Electrochemical Reduction of (U,Pu)O 2 in Molten LiCl and CaCl 2 Electrolytes Masatoshi IIZUKA 1; , Tadashi INOUE 1 , Michel OUGIER 2 and Jean-Paul GLATZ 2 1 Central Research Institute of Electric Power Industry, 2-11-1 Iwato-kita, Komae, Tokyo 201-8511, Japan 2 Institute for Transuranium Elements, Postfach 2340, D76125 Karlsruhe, Germany (Received August 15, 2006 and accepted in revised form February 2, 2007) The electrochemical reduction of UO 2 -PuO 2 mixed oxides (MOX) was performed in molten LiCl at 923 K and CaCl 2 at 1,123 K to evaluate the behavior of the plutonium quantitatively and to define the op- timum conditions for the electrochemical reduction of those materials. In LiCl, excess deposition of lithium metal can be avoided and the MOX was smoothly reduced at 0:65 V vs. Bi-35 mol% Li reference electrode. The reduction ratio calculated from the mass change of the samples taken during the electrochemical reduction and the ratio evaluated by gas-burette method were in good agreement. The cathodic current efficiency remained 30–50% mainly due to the deoxidation of tantalum cathode basket. Although dissolution of plutonium and americium into the electrolyte was found by the chemical analysis, the dissolved amount was negligible and had no immediate influence on the feasibility of the electrochemical reduction process. In CaCl 2 , reduction of the MOX occurred in whole range of the tested cathode potential (0:15 V to 0:40 V vs. Ca-Pb reference electrode). The cathodic current efficiency was around 30%. Although the MOX was completely reduced at 0:25 V, the reduction was interrupted by formation of the surface bar- rier made of the reduced material and the vacancy between the reduced and the non-reduced areas at 0:30 V. Plutonium and americium dissolved also into the CaCl 2 electrolyte to slightly higher concentra- tions than those observed in LiCl electrolyte. The analyses for the reduction products showed that the amount of those actinides lost from the cathode was much larger than that found in the electrolyte, prob- ably due to the formation of mixed oxide precipitate. KEYWORDS: molten salt, lithium chloride, calcium chloride, pyrometallurgical reprocessing, re- duction, electrochemistry, MOX, current efficiency I. Introduction Development of innovative nuclear fuel cycle technology is strongly required to serve for both the environmental sus- tainability and the increasing energy demand. Central Re- search Institute of Electric Power Industry (CRIEPI) propos- es the metallic fuel cycle as one of the most promising op- tions in the nuclear fuel cycle technologies for the next gen- eration. The metallic fuel cycle, a combination of the metal fuel fast reactor and the pyrometallurgical reprocessing, has excellent advantages in economic, safety and proliferation resistance aspects. 1) In order to supply the metallic fuel for start-up of the fast reactors, the oxide materials, either the LWR spent fuels or the UO 2 -PuO 2 mixed oxide (MOX) products from the light water reactor (LWR) fuel reprocessing, must be reduced to metals. The lithium reduction process using lithium metal reduc- tant in molten LiCl bath has been developed for that purpose. Reduction of UO 2 2) and simulated spent LWR fuel 3) was demonstrated. The thermodynamic conditions for reduction of transuranium elements (TRUs) 4) were also studied. The lithium reductant was successfully recovered by electro- chemical decomposition of the co-produced Li 2 O in the bath. 5) Although the lithium reduction process is thus already an experimentally proven technique, there are still some technical problems to be solved such as the handling of high- ly reactive and sticky lithium metal and the difficulties in the design of the lithium recovery equipment. Recently, the electrochemical reduction method has been studied for reduction of oxide nuclear fuels. Figure 1 shows the principle of the electrochemical reduction of spent oxide fuels. The cathode and anode reactions are as follows: Cathode: MO 2 þ 4e ! M þ 2O 2ð1Þ Anode: 2O 2! O 2 þ 4e ; ð2Þ where M denotes actinide metals When a carbon anode is used, CO or CO 2 is evolved in place of O 2 : ÓAtomic Energy Society of Japan Corresponding author, E-mail: [email protected]Journal of NUCLEAR SCIENCE and TECHNOLOGY, Vol. 44, No. 5, p. 801–813 (2007) 801 ORIGINAL PAPER
Transcript
Electrochemical Reduction of (U,Pu)O2 in Molten LiCl
and CaCl2 Electrolytes
Masatoshi IIZUKA1;�, Tadashi INOUE1, Michel OUGIER2 and Jean-Paul GLATZ2
1Central Research Institute of Electric Power Industry, 2-11-1 Iwato-kita, Komae, Tokyo 201-8511, Japan2Institute for Transuranium Elements, Postfach 2340, D76125 Karlsruhe, Germany
(Received August 15, 2006 and accepted in revised form February 2, 2007)
The electrochemical reduction of UO2-PuO2 mixed oxides (MOX) was performed in molten LiCl at923K and CaCl2 at 1,123K to evaluate the behavior of the plutonium quantitatively and to define the op-timum conditions for the electrochemical reduction of those materials.
In LiCl, excess deposition of lithium metal can be avoided and the MOX was smoothly reduced at�0:65V vs. Bi-35mol% Li reference electrode. The reduction ratio calculated from the mass changeof the samples taken during the electrochemical reduction and the ratio evaluated by gas-burette methodwere in good agreement. The cathodic current efficiency remained 30–50% mainly due to the deoxidationof tantalum cathode basket. Although dissolution of plutonium and americium into the electrolyte wasfound by the chemical analysis, the dissolved amount was negligible and had no immediate influenceon the feasibility of the electrochemical reduction process.
In CaCl2, reduction of the MOX occurred in whole range of the tested cathode potential (�0:15V to�0:40V vs. Ca-Pb reference electrode). The cathodic current efficiency was around 30%. Although theMOX was completely reduced at �0:25V, the reduction was interrupted by formation of the surface bar-rier made of the reduced material and the vacancy between the reduced and the non-reduced areas at�0:30V. Plutonium and americium dissolved also into the CaCl2 electrolyte to slightly higher concentra-tions than those observed in LiCl electrolyte. The analyses for the reduction products showed that theamount of those actinides lost from the cathode was much larger than that found in the electrolyte, prob-ably due to the formation of mixed oxide precipitate.
