Assessment of Progress
The Defense Threat Reduction Agency thermionics program sponsors a number of different efforts aimed at establishing a technology base for potential future users of this technology. As detailed in Chapter 2, the effort is spread among too many different projects within present funding constraints. However, even given the fairly large number of projects for the modest yearly funding available, the program does not sustain thermionic science and basic research related to electrode materials, plasma physics, and surface physics. The committee believes that the overall DTRA effort would be more successful if it were to emphasize a basic research program in electrode and materials processes relevant to thermionic energy conversion.
Recommendation 5. The sponsoring agency should reestablish an adjunct basic research program on electrode surface physics, plasma, and materials processes relevant to thermionic energy conversion. This program should be funded separately from the thermionics research program.
The term “adjunct” in the recommendation above refers to basic research that does not use DTRA contract funds. These non-DTRA contract funds might include funds from Small Business Innovation Research, the National Science Foundation, the Air Force Office of Scientific Research, NASA, and so on. Naturally, the future sponsor of thermionics research cannot dictate to the other agencies how to spend such funds: however, if they present the potential benefits of such a research program, the future sponsor may be successful.
The current funding mechanism, so-called congressional plus-up money, also negatively affects technology development, particularly at the university level. It is difficult for researchers to guarantee funding for graduate students for the duration of their degree program which spans multiple years. Because funding uncertainty, faculty and students often hesitate to participate in thermionic research.
Finding: Without stable, multiyear funding, university researchers hesitate to take part in thermionics research efforts, and the thermionics community misses the opportunity to leverage university-based research.
MATERIALS AND DEVICE RESEARCH
Single-Crystal Refractory Metals
Single-crystal refractory thermionic electrodes were tested in the early U.S. programs including the Solar Energy Technology program. The use of single-crystal refractory metal alloys was pioneered in the former Soviet Union (Gontar et al. 1996). According to research conducted there, single crystals offer three important advantages over their polycrystalline counterparts: (1) improved creep resistance, resulting in lower emitter cladding distortion and longer lifetime; (2) improved resistance against material diffusion, another factor important for long life; and (3) high bare work function, which produces a low cesiated work function, resulting in higher efficiency (Drake 1998). This advantage is primarily for planar converters.
Collaborative efforts involving Scientific Research Association (SRA) Luch, Auburn University, and General Atomics have tended to support these conclusions. Accordingly, the committee perceives that this collabo-
ration has significantly enhanced the U.S. state of the art in the area of high temperature metallurgy, as well as in thermionic device technology.
Finding: The collaboration between the Defense Threat Reduction Agency thermionics program and Russian research facilities is ongoing and has been successful.
Auburn University Materials Work
The DTRA has sponsored research at Auburn University through General Atomics that concentrates on the fabrication methods and underlying science in the fabrication, processing, and characterization of refractory alloy single-crystals. This work has proven to be successful and is one of the major success stories associated with the thermionics technology program. The committee learned in June 2001 that the U.S. Army will be providing funding to Auburn University to expand the single-crystal work discussed here to other non-thermionic applications.1 Auburn researchers are exploring three approaches: (1) electron beam float zone, (2) flow chemical vapor deposition (CVD), and (3) closed-cell CVD.2
The electron beam float zone method is considered a brute force approach to single-crystal fabrication, where the refractory alloy bar stock is melted and resolidified by sweeping a segment of the molten material up or down the specimen axis. There are several difficulties associated with this approach. The electron beam float zone method requires that the alloys be melted at high temperature (more than 3400°C for tungsten), which is difficult to do with the systems currently in place.
However, researchers at Auburn University have overcome many of the challenges and have successfully fabricated single-crystal alloys in small amounts of niobium, molybdenum, rhenium, and tungsten using this method. The growth facility at Auburn University is capable of processing the highest melting point tungsten alloys with diameters up to half an inch. Selected high temperature mechanical properties of the processed alloys have been examined and the effect of grain boundaries and solute additions identified.
CVD methods allow researchers to grow single-crystalline layers of refractory metals on a metal substrate at a much lower operating temperature than is needed with the electron beam float zone method. Conventional CVD technologies depend on flow-through methods, in which fresh vapor phase reactants are continuously passed over the surface to be coated.
CVD single-crystal alloy production was a true collaborative effort between Auburn University and General Atomics. Two chemical reduction reactions were used: one with the tungsten hexachloride and the other with tungsten hexafluoride. Researchers were able to grow layers of pure tungsten on molybdenum substrates using both reactions.
