ The construction and operation of accelerators used for radiation effects testing with heavy ions have been funded by the nuclear and high-energy particle physics communities. In 2015, they produced a report on the long range plans for these facilities. It can be found at under the reports heading at: https://science.energy.gov/np/nsac.

background (GCR and trapped radiation) and worst-case (SPE) mission environments. One area that has been a longstanding gap has been tracking the consequences of SEE modes from the transistor or gate level up through the circuit and system level. Recently, Model-Based Systems Engineering (MBSE) and Model-Based Mission Assurance (MBMA), while still at an early stage of development, have shown the potential for filling at least some of this gap.

Databases of radiation test results minimize duplication of test efforts and—if they contain data on parts similar to those of interest—can in some cases increase confidence in those parts prior to the availability of radiation test data. Databases of existing data can also help to validate the models. Radiation data are available from vendors, from NASA’s Jet Propulsion Laboratory and Goddard Space Flight Center, the European Space Agency, and in various data workshops and compendia associated with conferences on radiation effects, such as the Institute of Electrical and Electronics Engineers (IEEE) Radiation Effects Data Workshop. Although these sources provide data on large numbers of parts, use of these data poses challenges in that some results may be application specific and not valid for other applications; the product life cycles for commercial parts are so short that data may have little value soon after they are published; and the number of different commercial parts and technologies make it unlikely that a commercial part selected at random will have been tested previously. However, even given those

limitations, available databases are often an under-utilized resource, and their value to the community could be increased considerably if they were properly organized, the limitations of each data set made clear, and “big data”type models developed to extract as much information as possible to use on complex hardness assurance problems. While such efforts could be invaluable, especially for small satellite (SmallSat) builders with limited radiation testing budgets, money for maintaining existing databases continues to be a scarce resource.

Some parts are specifically designed with features and technologies that enhance radiation hardness. These radiation-hardened parts lag behind their commercial counterparts in performance, but they play a crucial role in radiation hardness assurance. Not only are such parts essential for critical functions in mission hardware, the fact that their radiation performance is guaranteed also keeps radiation analysts from being overwhelmed by too many time-consuming and labor-intensive radiation tests. If the majority of functions in the spacecraft can be fulfilled by radiation-hardened parts, the radiation analysts can concentrate their efforts on the state-of-the-art parts that will significantly enhance the capabilities of mission avionics and electronic systems. At present, vendors of radiation-hardened parts continue to produce new parts in new generations of radiation-hardened complementary metal–oxide–semiconductor (CMOS) and bipolar technologies. Although these parts are typically produced in small volumes (at least compared to the commercial microelectronics sector), most vendors have indicated intent to maintain production and development of these critical parts. At the same time, a wave of acquisitions and mergers has swept through the semiconductor industry in recent years, affecting commercial and radiation-hardened producers alike. It remains to be seen whether these changes in management will result in changed priorities as well.

A central element of the infrastructure is the workforce to conduct the needed tests. Experts conducting the tests must be familiar with radiation transport in materials, semiconductor physics, semiconductor device design, space radiation environments, and a variety of other fields. Given the rapid pace of change in electronics, continuing education through conferences, short courses, and workshops is critical to ensuring radiation analysts are familiar with the latest developments. However, in addition to keeping up with the latest developments, there are also skills that must be developed over time—for example, data analysis and testing techniques and how radiation engineering integrates into the spacecraft design process.

Although radiation engineering suffers from the same difficulties as most highly technical disciplines in finding qualified personnel, for the most part, experience and training are adequate to meet the current needs of the community. However, the workforce is aging and many are likely to retire before passing along their full skillset to younger workers unless extraordinary efforts are undertaken to avoid such a gap. In addition, the growth of commercial space companies means that many engineers are having to learn about radiation hardness assurance without the benefit of experienced colleagues at the same company—and not all of these companies are willing to send their engineers to radiation conferences or courses. As a result, there is a growing gap in the knowledge base of the growing workforce. Filling this gap is likely to require coordination of extraordinary efforts by interested parties—including NASA, DOE, the Department of Defense, and other government agencies, universities, and professional societies—to disseminate knowledge of best practices and the emergence of new threats, trends, and resources.

FUTURE INFRASTRUCTURE NEEDS

The space sector as a whole is evolving rapidly. In the coming decade, the number of spacecraft as well as the number of operators in space is expected to increase by one or two orders of magnitude. Commercial operators, especially those proposing smaller, less expensive satellites, are more inclined to use COTS parts (some of which may be sourced from other countries where their radiation testing history is unclear) and simultaneously less inclined to conduct high-cost radiation tests. There can be many reasons for this approach: they may be seeking reliability via greater numbers of cheaper satellites; they may not have the expertise, budget, or facilities to use radiation-hardened parts or test their COTS parts; or, if their satellite’s design lives are short enough, they may decide to accept the risks associated with operating in the space environment without testing or hardening.

It is not just components that would change; there are likely to be wholesale changes in systems based on emerging technologies—for example, use of optical communication devices instead of radio communication.

Functions are similarly going to change—there is already a growing focus on greater autonomy and capability of spacecraft, the need for more precision as we embark on tasks such as robotic assembly and manufacturing, and automated guidance, navigation, and control. Lastly, electronics are changing as well—CMOS technologies are likely reaching their limits—even current CMOS devices have different topologies than past planar CMOS generations. Devices will likely exhibit new SEE failure modes and require new rate estimation tools that will call for different types of radiation tests.