Development of innovative nuclear fuel cycle technologyis strongly required to serve for both the environmental sus-tainability and the increasing energy demand. Central Re-search Institute of Electric Power Industry (CRIEPI) propos-es the metallic fuel cycle as one of the most promising op-tions in the nuclear fuel cycle technologies for the next gen-eration. The metallic fuel cycle, a combination of the metalfuel fast reactor and the pyrometallurgical reprocessing, hasexcellent advantages in economic, safety and proliferationresistance aspects.1)
In order to supply the metallic fuel for start-up of the fastreactors, the oxide materials, either the LWR spent fuels orthe UO2-PuO2 mixed oxide (MOX) products from the lightwater reactor (LWR) fuel reprocessing, must be reduced tometals.
The lithium reduction process using lithium metal reduc-
tant in molten LiCl bath has been developed for that purpose.Reduction of UO2
2) and simulated spent LWR fuel3) wasdemonstrated. The thermodynamic conditions for reductionof transuranium elements (TRUs)4) were also studied. Thelithium reductant was successfully recovered by electro-chemical decomposition of the co-produced Li2O in thebath.5) Although the lithium reduction process is thus alreadyan experimentally proven technique, there are still sometechnical problems to be solved such as the handling of high-ly reactive and sticky lithium metal and the difficulties in thedesign of the lithium recovery equipment.
Recently, the electrochemical reduction method has beenstudied for reduction of oxide nuclear fuels. Figure 1 showsthe principle of the electrochemical reduction of spent oxidefuels. The cathode and anode reactions are as follows:
Journal of NUCLEAR SCIENCE and TECHNOLOGY, Vol. 44, No. 5, p. 801–813 (2007)
801
ORIGINAL PAPER
Anode: 2O2� þ C ! CO2 þ 4e� ð3ÞO2� þ C ! COþ 2e�: ð4Þ
The electrochemical reduction process eliminates the trou-blesome handling of lithium metal required in the lithium re-duction process. Another advantage of the electrochemicalreduction is that the oxide ions produced at the cathodeare simultaneously consumed at the anode. It means thatthe concentration of oxide ions in the electrolyte can bemaintained at a desired low level, while it gradually accumu-lates in the lithium reduction process. This nature of theprocess is very favorable because low oxide ion concentra-tion thermodynamically pushes the reaction towards com-plete reduction of the actinide elements. Since concentrationof oxide ion in the electrolyte does not increase, the amountof the electrolyte and the volume of the equipment, in con-sequence, can be reduced.
The fundamental behavior of UO2 in the electrochemicalreduction process has been investigated in both molten LiCland CaCl2 electrolytes.
6) In CaCl2 bath at 1,123K, a thin butdense layer has formed in the surface region of the oxidematerial at the cathode due to cohesion among the reduceduranium metal particles. Since this layer hampered the out-ward diffusion of oxide ions evolved by the reaction (1),the subsequent reduction was disturbed and the central re-gion of the cathode material remained original oxide. In LiClbath at 923K, on the other hand, the reduction product didnot cohere so much to form a dense layer, probably becauseof the lower temperature. Electrochemical reduction ofMOX in molten LiCl was also studied.7) Complete reductionof a small MOX piece of around two hundred milligramsin weight was confirmed by scanning electron microscope(SEM) observation. The reduction started at the grain boun-daries of the oxide material. In the course of the reduction
into the grain bulk, the reduction products formed thecoral-like structure.
The feasibility of the electrochemical reduction of UO2
and MOX has been experimentally well explained. Howev-er, some points still need to be clarified, such as the electro-chemically optimum conditions for the reduction and thecause for decrease of plutonium concentration in the reduc-tion product.7) In the next step, more detailed information onthe behavior of plutonium in the electrochemical reductionprocess should be studied in order to evaluate the materialbalance and establish the process flow sheet.
In the present study, the electrochemical reduction ofMOX was performed in molten LiCl at 923K and CaCl2at 1,123K to evaluate the behavior of plutonium quantita-tively and to define the desirable conditions for the electro-chemical reduction of those materials.
II. Experiments
1. ApparatusAll the electrochemical reduction experiments were car-
ried out in an argon atmosphere glove box, because bothLiCl and CaCl2 are highly hygroscopic and the reductionproducts might be easily re-oxidized even by low concentra-tion of oxygen. The concentration of water and oxygen in theargon atmosphere was kept lower than 1.0 ppm during thetests by the argon purification system.
Figure 2 shows a schematic view of the experimental ap-paratus. A MgO crucible (40mm in inner diameter, 75mmin depth and 5mm in thickness) containing LiCl or CaCl2was placed in the stainless steel vessel and heated to 923Kor 1,123K, respectively. The temperature was kept � 2Kby a PID controller and occasionally checked with a type-K thermocouple covered with a MgO tube.
Cathode :Actinide oxide isreduced to metal.
Anode :Oxygen or carbon oxide gas is evolved.
Spentoxide fuel
FPs (Cs,Sr,etc.) Molten salt
CO2/CO/O2
O2-
Fig. 1 Schematic diagram of the electrochemical reduction proc-ess
Ar glove box
N2 glove box
Heaters
Thermocouples forTemperature control
Cooling waterThermal insulation
A B C D
Motor/Lifter
A : Pt coil anodeB : Bi-Li reference electrodeC : Ta basket + MOX cathodeD : Thermocouple
Fig. 2 Schematic view of the experimental apparatus
802 M. IIZUKA et al.
JOURNAL OF NUCLEAR SCIENCE AND TECHNOLOGY
For the basic electrochemical measurement, a tantalumwire working electrode (1mm in diameter and about20mm in immersed depth) was used. In the electrochemicalreduction experiments, baskets of approximately 10mm di-ameter made from tantalum mesh and tantalum wire wereused to hold the small MOX pieces. In some cyclic voltam-metry measurements, alumina cathode containers of thesame dimension were used to minimize the change of thecathode surface area by the tantalum parts.
A platinum wire electrode was used as a counter electrodefor the electrochemical measurement or as an anode for theelectrochemical reduction experiments. The platinum wireof 1mm diameter was formed into a coil of 4mm in diame-ter and 20mm in length. The upper end of the wire was con-nected to a tantalum wire to make the electrical connection.The effective length and the surface area of the platinumwire were about 200mm and 6.3 cm2, respectively. Thisplatinum anode was then covered with a MgO shroud (6/10mm in inner/outer diameters) to avoid the decrease ofthe current efficiency by minimizing the migration of evolv-ing oxygen gas to the cathode.
