Nuclear and Emerging Technologies for Space (2012) 3054.pdf

REVIEW OF HISTORICAL TUNGSTEN CERMET FUEL DEVELOPMENT PROGRAMS AND LESSONS LEARNED. J A. Webb1, 1Idaho National Laboratory (1765 N. Yellowstone Hwy, Idaho Falls, ID 83415, [email protected])

Introduction: Starting in the 1950’s researchers particles were tungsten coated by CVD techniques have designed and tested many different types of space using WCl6 and WF6 as a pre-cursor [9], [10]. After reactors, all of which require fuels that can sustain very coating with tungsten the particle were consolidated by high temperatures for durations ranging from a few hot pneumatic compaction and by hot pressing tech- hours to years of continuous operations.[1],[2],[3]. niques. Hot hydrogen thermal cycling of both sets of As early as the 1950’s fuels engineers realized the specimens showed a vastly superior fuel retention to potential of uranium bearing tungsten fuels to sustain that of specimens consolidated by mixed particle fabri- high temperatures while retaining fission products[4]. cation techniques. The superior fuel retention proper- While the advantages of tungsten were realized early ties were attributed to the lower grain boundary in the 20th century, the fabrication techniques required misorientation of the CVD tungsten coating over that to produce W-UO2 fuels did not exist pushing the early of powder metallurgy tungsten and the segragation of National Aeronautics and Space Admininstration to fuel particles by the coating layers. Further studies at consider graphite based fuels as the primary fuel for General Electric seemed to indicate the mixing coated propulsion reactors. While the graphite fuels could be particles such that the larger particles are one order of fabricated, they were plagued with mid band corrosion magnitude larger than the smaller particles yielded problems and could not contain all of the fission prod- sintered kernels that were more symmetric and which ucts within the graphite matrix/substrate[3]. also had better fuel retention properties. Fuels studies Starting in the mid 1960’s the General Electric Co., also showed greater fuel retention for smaller fuel par- the Argonne National Laboratory and NASAs Lewis ticles than for larger particles during hot hydrogen test- Research Center embarked upon a path to develop fab- ing. rication method for forming W-UO2 fuels and for test- Stabilizers and Cladding: Studies were also con- ing different fuel-meat configurations and fuel element ducted to stabilize UO2 fuel kernels such that they have geometries[5],[6]. The combination of these programs a higher decomposition temperauture in a hot hydrogen proved the viability of W-UO2 fuels for ultra-high environment. Stabilizers such as ceria, yttria, and temperatue space reactors.. gadolinia were mixed with UO2 to produce a defect- Mixed vs. Coated Particles: Early W-UO2 fuels type fluorite lattice structure at amounts ranging from development was focused on establishment of fuel 2.5 mol.% to 10 mol.%. With respect to fast spectrum kernel specifications and how they affected the final reactors, gadolinia (Gd2O3) was found to be the opti- consolidated product. Early work conducted by the mum stabilizer for operations above 2700 K; however, Batelle Northwest Laboratory (BNWL) investigated the concentration is dependent on the number of en- the hot pneumatic impaction of compacts containing gine restarts required and the required engine thrust-to- mixed particles of tungsten and UO2[7]. In one set of weight-ratio. Gadolinia increases the fuel neutron ab- trials the UO2 particles were typical asymmetric ker- sorption cross section, requiring an increase in reactor nels used in traditional powder metallurgy studies. In mass with an increasing gadolinia concnetraiton [7]. the second batch of consolidation trials the fuel kernels Studies were also conducted where the tungsten were spherical in nature. Hot hydrogen thermal cy- matrix was mixed to approximately 2 wt.% ThO2 cling of the two fuel types showed a large fuel loss in (thoria) [7]. Thoriating the tungsten matrix seemed to both cases that exceeded 20% after a mere 10 cycles; increase fuel retention abilities of the fuel, likely by however, the fuel loss for the case with spherical parti- refining the tungsten grain size during the fabrication cle was far less than in the case with non-spherical process. [11] Surrounding the fuel in a tungsten or particles. Microstructural analysis showed that the tungsten-rhenium cladding layer also served to de- stress concentration produced at the asymmetric verti- crease fuel loss. ces of the UO2 kernals was causing excacerbated fuel Mechanisms and Recommendations: Fundamen- loss, leading to the requirement for spherical fuel ker- tal studies proved that the primary fuel loss mecha- nels. Analysis also showed that in both cases the UO2 nisms were mechanical and thermochemical in nature. kernel expanded at temperature faster than the tungsten The UO2 fuel has a higher thermal expansion coeffi- matrix and exerted a stress on the matrix that caused cient than the tungsten matrix and as such the UO2 crack development through creep mechanisms. exerts a stress on the tungsten matrix at elevated tem- In the next set of trials conducted by BNWL also perature. Combined with other studies that show an repeated by General Electric[8], spherically symmetric increasing tungsten creep strength with decreasing Nuclear and Emerging Technologies for Space (2012) 3054.pdf