On the testing side, there will be both new opportunities and challenges. For example, in the M&S community, future rate estimation methods will likely rely on Monte Carlo modeling and high-resolution computer-aided design (CAD) models of the device structures followed by validation testing using heavy ions. There will also be more opportunities for in situ testing on the International Space Station (ISS) or commercially available low Earth orbit (LEO)-based platforms. It is expensive. Ion fluxes for the ISS are ~30 ions/cm² per day (with a linear energy transfer [LET] value greater than 1 MeVcm²/mg), so to complete in 1 day the equivalent of a single heavy-ion test run would require ~1,000 devices. That is a lot of power in an environment where power is not inexpensive. It also requires launching the test hardware, which costs ~$10,000/lb. In situ testing is not a solution, but a good approach for validation. However, these tests would not be accelerated—hundreds or even thousands of device-years would be needed to accumulate fluence equivalent to a single heavy-ion accelerator test run. Further, hundreds of devices would have to be flown to accumulate statistics for on-orbit tests. Finally, LEO-based platforms are not suitable for electronics intended for deep space use.

Among the possibilities that could significantly alleviate many of the concerns discussed here are advanced ion acceleration concepts, likely driven by high-powered lasers. Such methods could reduce the size and perhaps the cost of ion accelerators, alleviating some of the fragility the field experiences with regard to access to heavy-ion accelerators. Unfortunately, such techniques are likely a long way from being practical for accelerating heavy ions (Z > 40) to sufficiently high energies for use in SEE testing. This development could be important, but is unlikely to be realized before 2030—the period with which this study is mainly concerned.

Another potential development on the testing side involves conducting radiation testing at higher levels of integration rather than at the level of the individual part. Although such a strategy can save testing costs and schedule through consolidation, these savings can only be realized if the test reveals no adverse results requiring a redesign, and if risk can be bounded despite the limited amount of information that can be ascertained from such a test.

OVERALL FINDINGS AND RECOMMENDATIONS

The committee has found that the radiation-testing infrastructure system is fragile; it is already experiencing long wait times and rising testing prices, and it could easily suffer major strains if even a single major facility closes down suddenly. The possibility of a sudden closure was realized in 2014 when the Indiana University Cyclotron Facility (IUCF) ceased operations (due to declining revenues from its primary business of treating cancer patients). The combination of this fragility and overloading of current beam-line facilities for space radiation testing, together with the growing complexity of commercially available microelectronic and optoelectronic systems that will further strain the system, and increasing requirements for accelerator testing by the private sector, all together project a growing shortage of available testing facilities to support future space missions among space agencies and industry. Overcoming roadblocks to sharing test procedures and data, including those related to confidentiality of proprietary data or restrictions related to classification of sensitive information, will be increasingly critical to avoid duplication of test efforts and to efficiently incorporate validated parts in space systems. The most important step to take is therefore to provide some high-level strategic coordination across the radiation testing community.

An apparent bimodal distribution in the radiation testing workforce exposes the risk that critical knowledge may not be transferring at a sufficient rate from mid-career to early career radiation engineers. Specialized training programs can be important means for building and sustaining critical skills of early career radiation engineers. Summer schools have played an effective means for training technical talent for careers of high national interest. The Nuclear and Space Radiation Effects Conference (NSREC) and Radiation and its Effects on Components and Systems (RADECS) short courses in particular have been a valuable resource for education of past generations of radiation engineers and would be equally, or more, valuable to the next generation. Graduates trained at universities that have beam-lines, plasma chambers, and high-energy laser test facilities and who have been trained in Monte Carlo modeling of radiation effects can be valuable for filling future losses in the current talent base. However, funding for specialized training appear not to be pre-planned.

Universities operate some of the large accelerators used in simulating space radiation for radiation hardening testing. Furthermore, there are additional universities that have sophisticated beam facilities employed in integrating radiation effects theory with research, modeling, simulation, and prototyping. These capabilities can contribute to (1) advancing space radiation testing, (2) discovering new mitigation strategies for improving system reliability and lifetime, and (3) sustaining the future workforce for this field.

One possible area of development involves the use of laser-driven plasmas to accelerate light and heavy ions. Such technologies could shrink ion accelerators to table-top size, allowing greater access to high-energy ion beams and reducing beam costs.²

The rapid development of semiconductor devices means that the body of knowledge for the field advances more rapidly than it can be accommodated in test standards. As a result, the standards are a mix of specific data and general guidelines, and significant experience and skill are required to adapt these resources to the requirements for testing a new part or technology.

Optoelectronic and photonic devices are susceptible to displacement damage and single-event transients. Often transients can be caused by direct ionization even by high-energy protons. Displacement damage studies are usually carried out using medium-energy protons.

The current system for radiation testing of electronics in the United States is working but increasingly more fragile. And fragile is not an acceptable state, given the importance of this type of testing to space electronics, which is irreplaceable in national security, civil space exploration, and the burgeoning commercial space sector. This report presents a way forward for the United States to restore the robustness of the radiation testing infrastructure before its fragility becomes a stumbling block to U.S. leadership in space.

___________________

² See, for example, D.B. Zhou, A. Pukhov, L.Q. Yi, H.B. Zhuo, T.P. Yu, Y. Yin, and F.Q. Shao, Laser-driven ion acceleration from plasma micro-channel targets, Nature Scientific Reports 7:42666, doi:10.1038/srep42666, 2017.