In LiCl electrolyte, Bi-Li alloy was used as a referenceelectrode. About 0.5 g of Bi-35mol% Li alloy was intro-duced in a MgO tube (4/6mm in inner/outer diameters)with one end closed. According to the Bi-Li binary phase di-agram, this alloy exists as a homogeneous liquid at a temper-ature higher than 673K. A tantalum wire of 1mm in diam-eter was inserted in the tube for the electrical connection. Asmall hole (0.7mm in diameter) was made in the wall of thetube at the interface between the molten LiCl and Bi-Li al-loy. This hole acts as a channel between inside and outsideof the MgO tube so that this electrode shows the potentialof Liþ/Li redox equilibrium. Theoretically this potential isaffected by the change of Li activity in the alloy due tothe dissolution into LiCl, migration of Li metal producedat the cathode, and so on. Practically, however, this electrodeshowed stable potential during the experimental period be-cause the migration of lithium through the hole was limitedto very small amount. In CaCl2 electrolyte, Ca-Pb alloy(37mol% Ca) reference electrode was used. The Ca-Pb bina-ry phase diagram shows that this alloy exists as a homogene-ous liquid at a temperature of higher than 903K. The struc-ture of Ca-Pb reference electrode was identical to that of theBi-Li electrode.
A potentio/galvanostat Model 263A from EG&G Prince-ton Applied Research and the software version 4.0 were usedfor both basic electrochemical measurements such as cyclicvoltammetry and electrochemical reduction experiments at aconstant cathode potential. In all the results in this report,cathodic current was expressed in minus direction.
2. ChemicalsLiCl and CaCl2 of 99.9% purity were purchased from Ap-
plied Physical Laboratory, U.S. It was used without any ad-ditional treatment before use since the amount of O2 andH2O impurities in them were extremely low. The concentra-tion of O2� in the molten salt electrolyte was adjusted byadding Li2O of 99.9% purity purchased from Johnson Mat-they Co, or CaO of more than 99% purity purchased from
Rare Metallic Co., Japan. MgO of 99.9% purity was usedas the crucible and the electrode material. The purity of allthe metals used as the electrode materials was higher than99.9%.
The MOX pellets of two different Pu/(U+Pu) ratios wereused in this study. One had a diameter of 6mm and Pu/(U+Pu) ratio of 9.45wt% (MOX-a). The content of americi-um yielded by �� decay from Pu241 in this material was0.756wt% of total amount of plutonium. Another MOX pel-let (MOX-b) was a little larger (8mm in diameter) and oflower plutonium content (Pu/(U+Pu) = 5.02wt%). Whenthese pellets were used in the electrochemical reductiontests, they were cut into slices of around 2mm in thicknessand loaded in the tantalum basket. The ceramograph ofMOX-a pellet before use in the electrochemical reductiontests (Fig. 3) shows that the average grain size in this mate-rial was 5–10 mm.
3. Analytical MethodsThe molten LiCl and CaCl2 samples were taken by dip-
ping an alumina tube at room temperature quickly into thecrucible and by quenching a small amount of the salt onthe surface of the tube. The salt samples were analyzed byICP-MS for their composition. The reduction product sam-ples were washed with methanol to remove the adheringchlorides and analyzed by SEM and energy dispersive X-ray analysis (EDX) for the microstructure and distributionof plutonium. When additional amount of sample was avail-able, it was submitted to ICP-MS and gas-burette analysisfor determination of its chemical composition and reductionratio, respectively.
III. Results and Discussions
1. Determination of the Desirable Cathode PotentialRange for Electrochemical Reduction of MOX by Cy-clic VoltammetryFigure 4 shows the cyclic voltammograms measured in
the LiCl electrolyte without and with a MOX-a sample inthe alumina cathode container. The difference between thecathodic current in these voltammograms, which is also
Electrochemical Reduction of (U,Pu)O2 in Molten LiCl and CaCl2 Electrolytes 803
VOL. 44, NO. 5, MAY 2007
shown in Fig. 4, corresponds to the current used in the re-duction of the MOX. Although a small current for MOX re-duction can be seen from about �0:45V, the current mainlyflows at lower than �0:60V. It is well known that the reduc-tion of lithium (Liþ þ e� ¼ Li) begins at lower than�0:70V. Excess amount of lithium metal deposit coveringthe MOX surface may hamper the diffusion of oxygen ionfrom the cathode and disturb the progress of the reduction.In the following electrochemical reduction tests, therefore,the cathode potential was chosen in the range between�0:45V and �0:70V.
In the cyclic voltammograms measured in the CaCl2 elec-trolyte with a tantalum wire working electrode (Fig. 5), it isobserved that the cathodic current due to the reduction ofcalcium rises at much higher potential compared with thecase in lithium chloride. This difference is caused by the ex-istence of the reduced calcium of lower activity due to itsrapid dissolution and its large solubility in the electrolyte.The dissolved calcium finally migrates to the anode and se-verely decreases the current efficiency in the electrochemicalreduction process. Figure 6 shows the cyclic voltammo-grams measured in the calcium chloride without and witha MOX-a sample in the tantalum cathode basket. There is
a clear difference in the cathodic scan of these two voltam-mograms at lower than �0:10V, indicating the large currentused in the reduction of the MOX. It is, however, difficult tospecify the optimal cathode potential for MOX reduction,since the difference between two voltammograms makes justa straight line to the potential and does not give any partic-ular information on desirable conditions for the reduction.Considering the influence of the reduced calcium metal,the cathode potential was chosen in the range between�0:10V and �0:40V in the following electrochemical re-duction tests.
2. Change of Cathodic Current during ElectrochemicalReductionTwelve electrochemical reduction tests in total, six in LiCl
and six in CaCl2, were carried out at various cathode poten-tials and oxide concentrations in the electrolyte. Major con-ditions and results of these tests are shown in Tables 1 and2, respectively. The concentrations of oxides (Li2O in LiCl,CaO in CaCl2) shown in this table are not chemically ana-lyzed but calculated from the amount of the oxides addedto the electrolytes. The expected reduced ratio in Table 2can be more than 100%, because it is calculated as the per-centage of the electric charge passed during the electrochem-ical reduction test to the charge theoretically equivalent tothe weight of the used MOX sample.
