Study of an ADS Loaded with Thorium and Reprocessed Fuel

Accelerator-driven systems (ADSs) are investigated for long-lived fission product transmutation and fuel regeneration. The aim of this paper is to investigate the nuclear fuel evolution and the neutronic parameters of a lead-cooled accelerator-driven system used for fuel breeding. The fuel used in some fuel rods was ThO2 for 233U production. In the other fuel rods was used a mixture based upon Pu-MA, removed from PWR-spent fuel, reprocessed by GANEX, and finally spiked with thorium or depleted uranium. The use of reprocessed fuel ensured the use of ThO2 without the initial requirement of 233U enrichment. In this paper was used the Monte Carlo code MCNPX 2.6.0 that presents the depletion/burnup capability, combining an ADS source and kcode-mode (for criticality calculations). The multiplication factor (keff) evolution, the neutron energy spectra in the core at BOL, and the nuclear fuel evolution during the burnup were evaluated. The results indicated that the combined use of ThO2 and reprocessed fuel allowed 233U production without the initial requirement of 233U enrichment.


Introduction
In recent years great interest has been given to acceleratordriven systems (ADSs).This is mainly because of their inherent safety features, their waste transmutation potential, and their capability to breed the required 233 U when the thorium fuel is used.ADSs are useful for recycling of americium, curium, neptunium, and plutonium.A great number of works on the ADS and the relative neutronics have been reported in the scientific literature [1][2][3][4][5][6][7].Pioneers in this revival have been Furukawa et al. [8], Bowman et al. [9,10], and Rubbia et al. [11,12].Very similar idea was first presented almost 60 years ago [13].In an accelerator-driven system, an accelerator is coupled to a subcritical core loaded with nuclear fuel.The particles accelerated are injected into a spallation target that produces neutrons, which are used in the subcritical core for the fission chain maintenance.
Both critical reactors and subcritical accelerator-driven systems are potential candidates as dedicated transmutation systems.Critical reactors, however, loaded with fuel containing large amounts of MA have safety problems caused by unfavorable reactivity coefficients and small delayed neutron fraction.Nevertheless, subcritical systems present criticality and safety advantages: they operate with a neutron source rate that can be increased to compensate for the negative reactivity or any variation during the operation.Therefore, research interest in spent nuclear fuel transmutation has focused on both accelerator-driven systems and fusiondriven systems.At DEN/UFMG we began about 2006 and some results were obtained [14][15][16].In this work, the focuses are 233 U production and transuranic transmutation possibility in accelerator-driven systems.
The main characteristic of ADS (i.e., subcriticality) is particularly favorable and allows a maximum transmutation rate while operating in a safe manner.An advantage of accelerator-driven systems is that, since there is no criticality condition to satisfy, almost any fuel composition can be used in the system [17].Moreover, ADS could transmute 99.9% of transuranics light-water-reactor-(LWR-) spent nuclear fuel [18].
Another fuel proposed for using in ADS is the thoriumbased fuel.There are many reasons for the resurgence of interest in the thorium fuel cycle nowadays.The thorium's abundance is about three times more than uranium abundance.Thorium-fueled reactor is also an attractive tool to produce long-term nuclear energy with low radiotoxicity waste [19].The main issue verified in the usage of this fuel in subcritical systems is the need of initial enrichment ( 233 U), since the use of natural thorium ( 232 Th) is not feasible due to the very low values of achieved criticality.
Based on those questions, this work presents a preliminary study of ADS for simultaneous use of thorium fuel and reprocessed fuel.The simultaneous usage of thorium fuel and spent fuel allows one to get appropriate values of criticality without the need of enrichment of thorium and still can reduce radiotoxicity of spent fuel by transmutation.The proposal is the simulation of a core with some fuel elements of 232 Th and other elements based on fuel spent from LWRs.There was simulated a cylindrical core of 12.0 m 3 filled with a hexagonal lattice formed by 156 fuel rods, the fuel used in 36 rods was 232 ThO 2 , and the fuel used in the other 120 rods was reprocessed fuel spiked with thorium or depleted uranium.The reprocessed fuel was obtained from spent fuel discharged from the Brazilian PWR ANGRA-I, reprocessed by GANEX processes, and spiked with thorium or depleted uranium.Different percentages of ThO 2 or UO 2 were added to the reprocessed fuel and analyzed, as the use of only reprocessed fuel generates high criticality values.The neutron spectra were evaluated at BOL (Begin Of Life) and the criticality and the fuel evolution were investigated during ten years with the system operating at steady state, at 515 MW of thermal power.During this time, the 233 U production and the reprocessed fuel evolution were quantified.
MCNPX 2.6.0 code [20] was used to simulate the geometrical and operational characteristics of the system.For the simulation there was used a combination of spallation source (external source-SDEF) and kcode-mode for the calculation of initial k eff and flux, describing in this way the real behavior of an ADS; that is, the initial criticality is the sum of neutrons produced by fission in the fuel and the neutrons produced by spallation in the target.During the burnup the code does not take into account the flux from external source, so the k eff and fuel evolution results obtained are just approximations.Moreover, the k eff results that were obtained during the burnup (kcode-mode) provide an insight of the necessary contribution of the external source during the operation period.
The isotopic composition of the reprocessed fuel used in this simulation was obtained using GANEX process (Group ActiNide EXtration) which is currently in testing phase [21].The GANEX process developed by CEA for the reprocessing of Generation IV spent nuclear fuels is composed of two extraction cycles following the dissolution of the spent fuel.Once the uranium is selectively extracted from the dissolution solution by a monoamide solvent, the transuranic elements (Np, Pu, Am, and Cm) are separated from the fission products in a second cycle prior to the co-conversion step [22].
The GANEX first cycle recovers more than 99.99% of the total amount of uranium.In a second cycle neptunium, plutonium, americium, and curium are recovered altogether in one liquid flow (actinide product) and the losses of transuranics in the different outputs and in the solvent were estimated at a value lower than 0.5% (neptunium essentially) at the end of the test, corresponding to a recovery yield of actinides higher than 99.5%.Nevertheless, the decontamination factors versus some lanthanides (especially Nd, Sm, and Eu) were much lower than expected and the mass of lanthanides in the actinide product was around 5% at the end [23].

