Proceedings of the HTR 2014 Weihai, China, October 27-31, 2014 Paper HTR2014-31254 Detection and Analysis of Particles with Failed SiC in AGR-1 Fuel Compacts John D. Hunn, Charles A. Baldwin, Tyler J. Gerczak, Fred C. Montgomery, Robert N. Morris, Chinthaka M. Silva, Paul A. Demkowicz1, Jason M. Harp1, Scott A. Ploger1 Oak Ridge National Laboratory (ORNL) P.O. Box 2008, Oak Ridge TN, 37831-6093, USA Phone: +001-865-574-2480, [email protected] 1Idaho National Laboratory (INL) P.O. Box 1625, Idaho Falls ID 83415-6188, USA Abstract – As the primary barrier to release of radioactive isotopes emitted from the fuel kernel, retention performance of the SiC layer in tristructural isotropic (TRISO) coated particles is critical to the overall safety of reactors that utilize this fuel design. Most isotopes are well-retained by intact SiC coatings, so pathways through this layer due to cracking, structural defects, or chemical attack can significantly contribute to radioisotope release. In the US TRISO fuel development effort, release of 134Cs and 137Cs are used to detect SiC failure during fuel compact irradiation and safety testing because the amount of cesium released by a compact containing one particle with failed SiC is typically ten or more times higher than that released by compacts without failed SiC. Compacts with particles that released cesium during the AGR-1 irradiation test or post-irradiation safety testing at 1600– 1800°C were identified, and individual particles with abnormally low cesium retention were sorted out with the ORNL Irradiated Microsphere Gamma Analyzer (IMGA). X-ray tomography was used for three-dimensional imaging of the internal coating structure to locate low-density pathways through the SiC layer and guide subsequent materialography by optical and scanning electron microscopy. All three cesium-releasing particles recovered from as-irradiated compacts showed a region where the inner pyrocarbon (IPyC) had cracked due to radiation-induced dimensional changes in the shrinking buffer and the exposed SiC had experienced concentrated attack by palladium; SiC failures observed in particles subjected to safety testing were related to either fabrication defects or showed extensive Pd corrosion through the SiC where it had been exposed by similar IPyC cracking. I. INTRODUCTION pyrolytic carbon (IPyC), a polycrystalline SiC, and a dense outer pyrolytic carbon (OPyC) [2]. These Development and qualification of tristructural layers were designed to contain fission products isotropic (TRISO) coated particle fuel is part of the throughout the life of the fuel particle. US Department of Energy's Advanced Reactor Post-irradiation examination (PIE) of AGR-1 Technologies Initiative to promote scientific irradiation test fuel has focused on evaluating the understanding and demonstrate the technical fuel performance during irradiation [3] and during viability of high-temperature gas-cooled reactor post-irradiation safety testing [4]. Retention of the technology. The AGR-1 experiment was the first in a many isotopes generated in the fuel as a product of series of US irradiation tests being performed on fission, neutron activation, and radioisotopic decay cylindrical compacts containing TRISO particles in a was one of the primary indicators of fuel graphite matrix with kernels made up of a mixture of performance that was studied. Rapid release of uranium carbide and uranium oxide (UCO) [1]. gaseous radioisotopes of krypton or xenon was used AGR-1 TRISO particles consisted of four concentric to monitor for TRISO coating failure in individual coatings on a 350-µm-diameter spherical UCO fuel particles. Detection of this type of coating failure kernel: a low-density carbon buffer, a dense inner was an extremely rare event, with no occurrences Proceedings of the HTR 2014 Weihai, China, October 27-31, 2014 Paper HTR2014-31254 during the three-year AGR-1 irradiation test of III. DETECTING SIC FAILURES DURING POST- approximately 298,000 particles [5] and only two IRRADIATION SAFETY TESTING suspect particles in one of the fourteen safety-tested compacts (a compact tested at 1800°C) [6]. After completion of the AGR-1 irradiation test, Release of a significant portion of an individual numerous compacts were subjected to post- particle's inventory of cesium was used to detect the irradiation safety testing in helium atmosphere to presence of particles with failed SiC. Particles with explore the fuel performance at maximum expected through-layer pathways in the SiC that allow cesium accident temperature