grain size, extra care should be taken to fabricate W- UO2 fuels with the minimum possible tungsten grain size. At elevated temperature it has been shown that UO2 will begin to volatize at around 1600 K, at which point liquid uranium can diffuse into grain boundaries. During the cool down portion of a thermal cycle, hy- drogent that has diffused into the grain boundaries can hydride the uranium and cause grain boundary separa- tion. Great care should be taken to minimize the tung- sten grain size so as to add more grain boundaries and impede hydrogen diffusion into the specimen. Smaller fuel kernels (below 50 µm) were shown to decrease fuel loss. Great care should be taken to fabri- cate and tungsten coat high purity UO2 fuel kernels with the smallest possible tungsten grain size. Effort should also be dedicated to coating UO2 particles such that the grain boundary misorientation in the tungsten coating ends up being below 35 degrees. In conjunction with coating particles with a fine grained tungsten layer, studies should also be con- ducted to determine the effects of gadolinia concnetrai- tons ranging from 0 to 10 mol.% and cladding layers of tungsten and tungsten-rhenium of varying thickness. References: [1] Angelo J.A. and Buden D. (1985) Space Nuclear Power, [2] Bhatacharyya S.K. et al. (1993) Space Exploration Initiative Fuels, Materials and Related Nuclear Propulsion Technologies Panel, NASA TM-105706. [3] Bhatacharyya S.K. (2001) An Assessment of Fuels for Nuclear Thermal Propulsion, Argonne National Laboratory, ANL/TD/TM01-22, [4] Rom F.E. and Ragsdale R.G. (1962) Advanced Con- cepts for Nuclear Rocket Propulsion, NASA SP20 [5] (1968) Nuclear Rocket Terminal Report, ANL-7236 [6] 710 High Temperature Gas Reactor Closeout Re- port , Vol. I and VI, GEMP-600, [7] Baker R.J. (1966) Basic Behavior and Properties of W-UO2 CERMETS, BWNL-394 [8] Marlowe M.O. and Kaznoff A.I., Dev- leopment of Low Thermal Expansion Tungsten-UO2 CERMET Fuel, GESP-9014, [9] MacInnis M.B. and Schulze H.O. (1965) Tungsten Cladding of Tungsten- Uranium Dioxide Composites by Chloride Vapor Deposition, NASA CR 54728 [10] Lamartine J.T. and Hoppe A.W. (1965) Tungsten Cladding of Tungsten- Uranium Dioxide Composites by Deposition from Tungsten Hexafluride, NASA CR-54426. [11] Tak- kunen P.D., Gluyas R.E. and Gedwill M.A. (1967) Thermal Cyclic Behavior of Tungsten-Uranium Diox- ide CERMETS Containing Metal Oxide Additives, NASA TM X-1446