and Thierry Caillat 7 and Jean-Pierre Fleurial 8 Jet Propulsion Laboratory, California Institute of Technology, Pasadena, CA 91109

Transcription

1 Enhancement of the Multi-Mission Radioisotope Thermoelectric Generator with Efficient Skutterudite Thermoelectric Couples: Current Status of the Skutterudite Technology Maturation Program Tim C. Holgate 1, Ying Song 2, Russell Bennett 3, Steven Keyser 4, Thomas Hammel 5 and Robert Sievers 6 Teledyne Energy Systems, Inc., Hunt Valley, MD and Thierry Caillat 7 and Jean-Pierre Fleurial 8 Jet Propulsion Laboratory, California Institute of Technology, Pasadena, CA The Multi-Mission Radioisotope Thermoelectric Generator (MMRTG) developed for NASA by Aerojet Rocketdyne and Teledyne Energy Systems, Inc. (TESI) has been providing reliable power to the Mars Curiosity Rover since its launch in 2011 (Mission on Mars began in August 2012). An improvement of the performance of the MMRTG is underway at TESI through a technology maturation program where higher efficiency skutterudite materials developed on the laboratory scale at NASA s Jet Propulsion Laboratory (JPL) are being further developed at the production level into flight-ready materials and components. The success of this project will result in an enhanced MMRTG (emmrtg) with an anticipated improvement of 25-30% in the beginning-of-life power and even greater improvement by end-of-life due to improved thermoelectric life properties. The status of the program, including engineering challenges, successes and current production capabilities will be presented. Nomenclature BOM = beginning-of-mission DCLT = Double Couple Life Tester MMLT = Mini-Module Life Tester EOM = end-of-mission HABI = hybridized aerogel-based insulation JPL = Jet Propulsion Lab emmrtg = enhanced Multi-Mission Radioisotope Thermoelectric Generator MMRTG = Multi-Mission Radioisotope Thermoelectric Generator SKD = Skutterudite TE = thermoelectric TESI = Teledyne Energy Systems, Inc. W E = Watts-electric = Watts-thermal W Th 1 Materials Engineer, Advanced Power Group 2 Project Engineer, Advanced Power Group 3 Principal Engineer, Advanced Power Group 4 Program Manager, Advanced Power Group 5 Engineering Manager, Advanced Power Group, AIAA Member 6 Director, Advanced Power Group, AIAA Member 7 Principal Engineer, Power and Sensor Systems Section, AIAA Member 8 Project Manager, Power and Sensor Systems Section, AIAA Member 1

2 I. Introduction he Multi-Mission Radioisotope Thermoelectric Generator (MMRTG) was designed to operate in a variety of Tenvironments, making it suitable for deep-space, orbital and terrestrial missions. The success of its first deployment, the Mars Science Laboratory s (MSL) rover, Curiosity, has resulted in the MMRTG being slated as the baseline power source for NASA s next planned mission to Mars in Based on the SNAP-19 radioisotope thermoelectric generators (RTGs) that provided decades of power to the Viking I & II rovers and the Pioneer 10 & 11 probes, respectively 1, Curiosity s MMRTG, designated F1, provided an electrical power output of just over 110 W E from a nominal thermal inventory of 2000 W Th at beginning-of-mission (BOM) 2. The MMRTG s reliable power is generated by the same TESI s (Teledyne Energy Systems, Inc.) proven heritage thermoelectric materials: n-type PbTe and p-type TAGS-85, with the latter being segmented with PbSnTe rather than SnTe (as in the SNAP-19s) to increase the thermoelectric efficiency of the couples 1. As the possible shortage of the required heat source fuel (238Pu) has threatened the future of space exploration 3, there is further need to improve the conversion efficiency of RTGs. In recent decades, thermoelectrics research has produced wellperforming high-temperature materials from several material classes (i.e., half-heuslers, skutterudites, Zintl phases, etc.). The Skutterudite Technology Maturation Program aims to increase the efficiency of the MMRTG by exchanging the heritage thermoelectric couples with ones consisting of n- and p-type skutterudites developed at NASA s Jet Propulsion Laboratory (JPL). Figure 1. CAD models of the MMRTG/eMMRTG and its components. The TE couples of the TE module assembly are to be upgraded from the heritage PbTe- TAGS to skutterudite-based couples. II. Program Description As a part of the Advanced Thermoelectric Couple (ATEC) task, with funding from NASA s Radioisotope Power Systems Project, JPL has demonstrated skutterudite (SKD) TE couples with conversion efficiencies on the order of 9% with a hot-side temperature of 600 C and a ΔT of 400 C across the couple 4. This improvement over the 7.5% demonstrated by MMRTG couples is largely due to the increased hot-side temperature compared to the 535 C of the MMRTG, which is made possible by the higher temperature stability of particularly the p-type SKD material, compared to that TAGS-85 segment of the p- leg in the MMRTG couple. An initial goal of the Skutterudite