In previous chapters, the committee evaluated the status of the current radiation testing infrastructure and future needs expected to be placed on it. This chapter reiterates the committee’s key findings and recommends steps needed to establish an effective infrastructure that reasonably minimizes the gaps identified in the preceding chapters, to provide effective stewardship for the foreseeable future. Recommendations focus on facilities, workforce and training, databases, and standards and guidelines.
For the foreseeable future, beam time is likely to be a scarce resource, so cooperation to limit wasteful duplication of effort and increase efficiency will be key to ensuring all players have access. Chapters 3 and 4 describe the need to improve access to beam lines, sharing test fixtures and technical support, and reducing costs. There are many ways to accomplish these goals that include creation of partnerships, collaborations, quid-pro-quo arrangements, and dedicated facilities. Each is discussed in turn below, followed by recommendations on how they can be initiated.
Block funding among government agencies and collaborations with commercial space industries could result in greater assured access to test facilities. Included in these arrangements would be joint planning and scheduling, fixture designs, technical support, and data sharing. Also to be considered are joint technical road mapping among chip suppliers and space users to forecast linkages between future needs and commercial developments.
Public–private partnerships could also apply to joint investments for expanding the capacity of existing facilities, facilitating upgrades, re-purposing facilities scheduled for phase-out, and building new facilities. Partnerships could upgrade capability of a facility in return for fuller access. These partnerships should also facilitate bringing together the commercial CubeSat and SmallSat companies to join with big prime contractors in identifying key commercial-off-the-shelf (COTS) parts for testing and sharing of data.
CRADAs (Cooperative Research and Development Agreements) among government, industry, and academic participants could also be pursued for joint research, modeling, and testing to verify and validate emerging device technologies for radiation hard applications. An objective of these CRADAs would be to achieve greater synergies among government, industry, and academia knowledge bases to implement these technologies.
A persistent roadblock to such partnerships that requires a solution is the tendency by industry and other entities to hold their test decisions, procedures, and data as proprietary. Compelling cases need to be made to demonstrate that the benefits from synergy and leverage through sharing can be important offsets to such “close hold” tendencies.
Quid-pro-quo arrangements with beam facility operators are proposed to share costs for expansions, refurbishments, and technical upgrades in return for assured testing schedules at reduced costs. This approach could also apply to medical proton beam facilities (both existing or planned).
A dedicated facility with mixed-beam capabilities would be a desirable approach for assuring access for radiation hardness testing to meet the specific needs of the space community. Such a facility would have tailored test fixtures and especially trained technical support for more adaptable and cost-effective testing. It could also be an important asset for attracting industry and academic partners under the CRADAs proposed above. A logical location for a dedicated facility would be at a university that already has an accelerator and has personnel with accelerator design and operations experience. A university could provide basic learning opportunities for students that would eventually work in the SEE testing community. A university could also provide the land for the facility and would be tax free if operating as a nonprofit entity.
Finding: Cooperative mechanisms for assuring more timely access to testing beam time are required to meet development milestones.
Finding: 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 incorporate effectively validated parts in space systems.
Recommendation: The Department of Energy, in collaboration with the Department of Defense and NASA, should establish a joint coordination body to define the usage needs for parts radiation testing and assure the adequacy and viability of radiation test facilities out to 2030. The joint coordination body should be inclusive and recognize the needs of the broader space community.
Recommendation: The joint coordination body or an equivalently empowered entity should accomplish the following:
- A review of testing under way at facilities across the country and internationally;
- An assessment of survey test equipment availability and needs at participating institutions to facilitate sharing and to avoid needless duplication of hardware critical to testing state-of-the-art electronics;
- A strategic forecast of both government and commercial satellite launches that will require radiation-hardened microelectronic and optoelectronic (M&O) components to include reliability and lifetime requirements;
- A joint roadmap developed by representatives from commercial (M&O) device suppliers and the radiation-hardening testing community to ensure test procedures and facilities are capable of testing the latest electronics technologies;
A facilities plan, updated periodically, which includes the following:
- A projection of testing time availability of current radiation testing facilities, planned upgrades, and new facilities, including cost-effective strategies for increasing testing capacity and technical support;
- A review of reliability issues for critical systems at accelerators under current use, which identifies potential threats to sustained operation and the means to mitigate these threats; and
- An assessment of the business models and financial stability of critical accelerator facilities, which can affect total testing capacity and costs, including the possibility of a dedicated facility for electronics testing; and
- Mechanisms for incentivizing modeling and simulation capabilities, data sharing, and collaborations that can reduce total testing burden.
The committee notes that shutting down a facility for a period of only a few months in order to upgrade it will cost the facility revenue, potentially putting it at risk.