To reduce the amount of waste associated with the corrosive vapors used in CVD processes, a closed-cell CVD concept was investigated. A closed-cell CVD process gives only small volumes of waste products, and these can be condensed within the CVD cell. Auburn has demonstrated the feasibility of employing a closed-cell CVD process for depositing pure tungsten into molybdenum substrates with the following attributes:
Environmental friendliness (no exhaust vapor),
Low cost and potentially significant raw material savings, and
Ease of operation owing to the low operating process temperature and the absence of exhaust.
The closed-cell approach is relatively simple in an experimental configuration. However, the thermodynamic characteristics of the system are significantly more complex than those of the flow CVD system. These complexities are due to significant variations in the temperature and pressure of the reaction gases in the system. Researchers are continuing to investigate the correlation between crystal growth characteristics, thermodynamic parameters, and properties in this closed system.
General Atomics informed the committee that some work has been reported in Russia where a single-crystal tungsten wire was fabricated that had a final-diameter to start-diameter ratio of 30:1. General Atomics is currently trying to replicate this large-diameter CVD work.
While General Atomics has made some progress on CVD single-crystal tungsten, it is working to develop CVD single-crystal tungsten alloys. Single-crystals have several advantages over polycrystalline metals
because grain boundaries are eliminated. High temperature creep strength is enhanced, grain boundary diffusion is eliminated, and the highest bare work function of the crystal can be exposed on the emitter and collector faces, producing optimum cesiated performance. These factors are all available in planar converter geometries. In cylindrical geometries, a single-crystal electrode will provide enhanced mechanical properties, but the work function will vary around the circumference. Russian CVD work indicates that some possible improvements in work function may be realized by using cylindrical geometries with single-crystals, but the exact amount of improvement has either not been determined, or the Russian researchers have not revealed their results.
Uniaxial creep test data were obtained during the first year of funding on single-crystal tungsten-tantalum. The DTRA anticipates using tungsten-tantalum nuclear fuel cladding for the thermionic converter heat source. If the single-crystal is used with a nuclear fuel heat source, the material will be subject to biaxial stresses. DTRA management proposed that biaxial verification experiments be conducted on single-crystals. As a result, the General Atomics subcontract to Auburn University includes a task to test the biaxial creep properties of tungsten-tantalum single-crystals at high temperature.
In biaxial creep testing, a closed-end tube made of the material to be tested is internally pressurized using a static, inert gas. The test assembly is heated to the desired temperature in a diffusion pump vacuum chamber using a 5 kilowatt heat source. The diameter of the tube is measured periodically using a noncontacting laser micrometer system. Auburn researchers will subject each sample to different stress levels from 6,000 to 20,000 pounds per square inch in a stepwise manner with a set temperature. Two axial measurements will be made periodically to determine creep rates. The tests are expected to be complete shortly after this report is published in late 2001.
Committee’s Assessment of the Single-Crystal Research
Single-crystal materials can improve thermionic converter performance. As Figure 7.1 shows, increased bare work function results in higher efficiency and higher optimum power density. A work function of 5.6 corresponds to single-crystal rhenium, and one of about 5.0, to poly crystalline tungsten.
As bare work function increases, cesiated work function decreases (Figure 7.2). Thus, to maximize work function differences between emitter and collector, the highest collector bare work function is desirable.
The research being conducted in single-crystal refractory metals is a critical technology area for any future thermionic program. The research also has the potential for significant spin-off applications into numerous other areas. Industry alone cannot sustain these technologies without recruiting new interested personnel from the university ranks. University research must be supported not only to provide technical knowledge continuity but also to gain advocates, now and in the future, for the complex and wide ranging technologies associated with thermionic energy conversion.
General Atomics has the primary experience in the cylindrical converter and trilayer technology. There are clear materials issues that must be addressed and solved in order to achieve the anticipated performance. In the United States, cylindrical CVD techniques used in the past produced textured polycrystalline structures with high work function grain exposure on the electrode surface. There are several ultimate tradeoffs that must be evaluated between the geometrically simple, demonstrated planar converters on the one hand and the more complex and untested cylindrical converters on the other.
Previous work established that oxygenation of emitter and collector surfaces in a thermionic converter can be a useful technique for improving thermionic performance (Figure 7.3). An oxygen layer on the surface creates a very high work function. A cesium mono-layer on the oxygenated surface will result in a very low work function surface. The work function is less than 1.3 electron volts at 900 K or about 0.2 electron volts lower than cesiated niobium.
