Co-reduction of aluminium and lanthanide ions in molten fluorides: Application to cerium and samarium extraction from nuclear wastes

Article abstract of DOI:10.1016/j.electacta.2009.01.074

This work concerns the method of co-reduction process with aluminium ions in LiF-CaF₂ medium (79-21 mol.%) on tungsten electrode for cerium and samarium extraction. Electrochemical techniques such as cyclic and square wave voltammetries, and po

Full text of DOI:10.1016/j.electacta.2009.01.074

Electrochimica Acta 54 (2009) 5300–5306
Contents lists available at ScienceDirect
Electrochimica Acta
journal homepage: www.elsevier.com/locate/electacta
Co-reduction of aluminium and lanthanide ions in molten fluorides:
Application to cerium and samarium extraction from nuclear wastes
M. Gibilaro, L. Massot∗, P. Chamelot, P. Taxil
Laboratoire de Génie Chimique UMR 5503, Département Procédés Electrochimiques, Université de Toulouse, 31062 Toulouse Cedex 9, France
a r t i c l e i n f o
a b s t r a c t
Article history:
This work concerns the method of co-reduction process with aluminium ions in LiF–CaF2 medium
(79–21 mol.%) on tungsten electrode for cerium and samarium extraction. Electrochemical techniques
such as cyclic and square wave voltammetries, and potentiostatic electrolyses were used to study the
co-reduction of CeF3 and SmF3 with AlF3. For each of these elements, specific peaks of Al–Ce and Al–Sm
alloys formation were observed by voltammetry as well as peaks of pure cerium and aluminium, and pure
samarium and aluminium respectively. The difference of potential measured between the solvent reduc-
tion and the alloy formation suggests expecting an extraction efficiency of 99.99% of each lanthanide by
the process.
Received 31 October 2008
Received in revised form 16 January 2009
Accepted 26 January 2009
Available online 3 February 2009
Keywords:
Fluorides
Electrochemical co-reduction
Alloys formation
Aluminium
Different intermetallic compounds were obtained for different potentiostatic electrolysis and were
characterised by Scanning Electron Microscopy with EDS probe.
The validity of the process was verified by carrying out cerium and samarium extractions in the form
of Al–Ln alloy; the extraction efficiency was 99.5% for Ce(III) and 99.4% for Sm(III).
© 2009 Elsevier Ltd. All rights reserved.
Lanthanides
1. Introduction
[5]. Reductive extraction is considered as the more efficient process
for the French option (fluoride/aluminium) and it is also considered
Since 1991, alternative solutions for radioactive wastes treat-
ment have been examined by several nuclear research programs.
The aim of Partitioning and Transmutation (P&T) is to reduce signif-
icantly the amount of radiotoxic nuclear wastes after fuel recycling,
by consuming the most radiotoxic elements (Np, Pu, Am, Cm). The
separation process for actinides (An) and lanthanides (Ln) has to
be completely achieved. Uranium, neptunium and plutonium are
actually extracted for further recycling with the hydrometallurgical
process PUREX.
The pyrochemical route is mainly envisaged for the reprocessing
of advanced fuels and transmutation blankets which are expected
to have a high content of Pu and minor actinide and high burn up.
The interest of pyroprocesses is based on the properties of the salt
phase: high radiation resistance, high thermal resistance and they
are almost transparent to neutrons. Pyrometallurgical processes are
also attractive for the dissolution of some of the proposed fuels or
blankets such as inert matrix, which are very difficult or impossible
to be dissolved in aqueous media. An alternative route would be
pyrochemical processes using molten salts for the separation of An
from Ln.
within the Japanese process in molten chloride with bismuth. Here,
lanthanides LnF3 and actinides AnF3 to be extracted are placed in a
molten solution in contact with a liquid metallic phase acting as a
reductive agent A.