Although much can be gained by efforts in modeling and simulation, dispensing with beam testing in favor of more modeling and simulation will not help significantly. In particular, for commercial parts, it is unlikely that users will have sufficient fidelity to simulate and eliminate the need to test. In addition, the technology of parts changes rapidly, requiring new testing to develop and validate new models. Simulation could be done a priori with multiple models and then validation tests could be run looking for predictions that distinguish the best model or set of models. Even if modeling cannot replace testing, it certainly has the potential to reduce it and increase understanding of test results.
Recommendation: The Department of Energy (DOE), NASA, the U.S. Air Force, and other interested parties should stabilize funding for proton and heavy-ion accelerator facilities in order to restore resilience in national testing capabilities.
- At the Lawrence Berkeley National Laboratory (LBNL) cyclotron, NASA, DOE, and the U.S. Air Force should determine a method to increase beam time availability to the community to meet projected needs and to provide resiliency. The prior joint-stewardship program at LBNL was a model for how to exploit this existing U.S. capacity for heavy-ion testing.
- At Texas A&M University, support efforts to bring the K150 accelerator online for proton and heavy-ion testing.
- Facilitate advanced purchases to guarantee minimum beam time to both the proton and heavy-ion testing community. This will provide greater financial stability to LBNL and proton test facilities in the near term while ensuring access to electronics testers over the coming years. Without such advance purchases, LBNL in particular may need to make staffing and development decisions that harm the interests of the electronics testing community.
The committee notes that Texas A&M University (TAMU) has had a SEE testing line on the K150 cyclotron for the past 2 years and is offering proton beams. To offer heavy-ion beams from the K150 cyclotron for SEE testing, TAMU needs support for a modern liquid helium (LHe) refrigeration system. LHe is needed to improve the internal vacuum of the accelerator in order to accelerate heavy ions.
Although speculation about the future development of such a rapidly changing field is a fraught proposition, it is possible to assess likely near-term evolution of SEE verification, testing, and qualification methodologies based on current trends and likely future needs of the space community. Even at present, SEE rates for an increasing number of advanced technology parts cannot be predicted accurately by the relatively simple rectangular parallelepiped models (e.g., CRÈME-96) currently in wide use. As microelectronic devices continue to evolve and parts become even more complicated, modeling and simulation are likely to play an increasingly important role in everything from rate estimation to error propagation from the transistor to the system level. CRÈME-MC, Monte Carlo Radiative Energy Deposition (MRED), and other Monte Carlo–based SEE rate estimation packages will likely become increasingly important, and SEE testing will have to be closely integrated with the needs of
these packages to ensure reliable rate estimation. Unfortunately, Monte Carlo methods are inherently more time-consuming and resource intensive than current methods, and analysts will need to make each run count. One possible approach would be for analysts to develop one or more models of the devices under investigation based on past experience and knowledge of the device technology and then carry out simulated tests for a range of test conditions. Observation of the behaviors predicted in the simulations would then serve as validation of the models that predicted them.
Model-Based Systems Simulation will also likely play an important role in understanding the error and failure modes of greatest importance for the mission, and will likely also play an important role in deciding which parts are tested and how.
In addition to these developments in the evolution of conventional test methods, the future may also offer new options for radiation testing, two of which are mentioned here. The International Space Station (ISS) and various satellite missions, including CubeSats, offer an opportunity to monitor the SEE performance of important technologies in real space radiation environments. Such platforms can be invaluable for validating SEE testing and rate-estimation methodologies and, in some cases, for qualifying critical technologies. The downside to such platforms is that the testing cannot be accelerated, and on the ISS, it would take more than 1,000 device-years to accumulate the heavy-ion fluence equivalent to a single, low-linear energy transfer (LET) terrestrial test run. The only way to make such testing practical would be to fly very large numbers of each test device—a strategy that would require large amounts of power and test hardware in an environment where weight and power are valuable and finite resources. For this reason, such on-orbit test platforms are likely to be used mainly for methodological validation rather than qualification.
Another technique that is already popular for SmallSat builders is testing at the board or box level rather than at the part level. Such a strategy can save testing costs by consolidating several part-level tests into a single board- or box-level test and eliminating the need for costly test hardware to interface to each part. However, such board- and box-level testing can usually only be carried out when the design reaches an advanced, near-flight-like stage. A radiation failure susceptibility discovered at such a late stage can be extremely disruptive and costly, with potential cost and schedule implications far exceeding any savings realized by omitting part-level testing. Also, heavy-ion testing is very difficult to carry out at the board level, and likely impossible for higher levels of integration. Moreover, the test conditions, particle fluxes, and other test conditions are often limited by the weakest part on the board. These limitations mean that board-level testing is less likely to reveal all the radiation susceptibilities in the system—especially for destructive SEE modes. Lastly, unlike part-level testing, board- and box-level testing yield little or no information on issues such as the following:
- Part-to-part variation and its possible effects on flight system performance;
- Design margins and failure tolerance of the parts in the system; and
- Susceptibilities for conditions, operating modes, or logical states that are not covered during testing but that may be realized during the mission.
