ORNL/SPR-2026/4746 Surrogate Process Development for Coated Particle Burnable Absorber Fuel Pellets August 2026 Alicia M. Raftery Henry Osborne Ryan Heldt Eddie Lopez Honorato Tyler Gerczak Surrogate Process Development for Coated Particle Burnable Absorber Fuel Pellets Alicia M. Raftery Henry Osborne Ryan Heldt Eddie Lopez Honorato Tyler Gerczak August 2026 Prepared for the U.S. Department of Energy Office of Nuclear Energy Page intentionally left blank SUMMARY A particle-dispersed burnable absorber fuel design could enable boron-based absorber materials to be distributed throughout an oxide fuel pellet while providing internal volume for helium accommodation and barriers between the absorber and the fuel matrix. Development of this concept requires successful coating of the absorber particles and consolidation of the coated particles within a dense oxide matrix. This work evaluated preliminary fabrication processes for both requirements using surrogate materials. First, cold-flow fluidization tests were performed using irregular zirconium diboride particles to identify a nozzle geometry suitable for future fluidized-bed chemical vapor deposition coating development. Five injector geometries were evaluated based on central spout stability, particle circulation along the chamber walls, channeling, and particle loss. A single-inlet, five-outlet nozzle produced the most favorable fluidization behavior while operating at approximately 6.5 standard liters per minute and was selected as the starting configuration for future coating studies. Oxide overcoating was then demonstrated using cerium dioxide powder as a surrogate fuel matrix and yttria-stabilized zirconia (YSZ) microspheres as surrogate coated particles. A normalized powder-to- binder-solution addition parameter was used to identify a viable overcoating process window. Stable overcoating behavior was obtained at a normalized ratio of 1.5. However, the resulting feedstock exhibited a bimodal particle size distribution and limited mechanical integrity after drying, indicating that further powder conditioning and binder development are required. Overcoated particles were separated into four size fractions and consolidated into pellets with different estimated YSZ packing fractions. Optical dilatometry showed that pellet shrinkage decreased as YSZ packing fraction increased. Optical microscopy showed progressively larger and more extensive matrix cracking at higher packing fractions. Scanning electron microscopy energy-dispersive x-ray spectroscopy analysis of the pellet with the lowest packing fraction showed a gradual decrease in zirconium signal across the YSZ–ceria interface, consistent with limited interfacial elemental redistribution during sintering. Overall, this work established preliminary fabrication conditions for the coated particle concept and identified feedstock integrity, matrix connectivity, high packing fraction cracking, and material compatibility as key areas for continued development. iii Page intentionally left blank iv CONTENTS SUMMARY..................................................................................................................................................iii ACRONYMS................................................................................................................................................ix 1. INTRODUCTION................................................................................................................................1 2. COATING PROCESS DEVELOPMENT ...........................................................................................1 2.1 Experimental Methods ..............................................................................................................1 2.2 Materials and Experimental Setup ............................................................................................2 2.3 Results and Discussion..............................................................................................................4 3. OVERCOATING .................................................................................................................................6 3.1 Experimental Methods ..............................................................................................................6 3.1.1 Process Window Development ....................................................................................6 3.1.2 Packing Fraction Sintering Tests .................................................................................7 3.2 Materials and Experimental Setup ............................................................................................7 3.3 Results .......................................................................................................................................9 3.3.1 Process Window Development ....................................................................................9 3.3.2 Packing Fraction Sintering Tests ...............................................................................11 4. CONCLUSION AND FUTURE WORK...........................................................................................15 5. REFERENCES...................................................................................................................................17 v FIGURES Figure 1. Injector nozzle geometries evaluated during cold-flow testing: (a) nine-inlet, nine- outlet; (b) single-inlet, eight-outlet; (c) single-inlet, five-outlet; (d) enlarged single- inlet, five-outlet; and (e) single-inlet, four-outlet. .........................................................................2 Figure 2. ZrB₂ feedstock characterization: (a) as-received material divided into trays; (b) irregular particle morphology; and (c) particle breakup during fluidization................................................3 Figure 3. Measured particle size distribution of the ZrB₂ feedstock. ............................................................3 Figure 4. Laboratory-scale fluidization rig used for cold-flow testing..........................................................4 Figure 5. Fluidization behavior obtained using the single-inlet, five-outlet nozzle: (a) fluidized particle bed; (b) central particle spout; and (c) top view of the bed..............................................5 Figure 6. Particle diameter distribution of the YSZ feedstock measured using shadow imaging. ................8 Figure 7. (a) Overcoating setup with main components labeled; and (b) the YSZ particle bed during rotation before powder and binder addition. ......................................................................9 Figure 8. Overcoating results for various PBAR* values..............................................................................9 Figure 9. YSZ particles overcoated with ceria using PBAR* = 1.5. ...........................................................10 Figure 10. Mass yields of the sieve fractions from the selected overcoating batch indicating a bimodal size distribution for the overcoated particles.................................................................10 Figure 11. Overcoated particles after vigorous shaking with evident separation of the ceria overcoat from the surface of the YSZ particles...........................................................................11 Figure 12. Diameter change measured during optical dilatometer sintering of pellets with different estimated particle packing fractions. ............................................................................12 Figure 13. Pellets after sintering in the optical dilatometer: (a) BA01; (b) BA02; (c) BA03; and (d) BA04. .....................................................................................................................................13 Figure 14. Optical images of polished cross sections from pellets (a) BA01; (b) BA02; (c) BA03; and (d) BA04. ..............................................................................................................................14 Figure 15. Representative cracking in the sintered pellets: (a) smaller cracks in BA01; and (b) larger interparticle cracks in BA04..............................................................................................14 Figure 16. SEM image and corresponding EDS line scan across the YSZ–ceria interface in pellet BA01 showing a gradual decrease in zirconium signal from the YSZ particle into the ceria matrix. .................................................................................................................................15 vi TABLES Table 1. Range of PBAR* values tested to identify stable oxide overcoating behavior. ..............................7 Table 2. Pressed pellets IDs and estimated particle packing fraction ranges based on the sieve fractions of the overcoated particles............................................................................................12 vii Page intentionally left blank viii ACRONYMS BA burnable absorber EDS energy-dispersive x-ray spectroscopy FBCVD fluidized-bed chemical vapor deposition HTGR high-temperature gas-cooled reactor PBAR powder-to-binder-solution ratio PBAR* normalized powder-to-binder-solution ratio SEM scanning electron microscopy slpm standard liters per minute TRISO tristructural-isotropic YSZ yttria-stabilized zirconia ix Page intentionally left blank x 1. INTRODUCTION Light-water reactors require the incorporation of burnable absorbers (BAs) to manage excess beginning-of-life reactivity. Various methods are currently used to incorporate BA materials into commercial reactor fuel systems, which can generally be grouped into non-fueled and fueled designs [1]. A non-fueled design is typically a BA rod consisting of a monolithic BA material fabricated into rod form and placed in specific locations throughout the reactor core. Fueled designs incorporate BA material directly into fuel rods. Common approaches include homogeneous mixing of BA material with uranium dioxide (UO2) fuel or applying a BA coating to the outer surface of fuel pellets. The fabrication route is highly dependent on the absorber material because each BA exhibits unique neutronic, chemical, and mechanical behavior during irradiation. Boron-based absorbers, for example, must account for the 10B(n,alpha)7Li reaction, which generates helium during reactor operation. The accumulated helium can result in swelling and degradation of the absorber material if it is not properly accommodated. Consequently, fueled boron-based absorber concepts have been implemented primarily as coatings rather than as homogeneously distributed absorbers. One commercial example is Westinghouse’s Integral Fuel Burnable Absorber design, which consists of a thin zirconium diboride (ZrB2) coating applied to the surface of a UO2 fuel pellet [2]. Despite the success of coating-based approaches, fueled designs that distribute the absorber material throughout the fuel volume are generally attractive because they provide a more uniform macroscopic absorber distribution and additional flexibility in fuel assembly design. However, incorporating