CONVERSION OF OXIDE INTO METAL OR CHLORIDE
FOR THE PYROMETALLURGICAL PARTITIONING PROCESS
Yoshiharu Sakamura, Masaki Kurata, Tsuyoshi Usami and Tadashi Inoue
Central Research Institute of Electric Power Industry (CRIEPI),
Iwadokita 2-11-1, Komae-shi, Tokyo 201-8511, Japan
Abstract
Oxide fuel is not directly treated in the pyrometallurgical partitioning process that consists of
electrorefining and reductive extraction in LiCl-KCl eutectic system. Therefore, some reduction and
chlorination techniques for removing oxygen have been developed by CRIEPI. Typical experimental
results and characteristics of the pretreatment techniques are summarized in this paper.
Introduction
In the pyrometallurgical partitioning process, uranium, plutonium and minor actinide are
dissolved into a LiCl-KCl eutectic salt and then are selectively collected by the electrorefining and/or
the reductive extraction using a liquid metal [1-5]. It is easy for metal and nitride fuels to be anodically
dissolved into the molten salt [6-8]. However, oxide fuel as well as the oxide obtained by denitration
of HLLW is hardly dissolved. Actinide cation in the salt (e.g. U3+, Pu3+, etc.) immediately reacts with
oxide ion to give a precipitate of oxide or oxychloride.
The products of the electrorefining and reductive extraction steps are distilled at a high
temperature to remove the adhering salt and cadmium from the actinide metal, when a carbon crucible
coated with ZrO2 or Y2O3 is used for containing the corrosive materials [9]. In the distillation step, a
part of actinide metal may react with crucible coating materials or oxygen impurities to give some
dross consisting of actinide oxide and oxychloride. The actinide in the dross has to be recycled to
attain a high actinide recovery ratio.
Since the pyrometallurgical partitioning process cannot directly accept oxides or oxychlorides, a
pretreatment step is necessary for removing oxygen. Therefore, CRIEPI has developed both reduction
techniques to convert oxide into metal and chlorination techniques to convert oxide into chloride.
They are listed as follows:
- Lithium reduction
- Electrochemical reduction
- Chlorination using chlorine gas.
- Chlorination using ZrCl4.
In this paper, typical experimental results and characteristics of the pretreatment techniques are
summarized.
Reduction technique
Lithium reduction
In the lithium reduction process, actinide oxides are converted into metals by adding lithium
metal reductant in a LiCl salt bath at 650 ˚C. The reaction is expressed as
MO2 + 4 Li (LiCl) → M + 2 Li2O (LiCl) ,
(1)
where MO2 is an actinide oxide and (LiCl) denotes the molten LiCl phase. It was experimentally
verified that UO2, NpO2, and PuO2 could be reduced into metal [10]. As for AmO2, the Li2O
concentration in the salt had to be less than 5.1 wt% for the complete reduction into metal [11]. Figure
1 shows the results of a lithuim reduction test for a simulated spent MOX pellet containing U, Pu, Np,
Am, Cm, Ce, Nd, Sm, Ba, Pd, Mo and Zr [12]. The cross section of the pellet after the reduction was
metallic, while oxide solid solutions of rare earths were observed microscopically. Rare earth oxides
could not be reduced to metals in this system. Moreover, a part of them were detected in the molten
LiCl phase and on the bottom of the crucible. The rare earths dissolved into the salt easier as the Li2O
concentration increased.
Ln2O3 + O2- (LiCl) → 2 LnO2- (LiCl)
(2)
U-Pu alloy
Rare-earth
BSE image
(a) MOX before reduction
(b) MOX after reduction
X-ray mapping for Ce
(c) Cross section of MOX
after reduction
Fig. 1
(d) Oxide solid solution
of rare earth in the
reduced MOX
The simulated spent MOX pellet reduced by adding lithium metal in a LiCl salt bath
at 650 ˚C. The MOX pellet was sintered from oxides of U, Pu, Np, Am, Cm, Ce, Nd,
Sm, Ba, Pd, Mo and Zr.
where Ln2O3 denotes rare earth oxide. Rare earths were supposed to dissolve into the salt from the
pellet where the Li2O concentration might be high, and then to precipitate in the bulk salt. Small
amounts of Pu, Am and Cm were also detected on the bottom of the crucible. These actinides might be
soluble in the salt but the solubility limit seemed much smaller than that of rare earths. Almost all of U
and Np remained in the reduced pellet. Barium was dissolved in the salt almost completely.
