Electrochemical formation of Sm-Ni alloy films in a molten LiCl-KCl-SmCl3 system

Article abstract of DOI:10.1016/S0013-4686(01)00470-4

The electrochemical formation of Sm-Ni alloys was investigated in a molten LiCl-KCl-SmCl₃ (0.5 mol%) system at 723 K. The cyclic voltammogram for a Mo electrode showed the reduction wave from Sm(III) to Sm(II) at 1.60 V (vs Li⁺/Li), but no reduction wave from Sm(II) to Sm metal. For a Ni electrode, small cathodic currents were observed at potentials more negative than 0.10 V, which indicated the formation of Sm-Ni alloy. The formation of an SmNi₂ phase was confirmed by XRD analysis of a sample prepared at 0.10 V for 72 h. The thickness of the SmNi₂ film was estimated to be approximately 100 nm. A much thicker SmNi₂ film (～20 μm) was obtained by cathodic galvanostatic electrolysis at 50 mA cm^-2 in a time period as short as 1 h. Since Li metal was codepositing during the electrolysis and the SmNi₂ film was rapidly formed, this electrochemical formation method was termed the 'Li codeposition method'. The formed SmNi₂ film was changed to other alloy phases by anodic potentiostatic electrolysis. The formation potentials of SmNi₅, SmNi₃ and SmNi₂ were found to be 1.20, 0.65 and 0.29 V, respectively.

Full text of DOI:10.1016/S0013-4686(01)00470-4

Electrochimica Acta 46 (2001) 2537–2544
www.elsevier.nl/locate/electacta
Electrochemical formation of Sm–Ni alloy films in a molten
LiCl–KCl–SmCl3 system
Takahisa Iida, Toshiyuki Nohira, Yasuhiko Ito *,1
Department of Fundamental Energy Science, Graduate School of Energy Science, Kyoto Uni6ersity, Sakyo-ku,
Kyoto 606-8501, Japan
Received 20 October 2000; received in revised form 12 February 2001
Abstract
The electrochemical formation of Sm–Ni alloys was investigated in a molten LiCl–KCl–SmCl3 (0.5 mol%) system
at 723 K. The cyclic voltammogram for a Mo electrode showed the reduction wave from Sm(III) to Sm(II) at 1.60
V (vs Li+/Li), but no reduction wave from Sm(II) to Sm metal. For a Ni electrode, small cathodic currents were
observed at potentials more negative than 0.10 V, which indicated the formation of Sm–Ni alloy. The formation of
an SmNi2 phase was confirmed by XRD analysis of a sample prepared at 0.10 V for 72 h. The thickness of the SmNi2
film was estimated to be approximately 100 nm. A much thicker SmNi2 film (ꢀ20 mm) was obtained by cathodic
galvanostatic electrolysis at 50 mA cm−2 in a time period as short as 1 h. Since Li metal was codepositing during the
electrolysis and the SmNi2 film was rapidly formed, this electrochemical formation method was termed the ‘Li
codeposition method’. The formed SmNi2 film was changed to other alloy phases by anodic potentiostatic electrolysis.
The formation potentials of SmNi5, SmNi3 and SmNi2 were found to be 1.20, 0.65 and 0.29 V, respectively. © 2001
Elsevier Science Ltd. All rights reserved.
Keywords: LiCl–KCl; Molten salt; Sm–Ni alloy; Electrolysis; Li codeposition method
1. Introduction
process has been investigated as a new formation
method of the rare earth–transition metal alloy films in
the authors’ laboratory [3–6]. This process has the
following advantages:
1. phases of the alloy films can be controlled by elec-
trochemical parameters (e.g. potential, current
density);
2. the alloy films can be formed even on a large and/or
complex shaped substrate;
3. mass production is easy.
From this background, the authors have been con-
ducting a series of formations of Sm–transition metal
alloy films by the molten salt electrochemical process.
This paper describes the results of electrochemical for-
mation of various types of Sm–Ni alloys in a molten
LiCl–KCl–SmCl3 system. In the LiCl–KCl–SmCl3
system, the Sm(III) ion is reported to be reduced to the
Recently, rare earth–transition metal alloys have
been attracting much research with their excellent mag-
netic, hydrogen absorbing and catalytic properties. In
particular, several Sm–transition metal alloys, e.g.
Sm2Fe17Nx, SmCo5 and Sm2Co17, are well known as
super permanent magnets [1]. If these alloys can be
formed as films with arbitrary shapes and sizes, various
applications will be possible, for instance this will con-
tribute significantly to micro machine technology [2].
On the other hand, the molten salt electrochemical
* Corresponding author. Tel.: +81-75-753-5827; fax: +81-
75-753-5906.
E-mail address: y-ito@energy.kyoto-u.ac.jp (Y. Ito).
1 ISE member.
0013-4686/01/$ - see front matter © 2001 Elsevier Science Ltd. All rights reserved.
PII: S0013-4686(01)00470-4

