Zirconium Behavior in Molten LiCl-KCL Eutectic

Article abstract of DOI:10.1149/1.1644605

Some oxidation states (0, +1, +2, and +4) of zirconium exist in a LiCl-KCl eutectic system over the temperature range 450-550°C, and the behavior is complicated. In cyclic voltammograms at 500°C, a cathodic peak was observed at about -1.2 V vs. Ag/AgCl reference electrode, which might be due to the reduction of Zr(IV) to ZrCl and zirconium metal. Two anodic peaks might correspond to the oxidation of ZrCl and zirconium metal, respectively. The electrolysis at a cathode potential of about -1.1 V yielded a nodular deposit identified as ZrCl, which appeared to be a metastable compound in this system. When the potential was sufficiently negative (i.e., <-1.35 V), zirconium metal was obtained. The deposited zirconium metal was fine black powder, and adhesion to the cathode wire was poor. In the presence of cadmium metal at the cathode, an intermetallic compound that might be Cd₃Zr was obtained. The collection efficiency of zirconium is improved using cadmium because the adhesion of the intermetallic compound was much better. Zirconium metal reacted with Zr(IV) to give Zr(II) whose solution was light brown, and Zr(II) was easily disproportionated into Zr(IV) and zirconium metal. The anodic dissolution test indicated that the zirconium metal primarily dissolved into the electrolyte salt as Zr(IV). The Zr(II)/Zr(IV) ratio seemed to be very low and to increase with increasing temperature.

Full text of DOI:10.1149/1.1644605

Journal of The Electrochemical Society, 151 ͑3͒ C187-C193 ͑2004͒
C187
0
013-4651/2004/151͑3͒/C187/7/$7.00 © The Electrochemical Society, Inc.
Zirconium Behavior in Molten LiCl-KCl Eutectic
Yoshiharu Sakamuraz
Central Research Institute of Electric Power Industry (CRIEPI), Komae-shi, Tokyo 201-8511, Japan
Some oxidation states ͑0, ϩ1, ϩ2, and ϩ4͒ of zirconium exist in a LiCl-KCl eutectic system over the temperature range
450-550°C, and the behavior is complicated. In cyclic voltammograms at 500°C, a cathodic peak was observed at about Ϫ1.2 V
vs. Ag/AgCl reference electrode, which might be due to the reduction of Zr͑IV͒ to ZrCl and zirconium metal. Two anodic peaks
might correspond to the oxidation of ZrCl and zirconium metal, respectively. The electrolysis at a cathode potential of about Ϫ1.1
V yielded a nodular deposit identified as ZrCl, which appeared to be a metastable compound in this system. When the potential
was sufficiently negative ͑i.e., ϽϪ1.35 V͒, zirconium metal was obtained. The deposited zirconium metal was fine black powder,
and adhesion to the cathode wire was poor. In the presence of cadmium metal at the cathode, an intermetallic compound that might
be Cd Zr was obtained. The collection efficiency of zirconium is improved using cadmium because the adhesion of the interme-
3
tallic compound was much better. Zirconium metal reacted with Zr͑IV͒ to give Zr͑II͒ whose solution was light brown, and Zr͑II͒
was easily disproportionated into Zr͑IV͒ and zirconium metal. The anodic dissolution test indicated that the zirconium metal
primarily dissolved into the electrolyte salt as Zr͑IV͒. The Zr͑II͒/Zr͑IV͒ ratio seemed to be very low and to increase with
increasing temperature.
©
2004 The Electrochemical Society. ͓DOI: 10.1149/1.1644605͔ All rights reserved.
Manuscript submitted April 21, 2003; revised manuscript received September 24, 2003. Available electronically February 5, 2004.
The metal fuel cycle, consisting of a metal fuel fast breeder
reactor, pyrochemical reprocessing, and fuel fabrication by injection
casting, is a promising option that satisfies the requirements of eco-
nomic advantage, environmental safety, and high proliferation
resistance.1 An electrorefining process in molten LiCl-KCl eutectic
has been developed for the pyrochemical reprocessing. A sche-
matic of the electrorefining process is shown in Fig. 1. Spent metal
fuels ͑i.e., irradiated U-Zr or U-Pu-Zr͒ contain fission products and
transuranium elements. The actinides in the spent fuels are anodi-
cally dissolved into the salt. Highly pure uranium is recovered onto
the solid cathode, and a mixture of plutonium, uranium, and the
other actinides are recovered into the liquid cadmium cathode. Ac-
tive fission products such as cesium, strontium, and rare earths ac-
cumulate in the electrolyte salt, while noble fission products such as
molybdenum, palladium, and ruthenium do not dissolve and remain
at the anode.
