Light Water Breeding with Nitride Fuel

Progress in Nuclear Energy 53 (2011) 862e866 Contents lists available at ScienceDirect Progress in Nuclear Energy journal homepage: www.elsevier.com/locate/pnucene Light Water Breeding with Nitride Fuel Bo Feng*, Eugene Shwageraus, Benoit Forget, Mujid S. Kazimi Center for Advanced Nuclear Energy Systems, Massachusetts Institute of Technology, 77 Massachusetts Avenue, 24-215 Cambridge, MA 02139, USA a r t i c l e i n f o a b s t r a c t Article history: Received 30 October 2010 Received in revised form 21 April 2011 Accepted 2 June 2011 This scoping study proposes using mixed nitride fuel in Pu-based high conversion LWR designs in order to increase the breeding ratio. The higher density fuel reduces the hydrogen-to-heavy metal ratio in the reactor which results in a harder spectrum in which breeding is more effective. A Resource-renewable Boiling Water Reactor (RBWR) assembly was modeled in MCNP to demonstrate this effect in a typical high conversion LWR design. It was determined that changing the fuel from (U,TRU)O2 to (U,TRU)N in the assembly can increase its fissile inventory ratio (fissile Pu mass divided by initial fissile Pu mass) from 1.04 to up to 1.17. Ó 2011 Elsevier Ltd. All rights reserved. Keywords: Nitride fuel Light water reactor Breeding RBWR 1. Introduction Breeder reactors have two potentially major advantages over existing light-water reactors in that they (1) produce more fuel than they consume by conversion of U238 or Th232 into Pu239 or U233, respectively, and (2) are more effective in fissioning certain actinides that pose waste management and/or proliferation challenges. The base-line breeder concept worldwide is the sodiumcooled fast breeder reactor (SFR). However, as presently considered, fast-spectrum breeder reactors have major disadvantages relative to existing LWRs operating on a once-through fuel cycle: (1) the SFR capital costs and cost uncertainty are substantially higher, (2) their reliability is substantially lowerdpartly due to limited operating experience, and (3) successful deployment requires a significant development effort of the reactor core, the reactor system, and the associated fuel cycle infrastructure. When the scarcity of uranium resources necessitates the transition toward fast reactors, it may be more economic to make such a transition with Pu-based breeders cooled by water. A watercooled breeder reactor can be more attractive to industry since it can use the (1) existing industrial technology, (2) reliable cost estimates, (3) and massive experience and operational base of LWRs. Commercial readiness would require development of the reactor core but not the entire nuclear reactor system. If properly Abbreviations: EFPD, effective full power days; MN, mixed nitride; MOX, mixed oxide; RBWR, resource-renewable boiling water reactor; SFR, sodium fast reactor; TD, theoretical density; TRU, transuranic elements. * Corresponding author. E-mail addresses: [email protected], [email protected] (B. Feng). 0149-1970/$ e see front matter Ó 2011 Elsevier Ltd. All rights reserved. doi:10.1016/j.pnucene.2011.06.001 designed, an existing LWR can be converted into a water-cooled breeder with minimal retrofitting. There have been multiple efforts over the years to increase the conversion ratio of LWRs. It was first suggested to reduce the moderator-to-fuel ratio by using a tight hexagonal pitch in a PWR (Edlund, 1976). Axially- and radially-heterogeneous zones alternating between fissile and fertile compositions (seed-blanket) were implemented in later conceptual high conversion PWR designs that allowed the conversion ratio to reach 0.96 as well as to maintain a negative moderator temperature coefficient by taking advantage of increased localized leakage (from fissile to fertile zones) upon spectrum hardening (Broeders, 1985; Ronen and Dali, 1998). The most recent conceptual designs, namely the Resource-renewable Boiling Water Reactor (RBWR) by Hitachi (Takeda et al., 2010) and the innovative water reactor for flexible fuel cycle (FLWR) by JAEA (Uchikawa et al., 2009) maintain the low moderator-to-fuel ratio via a tight lattice and increased core void fraction and