Lasers Simulate Space Radiation Effects

Susan Humphrey, Stephen LaLumondiere, and Steven Moss

Imagine that Serbian terrorists have assembled in preparation for an attack on nearby NATO troops stationed in Kosovo as part of a newly negotiated peace agreement. Unbeknown to the Serbs, their every movement and every communication have been followed by national intelligence assets —satellites—orbiting unseen in space. As the Serbs prepare to attack, allied commanders make use of intelligence information from these assets to direct their preparations and response.

At the most critical moment, however, the flow of information is suddenly and unexpectedly cut off. The normal operation of the satellites has been disrupted by intense solar flare activity. The resulting surge of extremely energetic particles rips through sensitive microelectronic devices onboard the satellites, producing single event phenomena (SEP).

SEP disrupt device and system performance and, in some cases, may even result in the destruction of the devices or systems affected. Whether a scenario like this ever occurs depends upon our ability to develop space systems containing microelectronic devices that are tolerant of SEP.

Single Event Phenomena

Microelectronic devices are susceptible to damage or interruption from exposure to radiation. Such devices contain structures made up of semiconductor materials that operate by regulating current flow or the amount of electrical charge held in a potential well. Interaction with radiation, including any energetic particle (electrons, protons, neutrons) or photon (gamma rays, X-rays), alters these precisely regulated conditions and produces SEP (see sidebar, Types of Single Event Effects).

Cosmic rays traversing a microelectronic device generate a track of charge that can alter precisely regulated current flow or charge storage. Single event phenomena can cause temporary or permanent damage to such devices.

SEP are classified by the type of effect that occurs in the device, known as a single event effect (SEE). Types of SEE include single event upset, latchup, burnout, gate rupture, and total dose.

Testing microelectronic devices for their degree of vulnerability to SEP is therefore integral to designing devices to be used in space systems. The Aerospace Corporation is developing a system to generate short-pulse X-rays for SEE testing of electronics that promises to overcome disadvantages of typical testing methods used today. Testing using this new method (explained in greater detail later in the article) is expected to begin in 2000.

The Radiation Environment

SEP occur both in the terrestrial and the space environment. SEP in the space environment are primarily induced by energetic protons and heavy ions, although space contains other energetic particles (e. g., electrons, X-rays, and gamma rays). The highest day-to-day fluxes of protons near Earth occur in the Van Allen radiation belts that surround our planet. Radiation belts are formed as a result of Earth's magnetic field, which traps energetic electrons and protons within the belt regions. The energetic proton belt extends from an altitude of about 500 kilometers to an altitude of 2.5 Earth radii (approximately 16,000 kilometers).

Spacecraft at altitudes below 500 kilometers experience few effects because potentially disruptive protons are trapped at higher altitudes; the Earth's magnetic field also provides some measure of protection from both galactic cosmic rays and solar protons. Spacecraft above the proton belt (for example, satellites in geosynchronous equatorial orbit at very high altitude or spacecraft on interplanetary missions) experience SEP caused by solar protons and galactic cosmic rays, which consist primarily of protons and alpha particles and a small number of heavier nuclei. However, spacecraft that traverse the proton belt are vulnerable to SEP from three sources: energetic protons in the belt, galactic cosmic rays, and solar protons.

Solar flares may also produce stressing environments for satellites and may temporarily or permanently degrade system performance. During and immediately after flares, energetic particle fluxes encountered by spacecraft may increase by orders of magnitude. Typical solar flares last only a few days. Nonetheless, some mission specifications require that systems be impervious to single event latchup (SEL) and only minimally responsive to other SEP throughout even a significant solar flare event.

The space radiation environment includes energetic protons, electrons, gamma rays, X-rays, and heavy ions. All of these can produce single event phenomena that have the potential to disrupt device and system performance—or even destroy the devices or systems affected.

The frequency of SEP occurrences can be reduced with shielding; that is, using dense metallic shields to cover electronic parts. However, the most energetic particles found in space will penetrate more shielding than most system designers will tolerate. Consequently, other methods of mitigating SEP must be used.

One widely used approach is to employ microelectronic devices that have been specially designed and engineered to be invulnerable to the space radiation environment or, at least, more tolerant of the environment. However, development of such devices is expensive, and their performance (processor speed, for example) generally lags the performance of commercially available nontolerant devices by 10 years or more.

A growing body of literature has recently been developed around SEP in the terrestrial environment. The most energetic cosmic rays penetrate the Earth's magnetic field and interact with nitrogen and oxygen nuclei in the upper atmosphere. These nuclear interactions yield neutrons that can penetrate to the ground. Anyone who uses a solid-state digital camera can observe these events simply by taking extended "exposures" in the dark. Frame-to-frame variations in the output in the form of intense, single-pixel spikes are due to cosmic ray effects.

Aircraft designers at Boeing, Airbus, and elsewhere are now taking account of SEP in systems design for aircraft that fly at higher altitudes (above 30,000 feet). Manufacturers of pacemakers for control of heartbeat irregularities worry about SEP effects on the performance of microcontrollers within the pacemaker. Finally, microelectronic device performance may be degraded by SEP due to alpha particle decay from radioactive isotopes of ions that are impurities in metals commonly used for the microelectronics themselves.

Regardless of whether radiation-tolerant or commercially available nontolerant devices are used in a system, the system designer will need to know their susceptibility to SEP. Consequently, tests of SEP susceptibility are typically required for all devices to be used aboard a spacecraft. Tests for SEP vulnerability in devices to be used terrestrially are becoming increasingly important.

