Electronic devices that never need to have their power source replaced and can function unattended for 100 years. Science fiction? No, science fact.
Raytheon's customers need compact, reliable and long-lived, high energy density power supplies for applications such as sustainable low-power electro-mechanical devices. One such application is unattended embedded stress monitoring devices using microelectromechanical systems (MEMS) that are located in inaccessible areas such as aircraft structures, bridges and buildings. These applications all beg for a robust, viable, cost-effective power supply that can satisfy the long- duration needs and sustainable power required for predicting the onset of a structural failure, and then conveying this information to allow for pre-emptive action and avoid catastrophe.
These isolated sensors are only practical if they are small, long-lived and unaffected by harsh environmental conditions. Typical chemical-based batteries may last a couple of years, whereas autonomous sensors require miniature power sources with much longer lifetimes. For sensor networks in harsh, inaccessible environments, battery replacement can be a practical impossibility or prohibitively expensive. There is also a need for MEMS technology that can overtly or covertly sense mechanical motion, temperature changes, chemicals and biological species. This requires long-term sources of compact energy. Applications include radio frequency identification (RFID) tags, autonomous sensors, and long-lived, miniature, wireless transmitters. The discriminator in all of these technologies is long-term, reliable, high-energy density that is addressed using devices that contain embedded betavoltaic power sources, referred to, in general terms, as "atomic batteries."
Atomic Battery Physics
Atomic batteries are power sources utilizing the emissive properties of a certain class of radioactive isotopes. These unstable isotopes are mostly man-made in nuclear reactors. They are a form of naturally occurring elements, where the normal distribution of protons and neutrons in the nucleus is disturbed, rendering it unstable. Over time it is destined to return to a stable state through an internal restructuring called transmutation. A consequence of transmutation is the emission of some nuclear constituents that convert the material into a different element. These constituents primarily include highly energetic particles such as alpha particles (helium nuclei) and beta particles (electrons).
Upon impact with matter, these constituents deposit their energy through ionization in a predictable manner, creating tracks of secondary charged particles, such as electrons and ions similar to electron-hole pairs in semiconductors. Eventually, the particles are captured by the encountered material when their kinetic energy dwindles down to zero. In all cases, the total energy content of each of these particles, initially in the form of energetic charged particles, eventually emerges as deposited energy in the form of heat. Some atomic battery technologies are based on capturing the copious amounts of charged particles created during ionization, while others utilize the resulting generated heat.
Radioisotope Thermoelectric Generator
Several applications have exploited the heat generation aspects of atomic batteries. One is the radioisotope thermoelectric generator (RTG), which has been used in numerous NASA deep-space missions that cannot implement photocells as power sources; e.g., missions to the outer regions of the solar system, where power generation by sunlight is ineffectual. The basis for these types of power generators consists of hundreds of Curies of the alpha particle-emitting isotope Pu-238 embedded in ceramics. This produces energy by heating the ceramic mass through alpha particle energy absorption, with subsequent thermoelectric conversion to useful electricity. These RTGs have no moving parts and have been the major source of power in at least 41 NASA missions on satellites expected to operate for more than 20 years. Another lower power application of this type of technology is found in pacemakers (see Figure 1). This application uses about three Curies of Pu-238. Weighing only about three ounces, the pacemaker produces approximately 1 milliwatt of power while contributing a generally acceptable typical radiation dose of 100 millirems per year to the patient. Although in use for a number of years, this application was replaced several years ago when improvements in pacemaker technology reduced energy requirements to the point where lithium battery technology became viable.
Betavoltaics, another form of atomic battery, are the little brothers to RTGs; the difference is that this energy source is not based on the heat generated, but on its ability to generate sufficient quantities of materialionizing beta particles. While betavoltaics are similar in concept to photovoltaic cells, there is a notable difference. Where photovoltaic cells harvest energy from interacting photons, betavoltaics function by capturing and converting the kinetic energy of energetic electrons, emitted from decaying radioactive isotopes, into large amounts of secondary electrons.