are able to achieve a breeding ratio of 1.01. Although this is much lower than the breeding ratio possible in a SFR, it will enable making use of the total energy potential of uranium. Such water-based breeder designs show promise in addressing fuel cycle challenges and economics, but the feasibility of safe operation with tight lattices and high void fractions needs to be verified. In particular, there are concerns about the ability to maintain a negative void coefficient in a hard-spectrum reactor over its entire cycle length, the high void fractions may lead to increased risk of flow instabilities, and the reduced water content may introduce more uncertainty about sufficient cooling under some transient conditions. However, increasing the core water content softens the spectrum and reduces breeding. Therefore, this study proposes an alternative design strategy: increasing the heavy B. Feng et al. / Progress in Nuclear Energy 53 (2011) 862e866 863 metal-to-moderator ratio by using uranium nitride (UN) with plutonium nitride (PuN) as a fuel with isotopically-enriched N15. The theoretical density (TD) of uranium is 9.67 g/cm3 in UO2 and 13.52 g/cm3 in UN so simply switching the fuel from oxide to nitride in current designs would provide a 40% increase in heavy metal given the same as-fabricated porosity. Therefore, using nitride fuel in breeder reactors such as the RBWR has the following potential advantages: 1) The breeding ratio can be significantly increased beyond 1.01. 2) The amount of water in these cores can be increased to provide additional cooling and safety while maintaining the 1.01 breeding ratio. 3) Breeding may be achieved with pressurized single-phase water cooling. To examine the breeding advantage gained in typical high converting LWR designs via nitride fuel, assembly models of the RBWR-AC (Takeda et al., 2007) (the breeder core of the RBWR concept) with oxide (TRU-O2) and nitride (TRU-N) fuels were created and benchmarked in MCNP-5 (Los Alamos National Laboratory, 2005). MCODE 2.2 (Xu et al., 2006), which couples the MCNP model with ORIGEN2 (Croff, 1980), was used for burnup (depletion) calculations. 2. Nitride fuel Nitride fuel has been considered for use with sodium fast reactors and space power reactors because of its compatibility with liquid metal coolant, higher thermal conductivity, and higher fissile density compared to oxide fuels (see Table 1). Unlike UO2, the thermal conductivity of UN increases with temperature. However, the fuel performance and property database for nitride fuel is not as extensive as that of oxide fuels. Currently, there is ongoing materials research on nitride fuels at institutions such as JAEA (Arai et al., 2006) but mainly with liquid metal coolants. Unlike liquid metals, water is reactive with nitride fuel: 2UN þ 4H2 O/2UO2 þ 2NH3 þ H2 þ Energy [1] One potential solution may be to introduce additives (small amounts of metals such as yttrium and titanium) into the nitride fuel matrix to reduce dissociation of the heavy metals and nitrogen (Alexander, 1986). Further investigations (including experiments) of this approach are required to demonstrate its potential for enhancing the stability of nitride fuel in the event of cladding failure and contact with the water coolant. Enrichment of N15 to 99.5%e99.9% is required to not only minimize neutron losses via capture by N14 but also to minimize the amount of radioactive C14 produced as a byproduct of the reaction: neutron þ N14 /C14 þ proton Fig. 1. MCNP model depicting radial geometry of RBWR assembly with 5 different pin enrichments (before homogenization) (Takeda et al., 2007). In the past, nitrogen enrichment has been costly but recent research in displacement chromatography using cation-exchange resins (Ding et al., 2008) show promise that it can be done economically on a large scale. Although it is possible to fabricate nitride fuel at 95% TD to maximize the heavy metal density for breeding purposes, it may be necessary to operate nitride fuel with as-fabricated porosities as high as 15% to reduce its hardness and swelling. Nitride fuel exhibits higher swelling rates (Tanaka et al., 2004) and hardness (Hayes et al., 1990) than oxide