Figure 7 shows a typical change of the cathodic currentduring the electrochemical reduction tests in LiCl. The blackline corresponds to Li-3 test, where �0:50V vs. Bi-Li wasapplied to the MOX-a cathode. Though the reduction currentincreased slightly during the test, the overall trend wassmooth. In Li-2 test where MOX-a was reduced at�0:65V, on the other hand, the cathodic current showedmore unstable behavior. Such variation of the cathodic cur-rent may come from interactive effects of the following threefactors arising during the progress in the reduction of theMOX samples. The first is the improvement of the electriccontact between the tantalum basket and the gradually re-duced MOX surface. The second is disturbance in the diffu-sion of oxygen ions out from the cathode by the growth ofthe reduction product layer at the surface of the MOX.The last is exhaustion of MOX to be reduced in the cathode.
-30
-25
-20
-15
-10
-5
0
5
10
-1.00 -0.75 -0.50 -0.25 0.00 0.25 0.50 0.75 1.00
potential (V vs. Bi-Li)
curr
ent (
mA
)
in LiCl - 0.105 wt % Li2OT = 923 KTa wire working electrodeScan rate : 20 mV/s
with MOX in Ta basket
without MOX in Ta basket
difference
Fig. 4 Cyclic voltammograms measured in LiCl electrolyte with/without MOX in cathode
-700
-600
-500
-400
-300
-200
-100
0
100
-1 -0.75 -0.5 -0.25 0 0.25 0.5 0.75
potential (V vs Ca-Pb)
curr
ent (
mA
)
in CaCl2T = 1123 KTa wire working electrodeScan rate : 100 mV/s
in CaCl2T = 1123 KTa wire working electrodeScan rate : 50 mV/s
with MOX in Ta basket
without MOX in Ta basket
Fig. 6 Cyclic voltammograms measured in CaCl2 electrolytewith/without MOX in cathode
804 M. IIZUKA et al.
JOURNAL OF NUCLEAR SCIENCE AND TECHNOLOGY
According to the result shown in Fig. 7, the variation of thecathode potential does not give much difference in the catho-dic current during the electrochemical reduction test. It isprobably due to the fact that a large part of the cathodic cur-rent is consumed by the deoxidation of tantalum used for thecathode container and the electrical connection.6)
The change of the cathodic current during the electro-chemical reduction tests in CaCl2 is shown in Fig. 8. Thecathodic current was larger at the cathode potential rangingfrom �0:25V to �0:30V than at �0:15V. Because there islittle difference in the cathodic current between �0:25V and�0:30V, the reduction current for the MOX is consideredsaturated within this potential range. The average cathodiccurrent at �0:40V of the cathode potential became muchlarger, certainly due to the increase of the reduction currentof calcium. The frequent bursts of the cathodic current are
Table 1 Major conditions of the electrochemical reduction tests
Run no. ElectrolyteOxide
concentrationa)
(wt%)Used MOX
Amountof MOX
(g)Cathode potentialb)
Appliedelectric charge
(C)
Duration(h)
Li-1 LiCl 0.103 MOX-a 0.286 �0:5V vs Bi-Li 82 7.34Li-2 LiCl 0.103 MOX-a 0.262 �0:65V vs Bi-Li 431 5.01Li-3 LiCl 0.213 MOX-a 0.343 �0:5V vs Bi-Li 303 6.77Li-4 LiCl 0.213 MOX-b 0.564 �0:5V vs Bi-Li 343 6.33Li-5 LiCl 0.272 MOX-a 0.301 �0:65V vs Bi-Li 463 9.60Li-6 LiCl 0.517 MOX-a 0.279 �0:65V vs Bi-Li 335 7.32
Ca-1 CaCl2 0.196 MOX-a 0.293 �0:15V vs Ca-Pb 638 6.06Ca-2 CaCl2 0.196 MOX-a 0.296 �0:25V vs Ca-Pb 846 2.91Ca-3 CaCl2 0.196 MOX-a 0.298 �0:4V vs Ca-Pb 523 1.21Ca-4 CaCl2 0.501 MOX-a 0.351 �0:25V vs Ca-Pb 1480 8.27Ca-5 CaCl2 0.501 MOX-a 0.252 �0:3V vs Ca-Pb 1551 6.49Ca-6 CaCl2 1.013 MOX-b 0.231 �0:3V vs Ca-Pb 483 2.16
a)calculated from the amount of oxides (Li2O in LiCl, CaO in CaCl2) added to the electrolyte.b)Bi-35mol% Li and Pb-37mol% Ca reference electrodes were used in LiCl and CaCl2, respectively.
Table 2 Major results of the electrochemical reduction tests
a)Bi-35mol% Li and Pb-37mol% Ca reference electrodes were used in LiCl and CaCl2, respectively.
-80
-70
-60
-50
-40
-30
-20
-10
0
0 2 4 6 8 10 12
Time (h)
Cur
rent
(m
A)
Li-3, at -0.50 V vs Bi-Li
Li-2, at -0.65 V vs Bi-Li
Fig. 7 Change of the cathodic current during the electrochemicalreduction tests in LiCl
Electrochemical Reduction of (U,Pu)O2 in Molten LiCl and CaCl2 Electrolytes 805
VOL. 44, NO. 5, MAY 2007
considered to be caused by excess amount of oxygen gasgenerated at the surface of the platinum anode, which de-creased the effective anode surface area. Because of the highsolubility of calcium metal in CaCl2 at 1,123K, reduced cal-cium would disperse throughout in the electrolyte, severelydecrease the current efficiency, and react with the insulatingmaterial in the apparatus. The cathode potential should,therefore, be kept higher than at least �0:40V.
3. Evaluation of Reduced Ratio(1) Evaluation by Mass Change of Samples
The cathode products were washed with 20�30ml ofmethanol, because they contained the solidified electrolytesand stuck in the tantalum baskets. Both LiCl and CaCl2 werequickly dissolved in methanol and completely removed fromthe surface of the reduction products and the tantalum bas-ket. Except for a few samples after the tests in LiCl, the sur-face of most cathode products was black probably due to ox-idation by methanol. While all the cathode products fromtests in LiCl electrolyte were easily removed, the productsfrom CaCl2 electrolyte were stuck strongly to the tantalumbasket even after washing. The cathode products from Ca-1 and Ca-6 tests were successfully taken out from the basket,but the others were brittle and broken when removed leavingsmall portions of them in the baskets.
The weight of each reduction product after the washingand separation from the tantalum basket is shown in Table 2.By comparison between the weight of the reduction productand the theoretical value assuming 100% reduction (i.e., thesubtraction of the weight of oxygen from MOX sampleweight), the reduction ratio of each sample can be evaluated.