500°C. Iizuka reported that the electrodeposition in a LiCl-KCl eu-
tectic containing zirconium yielded no zirconium metal but did yield
11
ZrCl at the cathode. Cyclic voltammograms ͑CVs͒ were measured
11
12
by Iizuka and Basile et al., but they did not explain the forma-
tion of ZrCl. They insisted that Zr͑IV͒ was first reduced to Zr͑II͒
followed by the reduction of Zr͑II͒ to zirconium metal and that most
of the Zr͑IV͒ was reduced to Zr͑II͒ by adding zirconium metal in the
salt at around 550°C. In other alkali metal chloride and fluoride
systems in the temperature range above 650°C, Kipouros et al., in-
vestigated ϩ1, ϩ2, ϩ3, and ϩ4 oxidation states of zirconium.1
Struss et al. reported that lower chlorides ͑ZrCl, ZrCl , and ZrCl )
,2
3
-5
3,14
2
3
were given by the reaction of ZrCl with zirconium metal in a sealed
4
1
5
tantalum tube.
According to the literature, zirconium has various lower oxida-
tion states in addition to the tetravalent state. Stable oxidation spe-
cies may strongly depend on temperature and the constituents of the
solvent. Even in the LiCl-KCl eutectic, one of the most popular
electrolytes, zirconium chemistry has not yet been well clarified.
Hence, in the present work, zirconium characteristics such as pre-
dominant species in the salt, anode, and cathode reactions and mor-
phology of cathode deposits were investigated over the temperature
range 450-550°C using CV, anode dissolution, and cathode deposi-
tion.
Zirconium accounting for 10 wt % of the metal fuel is expected
to behave in a complicated manner because zirconium is the next
easiest element to oxidize after uranium. The difference in the stan-
dard potential between zirconium and uranium is no more than 0.4
6
V, and uranium forms intermetallic compounds with zirconium and
noble metal fission products. Therefore, a high uranium recovery
ratio from the spent fuel is not attained without zirconium dissolu-
tion into the salt. Actually, it was reported that a portion of the
zirconium dissolved into the electrolyte during the electrorefining
for irradiated U-Zr fuels and that the uranium deposited on the solid
cathode contained some zirconium.7 Retention of zirconium at the
anode depends on the electrorefining conditions which are the anode
potential and the retention of uranium. Most of the dissolved zirco-
nium is believed to be reduced by uranium metal falling from the
solid cathode and remaining in the cadmium pool placed at the
bottom of the electrorefiner vessel. The accumulated zirconium has
to be recovered. Therefore, understanding the zirconium behavior in
a molten LiCl-KCl eutectic is important for the electrorefining pro-
cess.
Experimental
,8
Materials and equipment.—High-purity LiCl-KCl eutectic
͑59:41 mol %͒, LiCl-KCl eutectic containing 1 wt % AgCl, and
ZrCl with a purity of 99.9% were obtained from Anderson Physics
4
Laboratory. To avoid sublimation of ZrCl , when ZrCl4 is added
4
into molten salts for adjusting the Zr͑IV͒ concentration, a
LiCl-KCl-ZrCl mixture ͑29.4 wt % ZrCl ) was prepared by heating
4
4
a sealed quartz tube containing LiCl-KCl eutectic and ZrCl4 to
50°C. Once ZrCl dissolves into a molten LiCl-KCl eutectic, little
5
4
ZrCl4 escapes as vapor from the salt. Cadmium metal ͑99.999%
purity͒, zirconium wire and rods ͑99.5% purity͒, molybdenum wire
Baboian et al. measured the standard potential of zirconium in a
9
molten LiCl-KCl eutectic. They concluded that zirconium existed
͑
͑
99.95% purity͒, tantalum wire ͑99.95% purity͒, and silver wire
99.99% purity͒ used for the measurements were supplied from Rare
in the salt as Zr͑IV͒ at 450°C and predominantly as Zr͑II͒ at 550°C,
considering the slope of the lines for potential vs. log ͑zirconium
concentration͒ and the weight loss of the zirconium rods used for the
anodic dissolution. Suzuki also measured the zirconium potential in
Metallic Co., Ltd.
All the experiments using molten salts were conducted in a high-
purity argon atmosphere glove box in which the concentrations of
oxygen and moisture were controlled to be less than 2 ppm. The cell
was located in a stainless steel thermowell attached to the floor of
the glove box, which was heated externally with an electric furnace.
The potential data were acquired by a digital electrometer
TR8652 of Advantest. CV and electrolysis were performed using a
model 273A potentiostat/galvanostat of EG&G Princeton Applied
1
0
a molten LiCl-KCl eutectic over the temperature range 450-675°C.
He drew conclusions similar to those indicated by Baboian; how-
ever, they did not state which species is predominant in the salt at
z
E-mail: sakamura@criepi.denken.or.jp
Downloaded on 2015-03-08 to IP 128.226.37.5 address. Redistribution subject to ECS terms of use (see ecsdl.org/site/terms_use) unless CC License in place (see abstract).

C188
Journal of The Electrochemical Society, 151 ͑3͒ C187-C193 ͑2004͒
Figure 2. Electrochemical cell for CV.
Figure 1. Schematic of the electrorefining process for the metal fuel repro-
cessing.
agreement with those measured before.16,17 The concentration of
ZrCl4 was incrementally changed by adding the prepared
Research and an EG&G M270 computer software.