Another advantage of oxygenation is that the cesium pressure in the space between the emitter and collector is reduced. Lowering the pressure might lower plasma losses. In the conventional nonoxygenated converter, the output voltage is reduced by about 0.5 volts due to electron scattering losses in the plasma, thus diminishing the electrical output by 30 to 40 percent (Drake 1998, Begg 1998).
The committee, however, has heard concerns raised about the long-term stability of oxygenated electrode systems. Since oxygen is known to attack refractory metals, the long-term effects of even trace amounts of oxygen need to be carefully monitored over the operating lifetime of any thermionic device under test. Also, the benefits of oxygenation on system design, thermionic fuel element performance, and overall life have not been clarified. These issues must be addressed before benefits of oxygenation can be conclusively established, and the committee believes that the benefits are difficult to justify.
Finding: The benefits of oxygenation in enhancing converter efficiency are difficult to justify in view of the technical risks associated with system design, thermionic fuel element performance, and overall life.
Thus, oxygenation should not be incorporated into advanced technology development programs at this time, but it might be considered as a legitimate activity for basic research or possibly future applied research.
CLOSE-SPACED VACUUM CONVERTER
One of the features of the DTRA thermionics program is research on close-spaced thermionic converters being carried out by SRA Luch, the primary subcontractor in Russia.
In theory, the efficiency of the thermionic converter increases significantly as the gap approaches zero, because in an infinitesimally small converter space, there is no opportunity to build up space-charge. Hence, the barrier potential is nearly zero, and energy conversion efficiency is primarily limited by conduction and radiative losses.
However, once the gap size increases beyond a few microns, the electron barrier potential can no longer be ignored, and it is necessary to provide charge neutralization with a plasma such as cesium.
The physics behind this effect is not at all new. Vacuum mode converters have been investigated since the 1950s. In the past, such converters were judged to be impractical from an engineering point of view, owing to the difficulty of maintaining the extremely small tolerances that are required, substantially less than 1 micron.
The close-spaced converter effort has resulted in the manufacture of a thermionic converter with a 6 micron gap, which the committee feels was an impressive accomplishment. The converter apparently worked well during tests at SRA Luch. However, when it was shipped to New Mexico Engineering Research Insti-
tute, a cracked housing prevented the unit from being tested. After repair, poor performance was observed. This poor performance may have been due to partial shorting from contaminants in the gap region. The DTRA management’s intent now is to revalidate the test and to design and build a three-cell, close-spaced converter module.
It appears that the close-spaced converter project has been carried out with great competence and skill. However, the committee feels that a successful project will not provide enabling technology for a long-life nuclear thermionic reactor, because the close-spaced converter device cannot be easily integrated into a thermionic nuclear reactor core. Also, the project goals are directed toward conversion efficiency and performance rather than long life.
The close-spaced vacuum diode concept as presently envisioned must be a planar arrangement, so efforts on cylindrical thermionic converters are not applicable. Thus, the technology is not likely to be used with in-core thermionic fuel element type reactors, although it might be used with other core concepts (STAR-C or Romashka-derivative) or solar thermionic cells.
As mentioned above, the performance of the test converter that was transported from Russia to New Mexico was less than expected, probably owing to the presence of a foreign particle in the gap, which caused partial shorting. Such particles may have been liberated during transport, and launch vibration might cause similar problems for spaceborne versions of such a converter. However, the exact cause of the electrical short was never conclusively identified.
Also, there are significant concerns about mass transport and electrode distortion, either of which could be inherently life limiting. The vacuum converter must have very clean electrodes at all times. Even a small deposit could result in a short circuit in the converter. Evaporation and redeposition of even a few atomic layers could change the single-crystal nature of the surfaces. Thus, even if there are performance enhancements available from single-crystal surfaces, the long-term stability of closed-spaced converter performance is in doubt.
Thus the potential advantage of the close-spaced thermionic converter is its improved conversion efficiency. However, this improvement requires extraordinarily tight tolerances in machining the surface of the converters, similar to the requirements for machining laser mirrors, for example. In addition, the converter is subject to several failure modes that make long-term reliability questionable. Making such a converter practical for spacecraft use could be very difficult, and the benefits might not outweigh the risks.
THEORY AND THEORY VALIDATION
One of the three major tasks within the DTRA research and development program is thermionic device theory development and theory validation. The theory development and validation work is aimed at characterizing the effects of emitter and collector surface reflection effects, characterizing non-uniform surface work function effects (a.k.a. patch effects), and optimizing a thermionic system mass model.