The reduction ratio in the cathode products from the testsin LiCl ranged from 11.3% to 53.3%. For the tests carriedout in LiCl, with the exception of Li-1 test, the evaluatedcathodic current efficiency was around 10% at a cathode po-tential of �0:50V, and within a range of 35–50% at�0:65V, respectively. These low values are reasonable con-sidering that a large portion of the cathodic current is con-sumed in the deoxidation of tantalum especially at highercathode potential as shown in the cyclic voltammetry results(Fig. 4). It is uncertain why the cathodic current efficiency inLi-1 is estimated at over 100%. The most probable reason is
underestimation the product weight due to a loss of smallportions in the sample handling.
In the electrochemical reduction tests carried out in CaCl2,except for Ca-1 and Ca-6 tests, the reduction ratio evaluatedfrom the mass change was evaluated more than 100%. Thatis obviously due to the fact that those products were brokenand small portions of them could not be recovered when theywere removed from the tantalum basket, as described above.In Ca-1 and Ca-6 tests, where unbroken cathode productswere recovered, the evaluated cathodic current efficiencywas about 40% regardless of the cathode potential. This val-ue is slightly lower than that in LiCl electrolyte, certainlydue to the current losses through the reduction of calcium.(2) Evaluation by Gas-burette Method
For two cathode products from the electrochemical reduc-tion tests in LiCl, the reduction ratio was also estimated bygas-burette method.3) A small piece of the cathode productwas reacted with hydrochloric acid in a closed system, wherehydrogen gas is generated by the following reactions:
By measuring the volume of the evolved hydrogen gas, theamount of actinide elements in the metallic state containedin the cathode product can be determined. In this study,the reduction ratios of uranium and plutonium are assumedto be equal. Finally, comparison between this value andthe amount of the MOX loaded into the cathode gives the re-duction ratio. The estimated reduction ratios by this methodfor the products of Li-2 and Li-7 tests are 46.2% and 49.5%,respectively (Table 2). These values are in good agreementwith those evaluated from the mass change of the cathode,indicating that these two methods are both appropriate for re-duction ratio estimation.
4. Behavior of Actinide Elements in Electrochemical Re-duction—Change of Molten Salt Electrolyte Composi-tionThe change of the actinide concentration in the LiCl and
CaCl2 electrolytes during a series of electrochemical reduc-tion tests was investigated by ICP-MS analysis. The resultsare plotted in Fig. 9. In all the tests shown in this figure,MOX-a samples are used, except for the last test in CaCl2where MOX-b sample was used. The cumulative weight ofthe MOX immersed in the electrolyte was taken as the ab-scissa for the plot.
The higher actinide concentration in the first sample takenfrom LiCl electrolyte is supposed to be due to contaminationof the sample. It can be seen that the concentration of pluto-nium and americium significantly increases compared withthat of uranium according with the progress of the tests.The amounts of plutonium and americium found in the LiClelectrolyte at the end of all the tests were evaluated to be0.14% and 2.6%, respectively, in terms of the proportionto the total amount of each element in the MOX samples im-mersed in the electrolyte. It is also noted that their concen-trations largely increased when Li2O was added to the LiClelectrolyte. Since the MOX was not suspended in the elec-trolyte when Li2O was added, the following behavior of plu-
-300
-250
-200
-150
-100
-50
0
0 2 4 86 10 12
Time (h)
Cur
rent
(m
A)
CaO in CaCl2 : 0.196 - 0.501 wt %
T = 1123 K
Ca-4, -0.25 V vs. Ca-Pb
Ca-1, -0.15 V vs. Ca-Pb
Ca-3, -0.40 V vs. Ca-Pb
Ca-5, -0.30 V vs. Ca-Pb
Fig. 8 Change of the cathodic current during the electrochemicalreduction tests in CaCl2
806 M. IIZUKA et al.
JOURNAL OF NUCLEAR SCIENCE AND TECHNOLOGY
tonium and americium in the MOX under the electrochemi-cal reduction is presumed.(1) Since the Li2O concentration in the vicinity of the cath-ode is increased during the electrochemical reduction, solu-ble mixed oxide Li2MO3 (M: plutonium or americium) isformed and dissolves into the electrolyte.
(2) In the bulk LiCl electrolyte, where the Li2O concentra-tion is lower than that around the cathode, the mixed oxidedissociates to form an MO2 precipitate.(3) When Li2O is added to adjust its concentration in theLiCl electrolyte, a part of the precipitate forms the mixed ox-ide and dissolves into the electrolyte again.
The concentration of the actinides gradually increased al-so in the CaCl2 electrolyte. The large value in the uraniumconcentration and the high background in the plutonium con-centration must be due to contamination of the samples. Theproportion of the amount of plutonium or americium foundin the CaCl2 electrolyte to the total amounts in the MOXsamples is much larger than that of uranium. Because allthe samples from the CaCl2 electrolyte were taken wheneach electrochemical reduction test was finished, the burstof the actinide concentration was not found as it was withthe LiCl electrolyte. After compensation for the backgroundfor the plutonium concentration, the ratios between theamount of actinide found in the CaCl2 electrolyte at theend of all the tests and that in the MOX samples beforeuse were 0.37% for plutonium and 5.2% for americium, re-spectively, which are a little higher than those evaluated withthe LiCl electrolyte.
5. Behavior of Actinide Elements in Electrochemical Re-duction—Chemical Composition of Reduction Prod-uctsThe compositions of some electrochemical reduction
products analyzed by ICP-MS are shown in Table 3. Whilethe Pu/(U+Pu) ratios in the products of Li-1 and Li-6 didnot essentially decrease during the tests, these ratios in theproducts of Ca-2, Ca-3, and Ca-5 decreased by approximate-ly 19% in average. More noticeable difference in the electro-chemical reduction behaviors between those with the LiCland CaCl2 electrolytes is the variation of the Am/Pu ratios.Although the Am/Pu ratios in the products from the LiClelectrolyte are not so much different from those in theMOX-a samples, those ratios in the products from the CaCl2electrolyte drastically decreased by about 81% in average.