As oxide fuels are processed, Li2O accumulates in the salt bath. So, electrowinning for
decomposing the Li2O must be performed to keep the Li2O concentration lower and to recycle the
lithium metal. The reduction step and the electrowinning step are repeated one after the other.
Anode:
2 O2- (LiCl) → O2 + 4 e- ,
Cathode: Li+ (LiCl) + e- → Li .
(3)
(4)
It was experimentally demonstrated that the Li2O concentration was decreased from 3.0 wt% to 0.2
wt% with a current efficiency higher than 90% [13]. Since the solubility of lithium metal in LiCl is no
more than 0.6 at% at 650 ˚C, almost all of the lithium metal was collected at the iron cathode
surrounded by a MgO shroud.
Electrochemical reduction
Recently, an electrochemical reduction technique has been developed to make the reduction
process more efficient. By electrolysis with a cathode where actinide oxide is loaded, the oxide ion is
released into the salt and the actinide metal remains at the cathode. The oxide ion is transported to the
anode and oxygen gas is evolved at the anode. When a carbon anode is employed, CO2 or CO is
evolved.
Cathode: MO2 + 4 e- → M + 2 O2Anode:
2 O2- → O2 + 4 e- ,
2 O2- + C → CO2 + 4 e-, O2- + C → CO + 2 e- .
(5)
(6)
(7)
Advantages of the electrochemical reduction technique are as follows:
- The reduction and electrowinning steps are performed simultaneously.
- The concentration of oxide ion in the salt is almost constant and can be maintained at a low
value.
- The amount of the salt bath may be small because oxide ion do not accumulate.
An electrochemical reduction technique has been developed for an economical process to
produce titanium metal [14]. A CaCl2 salt bath having low oxygen potential is employed for the TiO2
reduction because the affinity of oxygen to titanium is very large. If the CaCl2 salt bath is used for the
oxide nuclear fuel, not only actinide oxides but also rare earth oxides will be reduced to metals.
However, the operating temperature for the CaCl2 system (melting point of CaCl2: 772 ˚C) is higher
than that for the LiCl system, which may make the design of a facility more challenging.
CRIEPI is now investigating the electrochemical reduction of UO2 and mixed oxides in both of
LiCl and CaCl2 salt baths. Figure 2 shows typical uranium metal products obtained in UO2 reduction
tests using about 0.2 g of sliced UO2 pellets. The sample in the LiCl salt bath shrank during the
reduction and the uranium metal having uniform porosity was obtained as shown in Fig. 2(a). In the
LiCl salt bath at 650 ˚C, the reduction progressed quite satisfactorily.
In case of the CaCl2 salt bath at 820 ˚C, the uranium metal product had a big cave in the center as
shown in Fig. 2(b). The reduced uranium metal condensed at the surface of the UO2 piece, which
might be due to the high operating temperature. The melting point of uranium metal is 1132 ˚C. The
dense uranium metal membrane formed at the surface often prevented the reduction from proceeding.
However, the dense uranium metal membrane was not observed in the reduction tests for MOX and
UO2-Gd2O3.
A large scale test was carried out in a LiCl salt bath. At the cathode, 100 g of UO2 (0.3-1.0 mm
grain) was loaded in a stainless steel basket. Figure 3 shows the uranium metal product. It was
demonstrated that UO2 was successfully reduced to uranium metal.
MOX fuel reduction tests were carried out and in result U-Pu alloys were obtained [15]. During
the reduction, uranium was reduced prior to plutonium and Pu-spots were observed in the product. It
was suggested that the reduction rate of MOX was higher than that of pure UO2.
1mm
(a) A product in a LiCl salt bath at 650℃
1mm
(b) A product in a CaCl2 salt bath at 820℃
Fig. 2 The cross section of sliced UO2 pellet (83%
of theoretical density) after reduction in LiCl
at 650 ˚C (a) and in CaCl2 at 820 ˚C (b).
10mm
Fig. 3 The product of 100g-UO2 reduction
test in LiCl at 650 ˚C.