during a depressurized to easily penetrate the layer will not rapidly release conduction cooldown event (1600°C) and at higher krypton or xenon if at least one pyrocarbon layer temperatures (1700 and 1800°C) to explore the fuel remains intact. Cesium release from four particles performance margin. During these safety tests, with failed SiC was observed in the PIE of the cesium release was monitored with water-cooled graphite holders that surrounded the compacts in the collectors that were periodically removed from the irradiation test capsules [7]; three of these particles furnace and gamma scanned to measure condensed were recovered for further analysis discussed below. cesium. Fig. 2 shows the measured release rate for Three particles released cesium during 1600°C one of the 1600°C safety tests. In this test, a burst of safety testing, and numerous particles released cesium was detected soon after reaching 1600°C cesium during safety testing at 1700 and 1800°C [8]; (presumably initiated by SiC failure in a single many of these were also recovered for analysis. particle). After the initial release, the rate of cesium condensing on the cups slowly decreased back to II. IDENTIFYING COMPACTS CONTAINING background levels and the total integrated release PARTICLES WITH FAILED SIC was 68.5% of an average particle's inventory. The AGR-1 irradiation test train consisted of six 8E-6 1600 independently-monitored capsules, each containing twelve compacts in three stacks inside a graphite hr) 67EE--66 Cs-134 11240000 C hfcrooomlmdpe roth;n eea nnctaoslm ypspriaso cvotisfd e[t9dh] e.d sCea toahl oliolmdne arrtsea ddai-ongidasm ootmtohpaei rcs ccraaenplnesiaunslgee se Rate (1/ 345EEE---666 CThs-e1rm37o couple 681000000 0 perature ° of the graphite holders also provided information on elea 2E-6 400 Tem the spatial distribution of this release. Two graphite R 1E-6 200 holders showed regions of elevated 134Cs in the 0E+0 0 vicinity of three compacts (e.g., Fig. 1). Additional 0 50 100 150 200 250 300 350 analysis described below was performed on these Elapsed Time (Hours) compacts to determine the source of the elevated Fig. 2. Fraction of the compact's cesium inventory cesium release. released per hour during 1600°C safety testing of AGR-1 Compact 4-1-2. 0.15 Cesium release was more complicated during 0.36 1800°C safety testing due to releases from multiple particles with failed SiC, with failures occurring at 0.25 5-­‐2-­‐2 varied times and each particle releasing a varied 0.00 fraction of its cesium inventory (Fig. 3). 0.00 0.12 9E-6 1800 323 Cs-137 8E-6 1600 0.00 hr) 7E-6 323 Cs-134 1400 C 5-­‐2-­‐1 5-­‐2-­‐3 010...605441 ase Rate (1/ 3456EEEE----6666 681100020000 00 mperature ° 0.72 Rele 2E-6 551133 CCss--113347 400 Te 1E-6 200 Thermocouple 0.89 0E+0 0 0 50 100 150 200 250 300 350 0.49 0.61 0.34 0.19 0.33 0.00 0.24 0.61 0.20 0.31 0.62 0.31 Elapsed Time (Hours) Fig. 1. Intensity map of 134Cs activity in one section Fig. 3. Fraction of the compact's cesium inventory of Capsule 5 graphite holder showing hot spots (red) released per hour during 1800°C safety testing of adjacent to Compacts 5-2-1 and 5-2-3 [9]. AGR-1 Compacts 3-2-3 and 5-1-3. Proceedings of the HTR 2014 Weihai, China, October 27-31, 2014 Paper HTR2014-31254 III. ISOLATING SUSPECT PARTICLES WITH IV. ANALYZING FAILED SIC IN LOW-CESIUM- LOW CESIUM INVENTORIES RETENTION PARTICLES As-irradiated and safety-tested compacts that After using IMGA to isolate particles with exhibited cesium release indicative of failed SiC significantly reduced cesium inventory, the particles were subjected to electrolytic deconsolidation and were subjected to a series of analyses to investigate acid leaching to separate the TRISO-coated particles their internal microstructure. Particles were first from the encapsulating graphite matrix. In most imaged in three dimensions (3D) by non-destructive cases, the particles were washed and dried for x-ray tomography, which was very effective at examination using the computer-automated ORNL identifying the presence of through-layer anomalies Irradiated Microsphere Gamma Analyzer (IMGA). in the SiC, as well as providing detailed internal Detailed descriptions of the ORNL equipment and structure data for all the other layers and interfaces. procedures relevant to IMGA and deconsolidation- Particles were then mechanically sectioned and leach-burn-leach (DLBL) are reported elsewhere [4]. polished for optical and scanning electron The IMGA was used to survey the cesium inventory microscopy (SEM), using the 3D x-ray images to in every recovered particle. Particles were placed in orient and guide the process; this approach offered a glass jars and individually extracted with the IMGA much higher probability that SiC failures could be vacuum needle for gamma analysis (Fig. 4). After a successfully exposed for further study. 