The amounts of the plutonium and americium decreasedfrom the cathodes immersed in CaCl2 are evaluated to be ap-proximately 1:96� 10�2 g and 1:08� 10�3 g, respectively.They are much larger than those found in the CaCl2 electro-lyte which can be calculated to be about 5:58� 10�4 g and5:73� 10�5 g, respectively, from the change of the concen-
0.0
2.0
4.0
6.0
0 0.5 1 1.5 2 2.5
0.05.0
10.015.020.025.0
0 0.5 1 1.5 2 2.5
0.0
0.5
1.0
1.5
0 0.5 1 1.5 2 2.5
Cumulative amount of processed MOX (g)
Con
cent
ratio
n in
CaC
l2-C
aO (
ppm
)
Pu
Am
U
0.0
2.0
4.0
6.0
0 0.2 0.4 0.6 0.8 1 1.2
0.01.02.03.04.05.0
0 0.2 0.4 0.6 0.8 1 1.2
0.00.10.20.30.40.5
0 0.2 0.4 0.6 0.8 1 1.2
Con
cent
ratio
n in
LiC
l-L
i2O
(pp
m)
Cumulative amount of processed MOX (g)
Pu
Am
ULi2O addition
Fig. 9 Change of actinide concentration in LiCl and CaCl2 elec-trolytes during electrochemical reduction tests
Table 3 Composition of electrochemical reduction products analyzed by ICP-MS
Ratio (wt%)Pu/(U+Pu) in reduction product 10.9 9.32 7.48 7.54 7.91
Pu/(U+Pu) in initial MOX 9.45Am/Pu in reduction product 0.811 0.732 0.167 0.400 0.140
Am/Pu in initial MOX 0.756
Electrochemical Reduction of (U,Pu)O2 in Molten LiCl and CaCl2 Electrolytes 807
VOL. 44, NO. 5, MAY 2007
trations shown in Fig. 9. Such large differences show thatmost part of the plutonium and americium lost from the cath-ode were not accounted for. In the lithium reduction of theMOX pellets in LiCl, it is reported that plutonium and ameri-cium eluded the pellets and precipitated at the bottom of thecrucible and that the proportion of the americium lost fromthe pellets to the initial loading was much higher comparedwith the plutonium.8) In the present study, it is likely that theplutonium and americium once dissolved into the CaCl2electrolyte and precipitated again during the electrochemicalreduction in the same manner as observed in the lithium re-duction.
Since many parameters such as the total duration of theMOX immersion and the oxide concentration in the electro-lytes were changed together with the electrochemical condi-tions during the present reduction studies, the influence ofeach parameter on the dissolution behavior of the actinide el-ements can not be evaluated separately. Although the disso-lution of the plutonium and americium from the cathode isnot significant in the LiCl electrolyte, a considerable propor-tion of those elements eluded the reduction products in theCaCl2 electrolyte. Fundamental studies focused on the clar-ification on the mechanism of the actinide dissolution espe-cially into the CaCl2 electrolyte are needed to understand thequantitative influence of the dissolution on the process per-formance and to adopt a proper measure for a high actiniderecovery rate.
6. SEM/EDX Analysis of Reduction ProductsSmall samples taken from some of the reduction products
were polished and submitted to the observation by the opti-cal microscope, Scanning Electron Microscope (SEM) andthe Energy Dispersive X-ray (EDX) analysis.(1) Products of Electrochemical Reduction in LiCl
Figures 10(a) and 10(b) show respectively the metallo-graphs for the total section view and the SEM image ofthe surface region of the reduction product of the Li-5 testat a cathode potential of �0:50V. From these observations,it is clear that the sample is not reduced at all. This result iswell related to the consideration that most part of the catho-dic current is used for the deoxidization of tantalum at thiscathode potential.
Figure 11 is the total section view of the reduction prod-uct from the Li-2 test at a cathode potential of �0:65V. It isseen that the reduction proceeds from the surface of theMOX and that the thickness of the reduced area reaches25–30% of the whole sample. The SEM images of the borderpart between the reduced and non-reduced areas in the sam-ple are shown in Fig. 12(a). They indicate that the reductionstarts from the grain boundary of the MOX and that the re-duced area forms coral-like structure exactly in the samemanner as described in the previous report.7) However, thecomplete network of the reduced grain boundary may notbe necessary for a further progress of the reduction, sincethere are isolated reduced areas even at the center of theMOX as shown in Fig. 12(b). When the reduced areas alongthe surface and the internal reduced areas are added together,the above visual observation is quite consistent with the totalreduction ratio of 43–46%, which was evaluated by the mass
change of the samples and by the gas-burette method.Figure 13 shows SEM and back scattering electron (BSE)images of the central region of the same sample, where thereduced and non-reduced areas coexist and the coral-likestructure just began to be formed. The Pu/(U+Pu) ratiowas compared between the reduced areas (point-A in theSEM image, brighter part in the BSE image) and the non-re-duced areas (point-B in the SEM image, darker part in theBSE image) by the EDX point analysis. The result is shownin Table 4. There is no difference between the plutoniumcontent in the reduced and non-reduced areas. In the previ-ous report, it has been mentioned that some part of the plu-tonium in the reduction product obtained in the LiCl electro-lyte is excluded from the coral-like structure and makes Pu-rich spots outside this structure.7) In the present study, how-ever, neither a decrease nor an inhomogeneity of the pluto-nium content in the reduction product was observed. The dif-ference between the content of the plutonium in the reduc-
(b) SEM image of the surface region
(a) Metallograph for the total section view
1 mm
Fig. 10 Section view of the reduction product from Li-5
1 mm
Fig. 11 Total section view of the reduction product from Li-2
808 M. IIZUKA et al.
JOURNAL OF NUCLEAR SCIENCE AND TECHNOLOGY
tion product and in the initial MOX may be due to the insuf-ficient accuracy in the quantitative data by EDX, but it is notessential to the reduction behavior.(2) Products of Electrochemical Reduction in CaCl2
Figure 14(a) shows the metallograph for the total sectionview of the reduction product from the Ca-1 test where theMOX-a sample was reduced at �0:15V. It indicates thatthe reduction began from the surface and reached up to15–20% of the thickness. The metallograph for the close-up near the surface of the sample is shown in Fig. 14(b).In the completely reduced area (upper part in Fig. 14(b)),the connection among the metal grains progressed andformed the structure of larger dimensions than that in the in-ner area. It should be noted that the reduced and non-reducedareas are in close contact with each other and that no disso-ciation is found between these areas. This result indicatesthat the electrochemical reduction would be completed with-out any disturbance if the operation was continued. TheEDX point analysis was made at the two points shown inthe SEM image (Fig. 15) to determine the Pu/(U+Pu) ratioin the reduction product. The result is shown in Table 5. Inthe completely reduced area near the surface of the product(point-A in Fig. 15), the Pu/(U+Pu) ratio was a little lowerthan that in the inner region (point-B in Fig. 15), where thisratio is almost same as that in the initial MOX. Although thisdifference could be attributed to the error in the EDX analy-
sis, it is also likely that the dissolution of the transuraniumelements into the CaCl2-CaO electrolyte, as described above,affects the composition of the surface region of the reductionproduct. Figure 16 represents the metallograph for the cen-tral region of the same sample. A considerable proportion ofthe MOX has been reduced from the grain-boundary. Al-though this behavior looks to be similar to that previously re-ported for the electrochemical reduction in LiCl,7) the prog-ress of the reduction in the central region starts much earlierin this case even without an enough connection among thereduced areas. At present, it is difficult to distinguish wheth-er the reduction in the central region was caused by the elec-trochemical reduction or by the chemical reaction with cal-cium metal generated by the reduction of the electrolyte.