LiCl-KCl-ZrCl salt. Salt samples were taken by inserting a Pyrex
4
Zirconium concentration in samples was determined by an induc-
tively coupled plasma atomic emission spectrometer ͑ICP-AES͒,
IRIS Advantage of Jarrell Ash, and potassium concentration by an
atomic absorption flame emission spectrophotometer, AA-680 of
Shimazu. Cathode deposits and precipitates were identified by X-ray
diffraction ͑XRD͒ using a RINT 2500V of Rigaku.
glass tube into the melt and were dissolved in dilute nitric acid for
the ICP-AES analysis. The temperature was varied over the range
450-550°C.
Cathode deposition.—An electrodeposition test was conducted
to identify the zirconium deposits as a function of the cathode po-
tential. A tantalum wire, 1 mm diam, was used as the cathode. Its
lower end was bent like a fishhook to prevent the deposits from
falling. A cathode having a small alumina crucible ͓outside diameter
͑o.d.͒5.2 mm͔ placed under a tantalum wire was also employed to
recover any deposit falling from the wire. The cadmium pool con-
taining zirconium metal, as shown in Fig. 2, was used as the counter
electrode. The experimental conditions are presented in Table I.
Tests C1 and C2 were carried out at low current. The cathode po-
tential was more negative than Ϫ1.1 V at the beginning of the elec-
trolysis and then shifted to Ϫ1.04 V. In tests C3 and C4, the current
was incrementally varied during the electrolysis in order to keep the
cathode potential in the range from Ϫ1.35 to Ϫ1.40 V. Because the
deposit obtained under this condition fell from the cathode wire very
easily, a catch crucible was employed. A liquid cadmium cathode
was used for test C5 in order to improve the collection efficiency of
the zirconium deposit. There was a small amount of cadmium metal
͑0.34 g͒ in the catch crucible. The collected products were identified
by XRD analysis. A significant amount of the salt electrolyte always
Cyclic voltammetry.—A schematic of the experimental cell for
CV measurement is shown in Fig. 2. The LiCl-KCl eutectic electro-
lyte was contained in a high-purity alumina crucible having an inner
diameter of 41 mm ͑99.5% Al O ; Nikkato Co.͒. The working elec-
2
3
trode was either a molybdenum or tantalum wire, 1 mm diam, which
had an alumina tube insulator for determining the surface area. The
electrode wire was exposed to the electrolyte by 10 mm. A molyb-
denum wire electrode with a spiral at its one end or a cadmium pool
containing zirconium metal located at the bottom of the alumina
crucible was employed as the counter electrode. The molybdenum
counter electrode was used for investigating the system in which no
zirconium metal existed. The reference electrode was an Ag/AgCl
electrode consisting of a silver wire immersed in the LiCl-KCl eu-
tectic mixture with 1 wt % AgCl, which was contained in a closed-
end Pyrex tube. Prior to the CV measurements, the Ag/AgCl elec-
trode potential was checked against a Li-Al ͑42 atom % Li͒
electrode. The potential value of 2.129 V at 499°C was in good
Table I. Test conditions and results for cathode deposition in LiCl-KCl eutectic containing ZrCl4 .
Current
͑mA͒
Coulombs
passed
Potential
͑V vs. Ag/AgCl͒
Catch
crucible
Test no.
XZrCl4a
Productb
C1
C2
C3
C4
C5
0.0026
0.0026
0.0026
0.0026
0.0022
15-20
10-20
40-85
40-90
50
200
223
200
220
205
Ϫ1.15 to Ϫ1.04
Ϫ1.12 to Ϫ1.04
Ϫ1.42 to Ϫ1.35
Ϫ1.42 to Ϫ1.35
Ϫ1.32 to Ϫ1.06
Not used
Not used
Used
ZrCl
ZrCl
Zr and a little ZrCl
Zr and a little ZrCl
Cd Zr and a little ZrCl
Used
Used with 0.34 g Cd
3
a
XZrCl4: mole fraction of ZrCl in LiCl-KCl eutectic.
Identified by XRD.
4
b
Downloaded on 2015-03-08 to IP 128.226.37.5 address. Redistribution subject to ECS terms of use (see ecsdl.org/site/terms_use) unless CC License in place (see abstract).

Journal of The Electrochemical Society, 151 ͑3͒ C187-C193 ͑2004͒
C189
Table II. Test conditions and results for anode dissolution of zirconium metal rods „3␾ ϫ 25 mm… in LiCl-KCl eutectic containing ZrCl4 .
T
͑°C͒
Zr rod
͑g͒
Current
͑mA͒
Coulombs
passed
Avg potential
͑V vs. Ag/AgCl͒
Decrease of
Zr rod ͑g͒
Number of
electrons, n
Test no.