Betavoltaics-powered devices may be engineered to be extremely robust. Since the source of power is electrons emitted from the isolated atomic nucleus, electron emission rates are immune from effects of stressful, harsh environmental conditions. Since this technology is based on feature sizes on the scale of an atom, betavoltaics show potential improvement in both energy density and total energy content, compared with conventional power sources such as AA batteries. This large energy density is attributed to the huge number of radioisotope atoms contained in a small amount of material (recall Avogadro's number), and each atom is primed to unload its energy-generating beta particle emission at a rate that is only dependent, in a statistical manner, upon the particular isotope's half life. This advantage in energy density is indicated in Table 1, which shows a relative comparison of capabilities for a notional betavoltaic battery design with those of a typical lithium AA battery.
Two betavoltaic manifestations are possible: the so-called direct conversion category, where secondary electron-hole pairs are generated in P-N semiconductor diodes, or the vibrating cantilever concept that converts mechanical energy to electrical energy using a piezoelectric-driven, energy- scavenging mechanical converter. Miniature, low-powered technology devices, based on either of these two general operational classes, hold the potential for the development and integration of tiny smart sensors that will never need their power supplies replaced. Specific designs based on atomic batteries are customized for their intended applications; some of the basics that help dictate the design are briefly discussed below.
Direct Conversion Betavoltaics
One unique rendering of betavoltaics is the direct conversion approach based on a P-N semiconductor diode such as gallium nitride (GaN) placed in direct contact with a source of beta particles. Figure 2 is a notional design for the P-N junction method. In the figure, the source adjacent to the semiconductor is a thin plated film layer of a beta particle-emitting isotope. A typical useful source for these applications is a 5-micron layer of the pure beta particle-emitting isotope Ni-63, providing an activity of roughly 0.25 milliCuries, that emit beta particles over a wide range of energies, with an average energy of 17 kiloelectronvolts (keV) and peaking at 67 keV. On average, half of all emitted beta particles are transported toward the semiconductor Player where, upon interacting with the material, some beta particles are backscattered from the interface and do not penetrate into the semiconductor.
Those beta particles that make it into the semiconductor begin losing energy quickly, primarily through ionization, generating electron-hole pairs that are captured once all their energy is dissipated. Beta-particle path lengths depend on initial beta-particle energy and the material through which it is transported; in general, they are in the range of a few tens of micrometers. For this energy transfer to be effective as a power source, beta particles should be able to reach deep enough into the semiconductor to deposit most of their energy, through ionization, in the P-N junction depletion region. Those electron-hole pairs generated in the depletion region — where the number of pairs depends on material band gap and beta energy — are swept across the junction by the generated electric field and are converted into useful electricity to power an attached load (Figure 2).
These types of betavoltaics generally develop power levels that can approach 1 milliwatt. Radiation-tolerant, wide band gap semiconductors are ideal candidates for direct-conversion betavoltaics. Several semi-conductors have been identified as ideally suited for these applications. They include GaN, aluminum gallium nitride, silicon carbide and diamond. Since electrons are rapidly absorbed as they emerge from the radioactive plated surface, the useful isotope plating thickness is limited to a few micrometers at best. Therefore, methods to scale up the output of these devices depend primarily on increasing direct contact surface area. Honsberg, et al., described a conceptualized approach to address this issue. It consists of mating GaN layers on each side of thin Ni-63 wafers in order to maximize output power. Using this GaN- isotope sandwich design to capture a large fraction of emitted beta particles, the ability to develop a 2.3 volt open circuit voltage with a short circuit current of 1.1 microamperes was reported.1 Raytheon currently produces GaN devices for high-power microwave applications and also has an established and demonstrated capability for growing thin-film chemical vapor deposition diamond. With this established presence in developing materials that are highly desirable for betavoltaics-based power sources, Raytheon is in a good position to drive this technology forward.