20 µm
A
B
20 µm
A
B
(a) SEM image
(b) BSE image
Fig. 13 SEM and BSE images of the central region of the reduc-tion product from Li-2
Table 4 Pu/(U+Pu) ratio in Li-2 reduction product determinedby EDX point analysis
(a) The border part between reduced and not reduced areas
Fig. 12 SEM images of the reduction product from Li-2
Electrochemical Reduction of (U,Pu)O2 in Molten LiCl and CaCl2 Electrolytes 809
VOL. 44, NO. 5, MAY 2007
The total section views of the reduction product from theCa-4 (a cathode potential of �0:25V) and Ca-1 tests ob-served by the optical microscope are shown in Fig. 17.The metallic shine of the Ca-4 product obviously tells thatwhole product was completely reduced. Metallograph forthe total section view is shown in Fig. 18. Since the reducedmaterial cohered to the surface of the product, its density at
the surface region has become quite high, while the consid-erable portion of the hollow space was made inside. The bentshape of the product was certainly formed by its shrinkagedue to the cohesion of the reduced material. Fragility ofthe reduction products recovered from the CaCl2 electrolyte,observed after the separation from the tantalum baskets,
100 µm
B
A
Fig. 15 SEM image of the reduction product from Ca-1
Table 5 Pu/(U+Pu) ratio in Ca-1 reduction product determinedby EDX point analysis
Fig. 14 Metallographs for the section view of the reduction prod-uct from Ca-1
100 µm
Fig. 16 Metallograph of the central region of the reduction prod-uct from Ca-1
Ca-1 Ca-4
Fig. 17 Total section view of the reduction product from Ca-4with optical microscope
1 mm
Fig. 18 Metallograph of the total section view of the reductionproduct from Ca-1 and Ca-4
810 M. IIZUKA et al.
JOURNAL OF NUCLEAR SCIENCE AND TECHNOLOGY
might be due to the large vacant space inside them as shownin Fig. 18.
Figure 19(a) shows the total section view of the reductionproduct from the Ca-6 test, at a cathode potential of�0:35V, observed by the optical microscope. Figure 19(b)shows the metallograph of the close-up view near the surfaceof the same product. It is distinctive that the surface region isselectively reduced and that the boundary between the re-duced/non-reduced areas is more discontinuous than thosein the other products. In the metallograph of the central re-gion (Fig. 20), reduction of the MOX has begun but the re-duction ratio appears much lower than that in the Ca-1 prod-uct (Fig. 16). This is considered to be caused by formation of
a surface barrier of the coherently reduced material that im-pedes the further progress of the reduction into an inner re-gion of the MOX. The total section view by SEM(Fig. 21(a)) shows that the density of the reduction productis higher at the surface and central regions than that at theintermediate region. A closer view of the near surface region(Fig. 21(b)) shows that the reduced material in the interme-diated part has become detached as parallel lamellar struc-tures. A decrease in the electrical conductivity due to thevoid in the lamellar structures, as well as the formation ofthe surface barrier would explain the interruption of the re-duction. The cause for such a variation in the morphologyof the reduction product should not be the difference in thecathode potential itself but the excessive deposition of thecalcium metal on the cathode which brings very high reduc-tion rate and consequently leads to the fast and localized out-ward shrinkage of the reduced material in the surface regionof the product.
Table 6 shows the results of the EDX point analysis of thereduction product from Ca-6. The analysis of the reducedpart was made in the dense surface region. It is found thatthe averaged Pu/(U+Pu) ratio in the reduced part is about14% lower than that in the not-reduced area. This behavior
500 µm
(b) Metallograph of the close-up view near the surface
(a) Total section view with optical microscope
Fig. 19 Section view of the reduction product from Ca-6
100 µm
Fig. 20 Metallograph of the central region of the reduction prod-uct from Ca-6
1 mm
100 µm
(b) Closer view of the near surface region
(a) Total section view
Fig. 21 SEM images of the reduction product from Ca-6
Electrochemical Reduction of (U,Pu)O2 in Molten LiCl and CaCl2 Electrolytes 811
VOL. 44, NO. 5, MAY 2007
shows a clear contrast with that observed in LiCl, where nochange in the plutonium content in the cathode was foundduring the electrochemical reduction (Table 4). On the otherhand, this EDX result is in accordance with the above-men-tioned decrease in plutonium in the reduction products andthe increase in plutonium concentration in the CaCl2 electro-lyte during the electrochemical reduction tests investigatedby ICP-MS.
7. Favorable Conditions for Electrochemical Reductionof MOXBased on the above experimental results, the advantage of
selection between LiCl and CaCl2 as the electrolyte for theelectrochemical reduction of the MOX is discussed in thissection.(1) Reduction Potential, Reduction Rate and Cathodic Cur-
rent EfficiencyFrom the results obtained by the CV measurement and the
electrochemical reduction tests at a constant cathode poten-tial followed by the SEM/EDX analysis, it was found thatthe MOX was reduced at �0:65V vs. Bi-35mol% Li refer-ence electrode in LiCl. Under this condition, the reduced ra-tio evaluated both from the mass change between the reduc-tion and by the gas-burette method was around 50%. Thecathodic current efficiency calculated from the reduced ratiowas around 40%.
In CaCl2, on the other hand, the MOX was reduced in thewhole range of the tested cathode potential (�0:15V to�0:40V), although the reduction was interrupted by forma-tion of the surface barrier made of the reduced material at thelower cathode potential (�0:30V). At �0:25V, the MOXwas completely reduced although the product was brittleprobably due to the internal porosity formed by the shrink-age of the reduced material. In the tests where the unbrokencathode products were recovered, the cathodic current effi-ciency was evaluated 40–50%.