The separation efficiency is highly dependant on the metallic
phase and the most promising liquid solvent for An–Ln separation
is the aluminium [6]. Conocar et al. [7] investigated the An–Ln sep-
aration in LiF–AlF3 at 830 ◦C and demonstrated experimentally that
more than 99.3% of Pu and Am initially present in the molten solvent
are recovered in the aluminium phase according to the following
scheme:
AnF3(solution) + Al(metal) = An(metal) + AlF3(solution)
(1)
Hence, this process allows the selective extraction of actinides
while lanthanides remain in the salt. The next reprocessing
step is the lanthanides recovery in order to recycle the fluoride
solvent.
Nourry [8] succeed in extracting three lanthanides (Sm, Gd, Nd)
in LiF–CaF2 using electroreduction on a reactive cathode (Cu or Ni)
in the form of alloys Cu/Ln or Ni/Ln. The alloy formation allows the
lanthanide deposition to be shifted in the anodic sense which is
called depolarisation. Nourry showed experimentally that, as well
as on copper as on nickel cathodes, the lanthanide extraction effi-
ciency was about 100%.
Several separation techniques are examined as precipitation
with oxide ions [1], electrorefining [2–4] or reductive extraction
∗
The present work is dedicated to the lanthanides extraction:
the co-reduction by the co-reduction of Ln ions with aluminium
Corresponding author. Tel.: +33 561 558 194; fax: +33 561 556 139.
E-mail address: massot@chimie.ups-tlse.fr (L. Massot).
0013-4686/$ – see front matter © 2009 Elsevier Ltd. All rights reserved.
doi:10.1016/j.electacta.2009.01.074

M. Gibilaro et al. / Electrochimica Acta 54 (2009) 5300–5306
5301
ions. The two different elements studied, cerium and samarium,
are relevant in the list of fission products. Their behaviour in
the molten fluorides is typical of Ln elements: three exchanged
electrons and one or two reduction steps.
compartmentalisation avoids gas and fluoride media mixing and
re-oxidation of Al–Ln compounds cathodically formed.
3- The working electrode was a tungsten plate with a large surface
area (3 cm2).
3. Results and discussion
2. Experimental
3.1. Co-reduction process overview
- The cell consisted of a vitreous carbon crucible placed in a cylin-
drical vessel made of refractory steel and closed by a stainless
steel lid cooled inside by circulating water. The inner part of
the walls was protected against fluoride vapours by a graphite
liner containing the experimental crucible. The experiments were
performed under an inert argon atmosphere (U-grade: less than
5 ppm O2), previously dehydrated and deoxygenated using a
purification cartridge (Air Liquide). The cell was heated using a
programmable furnace and the temperature was measured using
a chromel–alumel thermocouple. A more detailed description of
the device can be found in previous papers of our laboratory such
as the one referred in [9].
- The electrolytic bath consisted of the eutectic LiF/CaF2 (SDS
99.99%) mixture (79/21 molar ratio). Before use, it was dehy-
drated by heating under vacuum (3 × 10−2 mbar) from the room
temperature up to its melting point (762 ◦C) for 72 h. For pro-
viding aluminium and neodymium ions, aluminium fluoride AlF3
(SDS 99.95%) and lanthanide fluoride (SDS 99.99%) powders were
introduced into the bath through a lock chamber under argon gas
atmosphere.
The co-reduction consists of a simultaneous reduction of two
or more metallic ions on an inert electrode to form an alloy of the
two metals. This process described by Brenner in aqueous media
[12] and more recently proposed by Taxil et al. in molten salts
for nuclear wastes reprocessing [13] exhibits advantages for lan-
thanides extraction compared to the use of a reactive electrode
by avoiding intermetallic diffusion limitation and consumption of
the anode material. Moreover, better selectivity is expected by this
technique due to a possible correlation between the alloy compo-
sition and the cathode potential imposed during electrolysis.
If R is the more reactive precursor, N the less reactive one and
RxNy the alloy formed by co-reduction, the co-reduction process in
the case of two metallic ions, Rn+ and Np+ involves the following
reactions:
xRn+ + ne− = xR
yNp+ + pe− = yN
(I)
(II)
xR + yN = RxNy
(III)
- Electrochemical equipment: all electrochemical studies and
electrolyses were performed with an Autolab PGSTAT 30 poten-
tiostat/galvanostat controlled by a computer using the research
software GPES 4.9.
