H. Nakajima et al. / Electrochimica Acta 53 (2007) 24–27
25
addition, the electrodeposition of tungsten was applied to coat-
ing a contact probe pin manufactured by the conventional LIGA
process.
2
. Experimental
Experiments were conducted in an argon glove box with a
gas circulating purifier (MIWA MFG Co., Ltd.). ZnCl2 (99.9%,
Wako Pure Chemical Industries, Ltd.), NaCl and KCl (99.5%
each, WakoPureChemicalIndustries, Ltd.)werewellmixedinto
an eutectic composition (ZnCl2:NaCl:KCl = 0.60:0.20:0.20, in
◦
mole fraction) and dried in a furnace under vacuum at 130
C for 3 days or more. After that, it was melted at 250
◦
C
in a Pyrex beaker installed in a four-necked separable flask
on a heating plate. WO3 (99.5%, Wako Pure Chemical Indus-
tries, Ltd.) and KF (99%, Wako Pure Chemical Industries, Ltd.)
were added as tungsten and fluoride ion sources, respectively.
A schematic drawing of the experimental apparatus has been
shown elsewhere [7]. The working electrode was a nickel plate
Fig. 1. Cyclic voltammograms at nickel electrodes in ZnCl2–NaCl–KCl–KF
(
4 mol% added) melt (gray curve) and ZnCl2–NaCl–KCl–KF (4 mol%
◦
added)–WO3(0.54 mol% added) melt (black curve) at 250 C. Scan rate :
0
−
1
.05 V s
.
(
99.7%, 5 mm × 10 mm × 0.2 mm, Furuuchi Chemical Corp.)
that was fully immersed in the melt. Prior to the experiment, the
nickel plate was electrochemically polished in a sulfuric acid,
and then immersed in an acid cleaner (Kizai Corp., Kokeisan
B) containing NaHF2 to remove surface oxides. A zinc wire
deposits have been obtained by the addition of KF in our pre-
vious studies, in which WCl4, MoCl3 and MoCl were used
5
as refractory metal ion sources [7–9]. Before investigating the
(
KF + WO3) added melt, cyclic voltammetry was carried out
(
99.99%, 0.5 mm, Nilaco Corp.) immersed in the melt was
for the only KF added melt. A gray curve in Fig. 1 shows a
cyclic voltammogram obtained at the nickel plate electrode in
ZnCl2–NaCl–KCl–KF (4 mol% added) melt. The shape of the
cyclic voltammogram was identical to that obtained in the orig-
inal ZnCl2–NaCl–KCl melt. Cathodic and anodic current peaks
around 0 V versus Zn(II)/Zn are attributed to the deposition and
dissolution of metallic zinc, respectively. Meanwhile, a cathodic
current peak around 0.03 V and an anodic current peak around
used as the reference electrode. The counter electrode was a
glassy carbon rod (Ø3 mm, GC-20, Tokai Carbon Co., Ltd.).
A chromel–alumel thermocouple was used for the temperature
measurement. Cyclic voltammetry with positive feedback IR
compensation and potentiostatic electrolysis were performed
using an electrochemical measurement system (Hokuto Denko
Co., Ltd., HZ-5000). The obtained samples were immersed in
acetone, and then rinsed with distilled water to remove adherent
salts.
The deposits were characterized by scanning electron
microscopy (SEM, Hitachi, Ltd., S-2600H), scanning ion
microscopy (SIM) with focused ion beam (FIB, FEI Corp.,
FIB200) cross sectioning, energy dispersive X-ray analysis
0.12 V are attributed to the formation and dissolution of Ni–Zn
alloy [7], respectively.
In the (KF + WO3) added melt, it was observed that WO3
dissolved in the melt and the color of the melt changed from
colorless to gray. The dissolution of WO3 probably proceeded
by the formation of oxo-fluoro tungsten complex ions. A certain
portion of WO3, however, remained undissolved. Thus, the melt
is regarded as saturated with WO3. A black plot in Fig. 1 shows
a cyclic voltammogram measured for the nickel plate electrode
in the ZnCl2–NaCl–KCl–KF (4 mol% added)–WO3 (0.54 mol%
(
EDX, Horiba Co., Ltd., EMAX ENERGY EX-200), X-ray
diffraction (XRD, Rigaku Industrial Co., Ltd., Multi Flex) with
Cu K ␣ radiation and X-ray photoelectron spectroscopy (XPS,
Shimadzu Co., Ltd., ESCA-3200).
◦
added) melt at 250 C. A new cathodic current is observed in the
3
. Results and discussion
potential region of 0.05–0.45 V, suggesting electrodeposition of
metallic tungsten. Neverthless, the current peak which is often
observed for the electrodeposition of metals in high temperature
molten salts [11,12] did not appear in the present case. The shape
change of the voltammogram in the potential region of 0–0.03 V
is explained by the deposition of tungsten, suppressing both the
deposition of pure zinc and the formation of Ni–Zn alloy.
3
.1. Cyclic voltammetry
Cyclic voltammetry was performed using a nickel elec-
trode after the addition of WO3 (0.54 mol% added) in
ZnCl2–NaCl–KCl melt at 250 C. The shape of the voltammo-
gram did not change by the addition of WO3. It was observed
that WO3 hardly dissolved in ZnCl2–NaCl–KCl melt. Accord-
ingly, no tungsten deposit was obtained by electrolysis in the
ZnCl2–NaCl–KCl–WO3 melt.
◦
3.2. Potentiostatic electrolysis and characterization of
deposit
Since fluoride melts generally have higher solubilities of
metal oxides than chloride melts, KF was added as a fluoride
ion source to increase the solubility of WO3. Moreover, fine
Since electrodeposition of tungsten was speculated to occur
in the potential region lower than 0.45 V, potentiostatic elec-