boron- based absorber particles directly into the fuel matrix requires addressing both helium generation and chemical compatibility with the surrounding fuel. One potential solution is to use a dispersed coated particle architecture derived from tristructural-isotropic (TRISO) fuel technology. In this concept, boron carbide particles are encapsulated within a multilayer coating system consisting of a porous buffer layer and subsequent TRISO coating layers. The buffer layer could provide volume to accommodate helium generated during irradiation, and the surrounding coating layers could isolate the absorber from the UO2 matrix and retain gaseous reaction products. Such an approach could enable the incorporation of boron- based absorber particles directly within an oxide fuel matrix and also mitigate some of the challenges associated with conventional homogeneous designs. Two major fabrication advances are required to realize this concept. First, boron carbide particles must be successfully coated with representative TRISO layers. Second, the coated particles must be overcoated with oxide fuel powder and consolidated into a fuel form that has a density and microstructure suitable for reactor applications. The objective of this work was to demonstrate process development activities supporting both of these fabrication challenges using surrogate materials. Preliminary coating process development was performed using ZrB2 surrogate particles supplied by Bangor University and focused on cold-flow fluidization behavior. Oxide overcoating was evaluated using yttria-stabilized zirconia (YSZ) microspheres to represent coated absorber particles and ceria powder to represent the surrounding UO2 matrix phase. The overcoated particles were then sieved, pressed, and sintered to assess pellet fabrication and the effects of particle packing fraction on sintering shrinkage and microstructure. Process limitations were identified in both the overcoating and pellet fabrication stages, providing direction for future development. 2. COATING PROCESS DEVELOPMENT 2.1 Experimental Methods Before chemical vapor deposition coating experiments were attempted, cold-flow testing was performed to evaluate whether ZrB2 particles could be adequately fluidized. Stable particle circulation is required during fluidized-bed chemical vapor deposition (FBCVD) to promote uniform exposure of 1 particles to coating gases and reduce the likelihood of nonuniform coating deposition. The irregular morphology of the ZrB2 particles was expected to make fluidization more difficult than it is for the spherical particles conventionally used in TRISO fuel fabrication. Therefore, a series of injector nozzle geometries was evaluated using a laboratory-scale cold- fluidization rig. Five nozzle designs (Figure 1) printed from different materials were tested: 1. A nine-inlet, nine-outlet nozzle (PLA plastic) 2. A single-inlet, eight-outlet nozzle (nylon) 3. A single-inlet, five-outlet nozzle (resin) 4. An enlarged single-inlet, five-outlet nozzle (resin) 5. A single-inlet, four-outlet nozzle (nylon) Figure 1. Injector nozzle geometries evaluated during cold-flow testing: (a) nine-inlet, nine-outlet (PLA plastic); (b) single-inlet, eight-outlet (nylon); (c) single-inlet, five-outlet (resin); (d) enlarged single-inlet, five-outlet (resin); and (e) single-inlet, four-outlet (nylon). Particle motion was visually observed during each experiment. Nozzle performance was evaluated qualitatively based on three primary behaviors: formation and stability of the central particle spout, movement of particles along the conical chamber walls, and evidence of gas channeling or particle loss. A suitable fluidization condition was defined as one that produced a stable central spout while maintaining continuous particle circulation along the chamber walls. Channeling, stagnant regions, excessive particle entrainment, and loss of particles through the injector were considered undesirable because these behaviors could lead to nonuniform coating during subsequent FBCVD processing. The nozzle configuration that produced the most stable overall fluidization behavior was selected as the preferred design for future coating development. 2.2 Materials and Experimental Setup The coating development feedstock consisted of ZrB2 particles supplied by Bangor University. Approximately 520 g of material was received. Visual inspection showed that the particles were irregular and nonspherical, with angular surfaces and a tendency for some particles or agglomerates to break apart during handling and fluidization. The target range of particle sizes for coating was 400–500 µm. The received material was sieved to isolate this fraction, and approximately 14% of the original feedstock mass was recovered within the target range. Particle size distribution and morphology were further evaluated using a Malvern particle analysis system. The analysis confirmed a broad size distribution, with a substantial fraction of particles larger than 500 µm. Images of the as-received particles and the measured particle size distribution are shown in Figure 2 and Figure 3, respectively. 