As electronic components have decreased in size and increased in complexity, their enhanced sensitivity to the space radiation environment and its effects (especially single-event effects, or SEEs) have become a major concern for the spacecraft engineer. In response, the determination of how that radiation propagates through material and how radiation affects specific circuit components have become important considerations in the design of modern spacecraft. This chapter reviews the environments responsible for the effects of radiation on microelectronic components and describes the effects these environments have on microelectronics.
The spacecraft designer addresses these concerns by first defining the environments the mission is expected to encounter. In turn, these environments are used to determine, through direct testing or modeling, their potential effects on spacecraft performance. The results of these studies are then used to either modify the design or make use of mitigation techniques, such as radiation shielding to limit the impact of the environment.
Spacecraft must operate in an environment with pervasive radiation consisting of a mixture of particles that, depending on the location, can vary only negligibly in intensity over many years or change intensity by many factors in as fast as a few hours. Figure 2.1 summarizes the general types of space radiation and where they are relevant. The key point illustrated in Figure 2.1 is that space is not a single place when discussing radiation, but a broad range of environments and conditions. Even within the set of well-used Earth orbits, the radiation environment differs greatly with location and over time—not only the intensity, but also the dominant particles and energies of these particles. Low Earth orbit does not have the same radiation exposure as other Earth orbits, orbits around other planets, or interplanetary space. This diversity of the radiation environments significantly complicates the ability to qualify electronics for space missions.
SEEs—the prompt response of semiconductor devices to the passage of a single ionizing particle—are discussed in more detail below. The high-energy-particle radiation environment responsible for SEEs on electronics consists of protons with energies greater than 1 MeV and heavy ions with energies greater than 1 MeV/nucleon. The range of penetration of these particles increases significantly with incident particle energy. Particles with energy close to 1 MeV/nucleon are stopped in a fraction of a millimeter of aluminum, while particles with energies greater than 1,000 MeV/nucleon (>1 GeV/nucleon) could go completely through a spacecraft.
The remainder of this discussion of space environment radiation will be divided into the following three categories of radiation:
- Galactic cosmic rays (GCRs), which consist of interplanetary protons, electrons, and ionized heavy nuclei;
- Trapped radiation (especially Earth and Jupiter); and
- Solar particle events (protons and other heavy nuclei).
The first two sources are relatively constant or change on time scales of days to years (e.g., the 11-year solar cycle). The third is highly variable, on the scale of minutes to a few days, in response to events on the Sun.
Galactic cosmic rays are primarily protons and ionized heavy nuclei of galactic and/or extragalactic origin with energies ranging from ~1 MeV/nucleon to more than ~10,000 MeV/nucleon. The major components are hydrogen (89 percent) and helium (10 percent). The remaining 1 percent includes the rest of the commonly known elements. GCR electrons do not typically contribute significantly to radiation effects.
The flux of GCRs in interplanetary space increases slowly with distance from the Sun. The solar wind also modulates the GCRs inversely with the solar cycle. At the maximum of solar activity, GCR intensity is at a minimum, and vice versa. This modulation is caused by changes in the solar wind magnetic field. Turbulence in the field (such as that associated with solar maximum) makes propagation of GCRs into the inner solar system more difficult—the particles are scattered more effectively. As the solar wind field relaxes during solar minimum, GCRs can more easily reach the inner solar system.
Earth’s geomagnetic field provides shielding against incident GCRs (and solar energetic particles) because it can effectively deflect the lower-energy particles. Because of the approximate dipole nature of the geomagnetic field, particles with near vertical velocities in the polar regions are essentially parallel to the magnetic field, so
that they can gain direct access to regions near Earth’s poles. At lower latitudes, only particles with sufficiently high energy can penetrate through the magnetic shielding.