In light of the limited range of low-energy beta particles considered here, beta sources for direct conversion devices are considered to be relatively safe since they are literally stopped by the outermost dead skin layer. There are a number of radiologically-safe pure beta emitting isotopes with half-lives ranging from 2.6 to 100.3 years and with energy densities as high as 1011 kilojoules per cubic meter (kJ/m3). In comparison, diesel fuel has an energy density of approximately 4x107 kJ/m3, illustrating that betavoltaic sources can have very high energy densities and, consequently, long lives.
The choice of the appropriate isotope is dictated primarily by operational considerations, where the isotope is selected based in part by matching its half-life to the application's expected operational life. Use of a pure beta emitter is also preferred, since generating other decay products can lead to a significant dose to the operator and possible damage to the direct conversion device semiconductor when long term exposures are required. In general, a long half-life pure beta isotope provides prolonged battery life at the expense of generating low power, while a short half-life isotope provides higher power at a more limited sustainable life span. Table 2 shows some candidate beta-emitting isotopes and their relevant properties.
Another unique betavoltaic application is based on a vibrating piezoelectric cantilever concept, the self-reciprocating cantilever. This functions initially as a charge-to-motion conversion process followed by mechanical energy conversion to electrical energy. In this rendering, the device contains a beta particle-emitting isotope-coated surface designed to continuously deliver a negative charge to a nearby piezoelectric material-coated cantilever. This conceptual design is shown in Figure 3, where the self-reciprocating cantilever process begins by charging the opposing surface with a large fraction of the charge emitted by the isotope. Once a sufficient negative charge builds up on the cantilever surface, the resulting electrostatic force field begins to draw the cantilever to the fixed location, positively charged lower surface. When an adequate charge is accumulated, the cantilever bends to the point where it contacts the radioisotope-coated surface. Upon contact, electrons flow from the negatively charged surface, causing surface charge neutralization, collapsing the electrostatic field to zero and forcing the cantilever to spring back and oscillate around its initial equilibrium position.
This approach allows for the continuous transfer of energy from mechanical to electrical, generated by the vibrating piezoelectric-based cantilever. As the platform continues to oscillate, the piezoelectric-attached structure generates useful electricity that can be harvested for various applications. Its oscillation frequency can be fine-tuned by modifying the length of the cantilever and the strength of the radiation source. Depending on its intended application, the output of this device can be utilized as a source of tunable rapid current spikes, or filtered to produce a continuous DC output stream.
An approach to implementing radio frequency identification (RFID), for example, is based on a modification of the self-reciprocating cantilever design described above. Tin, et al.2, report the generation of 264 MHz wireless signals induced directly by vibration of the cantilever. Another RFID design employs a surface acoustic wave (SAW) resonator connected to and excited by the vibrating cantilever. This concept has been shown to produce a frequency modulation found useful as a CMOS-compatible wireless communications beacon. The authors report on the design of a SAW transponder that can transmit an RF signal at 800 microwatts, with a 10 microsecond pulse duration, every three minutes at a frequency locked to a 315 megahertz SAW resonator.3
In addition to RFID, Raytheon is pursuing applications such as power sources for autonomous sensors and long-lived, miniature, wireless transmitters. The benefit provided to these applications is long-term unattended, reliable operation achieved using devices that contain embedded high energy density betavoltaic power sources.
1 C. Honsberg, et al., "GaN Betavoltaic Energy Converters," IEEE Photovoltaics Specialist Conference. Orlando, Fla., 2005.
2 S. Tin, et al.,"Self-Powered Discharge-Based Wireless Transmitter," IEEE International Conference on MicroElectroMechanical Systems, 2008.
3 S. Tin and A. Lal, "A radioisotope-powered surface acoustic wave transponder," J. Micromech. Microeng., 19 (2009).