Thermoelectric power conversion in space.
Previous applications of thermoelectricity to power generation date back more than a century and a half to Seebeck, who showed in 1822 that a current is obtained when the junctions of dissimilar materials forming a thermocouple loop are maintained at different temperatures. Such a circuit is shown schematically in Figure 1. For example, two dissimilar materials, an n- and a p-type semiconductor, are joined at their ends by a metallic conductor. Heat is supplied to the hot junction from an external source; the other junction is maintained at a fixed lower temperature. As a result of the temperature difference, a current, 1, flows through the branches of the thermocouple. By allowing the current to flow through an external load resistance, inserted into the circuit between the cold junctions, the arrangement represents a direct conversion of heat into electrical energy. The figure of merit, Z, measures how good a material is for thermoelectric energy conversion and is given by:
Z = S[.sup.]2/p[Lambda]
where S is the Seebeck coefficient, p is the electrical resistivity, and A is the total thermal conductivity.
The best materials to use for thermoelectric applications are semiconductors. This can clearly be seen in Figure 2, where the Seebeck coefficient, electrical conductivity, thermal conductivity, and figure of merit (which measures how good a material is for thermoelectric energy conversion) are plotted versus carrier concentration. Insulators have a high Seebeck coefficient but low electrical conductivity (high resistivity) while metals have a high conductivity (low resistivity) but low Seebeck coefficient. Semiconductors with a carrier concentration in the 10[.sup.]19-to-10[.sup.]21-percm3 range have a reasonably high Seebeck coefficient and conductivity which results in a very good figure of merit. The figure of merit has a maximum for semiconductors with a carrier concentration in the 10[.sup.]20-to-10[.sup.]21-per-CM[.sup.]3 range.
A number of thermoelectric materials have been developed for various operating temperature ranges: bismuth telluride and bismuth selenide alloys for lower-temperature (approximately room temperature) applications, lead telluride alloys for intermediate 200[deg] to 500[deg] C) applications, and silicon germanium alloys for high temperature (400[deg] to 1000[deg] C). The figure of merit values range from approximately 3 x 10[.sup.]-3 K-[.sup.]1 for bismuth telluride alloys to 1 X 10[.sup.]-3 K[.sup.]-1 for the silicon-germanium alloys.
Finding new semiconductor materials with superior thermoelectric properties involves finding materials with a maximum value of S[.sup.]2/P and a minimum value of total thermal conductivity A. Both approaches to improving Z are being pursued to improve silicon germanium alloys and hence to increase their efficiency. An improvement of 20 to 30 percent in n-type silicon germanium was recently achieved by doping it with gallium phosphide, which decreases the electrical resistivity and increases S[.sup.]2/p[Lambda].
Thermoelectric Converter Technology
The current generation of radioisotope thermoelectric generators (RTG) uses a radiatively coupled unicouple configuration. Each converter consists of a single n- and p-leg of silicon germanium thermoelectric material. The heat is transferred by electromagnetic radiation from an encapsulated plutonium dioxide heat source to the hot shoe where it is absorbed and concentrated. The hot shoe also forms the electrical connection between the n- and p-legs of the unicouple. On the cold side of the unicouple an interconnecting strap is used to connect the unicouples into a series/parallel network. Each unicouple is mechanically connected to a metal radiator that conducts the heat away from the unicouple and radiates it to space.
A radiatively heated multicouple for use in the next generation of RTG is under development. The major improvement of the multicouple is the use of 20 individual couples within a single cell. Thus, 40 n- and p-legs are interconnected in series. The legs are all bonded together by glass, which also acts to electrically isolate each leg. At the hot end, the legs are electrically interconnected using silicon molybdenum, which is attached to a graphite heat collector. On the cold side the legs are interconnected using tungsten and the entire cell is then mechanically attached to a radiator, which conducts the heat away and radiates it to space.