XZrCl4
A01
A02
A03
A04
A05
A06
A07
A08
A09
A10
A11
500
500
500
450
450
450
550
550
550
500
500
0.0012
0.0012
0.0012
0.0012
0.0012
0.0012
0.0013
0.0013
0.0013
0.0025
0.0025
1.082
1.079
1.102
1.106
1.110
1.098
1.095
1.108
1.103
1.105
1.083
80
80
100
50
50
50
100
100
150
0
525
620
1000
400
480
490
900
910
900
Ϫ0.78
Ϫ0.89
Ϫ0.76
Ϫ0.74
Ϫ0.75
Ϫ0.73
Ϫ0.84
Ϫ0.85
Ϫ0.83
Ϫ0.95
Ϫ0.95
0.150
0.289
0.372
0.092
0.110
0.112
0.227
0.228
0.222
0.012
0.033
4.02
3.89
4.04
4.12
4.13
4.13
3.74
3.78
3.83
—
a
0
b
0
0
—
a
The zirconium rod was immersed in the salt for 19 h.
The zirconium rod was immersed in the salt for 38 h.
b
Ϫ
attached to the zirconium deposit. The salt content was evaluated by
determining the potassium concentration in the nitric acid solution
of the product.
Peak͑E͒: Zr → Zr͑IV͒ ϩ 4e
͓4͔
It can be seen in Fig. 3 that peak ͑D͒ disappeared when the ZrCl4
Anode dissolution.—CVs of a zirconium metal wire ͑1 mm␾͒
were measured to observe the dissolution behavior. The anode dis-
solution tests of the zirconium metal rods (3␾ ϫ 25 mmL) were
then conducted by constant current electrolysis to determine the
number of electrons for the anode reaction over the temperature
range 450-550°C. The cadmium pool was used as the cathode. The
experimental conditions are presented in Table II. During the elec-
trolysis, the anode potential gradually drifted to the positive side.
The dissolution progressed uniformly at the surface exposed to the
electrolyte. After the electrolysis, the zirconium rod sample was
rinsed with dilute nitric acid and the weight loss was measured. The
last two tests in Table II were the corrosion test of the zirconium
concentration was as low as 0.00027 and remarkably increased with
an increase in the ZrCl concentration. Furthermore, a CV obtained
4
at a higher scan rate had a smaller peak ͑D͒ and a larger peak ͑E͒, as
shown in Fig. 4. The height of peak ͑E͒ was approximately propor-
tional to the scan rate. Consequently, it was considered that zirco-
nium metal might be deposited on a molybdenum electrode surface
to form a thin layer before ZrCl was continually deposited on it.
Peak ͑B͒ is mainly attributed to Reaction 2 when the Zr͑IV͒ con-
centration is high and the scan rate is low.
The CVs shown in Fig. 5 were obtained by varying the potential
range of scanning, when XZrCl4 was 0.00123 and the other condi-
metal. The metal rods were immersed in a salt containing ZrCl for
4
19 and 38 h, respectively.
Results and Discussion
Cyclic voltammetry.—Measurements were mainly carried out at
500°C. The blank test for a molybdenum working electrode at a
potential scan rate of 0.1 V/s in a molten LiCl-KCl eutectic alone
showed that the residual current was less than 0.7 mA, and no cur-
rent peak was observed within the potential range of Ϫ0.4 to Ϫ1.7
V vs. the Ag/AgCl. This result indicated that any reactions of impu-
rities in the system within this potential range were negligible.
Figure 3 shows CVs of the molybdenum electrode at various
ZrCl4 concentrations in a LiCl-KCl eutectic at 500°C. Figure 4
shows the scan rate dependence when the mole fraction of ZrCl in
4
the salt, XZrCl4, was 0.00123. The working electrode potential was
first scanned from the rest potential to Ϫ1.3 V and then was re-
versely scanned to Ϫ0.4 V. The sharp cathodic peak ͑B͒ at about
Ϫ1.2 V might correspond to the reduction of Zr͑IV͒ into ZrCl and
zirconium metal. ZrCl is insoluble in the salt.
Ϫ
Peak͑B͒: Zr͑IV͒ ϩ 4e → Zr
͓1͔
͓2͔
Ϫ
Zr͑IV͒ ϩ 3e → ZrCl
This interpretation is supported by the results of the cathode depo-
sition tests. The electrodeposition at about Ϫ1.1 V gave a ZrCl
product, which is later described. The peak current of ͑B͒ was pro-
portional to the ZrCl concentration in the salt and the square root of
4
the scan rate, which indicated that the reaction was controlled by
Zr͑IV͒ diffusion mass transfer. There were two anodic peaks of ͑D͒
and ͑E͒, which might be ascribed to the oxidation of the deposited
ZrCl and zirconium metal, respectively
Figure 3. CVs of a molybdenum electrode (1␾ ϫ 10 mm) at various ZrCl4
concentrations in LiCl-KCl eutectic salt. Temperature 500°C; scan rate 0.1
V/s.