50–100 s counting time, the relative inventories of 137Cs and 144Ce were used to sort particles, with IV.A. Imaging SiC Failures with X-ray Tomography programmed criteria to isolate particles that had inventories outside the normal distribution. Detailed descriptions of the coated particle x-ray tomography equipment and method can be found in Reference [4]. Fig. 6 shows a typical AGR-1 TRISO particle, where the only evident effects of the three- year irradiation were a slight enlargement in kernel diameter and densification of the buffer layer, which caused it to detach from the IPyC. Fig. 4. Sequential images showing IMGA computer- OPyC SiC automated vacuum needle selecting a single TRISO IPyC particle for gamma counting and depositing it into a Buffer designated vial after analysis. Kernel As an example of the method, Fig. 5 shows the results from the IMGA survey of the Compact 3-2-3 particles recovered after 1800°C safety testing. Nine particles with cesium inventories from 13–80% of calculated were found to lie below the normal distribution and were automatically sorted out for further analysis. A few additional particles in the lower tail of the main distribution were also segregated for analysis. 800 Fig. 6. Composite x-ray image of a typical AGR-1 cy 600 fuel particle from Compact 4-1-2; a 2D tomograph n ue has been superimposed with a 3D visualization of q Fre400 9 particles with the whole particle. Bright features at the bottom of e low Cs retention the particle are from bubbles in the mounting epoxy. cl rti200 a P More than twenty particles with failed SiC have 0 been examined with x-ray tomography, and only one 0.0 F0.2 raction0.4 of Ca0.6 lculate0.8 d Cs-131.0 7 Inve1.2 ntory 1.4 failure mechanism has been observed in particles Fig. 5. Cesium inventory distribution determined with coatings that conform to the intended AGR-1 with IMGA for particles from 1800°C safety-tested TRISO design. Ironically, this failure mechanism AGR-1 Compact 3-2-3. was directly related to radiation-induced changes in Proceedings of the HTR 2014 Weihai, China, October 27-31, 2014 Paper HTR2014-31254 the buffer, which traditionally is viewed as the least a Spearhead fracture important layer for fission product retention. SiC failure in buffer and IPyC Densification of the buffer from fission product recoil and neutron irradiation is unavoidable at the high burnup and neutron irradiation doses experienced by the AGR-1 fuel particles [7]. How the buffer responded to dimensional changes and interacted with the IPyC layer often determined whether the IPyC layer remained intact throughout the irradiation test. Fig. 7 shows various ways that IPyC cracking in particles with failed SiC was either directly related to buffer fracture or to the shrinking buffer pulling away from the IPyC. Although buffer fracture or incomplete delamination from the IPyC only rarely resulted in IPyC cracking, evidence from materialography of AGR-1 particles suggests that strong bonding between the buffer and IPyC increased the likelihood for IPyC cracking and the b SiC failure IPyC crack spearhead-shaped fracture like that in Fig. 7a was observed in every particle that exhibited buffer fracture without buffer/IPyC debonding [10]. Overall observation of the buffer/IPyC interaction in AGR-1 fuel particles has indicated that low interface strength to enhance buffer/IPyC debonding would be preferable and could further minimize cracking in IPyC layers like those in the AGR-1 fuel, where pyrocarbon anisotropy and density were successfully tailored to reduce radiation-induced cracking [7]. Except for two particles with pre-existing, as- fabricated defects (discussed below), all observed SiC failures occurred where IPyC cracking exposed the inner surface of the SiC. The x-ray images of the safety tested particles in Fig. 7 clearly show low- density regions (darker regions) penetrating the SiC where the IPyC cracks reach the SiC interface (the pathway through the SiC is less evident in the as- c irradiated particle shown in Fig. 7c). Clusters of higher density material appear as bright spots in the images and are most prevalent in the areas around the degraded SiC. These clusters are identified in the SEM discussion below as predominantly Pd and U