Technology Maturation program (SKD Tech Mat) is to transfer the technical know-how associated with the synthesis and processing of the SKD materials from JPL to TESI where the material synthesis may be scaled up to a production-ready level, and the production of elements and couples demonstrated. Ultimately, the project will culminate with the demonstration by TESI of flightlike TE modules consisting of SKDs TE couples that are ready to be dropped in as a one-to-one replacement of the MMRTG modules. The general design of the MMRTG/eMMRTG is presented in Fig. 1 and is described in more detail elsewhere 1,5. The underlying concept of the SKD Tech Mat program is to replace the heritage MMRTG couples with the JPL developed SKD couples without any significant design changes to the generator. The TE module assembly shown in Figure 1 consists of two layers: an array of TE couples set in insulation and a spring-and-piston array set in a metal module bar. While the latter is to remain unchanged, the TE couples will be upgraded to SKD couples, and the module insulation (only the insulation in the TE module assembly) is to be upgraded to a hybridized aerogelbased insulation (HABI) to better suppress sublimation from the TE elements. 2

3 The current design of the MMRTG is able to physically accommodate SKD couples without modification to the generator, and being able to provide those couples with the hot-side temperature of 600 C needed to take full advantage of the SKDs capabilities with only minor modification (treatment of the liner surface to increase heat flow from the heat-source chamber to the heat distribution block). The SKD Tech Mat program therefor stands to efficiently provide the increased weight-specific power and radio-fuel efficiency critical to future space missions. III. Skutterudite Technology Transfer The advanced skutterudite (SKD) materials developed at JPL are intended to serve as the lower temperature segment of an advanced high-temperature segmented couple capable of operating with a hot-side temperature of up to 1000 C. This program (ATEC) is ongoing and, as a whole, has yet to reach the readiness level for a technology transfer program. The SKD materials themselves having been demonstrated to be stable for more than one year at 600 C, were deemed ready for technology transfer to be used as drop-in replacements for the MMRTG heritage materials, giving up to a 25% boost in power at beginning-of-mission (BOM) and up to a 50% increase at end-ofmission (EOM). The degradation data obtained from couples on test at JPL were obtained on couples that were encased in aerogel and placed under a dynamic vacuum, consistent with the conditions of deep space and the intended use of the ATEC couples. As the MMRTG was designed to have a sealed argon filled TE chamber, conventional wisdom suggests that the degradation of the SKD materials may be less in the static argon environment of the MMRTG compared to in vacuum due to decreased sublimation rates of the SKD. The expected lower degradation rates of the SKD in comparison to those of the heritage TESI materials indicate that the gain in end-of-mission power of an emmrtg would be even greater up to 50% than an MMRTG. The SKD Tech Mat program is a multi-phase program that involves three iterations of the SKD couples.. Keeping with the proven SKD materials themselves, each iteration will involve further development of all other aspects of the couple design geometry, bonding techniques, interconnects, etc. needed to address the particular issues that may arise from transferring the technology from JPL s ATEC program to implementation in an emmrtg. In preliminary tests the first-iteration SKD couples produced at TESI and tested in argon with a hot-side temperature of 600 C and a ΔT of 400 C degraded faster than expected as metallic components of the interfaces oxidized. This unexpected result is believed to be due to the presence of a relatively high concentration of oxygen that is not consistent with the internal generator atmosphere effectively an over-test. Additionally, there has been observed an unfavorable reaction between SKD sublimation products and the first-iteration interconnect material (the hotshoe). These two adverse reactions have directed the current developmental efforts towards experimentally and computationally understanding the precise details of the MMRTG/eMMRTG internal TE compartment atmosphere (in terms of both residual moisture and oxygen) and towards the development of couples having the same n-skd and p- SKD materials joined with a more robust metallization/hot-shoe scheme. A. Skutterudite Materials At present, TESI has successfully demonstrated the ability to synthesize the SKD materials (p-type Ce x (Fe,Co) 4 Sb 12 and n-type (Ba,Yb) x Co 4 Sb 12 hereon denoted as n-skd and p-skd, respectively) on a scale that is consistent with the of the production of an emmrtg. Roughly five kilograms of each