A novel portion of the research, specifically the patch effect investigation, has been tied closely to the Microminiature Thermionic Converter (MTC) program, as is discussed below. Unfortunately, this theory development is of little importance to the overall program.
Also, the system mass modeling does not add significantly to the overall understanding of thermionic devices or systems. The committee believes that this work has already been conducted by many others and is satisfactorily complete for the present. The committee does not understand the need to explore the underlying principles of thermionic theory any further since the existing theory base is complete enough to work experimentally on developing thermionic components and systems.
Other than as related to the MTC, the current modeling work being conducted at Sandia National Laboratories does not appear to extend theoretical understanding of the current thermionic theory. Also, the patch effect explanation has been hypothesized to explain certain experimental observations. The committee believes that this explanation is but one of many potential explanations for the experimental observations. If the program were well funded, the committee would support the high risk, relatively low return patch effect research exercise. However, given the limited amount of funding for the thermionics program as a whole, the committee recommends that all theory and theory validation activities be discontinued.
MICROMINIATURE THERMIONIC CONVERTER
The DTRA-sponsored MTC program is another one of the major program elements in the overall DTRA effort and is slated to receive a significant percentage
of the limited funds of the DTRA program. The effort is directed toward the development of a converter using semiconductor-scale fabrication technology with the hope that extremely small emitter-collector gaps can lead to economical, high efficiency conversion. Also, unlike the close-spaced converter concept, the MTC would be a very small chip-scale device. By manufacturing the converters on a micron scale or smaller using electronic device fabrication technology, they could potentially be very small. Just as millions of electronic devices can be fabricated on a single silicon wafer, it could conceivably be possible to place millions of thermionic converters on a small surface. The committee’s investigations have led it to conclude that the funding for this portion of the DTRA thermionics program should be redirected toward more basic research objectives as discussed elsewhere in this report. The rationale for this recommendation is discussed below.
Research is in progress at Sandia National Laboratories to develop scandate-based MTCs with high energy conversion efficiencies using semiconductor integrated circuit fabrication methods. These converters are of the vacuum type. Analysis shows that in theory such converters operating at emitter temperatures of about 1200 K, collector temperatures of 700 K, and interelectrode gaps of between 1 and 5 microns could produce attractive power densities and conversion efficiencies, but practical manufacturing methods and dimensional tolerances have never been demonstrated.
Extensive work on vacuum converters was conducted by Hatsopoulos and Kaye in the late 1950s, in which they used electrodes coated with barium-strontium oxides. The lowest work function achieved was 1.75 electron volts, and that for only very short periods of time. To offset evaporation loss of the barium and thus maintain an effective coating at the emitter surface, they used tungsten emitters impregnated with mixed barium and strontium carbonate These did produce 1 watt per square centimeter at about 1500 K for a few hours, having achieved an emitter work function of about 2.1 electron volts and a collector work function of about 1.8 electron volts. After that, however, converter performance deteriorated substantially because barium that evaporated from the emitter condensed on the collector (Hatsopoulos and Kaye 1958a,b). The discovery of surfaces with work functions of less than 1.6 electron volts that are stable over a useful period of time would greatly benefit not only MTC but all types of thermionic converters, whether vacuum or cesium based. Such a discovery, however, would not necessarily mean that vacuum converters, with or without the extremely close electrode spacing proposed for the MTC devices, are practical. The committee believes that beyond addressing the electrode work function issues, it would be extremely difficult to maintain, for any reasonable period of time, a temperature difference of nearly 1000 K between two surfaces held apart by a miniaturized spacer that is a few microns thick.
A number of advantages are claimed for the MTC concept. Relative to dynamic energy conversion systems, conventional thermionic systems, and other static conversion systems, MTCs claim the presumption of low maintenance, silent operation, long life, and compactness. In many cases, modularity and simplicity of assembly can also be expected. It is hoped that MTCs would be able to operate at high efficiency using a relatively low temperature heat source. And, very importantly, MTC devices might be manufactured inexpensively using integrated circuit chip manufacturing methods to achieve the extremely close tolerances needed.
The current program consists of several efforts, including the following (Rightley 1998a,b):
Development of electrode coatings to improve the emitter and collector properties, since the structural materials suitable for integrated circuit manufacturing methods are not those normally used for thermionic converters,
Analysis of MTC device configurations, and
Testing of MTC cells.