and presumed to be linked with localized corrosion of the SiC. Similar clusters are often observed in irradiated AGR-1 fuel at intact IPyC/SiC interfaces, but at lower concentration and without any obvious adverse impact on the SiC structure. Fig. 7 shows protrusion of kernel material into some of the gaps between the buffer fragments. This was not unusual behavior in particles with fractured buffer and, in general, did not appear to be SiC failure connected with failed SiC [10]. However, Fig. 8 and Metal clusters (Pd and U) Fig. 9 show a very unusual particle with failed SiC that was recovered after 1800°C safety testing where Fig. 7. X-ray tomographs of AGR-1 particles with the kernel protruded through a spearhead fracture failed SiC; (a) Compact 3-2-3 Particle 5 after and reached the SiC layer, resulting in enhanced 1800°C safety testing, (b) Compact 3-3-1 Particle 1 interaction. This is the only AGR-1 particle of this after 1700°C safety testing, and (c) Compact 5-2-3 type that has been observed. Particle 2 after completion of AGR-1 irradiation test. Proceedings of the HTR 2014 Weihai, China, October 27-31, 2014 Paper HTR2014-31254 Fig. 8a also shows delamination at the IPyC/SiC a interface extending away from the IPyC crack and leading to tangential cracks in the SiC. These SiC cracks ran circumferentially and did not result in a SiC failure. Similar structure has been observed in as-irradiated AGR-1 fuel particles without associated cesium release that would indicate that SiC cracks Kernel traversed the layer [10]; in this particle, the SiC protrusion failure during 1800°C safety testing did not occur in the region with preexisting SiC cracks. Delamination at the IPyC/SiC interface has been observed most often when the buffer/IPyC interface remained intact and spearhead fracture occurred [10]. As mentioned above, only two particles have been observed with SiC failure mechanisms different from the general mechanism of SiC exposure and SiC failure chemical attack due to dimensional changes in the buffer cracking the IPyC. Both particles were b recovered from compacts safety tested at 1600°C and exhibited SiC defects introduced during particle coating. Fig. 10 shows the particle that released cesium during safety testing of Compact 3-3-2; it had a malformed and very porous SiC layer that failed during safety testing. This fabrication defect was caused by momentary overfluidization of the particle between the IPyC and SiC coating steps, where the particle was ejected above the coating bed and picked up carbon soot from the chamber wall; subsequent SiC deposition infiltrated the soot and produced the abnormal structure [2]. The presence of soot inclusions in the SiC was monitored as part of the AGR-1 quality control process and Widespread fluidization conditions were optimized to minimize SiC corrosion the population and severity of these defects. Nevertheless, the fuel particle composite used in Fig. 8. Orthogonal x-ray tomographs showing kernel Compact 3-3-2 was known to have a small fraction protruding to the SiC in Compact 3-2-3 Particle 6 of these defects (≤10-3, and most of those less severe after 1800°C safety testing. than this unusually extreme example). Porous SiC Carbon soot Partially leached kernel Fig. 9. Three-dimensional visualization of kernel surface in particle shown in Fig. 8 showing shape of Fig. 10. Particle that released cesium during 1600°C kernel protrusion through buffer/IPyC fracture. safety testing of Compact 3-3-2. Proceedings of the HTR 2014 Weihai, China, October 27-31, 2014 Paper HTR2014-31254 Fig. 11 shows the particle that released cesium Neither of the particles with SiC failures related during safety testing of Compact 4-1-2; it also had a to as-fabricated defects released cesium during the carbon soot inclusion, this time between the buffer three-year irradiation test (even with the extremely and IPyC layers. The soot inclusion resulted in a compromised SiC layer evident in Fig. 10), and both noticeable dimple in the SiC layer. During particles retained cesium until after reaching 1600°C irradiation, the buffer and IPyC did not separate and during safety testing (e.g., Fig. 2). This indicates appear to have fractured as a unit. Spearhead cracks that, although most of the observed internal fracture, through the buffer and IPyC led to regions of delamination, and dimensional changes presumably IPyC/SiC delamination, and fractures were observed occurred during irradiation, SiC failure did not occur in the SiC at the edges of these delamination regions until