2538
T. Iida et al. / Electrochimica Acta 46 (2001) 2537–2544
to Sm(0) to form Sm–transition metal alloys on a
transition metal substrate.
2. Experimental
Fig. 2 shows the experimental apparatus. The LiCl–
KCl eutectic (LiCl:KCl=58.5:41.5 mol%; reagent
grade, Wako Pure Chemical Co., Ltd.) was contained
in a high purity alumina crucible (99.5 wt.% Al2O3;
SSA-S grade, Nikkato Co., Ltd.) and dried under vac-
uum for more than 72 h at 473 K to remove water.
Pre-electrolysis was carried out in order to remove
residual water and some metal impurities. All experi-
ments were performed under Ar atmosphere at 723 K.
Anhydrous SmCl3 (99.9 wt.%; High Purity Chemical
Co., Ltd.) was added directly to the melts as a Sm(III)
ion source.
Fig. 1. Principle of the molten salt electrochemical process for
The working electrode was a Ni plate (5 mm×20
mm×0.2 mmt; 99.7%, Nilaco Co., Ltd.) or a Mo plate
(5 mm×20 mm×0.2 mmt; 99.95%, Nilaco Co., Ltd.).
All working electrodes were electropolished in a 10 mol
l−1 H2SO4 aqueous solution before electrochemical
measurements. A vitreous carbon rod (ƒ 6 mm; Tokai
Carbon Co., Ltd.) was used as a counter electrode
when it was used as an anode. An Al plate (10 mm×20
mm×0.2 mmt; 99.2%, Nilaco Co., Ltd.) was used as a
counter electrode when it was used as a cathode. The
reference electrode was an Ag wire immersed in LiCl–
KCl eutectic containing 1 mol% AgCl in a Pyrex tube
provided with a diaphragm at the bottom. The poten-
tial of this reference electrode was calibrated with refer-
ence to that of a Li+/Li electrode, which was prepared
by electrodepositing lithium metal on Ni wire. All
potentials in this paper are given with reference to this
Li+/Li electrode potential. A potentiogalvanostat (HA-
501G; Hokuto Denko Corp.) and function generator
(HB-105; Hokuto Denko Corp.) connected to a per-
sonal computer (PC9801DA; NEC Corp.) were used
for cyclic voltammetry and chronopotentiometry.
The Sm–Ni alloy samples were prepared by galvano-
static electrolysis or potentiostatic electrolysis using the
same potentiogalvanostat. Since the Sm alloys are
highly oxidizable, they cannot be washed by water.
Therefore, after the electrolysis, the samples were
washed by 1,2-ethanediol (99.7%; Wako Pure Chemical
Co., Ltd.), which scarcely reacts with the Sm alloys, to
remove salts and lithium metal. These samples were
analyzed by XRD (JDX-8030; JEOL Ltd.) using the
Cu–Ka line at 40 kV 40 mA. The XRD patterns were
compared with patterns of the Sm–Ni alloys calculated
by the PowderCell software [8] using the reported crys-
tallographic data [9]. SEM and EPMA (Hitachi; S-
2300) were used to measure the thickness of the alloys
and the concentration profiles of Sm and Ni.
Sm–transition metal alloy formation.
Sm(II) ion at about 1.60 V (vs Li+/Li), and the reduc-
tion potential of the Sm(II) ion to Sm metal is thought
to be more negative than the Li deposition potential [7].
However, electrochemical formation of Sm–transition
metal alloys was expected because the alloy formation
potential was thought to shift to a more positive poten-
tial than the elementary Sm deposition potential. Fig. 1
shows the principle of the molten salt electrochemical
process for the expected Sm–transition metal alloy
formation. When SmCl3 is added to molten LiCl–KCl,
SmCl3 dissociates to form Sm(III) and Cl− ions. By
cathodic electrolysis, Sm(III) ions will be reduced to
Sm(II) ions, and Sm(II) ions are expected to be reduced
Fig. 2. Schematic drawing of the experimental apparatus. (1)
LiCl–KCl–SmCl3 (added 0.5 mol%), T=723 K, (2) Ar gas
inlet, (3) thermocouple, (4) working electrode (Mo or Ni), (5)
counter electrode (vitreous carbon), (6) reference electrode
(Ag+/Ag), (7) Li+/Li electrode (Ni wire) and (8) Ar gas outlet.