There are a number of technical issues associated with the MTC program. These are discussed in the following subsections.
As outlined above, the efficiency gains and lower operating temperatures postulated for the MTC configuration appear to depend on substantial improvements in both the emitter and collector work functions. Achievement of these advances via coatings such as barium is problematic since barium and other candidate materials tend to evaporate rapidly at operating temperatures above about 1000 K. This means that the barium surface coating would require resupply from within the solid coating. Further, deposition of the lost barium on the cool collector surface degrades the criti-
cal properties of the collector. The committee deems it unlikely that the MTC device would have a long enough life under these conditions. Scandium oxide cathodes produced by a sputtering process at Sandia do not yet meet the required work function properties. Electrode investigations using thin film layers of low vapor pressure materials may offer the best opportunity for achieving useful improvement in electrode properties (Zavadil et al. 1999).
Theoretical analyses of electron reflection from metal surfaces with and without adsorbed cesium or coadsorbed cesium and oxygen suggest that physical modifications of the electrode surface might allow the full, low work function capability of these coatings to be realized (Rasor 1998). However, the committee questions whether these theoretical benefits can be realized.
The MTC’s device configuration has been modeled but the data taken to date do not appear to support the analytical results. Vacuum converters of the conventional type were analyzed in the 1950s and, for the converters of that time, gave good agreement with experiment (Hatsopoulos and Kaye 1958a,b).
The current modeling work does not appear to extend theoretical understanding. With respect to model validation by experiment, while inexpensive manufacture using fully developed integrated circuit methodologies might be feasible, fabrication and testing of small numbers of experimental converters would be expensive. Given likely funding limits, the prospects are poor for obtaining adequate data with which to validate the MTC analysis from feasible experiments, even if this effort were to be the major funded effort of the entire DTRA thermionics program.
The conversion efficiency of individual converters is measured as the ratio of the electric output power to the heat actually delivered into the diode. Regardless of the conversion efficiency, the utility of an MTC-based conversion system depends on the way in which the heat available for conversion can be forced to feed primarily the thermionic converter portion of the device while minimizing the thermal losses from the external surfaces of the converter. These losses can occur by radiative transfer across the gap and by parasitic thermal conduction transfer around the converter edges from the hot side to the cold side of the system as a whole. As noted above, the necessary level of thermal transport control tends to become much more difficult for physically small systems such as MTC. More specifically, the MTC configuration has the two electrode surfaces, differing in temperature by approximately 500 K, separated by approximately 1 micron. This situation creates a temperature gradient of 5×105 K per millimeter in the connecting structure. Removal of the parasitic heat from the collector is expected to be difficult, given the thermal power density the collector will receive. In the opinion of the committee, sustaining such an enormous gradient with tolerable thermal conduction losses is not credible.
A diode with an extremely low power was demonstrated at Sandia (King et al. 1999). The device had a peak power of approximately 1.2 milliwatts per square centimeter with the temperature of the emitter at 1173 K and the temperature of the collector at 973 K. This power density value is less than conventional thermionic technology capability by a factor of approximately 1000. Power density values were minuscule compared to those reported for vacuum converters as long ago as 1956. The committee was told that the critical problem leading to these weak results was nonuniform emission from the emitter electrode such that only a small fraction of the total cathode area was contributing to the output power. Unfortunately, both the validation of this hypothesis and the development of methods to overcome the difficulty, if it is proven to be correct, are likely to be expensive given the limited funding for the thermionics program. While the performance of electrodes in complete converters is the ultimate test, diode manufacture at the MTC scale will be inexpensive only when it becomes standardized and the major benefits of integrated circuit production methods can be used. Producing and testing enough single converters on the scale needed to produce critical electrode performance data will be expensive.
It is the opinion of the committee that, at this time, program funding should not be spent on electrode analysis validation, particularly when no method to correct performance problems is available.
Finding: The device being developed in the microminiature thermionic converter (MTC) effort has low effi-
ciency, and the explanation and understanding of the surface physics are incomplete.
MTC Electrode Materials
The rationale for the MTC configuration depends not only on the presumed low manufacturing cost using integrated circuit manufacturing techniques, but also on very substantial advances in durable electrode work function properties. The search for such materials has been thoughtfully and extensively pursued in a number of laboratories over the years without the constraint of compatibility with integrated circuit fabrication methods and materials. Even if low work function electrode materials can be fabricated inexpensively and made compatible with the integrated circuit industry fabrication methods used to form the main structure of the MTCs, the same electrode materials should also be helpful in forming more conventional thermionic fuel elements. In the case of conventional thermionic fuel elements (TFEs), the space charge limitations of higher gap spacing are compensated for by the use of Cs vapor. Therefore, even if an MTC device became practical, the gains in low work function materials would probably allow conventional TFEs to outperform the MTC device.