2 Figure 2. ZrB2 feedstock characterization: (a) as-received material divided into trays; (b) irregular particle morphology; and (c) particle breakup during fluidization. Figure 3. Measured particle size distribution of the ZrB2 feedstock. Cold-flow testing was performed using a laboratory-scale fluidization rig designed to simulate particle motion within the conical section of the FBCVD coating furnace. The system consisted of a laboratory argon supply, an injector nozzle, a transparent conical fluidization chamber, a lower particle collection chamber, and supporting hardware. The transparent cone reproduced the general geometry of the graphite cone used during high-temperature coating while allowing particle motion to be directly observed. Several modifications were made to improve the stability, sealing, and interchangeability of the cold- flow system. The fluidization assembly was secured in a fixed vertical position, and the nozzle and collection chamber connections were redesigned to simplify component replacement and reduce leakage. The modified system is shown in Figure 4. 3 Figure 4. Laboratory-scale fluidization rig used for cold-flow testing. Approximately 18.33 g of ZrB2 particles were used for each fluidization test. This charge was selected to provide a particle bed volume equivalent to approximately 60 g of a standard reference material, a tungsten carbide–YSZ composite. Laboratory argon was supplied through each nozzle, and the flow rate was adjusted using mass flow controllers, until measurable particle circulation was established. The injector nozzles were manufactured from nylon, resin, or plastic using additive manufacturing methods (Figure 1). Differences in material and surface finish between nozzle designs may have affected leakage and particle retention and are therefore noted in the discussion of the fluidization results. 2.3 Results and Discussion The irregular shape and broad particle size distribution of the ZrB2 feedstock presented significant challenges during fluidization. The angular particles exhibited a greater tendency to interlock, form stagnant regions, and redirect gas flow than would be expected for spherical particles. Considerable fragmentation or breakup of loosely bound material was also evident during testing, affecting the ability to properly fluidize this powder feedstock. These characteristics increased the likelihood of channeling and limited the range of flow conditions over which stable circulation could be achieved. Fragmentation of material resulted in constant changes in particle size, especially the formation of smaller particles inherently prone to form agglomerates. Although not tested, these constant changes in particle size would have resulted in the formation of irregular and defective coatings as previously observed in TRISO fuel systems [3]. There were minor variations in particle bed mass between tests because excessive gas flow occasionally caused particles to be expelled from the top of the cone. Some particles were also lost through the nozzles into the lower collection chamber, particularly when designs with larger outlet openings were tested. The nine-inlet, nine-outlet nozzle required an argon gas flow rate of approximately 10 standard liters per minute (slpm) to initiate particle movement along the chamber walls. Although particle motion was achieved, the bed exhibited substantial channeling due to particle shape and the constant fragmentation of the powder feedstock. The central spout shifted repeatedly as the gas was redistributed through the different outlet holes. This redistribution produced nonuniform circulation and intermittent stagnant regions. The relatively large outer outlet holes also allowed particles to pass through the nozzle when the gas velocity was insufficient to support the bed. Based on these observations, the nine-inlet, nine-outlet configuration was not considered suitable for subsequent coating experiments. 4 The single-inlet, eight-outlet nylon nozzle improved particle retention because of its smaller outlet holes and common inlet. However, significant channeling was still evident at approximately 11 slpm. As individual outlet passages became partially obstructed, gas was redirected through the remaining openings. This caused the spout position to shift and eventually eliminated particle movement along portions of the cone wall. The resulting circulation was not considered sufficiently uniform for FBCVD coating. The single-inlet, five-outlet nozzle produced the most controlled fluidization behavior. At approximately 6.5 slpm, the particles formed a central spout that extended approximately halfway up the cone. Despite the overfluidization, slow but continuous particle movement was also evident along the chamber walls. This combination of central spouting and wall circulation provided the most uniform bed motion among the nozzle geometries evaluated. It is expected that powders with a better structural integrity will be able to be fluidized more adequately using this nozzle configuration. The resin- manufactured nozzle also exhibited less gas leakage than the earlier plastic nozzle design. Representative images of the fluidized bed, central spout, and particle distribution are shown in Figure 5. Figure 5. Fluidization behavior obtained using the single-inlet, five-outlet nozzle: (a) fluidized particle bed; (b) central particle spout; and (c) top view of the bed. An enlarged version of the five-outlet nozzle was also evaluated. The outlet hole diameters were increased slightly relative to the original design. The enlarged openings produced an excessively large central spout but did not provide sufficient particle circulation along the cone walls. The resulting bed motion was therefore less uniform than that produced by the five-outlet nozzle. The final configuration evaluated was a single-inlet, four-outlet nylon nozzle. This design also exhibited significant gas channeling. The particles did not form a stable, uniform spout, and limited movement was observed along the cone walls. The four-outlet nozzle was therefore not considered suitable for coating trials. Overall, the single-inlet, five-outlet nozzle provided the best balance between spout formation, wall circulation, particle retention, and operating flow rate. Although fully uniform fluidization was not achieved, the results established a viable starting configuration for future FBCVD coating development. Additional optimization is required to account for the irregular morphology and broad size distribution of the ZrB2 feedstock. 