fuel, so a lower density would allow greater control over the fuel’s fragmentation behavior and pellet clad mechanical interaction during power ramps. Nitride fuel behavior and limitations on its density are discussed in greater detail in (Feng et al., 2011). Therefore, the physics calculations in this study were performed with nitride fuel at both 85% and 95% TD to represent conservative and potential scenarios, respectively. Considering the aforementioned material behavior uncertainties, more nitride fuel materials research, including irradiation tests, is required and hopefully will be encouraged by the results of this study. 3. RBWR with nitride fuel A comparative study was done by modeling a RBWR assembly with mixed oxide (MOX) fuel (including transuranics) and then with mixed nitride fuel (MN) consisting of UN mixed with transuranic nitrides (TRU)N at 85% and 95% TD. The most recent RBWR assembly radial and axial geometries from open literature (Takeda et al., 2007, 2010) are shown in Figs. 1 and 2, respectively. The specific RBWR design modeled in this study is based on the “RBWRAC” core in (Takeda et al., 2007) whose core specifications are shown along with those of the ABWR in Table 2. The blanket zones were modeled as shown in Fig. 2 but the fissile zone lengths and enrichments were varied to produce the axial power shape shown [2] Table 1 Comparison of oxide and nitride fuel properties (Ma, 1983). Theoretical Density [g/cm3] HM Density [g/cm3] Specific Heat [J/kgK] Melting Point [ C] Thermal Conductivity [W/mK] Linear Thermal Exp. Coeff. [10 6 K 1 ] UO2 UN 10.96 9.67 330 (at 1000  C) w2800 7.19 (at 200  C) 3.35 (at 1000  C) 11.5 (at 1000  C) 14.32 13.52 230 (at 1000  C) w2700 12 (at 200  C) 20 (at 1000  C) 9.2 (at 1000  C) Fig. 2. Axial geometry of RBWR assembly as described in (Takeda et al., 2010). 864 B. Feng et al. / Progress in Nuclear Energy 53 (2011) 862e866 Table 2 RBWR-AC and ABWR core specifications (Takeda et al., 2007, 2010). Item RBWR-AC Thermal Power [MWt] Electrical Power [MWe] Number of Fuel Bundles Core Height [mm] Coolant Flow Rate [kt/h] Core Exit Quality [%] Average Void Fraction [%] Pressure Drop [MPa] Puf/HM in Fissile Zones [w/o] Burnup [GWd/t] Maximum q0 [kW/cm] MCPR Void Coeff. [Dk/k/%void] Breeding Ratio 3926 1356 720 1200e1400 22 41 56 0.12 17.1 45 47.2 1.3 1.4  10 1.01 Table 3 Isotopic composition of TRU vector in fissile zones. ABWR 4 3926 1356 872 3710 58 13 36 0.21 e 45 12 1.3 12  10 4 e in Fig. 3. The average discharge burnup of 45 MWd/kg corresponds to a cycle burnup of 9.75 MWd/kg or roughly 360 effective full power days (EFPDs). Reflective boundary conditions were imposed only radially; homogenized axial shields and reflectors were modeled above and below the active region, therefore axial leakage is accounted for in the reactivity calculations. There are 271 fuel pins per assembly each with a diameter of 10.1 mm and pin pitch of 11.4 mm. A Y-cruciform B4C control rod or graphite follower borders the 2 wide gaps of each assembly. For this study, the 5 different pin enrichment zones shown in Fig. 1 were averaged into a single average enrichment for each of the two fissile axial zones. The average fissile plutonim (Puf) enrichment for the fissile zones was assumed to be 17.1 wt%. The transuranic (TRU) composition, as shown in Table 3, was from cooled spent LWR fuel, with 49.15 wt% of the TRU as Puf (44 wt % Pu239 and 5.15 wt% Pu241). The TRU proportions should remain virtually constant in terms of isotopic composition during the equilibrium cycle of a reactor with a breeding ratio of 1.0. The RBWR assembly models were discretized into 15 axial zones to account for local buildup of Pu239 in the blanket regions adjacent to fissile zones. Modeling the assembly in MCODE with only 5 axial zones would reduce accuracy considerably since the calculated compositions at each burnup step are homogenized over each axial zone. The coolant surrounding the pins was also divided into 15 axial void fractions corresponding to the same 15 fuel zones to better represent the void distribution, as shown in Fig. 4. The average void fraction was 0.56 and exit void fraction was roughly 0.80 (Takeda et al., 2010). For