- Electrochemical techniques: cyclic voltammetry, square wave
voltammetry and potentiostatic electrolyses were used for the
investigation of the aluminium–lanthanide co-reduction process.
- Characterisation of reduction products: after electrolysis runs, the
cathode surface was examined by Scanning Electron Microscopy
(LEO 435 VP) equipped with an EDS probe (Oxford INCA 200) for
determining the composition of the alloys.
The overall process is:
xRn+ + yNp+ + (n + p)e− = RxNy
(2)
The equilibrium potential of the system R/RxNy can be expressed
as:
RT
ERn+
= ERn+
−
ln[aR(in RxNy)]
(3)
/R N
/R
x
y
nF
where ERn+ is the equilibrium potential of pure R element, T the
/R
absolute temperature in K, n the number of exchanged electron, F
the Faraday constant (96 500 C) and aR (in RxNy) the activity of R in
the intermetallic compound RxNy.
2.1. Analytical setting
As aR in RxNy is less than one, it can be deduced that ERn+
>
/R N
x
y
For investigations on electrochemical behaviour of the fluoride
baths, we used the following cell:
ERn+
.
/R
Here the co-deposition of R with a more noble metal allows Rn+
to be reduced at a potential more anodic than the pure metal depo-
sition: this “depolarisation effect” is similar to that is observed on
reactive electrodes [14].
Tungsten wire (1 mm diameter) was used as a working elec-
trode with a surface area determined from its immersion depth
measurement in the bath after withdrawing from the cell. The aux-
iliary electrode was a vitreous carbon rod (3 mm diameter) with a
large surface area. Potentials were referenced to a platinum wire
(0.5 mm diameter) immersed in the molten electrolyte, acting as a
quasi-reference electrode Pt/PtOx/O2− [10].
Co-reduction has been mainly used for boron base metal deposi-
tion. Lantelme et al. [15] studied the electrodeposition of TiB2 from
titanium and boron ions in NaCl–KCl–NaF. When both metallic ions
are present in the melt, only one reduction peak is observed on
the cyclic voltammogram corresponding to TiB2 deposit, which is
shifted towards more positive values, than that of the pure boron
deposition. This observation was confirmed by Ett et al. [16] and
Wendt et al. in molten fluorides [17,18].
The selectivity of the co-reduction process was demonstrated by
Cohen [19] for niobium–germanium alloys in LiF–KF. Several inter-
metallic compounds (NbGe2, Nb3Ge2. . .) were prepared by varying
correlatively the metallic ion concentration in the bath and the
electrolysis potential.
The present work is aimed at using the alloy formation technique
for lanthanides extraction (Ce, Sm) in molten fluorides and at opti-
mizing the technique by the selection of the electrolysis potential.
The first part is a brief review of the main results on the
electrochemical behaviour of aluminium and lanthanide fluorides
obtained in our laboratory using the experimental setup 2.2 called
2.2. Extraction setting
Compared with the previous setting, the extraction experiments
need specific arrangements:
1- The reference electrode is
a nickel 1 mm diameter wire
immersed in a mixture of LiF–CaF2–NiF2 (1 mass%), placed in
a boron nitride basket. This reference electrode was proved to
be reliable in molten fluoride baths [11]. Moreover, the insula-
tion from the electrolytic bath ensures the invariability of its
potential with the change of composition of the electrolyte.
2- The anode was made of vitreous carbon and was isolated in a
graphite compartment containing LiCl–LiF–CaF2 eutectic mix-
ture. The anodic reaction was the chlorine release; the electrode

5302
M. Gibilaro et al. / Electrochimica Acta 54 (2009) 5300–5306
Fig. 2. (a) Cyclic voltammogram of the LiF–CaF2–SmF3 (2.1 × 10−4 mol cm−3) system
at 100 mV s−1 and T = 840 ◦C [21]. Working electrode: Mo; counter electrode: vitre-
ous carbon; quasi-reference electrode: Pt. (b): Square wave voltammogram of the
LiF–CaF2–SmF3 system (2.1 × 10−4 mol cm−3) at 9 Hz and T = 840 ◦C [21]. Working
electrode: Mo; counter electrode: vitreous carbon; quasi-reference electrode: Pt.