An example of a power converter using the multicouple technique is the SP-100. Its thermoelectric module is a 4 x 8 couple matrix, each couple consisting of one n- and one p-type silicon germanium-based material. Electrodes spanning the four parallel rows on the hot and cold sides are bonded to the couples to form integral electrical connections of four couples in parallel and eight couples in series. The material chosen for the electrodes is a bilayer of tungsten and graphite that meets the requirements for high temperature, high electrical and thermal conductivity, and compatibility with the thermoelectric material. The electrode material has to closely match the coefficient of thermal expansion of silicon germanium, as well as be chemically compatible with it and other materials within the cell for a period of seven years at operating temperature. Tungsten meets all the requirements execpt chemical compatibility. Thus, a graphite layer is added to act as a diffusion barrier between the silicon and tungsten.
The compliant pads permit an all-bonded system. Each compliant pad is made up of millions of extremely fine fibers, making it very flexible, yet with a high thermal conductivity because the fibers are axially aligned. The compliant pads accommodate distortions due to bowing of the module caused by the temperature gradient at operating temperature, and the difference in thermal expansion between the hot-side and cold-side heat exchangers. The compliant pads are made of niobium fibers supported between thin face sheets of niobium and tungsten. The two different face sheet materials are needed to match the coefficients of thermal expansion of the modules and the high-voltage insulators and not cause high stress on any of the components. The modules are bonded to the compliant pads with a thin layer of glass, similar to that used in bonding the n- and p-legs. This provides low-voltage electrical insulation of the individual electrodes and individual thermoelectric couples.
The high-voltage insulators are thin wafers of single crystal alumina (sapphire). When bonded to the heat exchangers and the compliant pads, they provide the thermal, electrical, and mechanical integrity of the assembly during operation. Sapphire is used because it is compatible with niobium and has demonstrated stable performance at elevated temperatures and electric fields.
The thermoelectric converter is a layered assembly that starts with one hot-side heat exchanger that has feed and return core headers. An array of 60 thermoelectric cells, connected in a series/parallel ladder network, are bonded to both sides of the hot heat exchanger. A cold-side heat exchanger is bonded to each array of thermoelectric cells. The hot-side heat exchanger transfers the heat generated inside the core of a nuclear reactor to the array of thermoelectric cells. The cold-side heat exchanger transfers the heat rejected by the thermoelectric cells to a heat pipe radiator. The heat transfers occur by means of electromagnetically pumped loops of liquid lithium. The power-conversion system is then made up of six assemblies stacked together and many stacks distributed around the periphery of a cone.
Depending on the type of space mission, thermoelectric converters can take different configurations or have design differences. Missions can be classified into types that pertain to regions of space or planetary bodies that are explored.
Deep space applications use RTGs developed for vacuum operation. Thermoelectric converter power systems using a unicouple configuration have flown on several noteworthy missions, including Pioneers 10 and 11, which had lead telluride thermoelectric material converters, and Voyagers I and II, which used silicon geranium-based thermoelectrics. The Voyagers completed the solar system-exploration portion of their missions recently and have provided spectacular data on Jupiter, Saturn, Uranus, and Neptune and their associated moons. The main point to note from a power standpoint is that these spacecraft have been powered for well over 10 years with predictable degradation of power. In addition, the Voyager interstellar mission will have sufficient power to last well into the next century.
The recently launched Galileo spacecraft also uses unicouple technology, but has an improved heat source design that allows the power density to be increased. The General Purpose Heat Source (GPHS) configuration, used for Galileo, will also be used for the Ulysses mission, which is scheduled to be launched late this year. The Galileo spacecraft flew past Venus in February 1990 and will fly by the Earth in late 1990 and then again in 1992 to obtain the gravity assist necessary to fly to Jupiter. Once at Jupiter, Galileo will split into an orbiter spacecraft and a probe. The probe will take atmospheric and other measurements of Jupiter as far as gravitational forces will allow until it is crushed. The orbiter will spend a few years mapping and photographing a significant number of Jupiter's moons as well as the surface atmosphere of Jupiter. Galileo uses two RTGs that provide a total of 570 watts of electric power.
Assisted by Jupiter's gravity, the Ulysses mission will fly over the south and north poles of the sun. The poles of the sun are being investigated because there is very little information on the nature of space out of the plane of the ecliptic of the solar system. The RTG power system will use the GPHS design but will have only one unit providing 285 watts of electrical power at the beginning of a mission.