These limitations make it unlikely that board- or box-level testing will replace part-level testing, at least for mission-critical systems requiring high reliability.
Thus, in the foreseeable future, most testing of critical parts is likely to continue to be terrestrial, albeit with much closer integration to modeling and simulation.
Role of Career Planning and Training
Chapter 3 describes a current study of the distribution of numbers of NASA radiation engineers by age as a bimodal distribution, which results from a noticeable shortage of mid-career engineers. This raises a concern about transferring a full skillset of critical knowledge in testing, simulation, modeling, and design from experienced, later-career radiation engineers to early career engineers.
A growing need for knowledgeable radiation engineers in the private sector adds to these concerns, which call for exploring alternatives to “on-the-job” mentoring to impart expert knowledge through continuous education and sustained career planning and development. A major benefit from these workforce investments could be faster advances in closing loops among radiation-hardness device testing, modeling, simulation, design, and space performance.
The training of early career radiation engineers should entail more than testing devices based on procedural knowledge alone. These tests are only rough approximations of actual space environments. They do not adequately take into account the effects of spatial, temporal, and atomic mass variability in space particle fluxes and energy spectra, which in the case of solar particle ejecta can vary by orders of magnitude.
Currently, much of radiation-hardened device design and software mitigation strategy development is done in open-loop mode without the direct involvement of testing engineers with detailed knowledge of test protocols and data interpretation and analyses. Significant benefits could result from their direct involvement in validating and verifying LET and SEE failure-rate models through testing to improve their applicability. Closing the loops among radiation-hardness design, testing, simulation, and statistical failure analysis could result in improved reliability forecasts and decreased testing burdens.
A properly constructed early career training agenda would include the following:
- Career planning and implementation to increase job interest and retention;
- Summer schools tailored to impart mid-career knowledge and skills at an earlier career stage; and
- Continual education through attendance at technical conferences, summer schools, university certificate programs, and short courses (the use of the short course materials, available in the public domain, such as the Nuclear and Space Radiation Effects Conference and the Radiation and its Effects on Components and Systems conference, are recommended in Chapter 3).
Some of these training programs, sponsored by government laboratories, professional societies, and universities, currently exist and frequently involve senior space radiation engineers. However, the funds made available for younger engineers to attend them are too often limited.
Finding: 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.
Finding: Specialized training programs can be important means for building and sustaining critical skills of early career radiation engineers. However, funding resources are often not available.
Role of Summer Schools
Historically, summer schools have been effective in developing emerging workforces to address national needs. Examples include the following:
- Oak Ridge School of Reactor Technology (ORSORT). This school, which operated from 1950 to 1965, provided 976 graduates in reactor operations and hazards analysis at a time when university nuclear engineering programs were at a nascent stage.
- M.I.T. Reactor Safety Summer School. This summer school of 2-weeks’ duration conducted during the 1960s and 1970s provided a cadre of nuclear safety experts in critical need for the government, nuclear supply industry, and electric utility industry.
- LANL Information Science and Technology Institute. This institute currently conducts a number of specialized schools in such subjects as data science, computational science, cyber security, computer clusters, and networking to provide key talent for the “information age.” The U.S. Particle Accelerator School run by LANL and the Space Weather School are directly relevant to workforce development for radiation hardness beam testing.
- The NASA Space Radiation Summer School (NSRSS) at the U.S. Department of Energy’s Brookhaven National Laboratory. This summer school, which has now operated for more than a decade, trains top talent in understanding and tackling the effects of space radiation environments on human health during space exploration.
Finding: Summer schools have played an effective means for training technical talent for careers of high national interest.
Recommendation: The Department of Energy, NASA, and the U.S. Air Force should cooperate with professional organizations (e.g., the Institute of Electrical and Electronics Engineers) and other interested parties to accelerate career development of the younger testing and modeling scientists and engineers through summer schools, short courses, university certificate programs, and internal mentoring to enable them to more rapidly achieve mid-career proficiency levels.
Role of Universities
Today’s university graduate talent engaged in researching radiation effects to materials would be critical assets for building the radiation hardness technology base for the future. They must, however, be well grounded in space weather, radiation effects to microelectronics and optoelectronics, Monte Carlo LET modeling, and accelerator testing protocols.