T. Iida et al. / Electrochimica Acta 46 (2001) 2537–2544
2539
electrode was used as a reference to compare with the
Ni electrode because no alloys or intermetallic com-
pounds exist for the Mo–Sm binary system at 723 K
[11]. On the cathodic sweep, cathodic currents were
observed from approximately 1.60 V. They are thought
to correspond to the following reaction [7].
Sm(III)+e− X Sm(II)
(1)
After that, large cathodic currents were observed
from 0 V, which is considered to correspond to the
deposition of Li metal. After reversal of the sweep
direction at −0.07 V, an anodic current peak corre-
sponding to dissolution of the Li metal was observed.
On the other hand, the cyclic voltammogram of a Ni
electrode is shown as a solid line in Fig. 3. On the
cathodic sweep, cathodic currents based on reaction (1)
were also observed. After that almost similar behavior
to the Mo electrode was observed except the slightly
larger cathodic currents at a more negative potential
than 0.1 V. These cathodic currents are considered to
be associated with the formation of Sm–Ni alloy. On
the anodic sweep, anodic current peaks were observed
at 0.29 and 1.58 V in addition to the anodic current
peak of Li metal dissolution. These anodic current
peaks are thought to be associated with the dissolution
of Sm from the different Sm–Ni alloy phases.
Fig. 3. Cyclic voltammograms for Ni and Mo electrodes in a
molten LiCl–KCl–SmCl3 (0.5 mol%) system at 723 K. Scan
rate: 0.002 V s−1
.
3. Results and discussion
3.1. Cyclic 6oltammetry
3.2. Formation of Sm–Ni alloy
In order to investigate Sm–Ni alloy formation, cyclic
voltammetry was conducted in a molten LiCl–KCl–
SmCl3 (added 0.5 mol%) system at 723 K. The broken
line in Fig. 3 shows the cyclic voltammogram of a Mo
electrode at a scan rate of 0.002 V s−1. The Mo
Based on the result of cyclic voltammetry, an alloy
sample was prepared by potentiostatic electrolysis using
a Ni electrode at 0.10 V for 72 h. Fig. 4 shows the
XRD pattern of the surface of the sample. The ob-
served peaks were identified as SmNi2 and Ni substrate.
However, the intensities of the Ni peaks were predomi-
nant over those of the SmNi2 peaks, which shows that
the thickness of the SmNi2 film on the Ni substrate was
very thin (estimated to be approximately 100 nm).
Thus, it was found that the formation rate of SmNi2
was very small. It is worth noting the dependence of the
current on time. The cathodic current density was
around 40 mA cm−2 at the beginning. However, it
decreased rapidly to about 20 mA cm−2 in 5 min, and
then reached about 10 mA cm−2 after 1 h. It scarcely
changed until the end of the electrolysis. The current
density of 10 mA cm−2 is thought to correspond
mainly to current for reaction (1).
Another sample was prepared by cathodic galvanos-
tatic electrolysis at 50 mA cm−2 for 1 h. The electrode
potential was about −0.03 V during the electrolysis.
After electrolysis, the existence of liquid metal on the
surface was confirmed by direct observation. This liquid
metal is thought to consist mainly of lithium metal,
because potassium metal does not deposit from this
system and Sm metal is solid at this temperature.
However, a small amount of Sm may be contained in
Fig. 4. XRD pattern of the Sm–Ni alloy film formed at 0.10
V for 72 h. SmNi2: cubic unit cell with a=0.723 nm. Ni: cubic
unit cell with a=0.352 nm.