Chlorination technique
Chlorination using chlorine gas
A chlorination process using chlorine gas and carbon has been developed. When uranium,
plutonium and rare earth oxides are converted to chloride, the following chemical reactions occur:
UO2 + C + 2 Cl2 → UCl4 + CO2 ,
PuO2 + C + 2 Cl2 → PuCl4 + CO2 ,
2 Ln2O3 + 3 C + 3 Cl2 → 2 LnCl3 + 3 CO2 .
(8)
(9)
(10)
Two chlorination experiments were performed for the oxide obtained by the denitration of a
simulated high level liquid waste [16]. The oxide was loaded in a LiCl-KCl eutectic salt bath
contained in a pyro-coating graphite crucible. Then chlorine gas was introduced into the salt through a
graphite tube at 700 ˚C. The results were shown in Fig. 4. All of elements containing uranium, rare
earth, noble metal, alkali, alkaline earth, etc. could be converted to chloride, when some transition
elements such as molybdenum, iron and zirconium volatilized with ease. It was suggested that the
volatilized chloride should be collected because 5 % of the detected uranium was volatilized during
the chlorination. The salt product was supplied for the reductive extraction test and consequently the
uranium was collected in the liquid cadmium phase.
100
100
80
80
Mass balance (%)
Mass balance (%)
recovered ratio in molten salt bath
volatilized ratio
60
40
40
20
20
0
60
Na Ba Y Ce Nd Sm Ru Rh Pd Ni Cr Cs Zr Fe Mo Re Se Te
0
U
Na
Cs
Sr
Ce
(a) Run 1
Fig. 4
Nd
Ru
Pd
Cr
Zr
Fe
Mo
Re
(b) Run 2
Mass balance in chlorination tests using the oxide obtained by the denitration of
a simulated high level liquid waste.
Chlorination using ZrCl4
ZrCl4 has a high reactivity with oxygen, but is not corrosive to refractory metals such as steel.
The actinide and rare earth oxides were allowed to react with ZrCl4 in a LiCl-KCl eutectic salt at 500
˚C to give metal chlorides dissolving in the salt and a precipitate of ZrO2 [17]. When rare earth oxides
are converted to chloride, the following chemical reactions occur:
2 Ln2O3 + 3 ZrCl4 (LiCl-KCl) → 4 LnCl3 (LiCl-KCl) + 3 ZrO2
(11)
where (LiCl-KCl) denotes the molten LiCl-KCl phase. For chlorinating actinide dioxides, an addition
of zirconium metal was efficient as a reductant, because trivalent actinide ion is stable in the salt.
UO2 + 3/4 ZrCl4 (LiCl-KCl) + 1/4 Zr → UCl3 (LiCl-KCl) + ZrO2 ,
PuO2 + 3/4 ZrCl4 (LiCl-KCl) + 1/4 Zr → PuCl3 (LiCl-KCl) + ZrO2 .
(12)
(13)
Divalent zirconium in the salt phase, which is formed by the disproportionation of ZrCl4 and
zirconium metal, might accelerate the chlorination [18].
Figures 5 and 6 indicate that UO2 and PuO2 were completely chlorinated and dissolved into the
salt by adding an excess of ZrCl4. Thermodynamic considerations indicate that the minor actinide
oxides will be also chlorinated. When the oxides were in powder form, the reaction was observed to
progress rapidly. By keeping the system quite still, the solution settled so that the ZrO2 precipitate
could be separated.
Advantages of the ZrCl4 chlorination technique are as follows:
- Very simple.
- The reaction rate is sufficient.
- ZrCl4 is not corrosive to refractory metals such as steel.
- Flexibility for scale, molten salt composition and temperature.
- Actinide metals as well as oxides are chlorinated.
- The by-product of ZrO2 is quite stable.
0.10
0.10
Pu and Zr in the salt ( g )
U and Zr in the salt ( g )
0.15
U
0.05
Zr
0.00
0.0
0.1
0.2
0.3
ZrCl4 added ( g )
0.4
Fig. 5 Results of a UO2 chlorination test. The
amount of U and Zr dissolved in the salt
as a function of the amount of ZrCl4 added.
The solid lines are the calculated values
based on the mass balance.
Pu
0.08
0.06
0.04
0.02
0.00
0
Zr
5
10
15
Time / hr
20
25
Fig. 6 Results of a PuO2 chlorination test. The
amount of Pu and Zr dissolved in the salt
after the salt melted at 500℃. The dotted
lines are the calculated values based on
the mass balance.
Conclusions
Oxide fuel and dross are not directly treated in the pyrometallurgical partitioning and
reprocessing process which consist of electrorefining and reductive extraction in LiCl-KCl eutectic
system. Therefore, some reduction and chlorination techniques have been developed by CRIEPI. Their
characteristics are summarized in Table 1. Since each technique has its own advantages, the suitable
pretreatment techniques will be selected as the needs of the case demand.
Acknowledgments
A part of this work was supported by MEXT (Ministry of Education, Culture, Sports, Science,
and Technology of Japan), Development of Innovative Nuclear Technologies Program.