type have been produced at TESI. From each batch, a minimum of two quality assurance (QA) pucks are produced using with uniaxial hot-pressing. The room temperature density, electrical resistivity and Seebeck coefficient of each puck is measured in order to qualify the batch based on criteria established at JPL. The QA data is recorded in a database for tracking batch reproducibility and traceability back to raw materials and synthesis or processing parameters. In Fig. 2, the resistivity and Seebeck data of all QA pucks produced to date from all n-skd and p-skd batches is shown with the acceptance bands (shaded areas) demonstrating the progress Figure 2. QA puck data of all batches produced to date. The room temperature resistivity and Seebeck coefficient were measured on all QA pucks. The shaded areas represent the acceptance bands. 3

4 TESI has made toward reproducibility and quality of the batches. Additionally, several samples of each type have been sent to third-party laboratories for high-temperature TE properties measurements. The resulting data were consistent with that of TESI samples measured at JPL, as well as JPL s own sample data, within the uncertainties of the respective measurements. A current plan is being implemented to acquire high-temperature TE properties on a statistically meaningful number of samples from many SKD batches in order to construct a representative dataset that may be used to perform statistical studies of generator level performance predictions. B. SKD Thermoelectric Couples As previously mentioned, TESI in collaboration with JPL, is currently working on a second iteration of the SKD couple. Specifically, the zirconium based interfaces of the first-iteration couples has been replaced with ones that are refractory-based. Preliminary tests indicate the current design of the second-iteration couples to be significantly more robust than the first-iteration in terms of resistance to oxidation at the interfaces. These interfaces are preformed onto the SKD materials during the pressing of the pucks that are later diced into elements, resulting in rectangular bar-shaped elements Figure 3. Skutterudite Puck Production. (Left) Two different sizes of metallized skutterudite pucks next to a U.S. dime for reference. (Right) A metallized puck diced into elements that will be used in SKD couples. terminating with metallized ends that will serve as diffusion barriers as well as surfaces that may be easily bonded to an interconnecting hot-shoe by conventional brazing methods. Alternative hot-shoe materials are currently being explored to determine a candidate that does not react with the SKD sublimate. Several promising alloy and pure metal candidates have been identified and bonding trials are underway with some good initial results. So far, the same bonding procedures and methods that proved successful in bonding the first-iteration couples have been sufficient in bonding the second-iteration couples with the alternative metallization and interconnect materials. IV. Hybridized Aerogel-Based Insulation (HABI) JPL has demonstrated that the sublimation of the SKD materials in a vacuum can be significantly suppressed by encapsulating the SKD elements in a silica-based aerogel. When the aerogel is super-critically dried, the thermal conductivity can be lower than that of the common commercially available microporous insulations commonly used in RTGs. The super-critical drying process is difficult to achieve for aerogel cast in-place with components that may not be able to withstand the conditions of the process or simply fit into the apparatus required. While ambiently dried aerogel is much easier to achieve, the higher density results in a thermal conductivity that is at least twice that of the commercial insulations. At TESI, a process was developed resulting in a hybridized aerogel-based insulation (HABI) that forms in ambient conditions and has improved mechanical properties over the baseline ambiently dried aerogel. More importantly, the thermal conductivity of HABI is about half that of the pure ambiently dried aerogel and approximately the same as the commercial microporous insulation, such as Min-K A. Hybridized Aerogel-Based Insulation vs. TESI Heritage Module Insulation To test the effectiveness of the HABI to suppress couples degradation from material sublimation, couples were put on test in a custom test fixture (DCLT) described later. A side-by-side comparison of refractory-based couples, one set in HABI and the other protected by the same module insulation design ( heritage ) used in the MMRTG. Three pairs of couples were tested, with the first two having identical conditions of T Hot =650 C and the third couple at T Hot =600 C. In all cases, T Cold =235 C, which is consistent with conditions specific to a Mars mission. The power output of each couple was measured at ~83 mv, which corresponds to a total generator voltage of 32V for 768 couples in series-parallel pairs of the emmrtg. All curves were normalized to their respective values at 50 hours in order to get meaningful depiction of degradation rates after the initial settling of the couples as well as the text fixture the so-called burn-in time. Afterwards, outliers were removed and the data density considerably reduced for clarity. 4