2540
T. Iida et al. / Electrochimica Acta 46 (2001) 2537–2544
the former case (potentiostatic electrolysis at 0.10 V).
Since Li metal was codepositing for the latter case
(cathodic galvanostatic electrolysis at 50 mA cm−2
)
and the SmNi2 film was rapidly formed, this electro-
chemical formation method was termed the ‘Li codepo-
sition method’.
3.3. Phase control of Sm–Ni alloy film
According to the phase diagram of the Sm–Ni sys-
tem (Fig. 7 [11]), there may exist eight Sm–Ni inter-
metallic compounds at 723 K. When Sm is dissolved
anodically from an SmNi2 electrode, there is the possi-
bility that the SmNi2 phase changes to other Sm–Ni
alloy phases with lower Sm concentrations. In order to
confirm this, chronopotentiometry was conducted for
the SmNi2 electrode at an anodic current density of 1
mA cm−2. The SmNi2 electrode was prepared by ca-
thodic galvanostatic electrolysis at 50 mA cm−2 for 20
min. The obtained chronopotentiogram is shown in
Fig. 8. For 1300 s in the beginning, the potential kept
at 0 V, which is thought to be due to the presence of
the codeposited Li metal on the electrode. After this,
potential plateaus were observed at (1) 0.29, (2) 0.65
and (3) 1.20 V, respectively. These plateaus are consid-
ered to correspond to coexisting phase states of Sm–Ni
alloys. Based on this result, samples were prepared by
the following procedures. Firstly, SmNi2 electrodes
were made by cathodic galvanostatic electrolysis at 50
mA cm−2 for 1 h (Li codeposition method). Then,
anodic dissolutions of Sm were conducted for 2 h by
potentiostatic electrolysis at 0.1 (sample 1), 0.4 (sample
2), 0.8 (sample 3) and 1.3 V (sample 4), respectively.
These potential values were determined by taking into
account the potential plateaus before and after the
potential jump indicating the complete phase change.
The XRD pattern of sample 1 (obtained at 0.1 V) is
shown in Fig. 9. All peaks were identified as SmNi2,
which suggests that SmNi2 phase is stable at more
negative potentials than 0.29 V.
Fig. 5. XRD pattern of the Sm–Ni alloy film formed at −50
mA cm−2 for 1 h.
the lithium considering that the solubility of Sm in Li is
around 3 atm% [10]. The XRD pattern of the sample
shown in Fig. 5 was identified as SmNi2. The existence
of Li was not observed in the XRD pattern, which can
be explained by Li being removed together with the salt
during the washing treatment of the electrode. Fig. 6
shows a cross-sectional SEM image and the concentra-
tion profiles of Sm and Ni by EPMA line analysis for
the sample. The observed alloy layer with approxi-
mately 20 mm thickness is regarded as a uniform SmNi2
layer, considering together with the XRD result.
From these results, it is found that the formation rate
of SmNi2 film is much larger for the latter case (ca-
thodic galvanostatic electrolysis at 50 mA cm−2) than
Fig. 10 shows the XRD pattern of sample 2 (ob-
tained at 0.4 V). It was found that the SmNi2 phase
changed completely to an SmNi3 phase. Therefore, the
potential plateau at 0.29 V is considered to correspond
to the following reaction:
2SmNi3+Sm(II)+2e− X 3SmNi2
(2)
Fig. 11 shows the cross-sectional SEM image and the
concentration profiles of Sm and Ni for sample 2. The
thickness of the alloy layer was approximately 20 mm.
Since the original SmNi2 layer had about 20 mm thick-
ness, it changed little during phase transformation from
SmNi2 to SmNi3.
The XRD pattern shown in Fig. 12 identified sample
3 (obtained at 0.8 V) as SmNi5. Thus, the potential
plateau at 0.65 V is considered to be due to the
following reaction:
Fig. 6. Cross-sectional SEM image and concentration profiles
of Sm and Ni for the sample formed at −50 mA cm−2 for 1
h.