Trapped radiation on Earth consists principally of energetic protons and electrons, with lesser percentages of heavy ions such as O+, contained in toroidal belts by Earth’s magnetic field. This toroid is commonly known as the Van Allen belt(s) and consists of (at least) two zones: a low-altitude zone, or inner belt, and a high-altitude zone, or outer belt. The inner belt extends from hundreds of kilometers to ~6,000 km in altitude and is populated by high-energy protons (tens of MeV) and high-energy electrons (1–10 MeV), while the outer belt, up to 60,000 km in altitude, is predominately high-energy electrons. The inner belt zone, because of the dominance of Earth’s main field, is relatively stable. Most temporal variations in this population occur as the solar cycle proceeds and Earth’s neutral atmospheric density at a given altitude changes, causing variations in the altitude at which radiation particles are scattered.
In contrast, the outer belt, which is more influenced by Earth’s highly variable geomagnetic tail, experiences much greater temporal fluctuations. The electron concentration in the outer zone may vary as much as a factor of 1,000.
As in the case of Earth, many of the other planets in the solar system have been observed to have trapped radiation belts. The species, abundances, energies, and time variations of particles that are trapped in these radiation belts vary greatly depending on the planet and its magnetic field. Planetary magnetic fields influence the particle spectrum that is observed near a planet in two ways: First, the magnetic field of the planet shields the planet from solar particles and from GCRs and, second, it allows particles to be trapped near the planet in radiation belts. Earth, for its size, has proportionally one of the most intense magnetic fields in the solar system. Jupiter and Saturn are roughly 10 times the size of Earth, while their magnetic moments are ~2 × 104 and ~103 larger. Because magnetic fields scale roughly as the inverse of the cube of distance from the planet, Saturn’s magnetic field is comparable to Earth’s, while Jupiter’s magnetic field is 20 times larger than Earth’s. The maximum energy and fluxes of trapped particles on a magnetosphere are proportional to its magnetic field strength, so Jupiter has much higher particle energies than those in the magnetospheres of Saturn and Earth or in interplanetary space. Saturn’s environment is roughly equal to Earth’s (see Figure 2.2). As the United States continues exploration of the solar system, satellite design engineers and mission leaders need to keep in mind the unique radiation environments of the planets visited.
The Sun continually emits a “solar wind” of charged particles with kinetic energies largely between 1 eV to 10,000 electron volts (10 keV) that move radially away from the Sun at densities on the order of 10 particles per cubic centimeter superimposed on the background solar wind are high-energy particle bursts. In contrast, solar particle events (SPEs) are periods of significantly enhanced fluxes of particles (solar energetic particles, or SEPs) accelerated to high energies by solar electromagnetic fields and associated with periods of high solar activity. For periods of hours to days, the intensity of particles with kinetic energy well above the megaelectron volt range, and as high as a few gigaelectron volts is increased by many orders of magnitude over background solar wind levels, and also orders of magnitude more intense than galactic cosmic radiation at comparable energies. These particles are predominately protons, with a roughly 9 to 10 percent addition of alpha particles, but energetic particles with higher nuclear charge (Z) are also enhanced. The relative amounts are consistent with the abundances present in the solar atmosphere and interplanetary space (heliosphere), but the details of the abundances vary significantly from event to event. Because the particles are so energetic and numerous, they pose a substantial risk to electronics. Figure 2.3 shows the total fluence of some of the more significant SPEs.
The rate of SPEs varies with solar activity through the roughly 11-year solar cycle. While it is reasonable to say that the probability is lower during the 3 or so years centered on solar minimum, the probability does not
necessarily peak within a few years of solar maximum. SPEs are somewhat more likely during the declining phase of the solar cycle, and there is no correlation between event size (when one does occur) and its time within the cycle. One of the largest events on record, the August 1972 event that occurred between the Apollo 16 and 17 missions, was 3 and a half years out from the nearest solar minimum.
The peak flux from an SPE drops off with distance from the Sun, but not in any simple way. Impulsive events, being driven by a localized source on the Sun, drop off roughly between 1/r2 and 1/r3. Gradual events, driven by a rapidly evolving shock moving out from the Sun, have no easy characterization. The peak flux may even increase out to one astronomical unit (AU) before starting to decline at a low rate.