Fig. 1. (a) Cyclic voltammogram of the LiF–CaF2–CeF3 (2.5 × 10−4 mol cm−3) system
at 100 mV s−1 and T = 840 ◦C [20]. Working electrode: W; counter electrode: vitre-
ous carbon; quasi-reference electrode: Pt. (b) Square wave voltammogram of the
LiF–CaF2–CeF3 system (2.5 × 10−4 mol cm−3) at 9 Hz and T = 840 ◦C [20]. Working
electrode: W; counter electrode: vitreous carbon; quasi-reference electrode: Pt.
“analytical setting”. In the second part, the behaviour of the mix-
ture of aluminium and lanthanide ions in the molten salt has been
investigated in order to define the alloy composition as a function
of the potential. Finally, the results of Ln extraction by co-reduction
are presented.
successive steps:
Sm(III) + e− = Sm(II)
Sm(II) + 2e− = Sm
(5)
(6)
The second reduction step corresponding to the metal depo-
sition is not observed in our molten media. On the cyclic
voltammogram of Fig. 2a recorded at 100 mV s−1, only one reduc-
tion peak is observedat −1.25 V vs. Pt corresponding to Eq. (5). It was
established that the electrochemical reduction process is controlled
by the diffusion of Sm(III) ions in the solution.
This feature is confirmed by square wave voltammetry that
exhibits a single symmetric peak typical of a soluble/insoluble reac-
tion attributed to the Sm(III)/Sm(II) system at about −1.15 V vs.
Pt.
3.2. Previous results on solutes electrochemical reduction on inert
electrode
3.2.1. Cerium fluoride
The CeF3 electrochemical reduction was investigated in LiF–CaF2
on an inert tungsten electrode at 840 ◦C [20]. It has been shown
that its reduction is a one step process exchanging 3 electrons and
controlled by the diffusion of cerium ions in the melt.
Ce(III) + 3e− = Ce
(4)
On the cyclic voltammogram plotted at 100 mV s−1 on a W elec-
trode (Fig. 1a), the reduction peak at −1.8 V vs. Pt is attributed to
the cerium metal deposition and its re-oxidation peak is at about
−1.55 V vs. Pt. The solvent reduction, corresponding to lithium
deposit, takes place at −2 V vs. Pt.
3.2.3. Aluminium fluoride
The work performed in our laboratory [22] on tungsten elec-
trode at 860 ◦C concludes that the reduction of Al(III) in LiF–CaF2 is
controlled by the diffusion of aluminium ions in the melt and that
it follows a single step process exchanging 3 electrons, as observed
on the cyclic voltammogram presented in Fig. 3a where only one
wave at −1.25 V vs. Pt is present:
Fig. 1b shows a square wave voltammogram plotted under the
same conditions. The peak for cerium metal deposition is located
at −1.72 V vs. Pt.
Al(III) + 3e− = Al
(7)
3.2.2. Samarium fluoride
Massot et al. [21] studied the electrochemical behaviour of
samarium(III) ions in the eutectic mixture LiF–CaF2 on a molyb-
denum electrode. It was demonstrated that SmF3 is reduced in two
On the square wave voltammogram plotted at 9 Hz in Fig. 3b,
we can see one reduction peak at −1.2 V vs. Pt corresponding to
aluminium deposition.

M. Gibilaro et al. / Electrochimica Acta 54 (2009) 5300–5306
5303
Fig. 3. (a) Cyclic voltammogram of the LiF–CaF2–AlF3 (2.1 × 10−4 mol cm−3) sys-
tem at 100 mV s−1 and T = 860 ◦C [22]. Working electrode: W; counter electrode:
vitreous carbon; quasi-reference electrode: Pt. (b) Square wave voltammogram of
the LiF–CaF2–AlF3 system (2.1 × 10−4 mol cm−3) at 9 Hz and T = 860 ◦C [22]. Working
electrode: Mo; counter electrode: vitreous carbon; quasi-reference electrode: Pt.