5 3. OVERCOATING 3.1 Experimental Methods The experimental campaign was structured to first establish a viable process window for overcoating surrogate ceramic particles with an oxide powder. The resulting overcoated particles were then used to examine the effect of particle packing fraction on the final composite pellet microstructure. These manufacturing variables were selected because both were expected to strongly influence the success of the fabrication process. A new overcoating process window was required because the binder chemistry and material combination had not previously been evaluated for this application. In addition, particle packing fraction was expected to influence powder consolidation, sintering behavior, and the resulting composite microstructure. The experimental workflow therefore consisted of two stages: (1) overcoating process window development and (2) packing fraction sintering tests. The methods and motivation for each stage are described in the following sections. 3.1.1 Process Window Development Several processing variables influence the overcoating behavior of particles in a rotating drum system, as demonstrated in previous studies on graphite matrix overcoating for high-temperature gas- cooled reactor (HTGR) fuel fabrication [4]. In traditional overcoating, TRISO particles are rotated within a drum while graphite matrix and methanol are added, causing powder to build up on the perimeter of coated particles continuously. Variables that influence overcoating behavior include drum rotation speed, powder feed rate, binder spray rate, and particle/powder geometries. This study focused primarily on the effects of powder feed rate and binder spray rate because the powder and binder both differed from those used in traditional HTGR overcoating. The relative rates of powder and binder addition strongly influence overcoat formation and particle growth. Two primary limiting behaviors occur: excessive binder addition and excessive powder addition. Excessive binder addition causes the particle bed to become overly wet, resulting in particle agglomeration and material adhesion to the drum wall. Once this occurs, continued overcoating is not possible. In contrast, excessive powder addition results in poor adhesion of powder to the particle surface, leaving residual powder in the drum and producing mechanically weak overcoated particles. To describe the balance between powder and binder addition, a powder-to-binder-solution ratio (PBAR) was defined as follows: g powder addition rate [ min ] PBAR = (1) mL binder solution flow rate [ min ] PBAR is an empirical processing parameter with units of grams per milliliter. Because binder concentration was held constant throughout the experiments, PBAR provided a consistent measure of the relative powder and binder solution addition rates. Using this definition, low PBAR values correspond to binder-rich conditions that promote over-wetting, whereas high PBAR values correspond to powder-rich conditions that cause adhesion failure. Successful overcoating occurs within an intermediate region in which sufficient binder is available to support powder adhesion while free-flowing particle motion is still maintained within the rotating bed. A series of overcoating experiments was performed across the normalized PBAR (PBAR*) range shown in Table 1 to identify the region associated with stable overcoat growth for this material system. PBAR was normalized using an arbitrary reference value, PBARref, as follows: PBAR PBAR∗ = (2) PBARref 6 Only the normalized value, PBAR*, is reported in this study. The overcoating condition that produced the most stable particle motion and consistent overcoat growth was selected for subsequent pellet fabrication experiments. Table 1. Range of PBAR* values tested to identify stable oxide overcoating behavior. Overcoating batch PBAR* 1 0.38 2 1.00 3 1.13 4 1.25 5 1.50 6 1.75 7 2.25 3.1.2 Packing Fraction Sintering Tests One benefit of using the overcoating process to fabricate composite fuel forms is that the packing fraction of the composite pellet can be controlled by the overcoat thickness. Decreasing the overcoat thickness decreases the matrix ligament thickness between particles and increases the particle packing fraction within the composite. Because the particles have a high density and the overcoat is deposited onto the particle surface at a relatively low green density, differential shrinkage is expected during sintering after pellet pressing. During sintering, the ceria matrix must volumetrically shrink to densify, whereas the dense particles are expected to remain approximately the same size. It was hypothesized that this mismatch in shrinkage would generate internal stresses and cause cracking in the composite pellets. Cracking was expected to be more severe at higher particle packing fractions because less matrix material would be available to accommodate the shrinkage mismatch and maintain connectivity between particles. To evaluate this