Ϫ
Peak͑D͒: ZrCl → Zr͑IV͒ ϩ 3e
͓3͔
Downloaded on 2015-03-08 to IP 128.226.37.5 address. Redistribution subject to ECS terms of use (see ecsdl.org/site/terms_use) unless CC License in place (see abstract).

C190
Journal of The Electrochemical Society, 151 ͑3͒ C187-C193 ͑2004͒
Figure 6. CVs of a tantalum electrode (1␾ ϫ 10 mm) in LiCl-KCl eutectic
salt with various potential ranges scanned. Temperature 500°C; XZrCl4
ϭ 0.00123; scan rate 0.1 V/s.
posited ZrCl might be reduced to zirconium metal. Actually, zirco-
nium metal products were obtained in the cathode deposition tests
when the applied cathode potential was about Ϫ1.4 V. The anodic
peak ͑D͒ completely disappeared when the working electrode poten-
tial was scanned to Ϫ1.7 V. The cathodic and anodic areas were
calculated for each CV in Fig. 3 and 5. They were in good agree-
ment within 7%. It is suggested that insoluble species were depos-
ited on the working electrode during the cathodic scan and that they
were completely reoxidized into Zr͑IV͒ during the anodic scan.
There was a cathodic peak ͑A͒ at about Ϫ1.0 V, which was
probably related to the reaction
Figure 4. CVs of a molybdenum electrode (1␾ ϫ 10 mm) with various
scan rates in LiCl-KCl eutectic salt. Temperature 500°C; XZrCl4 ϭ 0.00123.
tions for the measurement were the same as for Fig. 3. It is obvious
that peak ͑D͒ became smaller but peak ͑E͒ grew larger as the scan
was moved to more negative potential. The cathodic peak ͑C͒ was
observed at Ϫ1.5 V in addition to peaks ͑A͒ and ͑B͒. Over the
potential range that was more negative than about Ϫ1.3 V, the de-
Ϫ
Peak͑A͒: Zr͑IV͒ ϩ 2e → Zr͑II͒
͓5͔
The Zr͑II͒ present at the cathode due to Reaction 5 should be re-
duced further at the more negative potential range. The ratio of
Zr͑II͒ to Zr͑IV͒ seemed to be so small that the current due to the
reduction of Zr͑II͒ might be hidden by the current peaks ͑B͒ and ͑C͒
due to the reduction of Zr͑IV͒.
After the series of CV measurements, a zirconium metal was
immersed into the electrolyte containing ZrCl . The colorless melt
4
then gradually turned light brown. Zr͑II͒ might form by the follow-
ing reaction
Zr͑IV͒ ϩ Zr → 2Zr͑II͒
͓6͔
It was commonly observed that alumina crucibles and quartz tubes
exposed to the molten salts turned black in the presence of zirco-
nium metal, which might be due to the corrosion with Zr͑II͒. The
zirconium metal immersed in a LiCl-KCl-ZrCl melt corroded. The
4
metal surface became rough, and a fine black powder was seen
attaching to the metal. The Zr͑II͒ yielded by Reaction 6 might dis-
proportionate into Zr͑IV͒ and the black powder of the zirconium
metal. The Zr͑II͒/Zr͑IV͒ ratio in the salt phase at 500°C seemed to
be very low, because the zirconium content in the salt dissolving
ZrCl4 remained nearly constant, and the measured CV did not
change in shape after immersing the zirconium metal into the salt.
CVs of a tantalum working electrode shown in Fig. 6 were ob-
tained immediately after the measurements with the molybdenum
electrode for Fig. 5 were finished. Therefore, the experimental con-
Figure 5. CVs of a molybdenum electrode (1␾ ϫ 10 mm) in LiCl-KCl
eutectic salt with various potential range scanned. Temperature 500°C;
XZrCl4 ϭ 0.00123; scan rate 0.1 V/s.
Downloaded on 2015-03-08 to IP 128.226.37.5 address. Redistribution subject to ECS terms of use (see ecsdl.org/site/terms_use) unless CC License in place (see abstract).

Journal of The Electrochemical Society, 151 ͑3͒ C187-C193 ͑2004͒
C191
Figure 9. XRD patterns of the cathode deposits obtained in test ͑a͒ C1, ͑b͒
C3, and ͑c͒ C5.
Figure 7. CVs of a molybdenum electrode in LiCl-KCl eutectic salt at 450,
00, and 550°C. XZrCl4 ϭ 0.0012; scan rate 0.2 V/s.
5
eutectic dissolving ZrCl4 and that no soluble oxychloride species
was observed in the system.