Methodology
2.1.Computational Tool.The MCNPX 2.6.0 code was used to simulate the geometrical and operational characteristics of the system.Such version is quite interesting for the ADS evaluation because it describes the nuclear fuel evolution during the operation.For the calculation of the initial k eff and flux there was used a combination of spallation source (external source-SDEF) and kcode-mode.The code does not take into account the flux from external source during the burnup, so the results obtained are just approximations.
The depletion/burnup capability is based on CINDER90.MCNPX depletion is a linked process involving steady-state flux calculations by MCNPX and nuclide depletion calculations by CINDER90.The code runs a steady-state calculation to determine the system eigenvalue, 63-group fluxes, energy-integrated reaction rates, fission multiplicity, and recoverable energy per fission (Q values).CINDER90 then takes those MCNPX-generated values and performs the depletion calculation to generate new number densities for the next time step.MCNPX takes these new number densities and generates another set of fluxes and reaction rates.The process repeats itself until after the final time step specified by the user [24]. 1 and 2 show schematic views of the simulated ADS.The basic geometry includes the spallation target, a subcritical core, and the reflector.The accelerator tube has a radius of 1.5 cm, and the axial position is in the center of the target.The spallation source is represented by a neutrons source with a spectrum characteristic of spallation reactions.Such spectrum was generated, in a previous simulation, using a beam of 1-GeV protons with a parabolic spatial profile.There was used Bertini intranuclear cascade model for the transport of protons, neutrons, and charged pions.The parameters of this simulation were described in [15].