additional stress was applied by the post- that tended to curve around the particle in a irradiation safety testing. The x-ray images of these circumferential direction but in some locations were particles show partially missing kernels. Fig. 11a connected to radial cracks through the layer (Fig. shows that the Compact 4-1-2 particle is missing its 11a). The 3D visualization of the SiC in Fig. 11b OPyC layer; exposed uranium detected during acid shows that one continuous, through-layer crack leaching indicated that the OPyC was initially intact circumscribes the particle and runs about halfway during deconsolidation (consistent with no 85Kr around the rim of the dimple; at the locations release during safety testing) but later broke off, marked A and B, secondary cracks branch off and allowing the acid to dissolve some of the kernel continue around the dimple but do not connect. material. Similar acid leaching of the kernel appears to have occurred in the particle from Compact 3-3-2, but no cracks in the pyrocarbon layers were found. a SiC failure IV.B. Sectioning of Particles with SiC Failures After x-ray imaging, particles were mounted in epoxy and polished planar sections were prepared that revealed a portion of the SiC failure for detailed Carbon Spearhead soot analysis. The 3D x-ray data was studied to determine fracture inclusions the optimum mounting orientation, and successive tomographs perpendicular to the grinding direction were compared to optical micrographs of the exposed section to periodically monitor progress and successfully arrive at the target depth by matching up features such as fractures and debonds (Fig. 12). SiC failure a Fracture 1 Fracture 1 b B Dimple Debond Debond b Fracture 1 Fracture 1 Fracture 2 Fracture 2 A SiC failure Debond SiC failure Debond Fig. 12. X-ray tomograph/optical micrograph pairs showing how x-ray imaging was used to guide materialographic preparation of the Compact 5-2-3 Fig. 11. Particle that released cesium during safety particle in Fig. 7c (grinding up from the bottom of testing of Compact 4-1-2; (a) x-ray tomograph and that image). Optical images (right) show in-progress (b) semitransparent 3D visualization of SiC surface. sectioning prior to final polish and cleaning. Proceedings of the HTR 2014 Weihai, China, October 27-31, 2014 Paper HTR2014-31254 Polished specimens were prepared in a dedicated IV.C. Imaging SiC Failures by Optical Microscopy materialography hot cell at ORNL that houses a Buehler Minimet® 1000 grinder-polisher with the Optical microscopy provided additional detail of optional precision thinning attachment and a Leica the region around the SiC failure, but was limited by DMI-5000 inverted optical microscope designed for the fact that the inherent three-dimensional structure remote operation (Fig. 13). Grinding and polishing could only be imaged in a single intersecting plane. was accomplished using metal-bonded-diamond Connecting pathways through the SiC layer were discs and diamond lapping films. To minimize rarely visible in a single plane, so presumption of movement and drop out of loose fragments during structure above and below the imaged plane must be grinding, void space was filled by vacuum inferred from the x-ray tomography data. Fig. 15 and impregnation with epoxy, typically during early Fig. 16 show planar sections that intersect the SiC grinding as soon as a buffer fracture or buffer/IPyC corrosion found in two of the particles that released gap (if present) was sufficiently exposed to allow cesium during the AGR-1 irradiation test. epoxy to flow down into the particle (e.g., Fig. 12a). Particle 1 from Compact 5-2-3 (Fig. 15) shows a crack through the IPyC layer; on one side of this crack the buffer is still intimately bonded, while on the other side the buffer debonded by separating away from itself (leaving some residual material attached to the IPyC). IPyC cracks have often been found to be located at a buffer/IPyC delamination boundary like this and their formation can be presumed to be related to the debonding process. Foreign matter in the IPyC crack is discussed in the next section, where SEM analysis showed the Buehler Optical presence of palladium which is presumed to have Minimet microscope been the cause of the local degradation of the SiC Fig. 13. Irradiated Fuel Examination Laboratory layer. Damage in the SiC only penetrates about one- materialography hot cell. third of the way through the layer in this observation plane and x-ray tomography could only marginally Most materialographic sections