1
8
Cathode deposition.—The electrodeposition tests were carried
out at 500°C. The results are presented in Table I, and the photo-
graphs of the obtained cathode products are in Fig. 8. All the prod-
ucts consisted of a large amount of the bulk salt and the zirconium
deposit. It was estimated by potassium analysis that C1, C2, and C4
products contained 82, 80, and 91 wt % of the bulk salt. Nodular
deposits were obtained by a low-current electrolysis of tests C1 and
C2. They were identified as ZrCl by the XRD analysis as shown in
Fig. 9a. No zirconium metal was detected. High-current electrolysis
of tests C3 and C4 caused more negative cathode potential and gave
pastelike deposits. Because the adhesion of the deposit to the cath-
ode wire was very poor, an alumina catch crucible ͑o.d. 5.2 mm͒
was placed under the wire cathode. The deposits consisted of zirco-
nium metal and a small amount of ZrCl, as shown in Fig. 9b. The
fine black powder of zirconium metal was obtained by washing the
deposits with dilute nitric acid. The dependence of the product ma-
terial on the cathode potential supported the Zr͑IV͒ reduction pro-
cess that was discussed in the CV study. Similar electrodeposition
tests were previously carried out by Iizuka, who found no zirconium
metal but did discover ZrCl in the cathode deposits.11 A high Zr͑IV͒
concentration and a less negative cathode potential might result in
the ZrCl deposition.
ditions for Fig. 5 and 6 were the same, except for the working
electrode material. Generally, the voltammogram of tantalum was
similar to that of molybdenum. However, the following slight dif-
ferences can be seen: When comparing the voltammograms obtained
over the scanning range of Ϫ1.3 to Ϫ0.4 V, the anodic peak ͑E͒ of
tantalum was larger than that of molybdenum. The cathodic peak
͑
C͒ of tantalum was smaller and located at a more negative poten-
tial. The large anodic peak ͑E͒ in Fig. 6 has a shoulder, which is
possibly due to the desorption of zirconium on tantalum. It is con-
sidered that the absorption of reduced zirconium metal on the tan-
talum electrode surface could promote the reduction of Zr͑IV͒ to
zirconium metal.
The CVs at 450, 500, and 550°C shown in Fig. 7 were measured
in the cell used for the anodic dissolution tests. A molybdenum wire,
1
mm diam, without an alumina insulator was employed as the
working electrode in this measurement. Thus, the exact surface area
was not certain. No significant difference in shape was observed
among the three voltammograms. The peak current increased with
increasing temperature which affects the diffusion coefficient of the
zirconium species in the salt, and the peak potential shifted against
the Ag/AgCl reference electrode. It is indicated that the zirconium
species existing in the salt did not change drastically over the tem-
perature range from 450 to 550°C.
During the electrodeposition, a decrease in the Zr͑IV͒ concentra-
tion in the salt was often observed. After the electrolysis, the con-
centration gradually increased. Oxidation of zirconium metal into
Zr͑IV͒ at the anode and deposition of ZrCl at the cathode might
Zr͑IV͒ is very sensitive to oxide ion. It was reported that ZrO2
was precipitated immediately after Li O was added into a LiCl-KCl
2
Figure 8. Cathode deposits obtained
in
a LiCl-KCl eutectic containing
ZrCl . A tantalum wire ͑1 mm␾͒ was
4
used as the cathode. The cathodes of
tests C3, C4, and C5 had a catch cru-
cible ͑o.d. 5.2 mm␾͒, and cadmium
metal was placed in the catch crucible
of test C5.
Downloaded on 2015-03-08 to IP 128.226.37.5 address. Redistribution subject to ECS terms of use (see ecsdl.org/site/terms_use) unless CC License in place (see abstract).

C192
Journal of The Electrochemical Society, 151 ͑3͒ C187-C193 ͑2004͒
and 550°C. Maximum current that could be passed without the an-
ode effect at 450 and 500°C was about 60 and 120 mA, respectively.
The number of electrons involved in the anode reaction, n, is com-
puted by the following equation
n ϭ qM/͑F⌬w͒
͓8͔
where M is the atomic weight of zirconium ͑91.224͒, F is the Fara-
day constant ͑96,485 C/mol͒, q is the quantity of electricity passed
for the electrolysis, and ⌬w is the weight loss of the zirconium rod.
The averages of n at 450, 500, and 550°C were 4.13, 3.98, and 3.78,
respectively. It is suggested that zirconium metal mainly dissolved
into the salt as Zr͑IV͒ over the temperature range 450-550°C and
that the ratio of Zr͑II͒/Zr͑IV͒ increases with increasing temperature.
The last two runs attempted to measure the corrosion rate of
zirconium metal in the LiCl-KCl-ZrCl system. The zirconium rods
4
were left in the melt for 19 and 38 h, respectively. The surface area
2
exposed to the melt was about 1.4 cm . During the test, the rest
potential of the zirconium rod was stable at Ϫ0.95 V. The samples
were pulled out from the salt and washed with distilled water for
weighing. It was observed that the metal surface became rough, and
fine black powder adhered to the surface. The weight losses of the
zirconium rods were 0.012 and 0.033 g, respectively. The corrosion
rate was far less than the anodic dissolution rate in this experiment.
To examine the black powder, an additional corrosion test was
conducted. A zirconium metal plate (50 ϫ 30 ϫ 1 mm) was left in
a LiCl-KCl-ZrCl4 melt (XZrCl4 ϭ 0.0016) at 500°C for 96 h. By
XRD analysis, no ZrCl was observed at the surface, and the col-
lected black powder was identified as zirconium metal.