effect, four particle packing fractions were produced by sieving the overcoated particles into different particle size ranges. Because the initial surrogate particle size was approximately constant, final overcoated particle size was used as an indirect measure of overcoat thickness. Smaller overcoated particles had thinner overcoats and therefore produced higher particle packing fractions. The sieved overcoated particles were pressed into pellets and sintered in an optical dilatometer. Dimensional changes were recorded during heating to evaluate pellet shrinkage behavior. After sintering, the pellets were mounted, polished, and examined using optical microscopy to evaluate cracking and the resulting composite microstructure. The pellet with the lowest particle packing fraction was further characterized using scanning electron microscopy (SEM) with energy-dispersive x-ray spectroscopy (EDS) to evaluate interaction or diffusion at the particle-matrix interface. 3.2 Materials and Experimental Setup Cerium dioxide (CeO2) was used as a surrogate for UO2. The powder was supplied by MSE Supplies and had a reported median particle size (D50) of 5 µm. The surrogate particles were 5 wt% YSZ microspheres with a reported average diameter of approximately 800 µm. The YSZ feedstock was characterized for particle diameter and average particle weight using methods established under the US Department of Energy Office of Nuclear Energy’s Advanced Gas Reactor Fuel Development and Qualification Program [5], [6]. The mean of the average diameter of the YSZ feedstock was 804.5 ± 23.7 7 µm, where the uncertainty represents one standard deviation. The YSZ particle distribution histogram is shown in Figure 6. The mean of the average weight per YSZ particle was 0.00168 g. Figure 6. Particle diameter distribution of the YSZ feedstock measured using shadow imaging. A water-soluble polymer binder was used throughout this study; binder solutions were prepared by mixing the binder with deionized water for a minimum of 1 h using a magnetic stirrer to promote complete dissolution. The binder concentration was fixed for all experiments. Each overcoating batch was initiated with 50 g of YSZ microspheres. The overcoating experiments were conducted using a laboratory-scale system previously developed for graphite matrix overcoating of TRISO particles. The system consists of a syringe pump, ultrasonic spray nozzle, and rotating drum, as shown in Figure 7. During each experiment, the surrogate particles were loaded into the drum and rotated at a fixed speed. The binder solution was delivered to the ultrasonic nozzle using the syringe pump at a controlled flow rate. The atomized binder solution was sprayed directly onto the rotating particle bed. The CeO2 powder was manually added to the particle bed as discrete, preweighed charges using a spoon. A timer was used to control the timing of each addition to maintain a consistent average powder addition rate throughout each overcoating run. 8 Figure 7. (a) Overcoating setup with main components labeled; and (b) the YSZ particle bed during rotation before powder and binder addition. 3.3 Results 3.3.1 Process Window Development The PBAR* values defined in Table 1 were tested, and the results are shown in Figure 8; images are included to demonstrate the particle bed overcoating behavior at various conditions. As expected, when the PBAR* value was too high, a lack of powder adhesion resulted. When the PBAR* value was too low, the result was the material becoming overly wet and sticking to the sides of the drum, ultimately ending the overcoating run. Satisfactory overcoating behavior was evident at a PBAR* value of 1.5. Figure 9 shows a batch of particles overcoated using this parameter. Figure 8. Overcoating results for various PBAR* values. 9 Figure 9. YSZ particles overcoated with ceria using PBAR* = 1.5. The resulting overcoated material was separated into four size fractions using standard sieve sizes from No. 12 to No. 18. The selected batch exhibited a bimodal size distribution for the overcoated particles, as indicated by the mass yield within each sieve fraction shown in Figure 10. The bimodal distribution may have resulted from the tendency of ceria powder to agglomerate under ambient conditions due to moisture adsorption. This behavior directly reduced the yield of particles within the desired size range. Drying the ceria powder immediately before overcoating may reduce agglomeration and improve particle size distribution; however, this approach was not evaluated in this study. Figure 10. Mass yields of the sieve fractions from the selected overcoating batch indicating a bimodal size distribution for the overcoated particles. The overcoated material was dried for 24 h. After being dried, the particles exhibited poor mechanical integrity and required careful handling during subsequent processing. The feedstock could be gently poured without visibly disturbing the overcoat. However, more aggressive handling caused the ceria powder to separate from the surface of the YSZ particles. For example, vigorous shaking of the material in a jar caused visible delamination of the powder overcoat, as shown in Figure 11. Despite this 10 limitation, the feedstock retained sufficient integrity for the subsequent packing fraction sintering tests when handled carefully. These tests are described in the following section. Figure 11. Overcoated particles after vigorous shaking with evident separation of the ceria overcoat from the surface of the YSZ particles. 