Table 1
Summary of the pretreatment techniques for applying pyrometallurgical partitioning
to oxide fuel
Molten salt bath
Temperature
Active material
Product
By-product
FP chemical form
・Noble metal
・Rare earth
・Alkali, alkaline earth
Reduction technique
Li reduction
Electrochemical reduction
LiCl
LiCl
CaCl2
650℃
650℃
∼800℃
Li metal
None
None
U, Pu metal
U, Pu metal
U, Pu metal
Chlorination technique
Cl2 method
ZrCl4 method
LiCl-KCl etc.
LiCl-KCl etc.
∼700℃
∼500℃
Cl2, C
ZrCl4, Zr
UCl3, PuCl3
UCl4, PuCl3
Li2O
O2,CO2
O2,CO2,CO
CO2
ZrO2
（○：Separated from actinides. ×：Accompany actinides ）
× Matal
× Matal
× Matal
× Chloride
○ Metal
× Oxide or oxychloride
× Oxide or oxychloride
× Matal
× Chloride
× Chloride
○ Chloride
○ Chloride
○ Chloride
× Chloride
× Chloride
References
[1] T. Inoue, M. Sakata, H. Miyashiro, M. Matsumura, A. Sasahara, N. Yoshiki, Nucl. Technol., 93,
206 (1991).
[2] Y. Sakamura, T. Hijikata, K. Kinoshita, T. Inoue, T.S. Storvick, C.L. Krueger, L.F. Grantham, S.P.
Fusselman, D.L. Grimmett and J.J. Roy, J. Nucl. Sci. Technol., 35[1], 49 (1998).
[3] K. Kinoshita, T. Inoue, S.P. Fusselman, D.L. Grimmett, J.J. Roy, R.L. Gay, C.L. Krueger, C.R.
Nabelek and T.S. Storvick, J. Nucl. Sci. Technol., 36[2], 189 (1999).
[4] K. Kinoshita, M. Kurata and T. Inoue, J. Nucl. Sci. Technol., 37[1], 75 (2000).
[5] K. Kinoshita, T. Inoue, S.P. Fusselman, D.L. Grimmett, C.L. Krueger, C.R. Nabelek and T.S.
Storvick, J. Nucl. Sci. Technol., 40[7], 524 (2003).
[6] R.W.Benedict, J.R.Krsul, R.D.Mariani, K. Park and G.M.Teske, in Proceedings of Future Nuclear
Systems: Emerging Fuel Cycles and Waste Disposal Options, Global'93, Vol. 2, Seattle, Washington,
Sept. 12-17,1993, p.1331.
[7] K. Kinoshita, T. Koyama, T. Inoue, M. Ougier and J.-P. Glatz, J. Physics and Chemistry of Solid,
to be submitted.
[8] O. Shirai, T. Iwai, K. Shiozawa, Y. Suzuki, Y. Sakamura and T. Inoue, J. Nucl. Mater., 277, 226
(2000).
[9] B.R. Westphal, D. Vaden, T.Q. Hua, J.L. Willet and D.V. Laug, in Proceedings of 5th Topical
Meeting, DOE Owned Spent Nuclear Fuel and Fissile Materials Management, Charleston, SC, Sept.
17-20, 2002.
[10] T. Usami, M. Kurata, T. Inoue, H.E. Sims, S.A. Beetham and J.A. Jenkins, J. Nucl. Mater. 300,
15 (2002).
[11] T. Usami, T. Kato, M. Kurata, T. Inoue, H.E. Sims, S.A. Beetham and J.A. Jenkins, J. Nucl.
Mater. 304, 50 (2002).
[12] T. Kato, T. Usami, M. Kurata, T. Inoue, H.E. Sims and J.A. Jenkins, CRIEPI Report T01003
(2001) in Japanese.
[13] A. Kawabe, T. Kato and M. Kurata, CRIEPI Report T02041 (2003) in Japanese.
[14] G.Z. Chen, D.J. Frey and T.W. Farthing, Nature, 407, 361 (2000).
[15] M. Kurata, T. Inoue, J. Serp, M. Ougier and J-P. Glatz, J. Nucl. Mater. 328, 97 (2004).
[16] M. Kurata, K. Kinoshita, T. Hijikata and T. Inoue, J. Nucl. Sci. Technol., 37[8], 682 (2000).
[17] Y. Sakamura, T. Inoue, T. Iwai and H. Moriyama, submitted for publication in J. Nucl. Mater..
[18] Y. Sakamura, J. Electrochem. Soc., 151 [3], C187 (2004).