Figure 10. CV of a zirconium electrode (1␾ ϫ 7 mm) in LiCl-KCl eutectic
salt at 500°C. XZrCl4 ϭ 0.00027; scan rate 0.1 V/s.
Conclusions
Some oxidation states ͑0, ϩ1, ϩ2, and ϩ4͒ of zirconium exist in
the LiCl-KCl eutectic system over the temperature range 450-
550°C. The zirconium behavior is summarized as follows: At the
anode, zirconium metal dissolved into the electrolyte salt as Zr͑IV͒,
which was the predominant species in the salt. At a higher tempera-
ture, a small amount of Zr͑II͒ was yielded in addition to Zr͑IV͒.
Zirconium metal reacted with Zr͑IV͒ to give Zr͑II͒, whose solution
was light brown. The Zr͑II͒/Zr͑IV͒ ratio seemed to be very low, and
Zr͑II͒ was easy to disproportionate into Zr͑IV͒ and the fine black
powder of the zirconium metal. At the cathode, not only zirconium
metal but also ZrCl was deposited. When the cathode potential was
sufficiently negative at probably ϽϪ1.35 V vs. Ag/AgCl electrode
at 500°C, zirconium metal was obtained. The deposited zirconium
metal was a fine black powder, and adhesion to the cathode wire was
poor. The mixture of the bulk salt and the zirconium metal re-
sembled a paste. The electrodeposition at about Ϫ1.1 V yielded a
nodular deposit of ZrCl on the cathode. ZrCl could not be obtained
by immersing a zirconium metal in the salt containing Zr͑IV͒, and
ZrCl seemed to disproportionate into Zr͑IV͒ and zirconium metal
very slowly. Therefore, ZrCl appeared to be a metastable compound
in this system. In the presence of cadmium metal at the cathode, an
decrease the Zr͑IV͒ concentration. A possible reason for the increas-
ing concentration is that the ZrCl which fell from the cathode dis-
proportionated into Zr͑IV͒ and zirconium metal. An additional test
was conducted. When the ZrCl obtained by test C2 was left in a
LiCl-KCl eutectic at 500°C for 6 days, the zirconium concentration
in the salt gradually increased. ZrCl could not be obtained by im-
mersing a zirconium metal into the salt containing ZrCl . Therefore,
4
the following reaction did not occur spontaneously
Ϫ
3
Zr ϩ Zr͑IV͒ ϩ 4Cl → 4ZrCl
͓7͔
Therefore, ZrCl is probably a metastable compound in this system
An attempt was made using cadmium metal at the cathode for
recovering zirconium metal efficiently. In test C5, a small amount of
cadmium metal was placed in the catch crucible. The XRD pattern
of the product is shown in Fig. 9c. There were some unknown peaks
in addition to the peaks of the electrolyte, cadmium metal and ZrCl.
According to the Cd-Zr binary phase diagram,19 they might be at-
tributed to Cd Zr. Because the intermetallic compound deposit en-
3
intermetallic compound that might be Cd Zr was obtained. The col-
larged and the adhesion was much better, the collection efficiency of
zirconium at the cathode was improved using cadmium. Cadmium is
easily separated from the recovered zirconium metal by distillation.
3
lection efficiency of zirconium deposit is improved using cadmium
because the adhesion of the intermetallic compound was much bet-
ter.
Anode dissolution.—A CV of a zirconium wire electrode
Zirconium is a nuisance in the electrorefining process for pyro-
chemical reprocessing. Zirconium metal dissolves into the salt over
a potential region more positive than Ϫ0.95 V vs. Ag/AgCl ͑1 wt %
AgCl͒ at 500°C. Therefore, zirconium in spent metal fuels dissolves
into the salt along with actinides when most of the actinides in the
fuels are stripped. The current efficiency for recovering actinides is
considerably decreased by the zirconium dissolution because the
number of electrons for the zirconium reaction is almost four. The
electrolysis at low current improves the retention of zirconium at the
anode, but the throughput of the electrorefining process is lowered.
Most of the dissolved zirconium accumulates in the cadmium pool
placed at the bottom of the electrorefiner vessel. It is possible to
electrotransport the zirconium from the cadmium pool to the cath-
(
1␾ ϫ 7 mm) is shown in Fig. 10. The electrode potential was
anodically scanned from the rest potential to Ϫ0.6 V and then re-
versed. The zirconium metal dissolved into the salt by the anodic
polarization. The dissolution current linearly increased up to 210
mA, at Ϫ0.75 V it suddenly dropped to about 80 mA, and then
remained almost constant. The electrode might be passivated by the
formation of K ZrCl at the surface. The solubility of K ZrCl in
2
6
2
6
the LiCl-KCl eutectic was reported to be 0.013 mole fraction at
2
0
5
00°C. The dissolution rate of K ZrCl into the bulk salt might
2 6
determine the limiting current.