3.3.2 Packing Fraction Sintering Tests The material collected from each sieve fraction was pressed into small pellets and sintered in an optical dilatometer to evaluate the effect of particle packing fraction on densification and cracking behavior. It was hypothesized that cracking within the ceria matrix would increase with increasing particle packing fraction because of differential shrinkage between the dense YSZ microspheres and the initially porous ceria matrix. During sintering, the ceria matrix was expected to undergo substantial densification, whereas the YSZ microspheres were expected to experience comparatively little dimensional change. This mismatch could generate internal stresses within the matrix and at YSZ–ceria interfaces. Four pellets were fabricated using overcoated feedstock from the four different sieve fractions. Because YSZ particle diameter was approximately constant, smaller overcoated particle diameters corresponded to thinner ceria overcoats and greater YSZ particle packing fractions. The four pellets therefore provided a range of particle packing fractions for evaluating the influence of particle volume fraction on pellet shrinkage and matrix cracking. The YSZ particle volume fraction within each green pellet was estimated using the overcoated particle size range, total feedstock mass loaded, and measured green pellet dimensions. The YSZ particles and overcoated particles were assumed to be spherical, with a mean YSZ core diameter of 804.5 µm. The mass of an individual overcoated particle was estimated from the theoretical density of YSZ and an assumed ceria overcoat density of 50% of the theoretical density of ceria. The estimated mass per overcoated particle was then used to calculate the number of particles loaded into each pellet. Total YSZ core volume was calculated from the estimated particle count and divided by the green pellet volume to determine the YSZ particle volume fraction. The packing fraction ranges calculated based on this methodology and the sizes of the sieve openings are listed in Table 2. Approximately 0.3–0.5 g of overcoated feedstock from each sieve fraction was loaded into a 5.95 mm diameter die. The material was uniaxially pressed at 150 MPa and held at pressure for 2 min. The resulting green pellets were heated in an optical dilatometer at 3°C/min to 1,500°C, held at 1,500°C for 4 h, and then cooled at 3°C/min to room temperature. All sintering experiments were performed in stagnant air. 11 Table 2. Pressed pellets IDs and estimated particle packing fraction ranges based on the sieve fractions of the overcoated particles. Pellet ID Sieved diameter range [µm] Packing fraction range [%] BA01 >1,700 <10.9 BA02 1,400–1,700 10.9–18.6 BA03 1,180–1,400 18.6–29.0 BA04 <1,180 >29.0 Figure 12 compares the pellet diameter change measured during sintering for the four estimated packing fraction ranges. The magnitude of pellet shrinkage decreased as YSZ particle packing fraction increased. This trend is consistent with the increasing volume of dimensionally stable YSZ particles and the decreasing volume of shrinkable ceria matrix within the pellets. Decreased shrinkage also suggests that final pellet density decreased as YSZ particle packing fraction increased. This behavior may have resulted from reduced connectivity of the ceria matrix at higher particle loadings, which limited matrix densification and increased the likelihood of residual porosity and cracking. Figure 13 shows the pellets after sintering. Macroscopic cracking became more pronounced as particle packing fraction increased, supporting the hypothesis that thinner ceria ligaments were less able to accommodate the shrinkage mismatch between the matrix and the YSZ particles. The pellets with lower packing fractions retained greater structural continuity, whereas the pellet with the highest packing fraction exhibited the most extensive visible cracking. Figure 12. Diameter change measured during optical dilatometer sintering of pellets with different estimated particle packing fractions. 12 Figure 13. Pellets after sintering in the optical dilatometer: (a) BA01; (b) BA02; (c) BA03; and (d) BA04. After sintering, the pellets were mounted and polished using a 9, 6, 3, and 1 µm diamond polishing sequence. Polishing was continued until a surface finish suitable for optical microscopy was obtained. The polished cross sections were then examined to evaluate crack morphology, matrix continuity, and the condition of the YSZ–ceria interfaces. Figure 14 shows the polished cross section of each pellet. Several trends were evident as YSZ particle packing fraction increased from BA01 to BA04. The ceria matrix appeared progressively less consolidated at higher packing fractions. This difference was also evident during sample preparation, because additional back-potting with epoxy was required to maintain the integrity of BA04 during polishing. Delamination between the ceria matrix and YSZ particle surfaces was evident in BA03 and BA04. Cracking was present in all four pellets, but apparent crack density and crack size increased with increasing particle packing fraction, as illustrated in Figure 15. The cracks occurred mainly within matrix ligaments between adjacent particles. The sample with the lowest packing fraction, BA01, was examined using a Hitachi SEM instrument. The primary objective of the SEM analysis was to evaluate whether interfacial interaction or elemental diffusion occurred between the YSZ particles and the ceria matrix during high-temperature sintering. Figure 16 shows an EDS line scan collected across the YSZ–ceria interface. The zirconium signal decreases gradually, and the cerium signal increases as the scan moves from the YSZ particle into the ceria matrix, suggesting that limited zirconium and cerium diffusion may have occurred during sintering. 