From the viewpoint of the cathode potential range for theelectrochemical reduction, the preferable condition is limitedsimilarly either in LiCl or CaCl2. Although the reduction rateseems a little higher in CaCl2 according to the change of thereduction current during the electrochemical reduction tests,the advantage is not yet quantitatively clear. There is no ob-vious difference in the cathodic current efficiency betweenthe both electrolytes.(2) Dissolution of Actinide Elements into Electrolyte
Dissolution of the plutonium and americium into the elec-trolyte was found both in the LiCl and CaCl2 electrolytes.The concentrations of these actinides at the end of the testswere higher in the CaCl2 electrolyte than those found in theLiCl electrolyte. The ICP-MS and EDX analyses for the re-
duction products also showed that the amounts of these ele-ments lost from the cathodes were much more than those in-creased in CaCl2, indicating the formation of the precipitatescontaining these actinides during the tests. From the view-point of process design for a high recovery ratio, LiCl is ob-viously a more favorable electrolyte.(3) Operating Temperature
Due to the high melting point of CaCl2 (1,045K), theelectrochemical reduction in the CaCl2 electrolyte shouldbe operated at around 1,123K, which is much higher than923K required for LiCl. This is a disadvantageous point interms of the material selection and the cost to design, fabri-cate and maintain the process equipment. Another problemarising from the high operation temperatures is the formationof the surface barrier formed on the reduction product due tothe progress of the connection among the reduced metalgrains.(4) Compatibility with the Pyrometallurgical Processes
Since the alkali and alkaline earth fission products are ex-pected to dissolve into the electrolyte during the electro-chemical reduction, the electrolyte should be consolidatedas a high level waste form. Development of the waste formfor water-soluble chlorides is the serious problem for the py-rometallurgical processes. Although the operation tempera-tures of the electrochemical reduction in the LiCl electrolyteare higher than those in the pyrometallurgical process for themetallic fuel (773K), the similar waste treatment processwith zeolite would be applicable. However, a completely dif-ferent waste treatment method should be developed for theCaCl2 electrolyte because of the much higher operating tem-peratures.
Taking the above discussions from the four viewpoints in-to consideration, LiCl is concluded to be advantageous as theelectrolyte used in the electrochemical reduction of theMOX at present.
IV. Conclusions
The electrochemical reduction of the UO2-PuO2 mixedoxides (MOX) was investigated in molten LiCl at 923Kand CaCl2 at 1,123K to evaluate the behavior of the pluto-nium quantitatively and to define the favorable conditionsfor the electrochemical reduction of those materials.
In LiCl, it was found that the MOX was reduced at�0:65V vs. Bi-35mol% Li reference electrode by the CVmeasurement and the electrochemical reduction tests at aconstant cathode potential followed by the SEM/EDX anal-ysis. Under this condition, the reduced ratio evaluated bothfrom the mass change between the reduction and by thegas-burette method was around 50%. The cathodic current
Table 6 Pu/(U+Pu) ratio in Ca-6 reduction product determined by EDX point analysis
efficiency calculated from the reduced ratio was around40%.
In CaCl2, on the other hand, the MOX was reduced in thewhole range of the tested cathode potential (�0:15V to�0:40V), although the reduction was interrupted by forma-tion of the surface barrier made of the reduced material at thelower cathode potential (�0:30V). At �0:25V, the MOXwas completely reduced although the product was brittleprobably due to the internal porosity formed by the shrink-age of the reduced material. In the tests where the unbrokencathode products were recovered, the cathodic current effi-ciency was evaluated as 40–50%.
Dissolution of the plutonium and americium into the elec-trolyte was found both in the LiCl and in CaCl2 electrolytes.The concentrations of these actinides at the end of the testswere higher in the CaCl2 electrolyte than those found in theLiCl electrolyte.
Taking the above experimental results and the compatibil-ity with the pyrometallugical processes into consideration,LiCl is concluded to be more advantageous as the electrolyteused in the electrochemical reduction of the MOX at present.
Acknowledgement
The authors gratefully appreciate Dr. J. Somers of Insti-tute of Transuranium elements (ITU) for his preparation ofthe MOX materials, Drs. T. Wiss, D. Cromboom, and M.Murray-Farthing of ITU for their analytical works, and Mr.
A. Rodorigues of ITU for his kind experimental support.Thanks are also due to Messrs. Y. Sakamura and K. Uozumiof CRIEPI for their helpful suggestions and comments.
References
1) T. Inoue, T. Yokoo, T. Nishimura, ‘‘Assessment of advancedtechnologies for the future nuclear fuel cycle,’’ Proc. Int. Conf.on Future Nuclear Systems (Global’99), Jackson Hole, WY,Aug. 29–Spt. 3, 1999 (1999).
2) T. Usami, M. Kurata, T. Inoue et al., ‘‘Pyrochemical reductionof uranium dioxide and plutonium dioxide by lithium metal,’’J. Nucl. Mater., 300, 15 (2002).
3) T. Kato, T. Usami, M. Kurata et al., Development of ReductionTechnology for Oxide Fuels—Lithium Reduction of a SimulatedSpent Oxide Fuel—, CRIEPI Report T01003 (2001) [in Japa-nese].
4) T. Usami, T. Kato, M. Kurata et al., ‘‘Lithium reduction ofamericium dioxide to generate americium metal,’’ J. Nucl.Mater., 304, 50 (2002).
5) A. Kawabe, T. Kato, M. Kurata, Development of ReductionTechnology for Oxide Fuels—Electrolysis of Li2O in MoltenLiCl—, CRIEPI Report T02041 (2003) [in Japanese].
6) Y. Sakamura, M. Kurata, T. Inoue, ‘‘Electrochemical reductionof UO2 in molten CaCl2 or LiCl,’’ J. Electrochem. Soc., 153,D31 (2006).
7) M. Kurata, T. Inoue, J. Serp et al., ‘‘Electro-chemical reductionof MOX in LiCl,’’ J. Nucl. Mater., 328, 97 (2004).
8) T. Usami, T. Kato, M. Kurata et al., ‘‘Lithium reduction of aMOX pellet,’’ J. Nucl. Sci. Tech., Supple. 3, 858 (2002).
Electrochemical Reduction of (U,Pu)O2 in Molten LiCl and CaCl2 Electrolytes 813