Figure 2.4 summarizes the possible impacts of space radiation on a spacecraft. Often the engineering and operating communities have learned about how radiation interacts with satellites to create these hazards the hard way—by flight experience with actual anomalies. Not every satellite anomaly arises from space radiation, and it is difficult to quantify exact likelihoods of radiation-induced anomalies in currently orbiting systems for a number of reasons, including reporting requirements, proprietary information restrictions, and the common practice of categorizing anomalies as known behavior only once root cause and mitigation have been established.
Figure 2.5 demonstrates the occurrence of the most important spacecraft anomalies caused by the on-orbit environment. Electrostatic discharge (ESD), single-event effects (SEEs), and cumulative radiation damage are the major concerns. Electrons are the primary driver of satellite charging that can lead to ESD. Total ionizing dose (TID) is mainly an issue at end of life for the spacecraft, and TID failures are often (although not always) preceded by gradual degradation. As a result, the main emphasis of this report is on SEE-related effects. SEEs are also more important to study because new electronics technologies can introduce new SEE susceptibilities that cannot be determined reliably without testing. In addition to size and density trends, the increased functional complexity of the devices and the dynamic nature of their operation have made it more difficult to test devices for SEE impact in all possible operating modes. For example, a single event upset at one location may be inconsequential while
executing one functional string and overwriting in the next cycle, but could be more significant, even catastrophic, while executing a different functional string.
High-energy particle interactions can destroy, degrade, or disrupt the electronic systems critical to spacecraft operations. SEEs are caused by a single particle impacting electronics in the spacecraft. This is in contrast to TID, which is a cumulative effect over the life of the spacecraft. A common aspect of SEEs and TID is that they both result from free charge left by ionizing particles. In the case of SEEs, there is free charge displaced by a single particle as it passes through matter. In the case of TID, there is a cumulative impact from collection of free charge from radiation passing though material over the duration of the mission.
SEEs can occur at any point in a mission and in all space environments when a charged particle traverses a sensitive volume in an active semiconductor device, depositing sufficient charge to disrupt the normal functionality of that device (see Box 2.1). The particle may be primary—originating from the space environment itself—or it may be secondary—generated via interactions between the primary knock-on particle and materials in and around
the sensitive part. What happens to this free charge determines whether the radiation effect is detectable in the operation of a device (e.g., a memory chip) or a subsystem (e.g., a camera payload on a satellite).
As Box 2.1 explains, when a single particle passes through an electronic component, it ionizes atoms and generates small amounts of free charge. If that charge collects within the component in a way that mimics or disrupts the normal operation of the device, then a spurious outcome is possible. For example, it may change the state of a memory or it may provide a random input to or output from a device. In more extreme instances, the charge may create a short circuit, and the subsequent current through the short may destroy all or part of a component. Depending on the electronic component’s vulnerabilities, the operation of the component, and the timing of the SEE, the results range from recoverable effects to catastrophic failure of the entire system. SEEs can occur at any time during the mission—they are just as probable at the beginning of the mission as at the end (known as Poisson processes).
For nondestructive SEEs, the consequences of the SEE depend on what the device was doing when the SEE occurred. If the overall logic of the processors is able to isolate, ignore, or repair the damage, these spurious results do not result in a significant or long-term impact on the spacecraft. The ability to handle SEEs depends significantly on the ability of the designers to anticipate what may happen to the spacecraft throughout its operational lifetime in all of the relevant operational environments. This is enabled when component radiation response is characterized well enough (either through heritage experience or testing) so that the effects of radiation-induced errors and failures can be modeled before incorporating the devices in the spacecraft.
Having discussed the space radiation environment and its effect on electronics, in Chapter 3 the focus is on how spacecraft designers and component manufacturers assure the radiation hardening of electronics via testing at various facilities throughout the United States and a discussion of those facilities.