Fig. 4. (a) Cyclic voltammogram of the LiF–CaF2–AlF3–CeF3 system (respectively
2.1 × 10−4 and 2.5 × 10−4 mol cm−3) at 100 mV s−1 and T = 840 ◦C. Working electrode:
W; counter electrode: vitreous carbon; quasi-reference electrode: Pt. (b) Square
wave voltammogram of the LiF–CaF2–AlF3–NdF3 system (respectively 1.8 × 10−4 and
2 × 10−4 mol cm−3) at 9 Hz and T = 860 ◦C. Working electrode: W; counter electrode:
vitreous carbon; quasi-reference electrode: Pt.
The same distribution of peak potentials is observed more
clearly in the square wave voltammogram of the mixture AlF3–CeF3
(Fig. 4b) on W electrode at 840 ◦C and a signal frequency of 9 Hz. The
presence of Al(III) and Ce(III) peaks together with new additional
peaks of co-reduction between them are observed.
These preliminary experiments lead us to suppose that the co-
reduction Al/Ce involves the formation of intermetallic compounds
at well defined potentials in a potential range located between the
deposition of each pure metal.
3.3. Co-reduction between aluminium and lanthanide ions
3.3.1. AlF3–CeF3 system
The Al–Ce phase diagram [23] exhibits 5 intermetallic com-
pounds Al11 Ce3, AlCe3, AlCe2, AlCe, AlCe3. The following part will
be focused on the determination of their potential formation.
3.3.1.1. Electrochemical behaviour. Experiments were carried out on
an inert tungsten electrode at 840 ◦C with the analytical setting
described in Section 2.1; the electrochemical behaviour of the Al–Ce
system was investigated by cyclic voltammetry and square wave
voltammetry. The same metallic ion concentrations as in each indi-
vidual study above were used: [Al(III)] = 2.1 × 10−4 mol cm−3 and
[Ce(III)] = 2.5 × 10−4 mol cm−3. The cyclic voltammetry of this mul-
tiple system on W electrode is presented in Fig. 4a.
3.3.1.2. Co-deposition products analysis. In this section, each peak
of the voltammogram is attributed to a specific composition of the
alloy compound.
Potentiostatic electrolysis runs were performed on a tung-
sten electrode in LiF–CaF2 by co-reduction of the respective ions.
The experimental conditions were: [Al(III)] = 2.1 × 10−4 mol cm−3
;
Four reduction peaks are observed:
[Ce(III)] = 2.5 × 10−4 mol cm−3; T = 840 ◦C; electrolysis time = 1200 s.
Just after each run, the electrode was quenched by a rapid with-
drawal from the cell. Then, an EDS analysis, joined to the SEM
observation of the sample cross section, allowed us to determine
the composition of the phases observed on the micrographs.
Electrolyses were performed with a cathode potential previously
defined by cyclic voltammetry.
•
Ep = −1.25 V vs. Pt: this peak seems to be superimposed on the
one of pure aluminium formation because the potential peaks
are quiet similar and the current density is appropriately higher
than for Al alone.
•
Ep = −1.79 V vs. Pt: a similar situation is observed, with a peak for
The micrographics presented in Fig. 5 show that:
the mixture at the same potential and with enhancement of the
current compared with pure cerium deposition.
•
Ep = −1.43 V vs. Pt and Ep = −1.6 V vs. Pt: the additional reduc-
tion peaks, comprised of the respective reduction potentials
of Al(III) into Al and Ce(III) into Ce, can be attributed to the
aluminium–cerium co-reduction.
(i) at each electrolysis potential, one intermetallic compound is
observed;
(ii) the new solid phase of the alloys is no adherent onto the
substrate and is released within the molten salts solution.

5304
M. Gibilaro et al. / Electrochimica Acta 54 (2009) 5300–5306
Fig. 6. (a) Cyclic voltammogram of the LiF–CaF2–AlF3–SmF3 system (respectively
1.8 × 10−4 and 5.2 × 10−4 mol cm−3) at 100 mV s−1 and T = 840 ◦C. Working electrode:
W; counter electrode: vitreous carbon; quasi-reference electrode: Pt. (b) Square
wave voltammogram of the LiF–CaF2–AlF3–SmF3 system (respectively 1.8 × 10−4
and 5.2 × 10−4 mol cm−3) at 9 Hz and T = 840 ◦C. Working electrode: W; counter elec-
trode: vitreous carbon; quasi-reference electrode: Pt.