the MCODE burnup calculations, the assembly power was fixed at 5.45 MW (3926 MWt core power divided by 720 Fig. 3. Relative axial power shape of RBWR in MCNP model. Isotope TRU wt% Isotope TRU wt% Np-237 Pu-238 Pu-239 Pu-240 Pu-241 Pu-242 Am-241 Am-242m 0.44 2.85 44 36 5.15 4.96 3.57 0.15 Am-243 Cm-243 Cm-244 Cm-245 Cm-246 Cm-247 Cm-248 Total 1.32 0.02 1.07 0.34 0.1 0.02 0.01 100 assemblies) and the control rods were assumed fully withdrawn (replaced by Y-cruciform graphite followers). The burnup time steps must be very small near the beginning of life (0.2, 1, 2, 4, 6 MWd/kg) in order to capture the initial reactivity increase due to the non-linear buildup of Puf in the blanket regions. The reaction rate of U238 capture is highest when the blankets have no plutonium and then levels off when sufficient Puf builds up and the competing effects of capture and fission in the blanket reach equilibrium. The small time steps also help to capture the equilibrium Xe worth although this is relatively small due to the RBWR’s hard spectrum as shown in Fig. 5. The composition of the fuel and blankets of the reference RBWR assembly were then changed from oxide fuel (RBWR-MOX) to nitride fuel (RBWR-MN-95), while still maintaining the 95% TD. This translates into a 40% increase in heavy metal. As seen in Fig. 5, the spectrum for assembly RBWR-MN-95 is slightly harder due to the higher heavy metal-to-moderator ratio. The third case, RBWRMN-85, uses 85% TD (25% increase in heavy metal) in the fissile regions but the density in the blanket regions is still maintained at 95% since the local burnup there is substantially lower. Having high density blankets increases the macroscopic capture cross section of U238 and in turn increases the breeding ratio. Proper normalization is required for a fair comparison of the breeding ratio between these 3 cases. Therefore, the single batch discharge burnup, B1 was fixed to that of the reference case, RBWRMOX. B1 was calculated based on linear reactivity theory in the equation below (Driscoll et al., 1991): B1 ¼ Bd ðn þ 1Þ=ð2nÞ [3] where n is the number of batches and Bd is the average discharge burnup. For a discharge burnup of 45 MWd/kg and 4.62 batches (4 batches of 156 assemblies and 1 batch of 96 assemblies), the single batch discharge burnup is 27.375 MWd/kg. This corresponds to a radially-reflected k-eff of about 1.004 for the reference RBWRMOX assembly. In order to have the reactivities of the nitride Fig. 4. Void fraction distribution of RBWR in MCNP model. B. Feng et al. / Progress in Nuclear Energy 53 (2011) 862e866 Fig. 5. Neutron energy spectrum comparison of RBWR with oxide and nitride fuels and PWR. 865 fixed, using the 95% TD nitride fuel increases the cycle length from 360 to 496 EFPDs. The corresponding fissile inventory ratios (FIR) are shown in Fig. 7. The FIR is defined as the mass of Pu239 and Pu241 at a given burnup divided by the initial mass of Pu239 and Pu241 (at zero burnup). The FIRs for the assembly models were all measured at the maximum fuel discharge burnup of 48.75 MWd/kg. The FIR for RBWR-MN-95 and RBWR-MN-85 are about 1.17 and 1.15, respectively, about 10% higher than the 1.04 FIR for RBWRMOX. Although significant for water-based breeders, this w10% potential increase in breeding ratio still does not match the breeding capabilities of SFRs. Alternatively, it may be more beneficial to alleviate safety concerns by increasing the coolant-to-fuel ratio by reducing the void fraction or increasing the pitch of an RBWR using nitride fuel. This modification to the RBWR is currently being investigated. 