Some universities are well equipped with proton and ion beam accelerators for radiation effects research (e.g., the Michigan Ion Beam Laboratory at the University of Michigan and the Center for Materials Under Extreme Environments at Purdue University).1 These facilities focus primarily on studying changes in physical states at surfaces and interfaces under irradiation by laser beams; plasmas; and electron, proton, and low-Z ion beams. Only a small number of universities investigate the effects of irradiation on microelectronic circuits. Furthermore, while some universities have strong capabilities in using Monte Carlo neutron codes for reactor safety analyses, only a few offer training in applying these codes to investigate radiation damage to microelectronic devices.
University beam lines are not comparable to high-energy accelerators currently employed for SEE testing of microelectronic devices over a range of megaelectronvolt per nucleon levels. However, they can be employed in combination with high-powered lasers for both simulating space displacement damage dose (DDD) and total ionizing dose (TID) damage and testing for soft errors and upsets in optical sensors, optoelectronic devices, and optical communication components. Defects of particular interest are color centers that can darken optical sensors, charge centers, and defect clusters that can reduce the efficiency of light-emitting diodes and charge-coupled device arrays, surface erosion by energetic ions, and photon scattering sites in optical communication components that can increase source and channel noise and reduce bandwidth.
DDD and TID effects are becoming evident in deep-submicron technologies. Such damage modes may become more problematical in “beyond Moore’s Law” technologies and during extended space missions. Forms of such degradation can include the following: strained-lattice relaxation, changes in state functions at ions and electrons, and impedance in the movement of domain and phase boundaries due to defect pinning.
Finding: 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.
1 For example, CMOS on insulator is an example of a technology that is inherently more radiation hard, Transistors are isolated and protected from SEL. The thinness of the silicon limits charge collection. Currently, nearly all deep submicron hardened designs start with SOI processes and implement circuit-level hardening for the cells. See http://mibl.engin.umich.edu and Center for Materials Under Extreme Environment (CMUXE) at http://engineering.purdue.edu.
Finding: Universities operate some of the large accelerators used in simulating space radiation for radiation hardness 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
Recommendation: The joint coordination body should assess and support university capabilities for improving space electronics testing and development infrastructure, including the following: the development of advanced accelerator concepts, improved testing strategies, improved radiation hardening solutions designs, and radiation mitigation techniques.
Validated databases, standards, and protocols are important requirements for an infrastructure for space radiation hardness testing, as follows:
- Support a supply base that can produce highly specialized devices in small lot sizes cost-effectively;
- Reduce test redundancy in order to improve beam time availability and reduce total testing costs;
- Improve quality and reduce risk uncertainties; and
- Support fault detection, isolation, and recovery on mission.
Currently, there are databases at the Jet Propulsion Laboratory, NASA Goddard Space Flight Center, and the European Space Agency that list tested parts. However, a need exists to provide outside users greater clarity in the specific nature of the parts, tests, findings, and applications. There is also a complication in accessing test data from commercial satellite builders due to their proprietary interests.
There also exist standards for TID, SEE, and single-event burnout/single-event gate rupture testing; however, applying these generalized test standards to a variety of specific device applications is problematic.3 In contrast, databases have more application-specific test results than they did in the past. The models and methodologies for applying such data to other applications do not exist at present (with the exception of some work of this nature for single-event transients at Vanderbilt University and the Naval Research Laboratory).
Finding: 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.
Recommendation: The joint coordinating body should engage with the commercial space sector to ensure testing norms meet the needs of this sector as well as the conventional satellite design and radiation testing communities.
2 See, for example, D.B. Zho, A. Pukhov, L.Q. Yi, H.B. Zhou, T.P. Yu, Y. Yin, and F.Q. Shao, Laser-driven ion acceleration from plasma micro-channel targets, Scientific Reports 7:42666.
3 See military standards for TID and displacement damage; JESD-57 and ASTM F1192 guidelines for SEE testing; MIL-750, Method 1080 (ESA has ESCC-25100) for SEB/SEGR testing.
Recommendation: The joint coordination body, in combination with existing working groups, should establish a mechanism to (1) assure the preservation and maintenance of existing modeling and simulation codes for the analysis of space radiation effects on microelectronic and optoelectronic components and (2) support basic research for the development of new codes.
The mechanism could be an informal coalition within the community or the creation of a multi-university center of excellence or a public–private partnership for private support, cooperative government basic research support through agency research offices, and CRADAs for joint technology development and transfer. It could also invite international space agencies and private companies to assist in supporting and maintaining this essential and indispensable resource.