were polished resolve a low-density pathway through the layer. with 1.0–0.1-µm-diamond lapping films prior to However, cesium was released from this particle in optical imaging. Fig. 14 shows a typical as-irradiated concentrations too high to be explained by passage particle with various common features often through intact SiC (based on the very low to non- observed in AGR-1 irradiated fuel particles. The existent cesium loss from particles and compacts in polishing process was very successful in maintaining the absence of failed SiC). Presumably, there was a flat surface and revealing the nature of the layer damage deeper into the SiC layer in a plane above or interfaces (intimate stitching between the IPyC/SiC below the image plane or a through-layer failure and OPyC/epoxy was observed). Prior to SEM, existed elsewhere in the particle. During grinding, a optically-polished samples were further polished similar IPyC crack with clustered foreign matter and with a water-based, alkaline suspension of 0.04-µm- the onset of SiC attack was observed in Particle 2 colloidal silica to remove residual scratches and from Compact 5-2-3 (Fig. 7c), but additional perform a mild chemical etch of the SiC. grinding did not unveil the suspected penetration. Detached Foreign matter Detached buffer in IPyC crack OPyC SiC damage Porous Residual buffer Epoxy in gaps kernel attached to IPyC Fig. 14. Typical as-irradiated Compact 5-2-1 particle Fig. 15. Compact 5-2-3 Particle 1 with SiC that showing common features. failed during irradiation testing. Proceedings of the HTR 2014 Weihai, China, October 27-31, 2014 Paper HTR2014-31254 The particle from Compact 5-2-1 that released from the start of the AGR-1 irradiation test, yet the cesium during the AGR-1 irradiation test showed a particle did not release cesium during the three-year more obvious through-layer penetration in the x-ray experiment. This implies that the porous structure in imaging (Fig. 16a) and a large degraded area in the the defective region did not present a connected SiC was clearly evident while sectioning the pathway through the SiC layer and the abnormally- particle. As grinding progressed through the thin SiC acting to seal this area remained intact until damaged region, the position of the degraded area in failure after heating to 1600°C. The crack shown in the observed plane moved from the outer edge of the Fig. 17 was at least one through-layer defect that SiC toward the IPyC/SiC interface, indicating the could have been responsible for the observed cesium corrosion pathway through the layer was oriented at release. an angle to the grinding plane. Sectioning was halted at the plane shown in Fig. 16b so that SEM analysis could be performed prior to risking further material SiC removal to reveal the initial point of attack at the crack IPyC/SiC interface where the IPyC crack extended to the SiC and corrosion presumably initiated. a IPyC crack Fig. 17. Compact 3-3-2 particle with soot inclusion defect that failed during 1600°C safety testing. A crack through intact SiC covering the porous region produced a pathway through the layer. IV.D. Analyzing SiC Failures with SEM SiC failure Scanning electron microscopy was performed using a JEOL JSM-6390L SEM fitted with an Oxford INCA Energy 250 energy dispersive spectrometry (EDS) system. Imaging with backscattered electrons provided contrast as a SiC b damage function of atomic number (Z); metallic fission products clustered within the lower-Z TRISO layers appeared bright and could be targeted for elemental analysis with EDS. Fig. 18 shows two parallel planar sections through one of the two particles recovered from Compact 3-3-1 that released cesium during 1700°C safety testing; these two planes are slightly offset in the grinding direction and reveal the same degraded Foreign matter pathway through the SiC at two different positions, at SiC interface similar to what was observed in the particle from Compact 5-2-1 (Fig. 16) as it was ground down due Fig. 16. Particle from Compact 5-2-1 that failed to the corrosion pathway being oriented at an angle during irradiation testing; (a) x-ray tomograph to the grinding plane. The IPyC layer was decorated oriented to show low-density pathway through SiC with high-Z clusters that EDS identified as mostly and (b) optical micrograph showing degraded area. uranium, except in the region surrounding the IPyC crack, where there appeared to be a depletion of Fig. 17 shows a location on the outer edge of the these clusters. The corrosion pathway through the soot