4. Conclusions Fig. 6. Reactivity versus burnup for RBWR assembly using oxide fuel (RBWR-MOX), 85% TD nitride fuel (RBWR-MN-85), and 95% TD nitride fuel (RBWR-MN-95). assemblies intersect at this point, as shown in Fig. 6, the average Puf enrichments were lowered to 13.2 and 14.7 wt% HM for the RBWRMN-95 and RBWR-MN-85 cases, respectively. The Puf was unevenly distributed in the lower and upper fissile zones in order to keep the same axial power profile as that of the RBWR-MOX assembly (from Fig. 3). Since the power density and assumed flow rate are the same for all 3 assemblies, keeping the same axial linear heat rate distribution would result in the same axial void distribution. Since the discharge burnup [MWd/kgHM] and power rating [MW] were This scoping study investigated the potential breeding gain of using high density nitride fuel to further harden the spectrum in a high converting water reactor design. Assembly models of Hitachi’s RBWR were depleted via MCODE to determine the gain in breeding ratio when switching from the reference (U,TRU)O2 fuel to the higher density (U,TRU)N. For a fixed discharge burnup, 95% and 85% TD nitride fuels yielded fissile inventory ratios of 1.17 and 1.15, respectively, a 10% increase from the reference value of 1.04. However, further work is recommended to investigate the stability of the nitrides in water, its fuel behavior under irradiation, the economics of production and N15 enrichment, and the safety of hard spectrum BWRs with high void fractions. If a similar increase (þ10%) in breeding ratio can be achieved by using nitride fuels in high converting PWRs, then it may even be feasible to develop a Pu-based breeder (FIR slightly >1.0) using single-phase liquid water as the coolant. A detailed analysis of such a concept is also recommended for future work. Acknowledgments This research was part of the MIT study on “The Future of Nuclear Fuel Cycle” sponsored by a consortium of industry and laboratories, including AREVA, EPRI, NEI, and INL. The authors would also like to thank Dr. Aydin Karahan and Dr. Ed Pilat for their help and contributions. References Fig. 7. Fissile inventory ratio versus burnup for RBWR assembly using oxide fuel (RBWR-MOX), 85% TD nitride fuel (RBWR-MN-85), and 95% TD nitride fuel (RBWRMN-95). Alexander, Carl A., 1986. Metal-actinide nitride nuclear fuel, US Patent 4,624,828. Arai, Y., Akabori, M., Minato, K., 2006. JAEA’s Activities on Nitride Fuel Research for MA Transmutation, Presentation, 9th IEM on Actinide and Fission Product Partitioning and Transmutation. Broeders, C.H.M., 1985. Conceptual design of a (Pu, U)O2 core with a tight fuel rod lattice for an advanced pressurized light water reactor. Nucl. Technol. 71, 82e95. Croff, A.G., 1980. A User’s Manual for the ORIGEN2 Computer Code. ORNL/TM-7175. Oak Ridge National Laboratory. Ding, Xing-Cheng, et al., 2008. High enrichment of 15N isotope by ion exchange for nitride fuel development. Prog. Nucl. Energy 50, 504e509. Driscoll, M.J., Downar, T.J., Pilat, E.E., 1991. The Linear Reactivity Model for Nuclear Fuel Management. American Nuclear Society. ISBN# 0894480359. Edlund, M.C., 1976. Physics of uranium-plutonium fuel cycles in pressurized water reactors. Trans. Am. Nucl. Soc. 24, 508. Feng, B., Karahan, A., Kazimi, M.S., 2011. Steady-State Nitride Fuel Behavior Modeling with FRAPCON-EP and its Application to PWRs, Proceedings of ICAPP 2011, paper 11340, Nice, France. Hayes, S.L., Thomas, J.K., Peddicord, K.L., 1990. Material property correlations for uranium mononitride-II. Mechanical properties. J. Nucl. Mater. 171, 271e288. Los Alamos National Laboratory, 2005. MCNP-A General Monte Carlo N-Particle Transport Code, Version 5, LA-UR-03-1987 (revised). 866 B. Feng et al. / Progress in Nuclear Energy 53 (2011) 862e866 Ma, Benjamin M, 1983. Nuclear Reactor Materials and Applications. Van Nostrand Reinhold Company. Ronen, Y., Dali, Y., 1998. A high-conversion water reactor design. Nucl. Sci. Eng. 130, 239e253. Takeda, et al., 2010. Core of Light Water Reactor and Fuel Assembly, US Patent Application Publication No. 2010/0034336. Takeda, R., Miwa, J., Moriya, K., September 9e13, 2007. BWRS for Long-Term Energy Supply and for Fissioning Almost All Transuraniums. GLOBAL 2007, Boise, Idaho, pp. 1725e1732. Tanaka, K., et al., 2004. Fission gas release and swelling in uranium-plutonium mixed nitride fuels. J. Nucl. Mater. 327, 77e87. Uchikawa, S., Okubo, T., Nakano, Y. 2009. Breeder-type Operation based on the LWR-MOX Fuel Technologies in Light Water Reactors with Hard Neutron Spectrum (FLWR). Proc. of ICAPP 2009, Tokyo, Japan, paper 9022, May 10e14, 2009. Xu, Z., Hejzlar, P., Kazimi, M., 2006. MCODE, Version 2.2-an MCNP-ORIGEN Depletion Program, MIT-NFC-TR-104. MIT Center for Advanced Nuclear Energy Systems.