13 Figure 14. Optical images of polished cross sections from pellets (a) BA01; (b) BA02; (c) BA03; and (d) BA04. Figure 15. Representative cracking in the sintered pellets: (a) smaller cracks in BA01; and (b) larger interparticle cracks in BA04. 14 Figure 16. SEM image and corresponding EDS line scan across the YSZ–ceria interface in pellet BA01 showing a gradual decrease in zirconium signal from the YSZ particle into the ceria matrix. 4. CONCLUSION AND FUTURE WORK This study demonstrated initial progress toward developing a particle-based BA fuel architecture using surrogate materials. Cold-flow testing showed that a single-inlet, five-outlet nozzle provided the most favorable fluidization behavior among the nozzle geometries evaluated. This design produced the best combination of spout formation and particle circulation despite the fragmentation of the powder used and will serve as the starting configuration for future coating studies using boron carbide feedstock once it becomes available. Oxide overcoating of YSZ microspheres with ceria was successfully demonstrated using the polymer binder solution. However, the resulting feedstock exhibited a bimodal particle size distribution and limited mechanical integrity after drying. Although it was possible to process the particles with careful handling, the overcoat was susceptible to separation from the particle surface during more aggressive handling. These results indicate that additional development is needed for the binder chemistry and powder conditioning procedure to improve coating uniformity, particle size distribution, and green strength. The packing fraction sintering tests showed that pellets with lower YSZ particle packing fractions underwent greater dimensional shrinkage during sintering. This behavior suggests that the ceria matrix 15 retained greater connectivity and densification potential at lower particle loadings. These results are encouraging for applications that require relatively low particle packing fractions, including the BA- containing pellet designs considered in this work. In contrast, cracking became more pronounced as YSZ particle packing fraction increased. Pellets with higher packing fractions exhibited larger cracks, decreased matrix continuity, and increased interfacial separation between the ceria matrix and YSZ particles. These results indicate that fabrication of ceramic matrix composites using conventional pressing and sintering may become increasingly difficult when high particle packing fractions are required. Alternative matrix compositions, consolidation methods, or sintering approaches may therefore be necessary for applications that require higher particle loadings. SEM-EDS characterization provided evidence of zirconium diffusion from the YSZ particles into the ceria matrix. The zirconium signal decreased gradually across the particle–matrix interface rather than exhibiting a sharp transition, suggesting the formation of a compositionally graded interfacial region during sintering. However, a distinct reaction layer was not clearly resolved under the characterization conditions used in this study. These results demonstrate that interfacial interactions can occur in the surrogate system and reinforce the need to evaluate interactions between a UO2 matrix and SiC- terminated coated particles under representative sintering conditions. Future work should focus on three primary areas. First, coating development should continue using boron carbide particles once suitable feedstock becomes available. Second, the binder formulation and overcoating procedure should be optimized to improve the mechanical integrity and handling characteristics of the overcoated feedstock. Sufficient green strength will be necessary for realistic transfer, storage, and pellet fabrication in a manufacturing environment. Finally, dedicated interaction studies should be performed using a UO2 matrix and the material selected for the outermost layer of the coated particles at temperatures and hold times representative of the intended fuel fabrication process. 16 5. REFERENCES [1] J. A. Evans, M. D. DeHart, K. D. Weaver, and D. D. Keiser, “Burnable absorbers in nuclear reactors – A review,” Nucl. Eng. Des., vol. 391, p. 111726, May 2022, doi: 10.1016/j.nucengdes.2022.111726. [2] R. L. Simmons, N. D. Jones, F. D. Popa, D. E. Mueller, and J. E. Pritchett, “Integral Fuel Burnable Absorbers with ZrB2 in Pressurized Water Reactors,” Nucl. Technol., vol. 80, no. 3, pp. 343–348, Mar. 1988, doi: 10.13182/NT88-A34058. [3] D. W. Marshall, “AGR-5/6/7 Fuel Fabrication Report,” Idaho National Laboratory, INL/EXT-19- 53720, Jun. 2020. [4] X.-W. Zhou, Z.-M. Lu, J. Zhang, Y.-P. Tang, and Y.-W. Zou, “Research on the overcoating process in the manufacture of spherical fuel elements for HTGR,” Nucl. Eng. Des., vol. 271, pp. 154–157, May 2014, doi: 10.1016/j.nucengdes.2013.11.024. [5] J. D. Hunn, R. Morris, J. Miller, and R. Hunt, “Overview of Key Issues and Guidelines for Regulatory Oversight and Inspection of High Temperature Gas Reactor Fuel Fabrication and Quality Control Activities,” ORNL/TM-2009/041, May 2009. [6] A. K. Kercher, J. D. Hunn, J. R. Price, and P. Pappano, “Automated optical microscopy of coated particle fuel,” J. Nucl. Mater., vol. 380, no. 1–3, pp. 76–84, Oct. 2008, doi: 10.1016/j.jnucmat.2008.07.011. 17