5 Figure 4. Normalized Couple Output Power. The couple power versus time data were normalized to the values at 50 hours to allow for a burn-in time and then reduced and smoothed for clarity. Figure 5. Diffusivities and Diffusivity-times- Density of Baseline and Neutron-Irradiated HABI. Before and after irradiation results are indistinguishable within the uncertainty of the measurement. The results, presented in Fig. 4, demonstrate that the HABI offers superior sublimation protection compared to TESI s heritage module insulations at the conditions tested, which are higher than the design T Hot of the MMRTG (510 to 535 C) 6. While the slightly higher thermal conductance of a TE module insulated with HABI rather than the TESI heritage module insulation may cost a couple of Watts at the generator level at BOM, the EOM power output would be greatly increased due to suppressed sublimation of the SKD materials. B. Irradiation Tests on Hybridized Aerogel-Based Insulation In order to determine whether or not the radiation that an emmrtg would likely experience during its mission would have any adverse effects on the thermal conductivity of the HABI, multiple specimens were cut from a single sample of HABI. Enough specimens were cut to provide for the measuring of four baseline samples to serve as a reference in addition to the samples sent for irradiation. A total of one Ø25.4mm x ~8mm and five Ø12.7mm x ~10mm cylindrical samples were sent to the University of Dayton Research Institute for irradiation at the Ohio State Research Reactor. The five Ø12.7mm samples were stacked vertically and sealed in a single evacuated quartz ampoule with the Ø25.4mm sample sealed in a separate ampoule. Both ampoules were subjected to a fast neutron fluence of 2.6x10 13 n/cm 2 an exposure equivalent to twice that which is expected from the 238 Pu source over a 17 year mission plus that anticipated from a Europa flyby mission. The diffusivities of the baseline (BL) and neutron irradiated (NI) samples were measured using the laser flash analysis technique (LFA) with a TA Instruments DLF in an argon atmosphere. The diffusivity is arrived at through a fitting parameter of the temperature versus time signal, which is scaled by the square of the thickness of the sample: D measured L 2. In Fig. 5, the main plot shows the diffusivity of the four baseline samples and four samples taken from four of the five Ø12.7mm irradiated samples. It is important to note that the thickness of the sample is a key parameter in calculating the both the diffusivity and the density: D d. Since thermal conductivity is calculated from the product of diffusivity, density and heat capacity (κ=d d c v ), the D d product lessens the effect of the uncertainty of thickness. The inset of Fig. 5 gives the averages of this product for BL and NI, respectively, and the error bars represent the respective standard deviations. These data indicate that the irradiation has no appreciable effect on the thermal conductivity of the HABI since c v is independent of structure on any scale, and approaches the Dulong-Petit limit of 3 kb per atom at high temperatures 7, regardless of atomic specie. 5

6 Figure 6. TESI Couple and Module Test Fixtures. The cut-away views of the DCLT and Full Module fixtures reveal the inner details of the designs. The DCLT accommodates two couples with independent hot-side temperature control and a common cold-side temperature. The MMLT fixture contains two center hot-side heater plates for two pairs of facing modules with common hot-side temperatures. The compensating strap heaters outside the MMLT housing provide the cold-side temperatures. V. Test Plan and Procedures The success of the MMRTG was built on the confidence and validation that came from a great deal of testing at every level materials, couples, modules and generator level. The same rigorous testing is planned for the emmrtg, especially at the sub-generator levels. A. TE Properties of SKD Materials JPL is currently evaluating the degradation of the intrinsic TE properties (electrical resistivity, Seebeck and thermal conductivity) of SKD materials in argon. This follows the more than 1.5 years of data on JPL produced n- SKD and p-skd test coupons on test at JPL in a dynamic vacuum at temperatures consistent with the hot-side of a couple that shows no degradation in the TE properties. At present, the JPL produced SKD samples have been on test in argon for one year with no observable change in the TE properties. TESI produced samples have arrived at JPL and will be put on test in argon with