Future missions in this class that will likely use RTGs include flyby probes and rendezvous vehicles. The Cassini Saturn orbiter/Titan probe flyby and Pluto flyby are examples of missions in this class. The comet rendezvous asteroid flyby (CRAF) is the first mission to use the Mariner Mark 11-class spacecraft. CRAF will cruise until closely approaching Hamburga 449, a main belt asteroid, where it will make a compositional analysis. The mission will continue with a rendezvous within 25 kilometers of the comet Kopff; the comet's nucleus will be investigated for composition, physical state, and processes.
The Cassini mission, using a Mariner Mark 11-class spacecraft, will carry an atmospheric probe to Titan, Saturn's largest moon. After the probe descends onto Titan, there will be a period of two hours in which compositional data are relayed back to Earth, after which the orbiter will continue on a three-year mission to investigate the other moons and rings of Saturn.
So little is known about the Pluto/ Charon system that it has been recommended by the Solar System Exploration Committee that a flyby of the system be scheduled after the year 2000. A Neptune flyby and probe were also suggested. These deep space missions will likely use RTGs of the new modular multicouple design.
Thermoelectric converters appropriate for planetary rovers and sample return missions require extra ruggedness, including the ability to withstand frequent vibration and large periodic fluctuations in cold-side temperature. A Martian rover RTG would have the additional problem of dealing with a reactive atmosphere with metal oxide dusts and oxides of carbon, which may interfere with normal operation of the RTG. A filtering system and a one-way valve may be required to keep adverse material from entering the RTG chamber while allowing buildup helium from the radioactive source material to escape to the environment. Alternatively, insulative materials unaffected by the Martian atmosphere and dust might be used.
RTGs would be ideal for extended roving missions where night and day considerations or dust conditions may seriously hamper other types of power-conversion systems. Maneuverability will not be affected by RTGs since they are compact firm packages that do not contribute to vehicle instability. Solar photovoltaic collectors may require a large area array, which could cause locomotion problems, particularly in areas of high rockiness or steep hills. Furthermore, solar collectors may be subject to dust buildup (on Mars particularly).
The Solar Probe mission will have severe thermal environmental requirements for its instruments and power system. The probe will use Jupiter's gravity to help plummet it towards the sun to within about four solar radii. A thermal shield is needed to protect the spacecraft and instruments. Additionally, the RTGs need to conform to the configurational requirements of this craft, which has a conical shield facing the sun.
An example of a nuclear electric propulsion mission is the thousand astronomical unit (TAU) mission. An electric propulsion unit powered by a nuclear reactor coupled to a thermoelectric or other type of converter has been proposed for the TAU mission. To traverse a physical distance of some 93 billion miles will require many years. A reactor to provide long-term hi h heat energy for conversion to electrical power and electric propulsion has been determined to be the best near-term option for propulsion. Reliability and redundancy favor modular-type power conversion units. The multicouple RTGs described earlier seem to fill the need for high reliability and steady performance for this length of time.
Lunar and Martian manned-base applications will ultimately require large power stations to meet life support needs. Power stations in the multikilowatt and -megawatt category are needed to provide power for various activities such as oxygen production for propellants and life support, power for recycling waste products, electrical production, air handling, and the like. An SP-100-type reactor-based thermoelectric conversion power system can meet this need. The reactor portion would probably be buried in the lunar soil to provide radioactivity exposure protection to base personnel. In this case, the heat rejection subsystem must be exposed and facing space. a Acknowledgments
The writing of this paper was supported by the Jet Propulsion Laboratory, California Institute of Technology, under contract with NASA. We would like to thank Dr. P. Bankston for many helpful comments.
|Printer friendly Cite/link Email Feedback|
|Author:||Awaya, Henry I.; Ewell, Richard; Nesmith, Bill; Vandersande, James|
|Date:||Sep 1, 1990|
|Previous Article:||Thermal energy storage.|
|Next Article:||Calibrating pressure transducers.|