System Parameters. Figures
Due to its high neutron yield by spallation reactions there was used lead as spallation target.The Pb cylindrical target has 9.5 cm radius and 39 cm height.Lead was also used as coolant and as reflector.The use of lead as coolant offers many advantages like convective cooling, passive safety and small neutron absorption cross section [25].
Another advantage of the lead coolant is that it is not a neutron moderator.This is important because the protactinium effect, which limits the achievable values of k eff , is less severe for harder spectra.For solid fuels, systems without moderator and based on thorium, smaller values of capture    cross sections of fission products will reduce the k eff variation and produce a hardening (shift to higher energies) in the neutron energy distribution.So, the inventory of 233 U is much larger in fast reactors (about 7 times), with the associated larger breeding times and inventory radiotoxicity [26].
The simulated core is a cylinder of 12.0 m 3 filled with a hexagonal lattice formed by 156 fuel rods.The fuel rod diameter is 6 cm, the pitch is 12 cm, and rod length is 400 cm.The fuel used in 36 rods is 232 ThO 2 .The fuel used in the other 120 rods is spent fuel discharged from reactor ANGRA-I, with initial enrichment of 3.1%.This fuel was burned in ORIGEN 2.1 code [27] for three cycles, with the burnup of approximately 11.000 MW d/t in each cycle, following the ANGRA-I power historic of real cycles 1, 2, and 3.The composition of the spent fuel used is shown in Table 1 [28], after remaining 5 years in a cooling pool.
After cooling by five years, the spent nuclear was submitted to GANEX reprocessing and the isotopic composition after the reprocessing is presented in Table 2.The amount of uranium after the reprocessing is 0.01% of the total amount of uranium in the burned fuel.In terms of actinides, neptunium, plutonium, americium, and curium are recovered with a loss of only 0.5% (neptunium essentially).The lanthanides contamination in the reprocessed fuel is around 5%.The isotopic composition presented in Table 2 was adjusted (with a computational program) for several percentages of thorium or depleted uranium added to the reprocessed fuel and the final composition (ThO 2 or UO 2 + reprocessed fuel) was normalized.The ratio between the number of thorium fuel rods and reprocessed fuel rods and the positions of the thorium fuel rods were determined to maximize the amount of reprocessed fuel and achieve k eff values close to 1.The isotopic compositions used in the simulations with several percentages of thorium or depleted uranium added to the reprocessed fuel are shown in Tables 3 and 4.

Calculational Procedure.
For this simulation there was used a combination of ADS source (SDEF) and kcode-mode   for calculation of initial k eff and flux.The SDEF source was positioned on the center of the target.The burnup calculations were performed in kcode-mode using 53 time steps; the total simulated time was 10 years and the thermal power of operation throughout this period was 515 MW t. were performed simulations using reprocessed fuel spiked with thorium in different percentages and simulations using reprocessed fuel spiked with different percentages of depleted uranium.

Neutronic Evaluation.
In Figure 3 is shown the normalized neutron energy distribution (for the systems spiked with 62% of thorium and 62% of depleted uranium) in the BOL (beginning of life).Both spectra in the thorium fuel present their maximum around 150 keV and around 250 keV in the reprocessed fuel.Therefore, the reprocessed fuel neutron spectrum is harder than thorium fuel spectrum.That is result of the plutonium presence that emits fission neutrons with a harder spectrum.The spectra for the system loaded with fuel spiked with thorium are very similar to spectra for the fuel spiked with depleted uranium.The spectrum in the lead of the core (coolant) presents a similar neutrons energy distribution to fuel spectra, but the number of neutrons available in the fuel is larger due to neutron production by fission reactions.Figure 4 shows the multiplication factor (k eff ) as a time function for the systems loaded with 232 ThO 2 (24 rods) and reprocessed fuel (132 rods) spiked with thorium (Figure 4(a)) and spiked with depleted uranium (Figure 4(b)).The percentages of thorium and depleted uranium in the reprocessed fuel were varied and the k eff values obtained are represented in Figure 4.The percentages of thorium or depleted uranium were chosen in such a way that the values for the initial k eff were close to 1, since a large variability on those percentages could lead to considerably low initial k eff values.
From Figure 4 can be observed that after some years of operation the decrease in the k eff value was smaller in the system which used reprocessed fuel spiked with thorium.That behavior indicates that the use of reprocessed fuel spiked with thorium allows an extension of the burnup without reloading fuel.
For the same percentage of thorium and depleted uranium (62%), the k eff values for the system spiked with depleted uranium (k eff variation of 0.991 to 0.761) were significantly higher than these for the system spiked with thorium (k eff variation of 0.964 to 0.782), but the decrease in the k eff for the system with depleted uranium was also higher.
The drop of the k eff value observed in two cases is mainly due to the increased poisoning caused by the accumulation of fission products having large neutron capture cross sections and, of course, the constant energy generation by the system.
3.2.Thorium Fuel Evolution.Figure 5 describes the 232 Th, 233 Pa, and 233 U concentrations in the thorium fuel rods during the 10 years of operation for the systems spiked with 62% of thorium and 62% of depleted uranium.It can be verified that for the system spiked with thorium the concentration of the isotope 232 Th was reduced by about 17% (623 kg), for the system spiked with depleted uranium that was about 18% (644 kg).It is the result of the considerable capture cross section of that isotope, which allows the 233 U production. 233U is formed when 232 Th captures a neutron, and it soon undergoes two beta decays: ( A relevant question in the neutron spectrum choice is the protactinium effect.Protactinium captures neutrons (due to its large capture cross section) and, thus, decreases the reactivity and fuel regeneration.From Figure 5 can be observed that this isotope is formed in small scale in both systems.Furthermore, the protactinium effect is less severe for harder spectra like that. Figure 5 also presents the 233 U concentration during the burnup, which is formed in large scale (∼282 kg) for the system with reprocessed fuel spiked with thorium and (∼284 kg) for the system spiked with depleted uranium.That 233 U production was due to 232 Th regeneration.Then these results show that the combined use of 232 ThO 2 and reprocessed fuel allowed 233 U production without the initial requirement of 233 U enrichment.