The experimental conditions and results of the anode dissolution
tests are presented in Table II. The tests were carried out at 450, 500,
Downloaded on 2015-03-08 to IP 128.226.37.5 address. Redistribution subject to ECS terms of use (see ecsdl.org/site/terms_use) unless CC License in place (see abstract).

Journal of The Electrochemical Society, 151 ͑3͒ C187-C193 ͑2004͒
C193
ode, which was demonstrated in this work. Using cadmium at the
cathode might be effective to improve the zirconium collection ef-
ficiency. The salt and cadmium adhering to the product metal can be
removed by distillation at high temperature.21 ZrCl deposits may
cause no problem, because ZrCl disproportionates into zirconium
5. M. Iizuka, K. Uozumi, T. Inoue, T. Iwai, O. Shirai, and Y. Arai, J. Nucl. Mater.,
99, 32 ͑2001͒.
2
6
. J. A. Plambeck, Fused Salt Systems, Encyclopedia of Electrochemistry of the Ele-
ments, Vol. X, A. J. Bard, Editor, Marcel Dekker, Inc., New York ͑1976͒.
. T. C. Totemeier and R. D. Mariani, J. Nucl. Mater., 250, 131 ͑1997͒.
7
8. S. X. Li, in The 8th International Conference on Nuclear Engineering (ICONE-8),
Baltimore, MD, April 2-6, 2000.
metal and ZrCl in the distiller. From a waste point of view, it seems
4
9
. R. Baboian, D. L. Hill, and R. A. Bailey, J. Electrochem. Soc., 112, 1221 ͑1965͒.
0. T. Suzuki, Denki Kagaku oyobi Kogyo Butsuri Kagaku, 39, 864 ͑1971͒.
1. M. Iizuka, CRIEPI Report T98001 ͑1998͒, in Japanese.
desirable to recover zirconium from the spent fuels. For that pur-
pose, high throughput of the collecting zirconium is required.
1
1
1
1
2. F. Basile, E. Chassaing, and G. Lorthioir, J. Appl. Electrochem., 11, 645 ͑1981͒.
3. G. J. Kipouros and S. N. Flengas, J. Electrochem. Soc., 132, 1087 ͑1985͒.
Acknowledgments
The author is grateful to M. Iizuka and T. Inoue of CRIEPI for
their helpful suggestions. I thank K. Kugimiya and N. Yahagi of
Techno Service Company, Limited, for the ICP-AES and atomic
absorption analyses.
14. G. J. Kipouros and S. N. Flengas, in Molten Salts, C. L. Hussey, S. N. Flengas, J.
S. Wilkes, and Y. Ito, Editors, PV 90-17, p. 626, The Electrochemical Society
Proceedings Series, Pennington, NJ ͑1990͒.
1
1
5. A. W. Struss and J. D. Corbett, Inorg. Chem., 9, 1373 ͑1970͒.
6. Y. Sakamura, O. Shirai, T. Iwai, and Y. Suzuki, J. Electrochem. Soc., 147, 642
͑
2000͒.
Central Research Institute of Electric Power Industry (CRIEPI) assisted
in meeting the publication costs of this article.
1
1
7. Y. Sakamura, O. Shirai, T. Iwai, and Y. Suzuki, J. Alloys Compd., 321, 76 ͑2001͒.
8. Y. Sakamura, T. Kato, M. Kurata, and H. Moriyama, CRIEPI Report T96073
͑
1997͒, in Japanese.
References
1
2
9. T. B. Massalski, Binary Alloy Phase Diagrams, Vol. 1, American Society for Met-
als, Metals Park, OH ͑1986͒.
0. Y. Sakamura, T. Inoue, and H. Moriyama, in The 6th International Conference on
Molten Salt Chemistry and Technology (MS6), C. Nianyi and Q. Zhiyu, Editors, p.
292, Shanghai, China, Oct 2001.
1
2
. Y. I. Chang, Nucl. Technol., 88, 129 ͑1989͒.
. T. Inoue, T. Yokoo, and T. Nishimura, in International Conference on Future
Nuclear Systems, GLOBAL’99, Jackson Hole, WY, Aug 29-Sept 3, 1999.
. Z. Tomczuk, J. P. Ackerman, R. D. Wolson, and W. E. Miller, J. Electrochem. Soc.,
3
4
1
39, 3523 ͑1992͒.
21. B. R. Westphal, D. Vaden, T. Q. Hua, J. L. Willit, and D. V. Laug, in The 5th
Topical Meeting, DOE Owned Spent Nuclear Fuel and Fissile Materials Manage-
ment, Charleston, SC, Sept 17-20, 2002.
. T. Koyama, M. Iizuka, Y. Shoji, R. Fujita, H. Tanaka, T. Kobayashi, and M. Toki-
wai, J. Nucl. Sci. Technol., 34, 384 ͑1997͒.
Downloaded on 2015-03-08 to IP 128.226.37.5 address. Redistribution subject to ECS terms of use (see ecsdl.org/site/terms_use) unless CC License in place (see abstract).