T. Iida et al. / Electrochimica Acta 46 (2001) 2537–2544
2541
Fig. 7. Phase diagram of the Sm–Ni system [11].
3SmNi5+2Sm(II)+4e− X 5SmNi3
(3)
ever, they were not identified in any samples under our
experimental conditions. This might be explained by the
fact that the formation rates of these alloys were very
low.
According to the cross-sectional SEM image and the
EPMA line analysis for sample 3 (Fig. 13), the thick-
ness of alloy layer was approximately 20 mm.
The XRD pattern of sample 4 (obtained at 1.30 V) is
shown in Fig. 14, in which only the pattern of a-Ni was
observed. Accordingly, the potential plateau at 1.20 V
is regarded to correspond to the following reaction:
5a-Ni+Sm(II)+2e− X SmNi5
(4)
Fig. 15 shows the cross-sectional SEM image and the
concentration profiles of Sm and Ni for sample 4. From
the EPMA line analysis, it was found that no Sm
existed in the film.
Additionally, samples were prepared from SmNi2
electrodes at 0.30, 0.50, 0.70 and 1.00 V, respectively,
for 2 h. XRD analysis of the samples showed that
SmNi3 was formed at 0.30 and 0.50 V, and that SmNi5
formed at 0.70 and 1.00 V. These results are consistent
with the speculated reactions (2)–(4) and the estimated
corresponding potentials. The above experimental re-
sults can be summarized as in Table 1. Accordingly,
formation reactions of Sm–Ni alloys and the corre-
sponding potentials were clarified as is summarized in
Table 2. On the other hand, the phase diagram of the
Sm–Ni system shown in Fig. 7 suggests the possibility
of formation of Sm2Ni7, Sm5Ni19 and Sm2Ni17. How-
Fig. 8. Chronopotentiogram for the SmNi2 electrode at 1 mA
cm−2 in a molten LiCl–KCl–SmCl3 (0.5 mol %) system at
723 K. The SmNi2 electrode was prepared by cathodic gal-
vanostatic electrolysis of a Ni electrode at 50 mA cm−2 for 20
min.

2542
T. Iida et al. / Electrochimica Acta 46 (2001) 2537–2544
Fig. 11. Cross-sectional SEM image and concentration profiles
of Sm and Ni for sample 2.
Fig. 9. XRD pattern of sample 1 prepared by anodic electrol-
ysis of the SmNi2 electrode at 0.1 V for 2 h. The SmNi2
electrode was prepared by cathodic galvanostatic electrolysis
of a Ni electrode at −50 mA cm−2 for 1 h.
4. Conclusions
Electrochemical formation of Sm–Ni alloys was
studied in a molten LiCl–KCl–SmCl3 (0.5 mol%) sys-
tem at 723 K. The results obtained through this study
can be summarized as follows.
The cyclic voltammogram for a Mo electrode showed
the reduction wave from Sm(III) to Sm(II) at 1.60 V (vs
Fig. 12. XRD pattern of sample 3 prepared by anodic electrol-
ysis of the SmNi2 electrode at 0.8 V for 2 h. SmNi5: hexagonal
unit cell with a=0.493 nm, c=0.396 nm.
Fig. 10. XRD pattern of sample 2 prepared by anodic electrol-
ysis of the SmNi2 electrode at 0.4 V for 2 h. SmNi3: hexagonal
unit cell with a=0.501 nm, c=2.46 nm.
Fig. 13. Cross-sectional SEM image and concentration profiles
of Sm and Ni for sample 3.