Traditional Testing Methods

The level of exposure a given device can survive determines the "hardness" of the device. In order to classify devices according to hardness, it is necessary to test them for SEE using calibrated methods that can be correlated to the different types of radiation environments. Traditionally, this is achieved by exposing the electronic device to radiation from particle accelerators such as Van de Graaff generators, cyclotrons, or synchrotrons, from radioactive elements such as Californium, and from laser sources.

Based on the results of SEE testing by one or more of these methods, the hardness of a device can be assessed. The probability of encountering a SEE during the required performance lifetime can be predicted as well. In the event that the device or part of the device is susceptible to SEE, then it may require special shielding or may need to be redesigned to withstand the radiation environment.

Ablation of metal targets with intense femtosecond laser pulses produces a dense plasma that emits ultrashort bursts of X-rays. These X-ray pulses can be guided to precisely determined locations on a device under test using monocapillary optics.

Particle accelerator testing is the most widely accepted method for measuring SEE susceptibility. Circuits are characterized for SEE sensitivity by measuring the cross section as a function of accelerator ion linear energy transfer (LET) for a variety of particles at a range of energies.

Californium (²⁵²Cf) is a radioactive isotope that spontaneously decays to produce an array of high-energy particles, alpha particles, and neutrons, which are used to simulate the space radiation environment. However, Californium emission does not provide as wide a range of particle energies as is available with particle accelerator techniques and thus does not provide a sufficient range of LET data to correlate to most space environments.

Most recently, laser testing methods have been developed that use a short laser pulse, less than 10 picoseconds, to probe a device for SEE vulnerability. It has been shown that the phenomena in the semiconductor device caused by an ultrashort laser pulse and a charged particle are similar. Devices are probed with a single laser pulse for SEE, and thresholds for SEE are determined by the absorbed laser energy.

Laser probing of device susceptibility has several advantages over particle beam techniques:

Laser beams can be focused to small spots, allowing location of sensitive nodes in complex integrated circuits with submicron precision.
Repeated laser excitation does not damage the material, whereas particle beam excitation produces permanent damage.
Lasers are less expensive and more convenient to operate than particle beam facilities.

The principal disadvantage of laser-based techniques for simulating SEE is that the laser light cannot penetrate metal. Thus, a sensitive node completely covered with metal cannot be interrogated using laser-based techniques. Modern high-performance integrated cicuits have smaller features, higher levels of integration, and multilevel metalization architectures compared with previous versions, leaving many complex integrated circuits essentially covered with metal and impenetrable to laser light.

Short-Pulse X-Ray Generation

Using short pulses of X-ray radiation capable of penetrating layers of metalization overcomes the disadvantages of laser testing of device susceptibility to SEE while preserving its advantages. To be reliable for hardness assurance testing, such a system must meet special criteria. In order to generate SEE in electronic devices, the X-rays must be produced in the 0.8 to 5 kilo-electron-volt energy range to be transmitted through metalization (for example, aluminum, polysilicon, copper) and oxide layers and be absorbed in the active region of the device. Additionally, charged particles pass through semiconductor devices in less than 1 trillionth of a second. Thus, the X-ray pulse duration must be ultrafast.

Recently, ultrashort X-ray pulses have been produced using ultrashort laser pulses via high harmonic generation, generation of ultrashort electron bunches that produce X-rays when launched into metal surfaces, and laser ablation of metals to generate dense plasmas that emit X-ray pulses. Of these techniques, the latter is the most feasible for producing pulses in the required energy range and for focusability of the X-ray pulses. In this method, ultrashort bursts of X-rays are generated through the ablation of metal targets illuminated with intense femtosecond laser pulses. The dense plasma produced by ablation generates X-ray pulses with ultrashort pulse durations.

Testing microelectronic devices for their degree of vulnerability to single event phenomena is integral to designing devices to be used in space systems. Aerospace researchers have designed a laser amplifier system that produces extremely intense laser pulses. These pulses produce short bursts of X-rays that can be directed at microelectronic devices for testing.

X-rays generated by this technique are emitted in all directions and need to be focused to be usable for SEE testing. We use tapered monocapillary optics to focus X-rays onto a sample. These monocapillary optics are hollow glass tubes that work as waveguides for X-rays via total external reflection. Tubes with input end openings of 25 millimeters have been used to focus X-rays to diameters as small as 3.5 millimeters; nothing, in principle, precludes production of tubes with exit openings of 50 nanometers or less.

SEE Testing at The Aerospace Corporation

We are in the process of developing a system to generate short pulse X-rays to be used for SEE testing of electronics. The X-ray pulses are generated by the laser ablation technique. We have designed a laser amplifier system to produce the extremely intense laser pulses necessary. The laser amplifier is based on a mode-locked titanium:sapphire laser that produces 100-femtosecond pulses with tunable wavelength ranging from 700 to 900 nanometers. The laser output is amplified in two stages. First, a regenerative amplifier increases the oscillator pulse energy from the nanojoule to the millijoule level. A second stage multipass amplifier boosts the pulse energy to approximately 200 millijoules. Before reaching the first amplifier, the 100-femtosecond pulses from the oscillator are temporally stretched to 200 picoseconds in order to avoid laser induced damage of the optics through the amplification process. The output of the multipass amplifier is spatially expanded, recompressed back to 100 femtoseconds, and then focused onto the target.

The focusing optics and the target are enclosed in a vacuum chamber to prevent damage to the focusing optics, ionization of the air and absorption of the X-rays by air. A portion of the X-rays generated is collected and focused onto the device to be tested.

Testing of microelectronic devices for SEP susceptibility using ultrashort X-ray pulses will begin in 2000.