Even if it is assumed that the several major technical hurdles identified above can be overcome, very small scale (on the scale of a chip or radioisotope heater unit) MTCs still face stiff competition from thermal electric generators. At 1 to 100 watts, MTCs also face strong competition from AMTEC and free-piston Stirling, which appear to have fewer material problems at the temperature levels proposed for the Sandia converters.
Ultimately, the experimental data from the MTC program do not support the theoretical predictions. Not only are postulated low work function emitters not yet functional, they also are not expected to be so in the near future. Finally, while it is assumed that integrated circuit fabrication methods will lead to low cost production, the fact remains that fabrication of stable, small gap converters has not been demonstrated for near-term experiments, and these devices do not have the right characteristics for long-term, low cost production given material constraints such as directional etching, compatible layer chemistry, and so on.
Recommendation 6. The sponsoring agency should discontinue the microminiature thermionic converter (MTC) program, the close-spaced vacuum converter tasks, the oxygenation effects research, and all current theory and theory validation work.
Begg, Lester L. 1998. “In-Core Thermionic Technology Development Program,” Paper IECEC-98–402, Proceedings of the 33rd Intersociety Engineering Conference on Energy Conversion, American Nuclear Society, Albuquerque, N.Mex., August 2–6, 1998.
DOE (Department of Energy). 1992. Report of the United States Space Nuclear Power and Propulsion Team: An Examination of the Space Nuclear Power and Propulsion Activities of the Commonwealth of Independent States. DOE, Washington, D.C., October 6.
Drake, T.R. 1998. “DoD’s Advanced Thermionics Program: An Overview,” Paper IECEC-98–404, Proceedings of the 33rd Intersociety Engineering Conference on Energy Conversion, American Nuclear Society, Albuquerque, N.Mex., August 2–6, 1998.
Gontar, A.S., et al. 1996. “Fuel Elements of Thermionic Converters,” Special Issue on Technology, Journal of the Franklin Institute 333A:(2–6).
Hatsopoulos, G.N., and J.Kaye. 1958a. “Analysis and Experimental Results of a Diode Configuration of a Novel Thermo Electron Engine,” Proc. I.R.E. 46(9):1574–1579.
Hatsopoulos, G.N., and J.Kaye. 1958b. “Measured Thermal Efficiencies of a Diode Configuration of a Thermo Electron Engine,” Journal of Applied Physics 29(7):1124–1125.
King, D.B., J.R.Luke, and F.J.Wyant. 1999. “Results from the Microminiature Thermionic Converter Demonstration Testing Program,” Proceedings of the Space Technology and Applications International Forum, CP458. American Institute of Physics, Melville, N.Y.
Rasor, N.S. 1998. “The Important Effect of Electron Reflection on Thermionic Converter Performance,” Paper 211, Proceedings of the 33rd Intersociety Engineering Conference on Energy Conversion, American Nuclear Society, Albuquerque, N.Mex., August 2–6, 1998.
Rightley, Gina. 1998a. “Summary Results of MTC Diode Studies,” Sandia National Laboratories memo, October 13.
Rightley, Gina. 1998b. “Summary Results of RHU/MTC System Studies,” Sandia National Laboratories memo, September 29.
Zavadil, K.R., J.H.Ruffner, and D.B.King. 1999. “Characterization of Sputter Deposited Thin Film Scandate Cathodes for Miniaturized Thermionic Converter Applications,” Proceedings of the Space Technology and Applications International Forum, CP458. American Institute of Physics, Melville, N.Y.
Marshall, A.C. 1998. “An Advanced Thermionic Theory for Development of High Performance Thermionic Energy Conversion Diodes,” Paper 212, Proceedings of the 33rd Intersociety Engineering Conference on Energy Conversion, American Nuclear Society, Albuquerque, N.Mex., August 2–6, 1998.
Marshall, A.C. 1998. “An Advanced Thermionic Theory: Recent Developments,” Proceedings of the Space Technology and Applications International Forum. American Institute of Physics, Melville, N.Y.
Marshall, A.C. 1998. “An Equation for Thermionic Currents in Vacuum Energy Conversion Diodes,” Applied Physics Letters 73:2971–2973.