inclusion in the cesium-releasing particle from SiC was surrounded by numerous clusters of high-Z Compact 3-3-2 where a crack extends from the elements ranging in diameter from <1 to 5 µm and defective SiC region through an intact overlayer. identified by EDS to be predominantly palladium This particle was an interesting anomaly because the and uranium (these clusters may also contain silicon gross soot inclusion shown in Fig. 10 was present and carbon, which could not be resolved due to Proceedings of the HTR 2014 Weihai, China, October 27-31, 2014 Paper HTR2014-31254 signal from surrounding material included in the densely filling their resident volumes (except for the electron excitation volume). Palladium and uranium largest carbon-rich feature, which may have were also observed at the IPyC/SiC interface around experienced some pullout during grinding and the entire circumference of the particle, but away polishing). from the degraded area, the IPyC/SiC interface remained intact and no significant SiC corrosion was a Carbon-rich features observed. The predominance of palladium and uranium surrounding the degraded area, in conjunction with the free carbon that was observed within the area, suggests that these metals may have IPyC reacted with the SiC to form silicides that migrated away leaving a carbon filled pathway through the SiC that would not retain cesium. Predominantly Pd and U a 3 kV Predominantly U b Predominantly Pd and U 20 kV 20 µm 20 kV Fig. 19. Polished planar section through degraded b SiC area in Compact 3-3-1 Particle 1 (Fig. 7b) after 1700°C safety testing; (a) scanning electron (SE) and (b) backscattered electron (BE) images that Denuded zone show numerous clusters of palladium and uranium and other areas that were predominately carbon. The SE image shows bright spots where Pd and U are embedded close to the surface, while the BE image Carbon-rich zone shows additional Pd and U due to greater sensitivity and analysis depth. V. ENUMERATING PARTICLES IN EACH 20 kV 20 µm COMPACT WITH FAILED SIC Fig. 18. Successive polished planar sections showing corrosion progression through SiC of Compact 3-3-1 Determination of the number of particles in each Particle 2 after 1700°C safety testing. compact with failed SiC was based on a combination of data on the amount of cesium released from the Fig. 19 shows secondary and backscattered compact, the amount of exposed cesium still in the electron images of the degraded area in the SiC compact outside of intact SiC layers (measured by exposed by sectioning the other particle recovered DLBL), and the amount of cesium retained by the from Compact 3-3-1 that released cesium during particles with failed SiC that were separated out and 1700°C safety testing. The exposed plane in this measured by IMGA. Supporting data for counting figure was prepared by grinding down from the top particles with failed SiC involved enumeration of the of the particle as shown in Fig. 7b and stopping at low-cesium retaining particles found with IMGA the edge of the corrosion site before the center of the (with examination by x-ray to verify the presence of low-density feature was exposed. Similar to the left failed SiC) and additional DLBL data of uranium side of the degraded SiC area in Fig. 18a, a dense from exposed kernels indicative of the loss of population of high-Z clusters was observed in this particles with failed SiC during deconsolidation, area. The secondary electron image (Fig. 19a) shows leaching, and sieving prior to the IMGA survey. that the carbon-rich (gray) and metal-rich (white) Table 1 and Table 2 present results using these two features are flush with the polished surface and approaches. Proceedings of the HTR 2014 Weihai, China, October 27-31, 2014 Paper HTR2014-31254 Table 1. Estimation of the number of particles with failed SiC based on recovery of 134Cs As-irradiated Compacts 1600°C Safety Testing 1700°C 1800°C Safety Testing 5-2-1 5-2-3 6-3-2 3-3-2 4-1-2 6-4-1 3-3-1 3-2-3 4-4-1 5-1-3 Number of particles worth of 134Cs 0.28 0.27 0.96 0.89 0.69 0.49 1.49 6.62 1.49 4.87 detected outside compact Number of particles worth of 134Cs 0.20 0.20 0.30 0.01 0.02 NA 0.92 0.33 0.37 0.10 detected by DLBL Number of particles worth of 134Cs 0.35 1.37 NA 0.03 0.01 NA 1.20 3.93 0.11 2.18 retained in particles found with IMGA Total number of particles worth of 134Cs presumed to be associated with particles 0.83 1.84 1.27 0.93 0.71 0.49 3.61 10.89 1.97 7.15 that released 134Cs through failed SiC Estimated number of particles with failed SiC based on 134Cs