•
•
E = −1.43 V vs. Pt and E = −1.6 V vs. Pt: Al3Ce deposit (Fig. 5b).
3Al(III) + Ce(III) + 12e− = Al3Ce
(9)
E = −1.79 V vs. Pt: AlCe3 and AlCe deposits (Fig. 5c).
Al(III) + 3Ce(III) + 12e− = AlCe3
(10)
(11)
Al(III) + Ce(III) + 6e− = AlCe
3.3.2. Characterisation of AlF3–SmF3 system
3.3.2.1. Electrochemical behaviour. The incomplete Al–Sm phase
diagram [24] exhibits 4 intermetallic compounds (Al3Sm, Al2Sm,
AlSm and AlSm2) which could be obtained by co-reduction.
The simultaneous presence of aluminium and samarium ions
produces a reasonable change in the individual electrochemical
signal for each element as shown in Fig. 6a recorded by cyclic
Fig. 5. (a) Micrograph of the cross section of a tungsten wire after electrolysis. Exper-
imental conditions: T = 840 ◦C, E = −1.25 V vs. Pt, t = 1200 s. (b) Micrograph of the cross
section of a tungsten wire after electrolysis. Experimental conditions: T = 840 ◦C,
E = −1.43 V vs. Pt, t = 1200 s. (c) Micrograph of the cross section of a tungsten wire
after electrolysis. Experimental conditions: T = 840 ◦C, E = −1.79 V vs. Pt, t = 1200 s.
voltammetry on W electrode at 100 mV s−1
.
This adherence defect is presumably due to a specific nucle-
ation phenomenon at the germination step that we still cannot
explain.
Only shoulders and not well defined peaks are observed in
Fig. 6a; a cathodic current, plotted at E < −1.25 V vs. Pt (pure alu-
minium deposition) is certainly due to Al–Sm alloy formation.
The square wave voltammetry, which is known as a more sen-
sitive method than cyclic voltammetry, produces with a better
characterisation of the system. In Fig. 6b, the square wave voltam-
mogram plotted on a W electrode of the metallic ions coexisting in
the solution exhibits the following reduction peaks or shoulders:
The formation of intermetallic compounds at each electrolysis
potential is reported below:
•
E = −1.25 V vs. Pt: Al11 Ce3 deposit (Fig. 5a).
11Al(III) + 3Ce(III) + 42e− = Al11 Ce3
•
E > −1.2 V vs. Pt: the current seems to be the sum of the current
(8)
of Sm(III)/Sm(II) and Al(III)/Al. The current signal superimposes

M. Gibilaro et al. / Electrochimica Acta 54 (2009) 5300–5306
5305
Fig. 8. Current density evolution during electrolysis measured by square wave
voltammetry. Experimental conditions: T = 840 ◦C, f = 9 Hz; working electrode: W;
counter electrode: vitreous carbon; reference electrode: NiF2/Ni.
Fig. 7. Micrograph of the cross section of a tungsten wire after electrolysis. Experi-
mental conditions: T = 840 ◦C, E = −1.45 V vs. Pt, t = 1200 s.
the reduction peak of Al(III) reduction in the decreasing part of
Sm(III) reduction.
E < −1.2 V vs. Pt: new reduction peaks or shoulders (−1.45, −1.53
and −1.66 V vs. Pt) are comprised of the reduction of Al(III) in
pure aluminium and the solvent reduction being specific of the
co-reduction process.
The tungsten electrode was replaced every 3 h and periodically
the progress of the reaction was measured by recording on a tung-
sten wire cyclic and square wave voltammograms.