the same conditions. Ultimately, all testing of SKD intrinsic TE properties will be performed for a minimum of two years. B. Couple Degradation and Power Output The power output and degradation of couples produced at TESI are evaluated using an in-house designed fixture that can accommodate two couples simultaneously with a common cold-side temperature and independent hot-side temperatures the Double-Couple Life Tester (DCLT). At present, four DCLTs have been manufactured and are in use. These systems periodically measure the open-circuit voltage and perform a V-I power sweep while intermediately measuring the power at a specific load voltage. The all-stainless housing and copper knife-edge seals result in a clean environment where moisture and oxygen levels can be monitored periodically through sampling loop. The DCLTs will be used to assess the sublimation rates of the SKD materials and changes in the interfacial contact resistance of the couple as a function of time and temperature. Once the final configuration of the couples has been down-selected from the second-iteration couple development, the couples will go on test for a minimum of one year. C. Mini-Module and Full-Module The MMRTG/eMMRTG generators contain sixteen 48-couple modules. As an intermediate step towards testing a full module, mini-modules consisting of three couples set in insulation will be tested. A Mini-Module Life Tester (MMLT) has been designed, and the first of four planned for the current phase of the program has been manufactured. Each MMLT will hold four 3-couple mini-modules with the hot-side of all modules at the center of the fixture in a two-by-two module layout. Two facing modules will share a heater plate and hot-side temperature, while two adjacent modules will share a cold-side temperature. The cold-side temperature will be provided by external strap heaters affixed to the housing. A preliminary outgassing test will be performed to assess the proper 6

7 procedure for acquiring a prototypic environment of a generator, and for the actual mini-module test the housing will be welded shut. The planned duration of each MMLT test is one year, from which the module insulation thermal conductance and overall degradation rates of the mini-modules will be assessed. The full-module tester is currently being designed and is planned for the next and final phase of the program. Based on a design used for MMRTG type heritage modules, the fixture will test the thermoelectric efficiency of the module and the overall degradation rate. VI. Conclusion The MMRTG has been providing reliable power to the Mars Science Laboratory rover, Curiosity, for four years. The high performing skutterudite materials developed at JPL have shown promise for long-term operation at temperatures consistent with the capabilities of the MMRTG. The realization of a drop-in upgrade of the TE couples from TESI heritage materials to the skutterudites will result in an enhanced-mmrtg capable of providing 25% more power at BOM and up to 50% at EOM. Current efforts at JPL and TESI to bring the SKD technology up to a flight-ready status are progressing towards that goal, following a rigorous multi-level characterization and validation plan designed to ensure the success of these new materials. Acknowledgments The authors would like to acknowledge all the contributing members of the following organizations: JPL, NASA, DOE, TESI and Aerojet Rocketdyne. References 1 Hammel, T., Bennett, R., Otting, W., and Fanale, S., Multi-Mission Radioisotope Thermoelectric Generator (MMRTG) and Performance Prediction Model, 7th International Energy Conversion Engineering Conference, Denver, CO: AIAA, Dudzinski, L., Hamley, J., McCallum, P., Sandifer, C., Sutliff, T. J., and Zakrajsek, J. F., NASA s Radioisotope Power Systems Program Status, 12th International Energy Conversion Engineering Conference, Cleveland, OH: AIAA, Gwynne, P., Space missions threatened by shortage of plutonium, Physics World, vol. 24, 2011, p Caillat, T., Firdosy, S., Li, B. C.-Y., Huang, C.-K., Uhl, D., Smith, K., Paik, J., Fleurial, J.-P., Bennett, R., and Keyser, S., Skutterudite-Based Advanced Thermoelectric Technology for Potential Integration Into an Enhanced MMRTG, Proceedings of NETS-2014, Stennis Space Center, MS: American Nuclear Society, Holgate, T. C., Bennett, R., Hammel, T., Caillat, T., Keyser, S., and Sievers, B., Increasing the Efficiency of the Multimission Radioisotope Thermoelectric Generator, Journal of Electronic Materials, Dec Hammel, T. E., Bennett, R., Sievers, R. K., Keyser, S., Otting, W., and Gard, L., Multi Mission Radioisotope Thermoelectric Generator (MMRTG) Performance Data and Application to Life Modeling, 11th International Energy Conversion Engineering Conference, San Jose, CA: AIAA, Petit, A.-T., and Dulong, P.-L., Recherches sur quelques points importants de la Théorie de la Chaleur, Annales de Chimie et de Physique, vol. 10, 1819, pp