T. Iida et al. / Electrochimica Acta 46 (2001) 2537–2544
2543
Table 2
Formation reactions and the corresponding potentials for the
Sm–Ni alloys in a molten LiCl–KCl–SmCl3 (0.5 mol%) system
at 723 K
Reaction
Potential/V vs. Li+/Li
2SmNi3+Sm(II)+2e− X 3SmNi2 0.29
3SmNi5+2Sm(II)+4e− X 5SmNi3 0.65
5a-Ni+Sm(II)+2e− X SmNi5
1.20
Li+/Li), but no reduction wave from Sm(II) to Sm
metal. For a Ni electrode, small cathodic currents were
observed at potentials more negative than 0.10 V,
which indicated the formation of Sm–Ni alloy.
The formation of an SmNi2 phase was confirmed by
XRD analysis of the sample prepared at 0.10 V for 72
h. The thickness of the SmNi2 film was estimated to be
approximately 100 nm. A much thicker SmNi2 film
(ꢀ20 mm) was obtained by cathodic galvanostatic elec-
trolysis at 50 mA cm−2 for 1 h. Since Li metal was
codepositing during the electrolysis and the SmNi2 film
was rapidly formed, the method was termed the ‘Li
codeposition method’.
Fig. 14. XRD pattern of sample 4 prepared by anodic electrol-
ysis of the SmNi2 electrode at 1.3 V for 2 h.
In the chronopotentiogram for the SmNi2 electrode
at an anodic current density of 1 mA cm−2, potential
plateaus were observed at 0.29, 0.65V and 1.20 V,
respectively, which were considered to correspond to
coexisting phase states of Sm–Ni alloys. The SmNi2
electrodes changed to SmNi3, SmNi5 and a-Ni phases
by anodic potentiostatic electrolysis depending on the
potential. Formation reactions of Sm–Ni alloys and the
corresponding potentials have been clarified.
Fig. 15. Cross-sectional SEM image and concentration profiles
of Sm and Ni for sample 4.
Acknowledgements
This work was carried out with support from a
Grant-in-Aid from the Japanese Ministry of Education,
Science, Sports and Culture. We greatly appreciate Dr
Gert Nolze for sending the PowderCell program.
Table 1
Potentials and the identified phases for the samples prepared
by potentiostatic electrolysis of the SmNi2 electrodes for 2 h in
a LiCl–KCl–SmCl3 (0.5 mol %) system at 723 K
Potential/V vs. Li+/Li
Identified phase
References
0.10
0.30
0.40
0.50
0.70
0.80
1.00
1.30
SmNi2
SmNi3
SmNi3
SmNi3
SmNi5
SmNi5
SmNi5
a-Ni
[1] J.M.D. Coey, Solid State Commun. 102 (1997) 101.
[2] S. Yamashita, J. Yamasaki, M. Ikeda, N. Iwabuchi, J.
Appl. Phys. 70 (1991) 6627.
[3] G. Xie, K. Ema, Y. Ito, M.S. Zhao, J. Appl. Electrochem.
23 (1993) 753.
[4] K. Tateiwa, M. Tada, Y. Ito, Hyomen Gijutsu (J. Finish-
ing Soc. Jpn.) 46 (1995) 673 in Japanese.
[5] Y. Ito, T. Nohira, Hyomen Gijutsu (J. Finishing Soc.
Jpn.) 49 (1998) 336 in Japanese.

2544
T. Iida et al. / Electrochimica Acta 46 (2001) 2537–2544
[6] Y. Ito, T. Nohira, Electrochim. Acta 45 (2000)
2611.
lographic Data for Intermetallic Phases, ASM Interna-
tional, 1991.
[7] K.E. Johnson, J.R. Mackenzie, J. Electrochem. Soc. 116
[10] I.N. Ganiev, A.D. Shamsiddinov, Kh.M. Nazarov, M.D.
Badalov, Metally 4 (1996) 163 in Russian.
[11] T.B. Massalski, J.L. Murray, L.H. Benett, H. Baker,
Binary alloy Phase Diagrams, American Society for
Metals, 1990.
(1969) 1697.
[8] W. Kraus, G. Nolze, J. Appl. Crystallogr. 29 (1996)
301.
[9] P. Villars, L.D. Calvert, Pearson’s Handbook of Crystal-
.
.