•
3.4.2. Aluminium cerium extraction process
The electrolysis potential was chosen from the square wave
voltammogram (Fig. 4b) at −1.43 V and the initial concen-
3.3.2.2. Co-reduction products analysis. The same methodol-
ogy as previously for Al–Ce alloy characterisation was used;
potentiostatic electrolysis was performed at the following exper-
trations in metallic ions were: [Al(III)] = 1.8 × 10−4 mol cm−3
;
[Ce(III)] = 5.6 × 10−5 mol cm−3
.
As mentioned above, the process was controlled by periodically
recording square wave voltammograms of the solution. Typical
square wave voltammograms are presented in Fig. 8. A signifi-
cant decrease of the current density is observed, correlated to the
decrease of Al(III) and Ce(III) ion concentration in the bath dur-
ing the extraction process. We noticed that the current density
decreased more quickly at the beginning of the electrolysis than
at longer times.
imental conditions: [Al(III)] = 2.1 × 10−4 mol cm−3
;
[Sm(III)] =
2.5 × 10−4 mol cm−3
;
T = 840 ◦C; electrolysis time = 1200 s. The
electrode was quenched after each run and the deposited product
was analysed by an EDS probe. The electrolysis potential was
determined from the square wave voltammogram of Fig. 6b. The
results are indicated below:
•
E < −1.15 V vs. Pt.
Electrolyses were performed for 120 h to obtain a cathodic
current density value approaching zero on the square wave voltam-
mogram. The extraction efficiency was calculated with ICP analysis
of the bath after electrolysis and was equal to 99.5% for cerium(III).
At these potentials, the electrochemical reduction of Sm(III)
into Sm(II) occurs. None of the electrolyses lead to Al–Sm alloy
deposition.
•
E > −1.2 V vs. Pt.
3.4.3. Aluminium samarium extraction process
The same electrolysis conditions have been applied for
E > −1.2 V vs. Pt and only one alloy was observed on the micro-
graph presented in Fig. 7: Al3Sm.
The same methodology was applied to the aluminium–
samarium co-extraction process. The initial concentrations were
[Al(III)] = 2.2 × 10−4 mol cm−3
;
[Sm(III)] = 6 × 10−5 mol cm−3 and
3Al(III) + Sm(II) + 11e− = Al3Sm
(12)
the electrolysis potential was −1.45 V.
The Al–Sm alloy of uniform composition is released in the salt
and seems to coalesce in 100 ␮m balls, each of them exhibiting a
noticeable porosity.
3.4. Co-reduction application to lanthanide extraction
3.4.1. Experimental procedure
The extraction was realised on tungsten electrode in LiF–CaF2
media at T = 840 ◦C.
The experimental setup is described in the Experimental Section
2.2 with a specific reference electrode and anodic compartment.
The use of a reliable reference electrode Ni(II)/Ni insulated from
the electrolytic bath ensures the stability of the potential reference
which is an essential condition for lengthy potentiostatic electrol-
yses runs. So, in the electrolysis part of the article, potentials are
referenced to the NiF2/Ni electrode.
Extraction electrolyses were performed on a W plate at the
potential predefined on previous square wave voltammogram cor-
responding to Al3Ln formation.
Fig. 9. Current density evolution during electrolysis measured by square wave
voltammetry. Experimental conditions: T = 840 ◦C, f = 9 Hz; working electrode: W;
counter electrode: vitreous carbon; reference electrode: NiF2/Ni.

5306
M. Gibilaro et al. / Electrochimica Acta 54 (2009) 5300–5306
Fig. 9 represents the electrochemical signal recorded by square
National Centre For Scientific Research) for financial support of this
work.
wave voltammetry for successive electrolysis time: 0, 30, 60 and
130 h. An important decrease of the current density can be noted,
showing that the metallic ions concentration decreases during the
extraction.
After 130 h, the signal recorded is approaching zero. An ICP
analysis indicated that 99.4% of the samarium fluoride has been
extracted.
References
[1] D. Lambertin, J. Lacquement, S. Sanchez, G. Picard, Proc. ATALANTE, Nimes,
2004.
[2] M. Iizuka, T. Koyama, N. Kondo, R. Fujita, H. Tanaka, J. Nucl. Mater. 247 (1997)
183.