accounting 1 2 1 1 1 1 4 11 2 7 NA denotes that data is not available because analysis was not performed. For compacts safety tested at 1800°C, 134Cs detected in post-burn leach of particles analyzed after IMGA survey was not included because a few particles with normal cesium retention were broken by the process and particles with failed SiC had already been removed from sample and accounted for with IMGA. No low-cesium particles from Compact 4-4-1 were found with IMGA, but fragments from at least two particles, including one buffer-coated kernel, were manually recovered after pre-burn leaching and gamma counted to measure 134Cs inventory (SiC in the recovered fragments showed signs of corrosion failure). Table 2. Estimation of the number of particles with failed SiC based on identification by the IMGA and DLBL As-irradiated Compacts 1600°C Safety Testing 1700°C 1800°C Safety Testing 5-2-1 5-2-3 6-3-2 3-3-2 4-1-2 6-4-1 3-3-1 3-2-3 4-4-1 5-1-3 Number of particles with failed SiC 1 2 NA 1 1 NA 2 9 0 5 found with IMGA Number of particles worth of 235U 0.27 0.24 1.13 0.82 1.13 NA 1.93 1.91 3.26 1.25 detected by DLBL Estimated number of particles with 0 0 1 1 1 NA 2 2 3 1 failed SiC detected by DLBL Estimated total number of particles 1 2 1 1 1 NA 4 11 3 6 detected by the IMGA and DLBL NA denotes that data is not available because analysis was not performed. Compacts 3-3-2 and 4-1-2 each had one particle with failed SiC; uranium was leached prior to IMGA, but particles remained in one piece (Fig. 10 and Fig. 11) and were found during IMGA survey. For compacts safety tested at 1800°C, 235U detected in post-burn leach of particles analyzed after IMGA survey was not included because a few particles with normal cesium retention were broken by the process and particles with failed SiC had already been removed from sample and accounted for with IMGA. While cesium and uranium measurements rarely located in the OPyC of the main particle sample, summed to an integer value due to analysis which didn't include the particles with failed SiC. uncertainty (typically ~10%), hot cell contamination, This uranium was presumably released during variation in actual isotopic content, and loss of irradiation and represents a compact fractional volatile cesium during analysis, enumeration using release of about 6.5×10-5. Compact 6-3-2 was the two methods in Table 1 and Table 2 agreed very examined at INL and DLBL was used, in lieu of well. Only two of the compacts safety tested at IMGA survey, to determine that only one particle 1800°C yielded different estimated totals. Possible with failed SiC was responsible for most of the explanation for the disagreement for Compact 4-4-1 cesium release from that compact during irradiation. was that one particle broke during the sieving Slightly more than one particle's average inventory operation (indicated by uranium detected in the of cesium and uranium were detected, but these burn-leach of the matrix debris remaining after values would also be impacted by analysis separating out the particles); some of the cesium uncertainties and low-level releases from particles from this particle would have been lost during the with intact SiC. 750°C burn and not included in Table 1, resulting in Only three particles with failed SiC were a lower total count compared to Table 2. In addition, detected in compacts safety tested at 1600°C, one the SiC in this particle may not have failed during from each of three compacts; a total of eight safety testing (it may have been broken by handling compacts (~33,100 particles) were tested at 1600°C, during the sieving process), resulting in an with no failed SiC in the other five compacts. As overestimation of the number of particles with failed shown in Fig. 10 and Fig. 11 and discussed in SiC in Table 2. It is unclear why the estimated totals Section IV.A, the particles with failed SiC from for Compact 5-1-3 disagree; uncertainty in the 134Cs Compacts 3-3-2 and 4-1-2 both had material missing analysis could be responsible. from the kernel. This is in agreement with the A total of four particles with failed SiC were uranium detected during DLBL of these compacts identified out of 298,000 particles included in the (Table 2). Compact 4-1-2 was missing its OPyC AGR-1 irradiation experiment. The small amount of layer when recovered for IMGA and this presumably uranium detected in Compacts 5-2-1 and 5-2-3 was was fractured during the third 24-hour nitric acid not from exposed kernels, and more than 75% was leach performed on the deconsolidated particles,