Reprocessed Fuel Evolution
. The long-term potential radiotoxicity of the spent nuclear fuel is mainly associated with transuranics (TRU) particularly Pu and MA: Am, Cm, and Np.These constitute a significant radiological source term over a very long period within a spent fuel [29].Figures 6-9 show the concentration of those isotopes during the burnup.
In a standard uranium-uranium ( 235 U x , 238 U 1−x ) fuel used in nuclear reactors, the isotope of neptunium, 237 Np, originates from 235 U, and the isotopes of americium, 241 Am and 243 Am, and curium, 242 Cm, 244 Cm, and 245 Cm, from 238 U as a result of nuclear reactions [30].
Plutonium is the major element for radiotoxicity and mass in the storage [31]. 238Pu, which is generated from 237 Np, was produced in large scale in the system simulated (Figure 6), while 239 Pu and 241 Pu mass were reduced mainly due to fission reactions. 240Pu is formed when 239 Pu captures a neutron that justifies the high production of that isotope due to the reduced in 239 Pu. 242 Pu was formed in small scale.
The greater plutonium production in the system spiked with depleted uranium was due to the greater uranium availability in this system that allows the 237 Np formation and consequently plutonium production.
Long-lived isotopes of americium produced are 241 Am and 243 Am.Americium is produced by β-decay of 241 Pu ( 241 Pu → 241 Am) and (n, γ) reaction of shorter lifetimes with intermediate processing and addition of new plutonium.Since 244 Cm half-life makes 18.1 years, one can organize a controllable storage of extracted curium [32].The Cm mass evolution in the two cases was very similar.
Neptunium is produced by (n, γ) reaction of 235 U, (n, 2n) reaction of 238 U, and α-decay of 241 Am.The most important Np isotope is 237 Np (half-life 2.14 × 10 6 years).The decrease observed in the 237 Np mass (Figure 9) is mainly due to 238 Np production by neutron capture.The very low concentration of 238 Np (Figure 9) is justified by the quick decay (2.117 days) to 238 Pu through β decay.Once more was not observed significant difference between the mass variations for each isotope in the two cases.
The large amount of 232 Th in the fuel spiked with 232 ThO 2 results in a high 233 U production (∼258 kg), as can be observed in Figure 10.That fuel regeneration potential appears with an advantage of the reprocessed fuel spiked with thorium.Figure 10 also shows the 235 U and 238 U mass variations.

Conclusions
The main conclusions indicated by the simulation presented in this work were as follow.
(i) It is noteworthy that the code does not take into account the flux from external source during the burnup, so the k eff and fuel results obtained are just approximations.Nevertheless, such approximated results are reasonable ones as long as the external source would not be sufficient to drive the system towards criticality, because the k eff values obtained were substantial low during the operation period.The initial k eff obtained (k eff(initial) = neutrons fission + neutrons spallation ) were appropriate for an ADS and the k eff obtained during the burnup (k eff = neutrons fission ) provide an insight of the necessary contribution of the external source on the operation period.(ii) The insertion of reprocessed fuel in a 232 ThO 2 system leads to harder neutron spectra due to the plutonium presence, facilitating the 233 U production.This is result of the protactinium effect reduction by hardening of the neutron spectrum.(iii) The reprocessed fuel evolution was very similar when that fuel was spiked with thorium or depleted uranium.However, the 233 U production appears with an advantage of the reprocessed fuel spiked with thorium.Furthermore, the decrease in the criticality was smaller in the system spiked with thorium that allows an extension in the burnup period.
of the tank Lead of reflector Reprocessed fuel Lead of target and core Thorium fuel

Figure 1 :
Figure 1: Horizontal cross section of the ADS simulated.

Figure 2 :
Figure 2: Vertical cross section of the ADS simulated.

Table 3 :
Fuel composition (normalized) for the reprocessed fuels spiked with thorium.

Table 4 :
Fuel composition (normalized) for the reprocessed fuels spiked with depleted uranium.