[3] J. Serp, M. Allibert, A. Le Terrier, R. Malmbeck, M. Ougier, J. Rebizant, J.P. Glatz,
J. Electrochem. Soc. 152 (3) (2005) C167.
[4] L. Cassayre, R. Malmbeck, P. Masset, J. Rebizant, J. Serp, P. Soucek, J.P. Glatz, J.
Nucl. Mater. 360 (2007) 49.
4. Conclusion
[5] L. Rault, M. Heusch, M. Allibert, F. Lemort, X. Deschanel, R. Boen, Nucl. Technol.
139 (2002) 167.
[6] O. Conocar, N. Douyere, J. Lacquement, J. Nucl. Mater. 344 (2005) 136.
[7] O. Conocar, N. Douyere, J.P. Glatz, J. Lacquement, R. Malmbeck, J. Serp, Nucl. Sci.
Eng. 153 (2006) 253.
[8] Christophe Nourry, PhD thesis, 2007, Université Paul Sabatier, Toulouse,
France.
[9] P. Chamelot, P. Taxil, B. Lafage, Electrochim. Acta 39 (17) (1994) 2571.
[10] A.D. Graves, D. Inman, Nature 208 (1965) 481.
[11] P. Taxil, Z. Qiao, J. Chim. Phys. 82 (1) (1985) 83.
[12] A. Brenner, Electrodeposition of Alloys, vol. 1, Academic Press, New York,
1963.
[13] P. Taxil, P. Chamelot, L. Massot, C. Hamel, J. Miner. Met. 39 (2003) 177.
[14] C. Nourry, L. Massot, P. Chamelot, P. Taxil, J. New Mater. Electrochem. Syst. 10
(2007) 117.
Cyclic and the square wave voltammograms in Al–Ce and Al–Sm
fluoride solutions showed evidences. Specific electrochemical sys-
tems for the co-reduction of aluminium and lanthanide ions were
identified in the potential range between pure Al and Ln metal
deposition. Each reduction waves observed was correlated to Al–Ln
intermetallic compounds, further identified on the cathode samples
with SEM observation and EDS analysis after a short electrolysis
run. For the Al–Ce system, when increasing cathodic potential, the
cerium content in the alloy increases. For the Al–Sm co-reduction,
only one intermetallic compound was obtained: Al3Sm.
These preliminary experiments allowed the expected forma-
tion of defined compounds to be correlated with specific cathode
potentials.
[15] F. Lantelme, A. Barhoun, E.M. Zahidi, J.H. von Barner, Plasmas Ions 2 (1999)
133.
[16] G. Ett, E.J. Pessine, Electrochim. Acta 44 (1999) 2859.
[17] H. Wendt, K. Reuhl, V. Schwarz, Electrochim. Acta 37 (2) (1992) 237.
[18] H. Wendt, K. Reuhl, V. Schwarz, J. Appl. Electrochem. 22 (1992) 161.
[19] U. Cohen, J. Electrochem. Soc. 130 (7) (1983) 1480.
[20] M. Gibilaro, PhD thesis, “Co-réduction électrochimique de l’aluminium et des
lanthanides en milieu fluorures fondus; application au traitement pyrochim-
ique des effluents nucléaires”, Toulouse, France, 2008.
[21] L. Massot, P. Chamelot, P. Taxil, Electrochim. Acta 50 (2005) 5510.
[22] M. Gibilaro, L. Massot, P. Chamelot, P. Taxil, J. Alloys Compd. (2008), doi:10.1016/
j. jallcom.2008.03.115.
Extraction electrolyses using these potential data were success-
fully performed with extraction efficiencies of 99.5% for cerium
and 99.4% for samarium. The co-reduction technique is promis-
ing for other lanthanide elements; thus this work opens the way
to extract other fission products from nuclear wastes by using the
co-reduction technique with aluminium.
Acknowledgements
[23] K.H.J. Buschow, J.H.N. Van Vucht, Philips, Res. Rep. 22 (1967) 233.
[24] K.H.J. Buschow, A.S. Van der Goot, J. Less-Common Met. 24 (1971) 117.
The authors express their thanks to the PCR RSF Thorium
and GDR Paris from the PACEN program (French program from