Initial stages of electrodeposition of metal nanowires in nanoporous templates

Article abstract of DOI:10.1016/j.electacta.2007.04.122

This paper describes the initial stages of electrochemical growth of Cu and Ni nanowires in polycarbonate (PC) track-etched membranes as the template. The diameters of the wires ranged from 50 to 200 nm. A thin Pt-Pd layer (～30 nm thick) was sputter-deposited onto one side of the membrane and used as the cathode. The layer was not thick enough to seal the pore mouths. Cathodic current in the early stages of both Cu and Ni electrodeposition abruptly decreased after a period proportional to the pore diameter. Growing Cu grains plugged the pore openings causing the current to decrease, while the Ni deposition initially yielded a hollow tube in each pore resulting in a nanostructure transition of the tube to the wire at the growth front and a decrease in the current.

Full text of DOI:10.1016/j.electacta.2007.04.122

Electrochimica Acta 53 (2007) 205–212
Initial stages of electrodeposition of metal
nanowires in nanoporous templates
M. Motoyamaa, Y. Fukunakaa,∗, T. Sakkab, Y.H. Ogatab
a Department of Energy Science and Technology, Kyoto University, Yoshida-honmachi, Sakyo, Kyoto 606-8501, Japan
b Institute of Advanced Energy, Kyoto University, Uji, Kyoto 611-0011, Japan
Received 28 June 2006; received in revised form 24 April 2007; accepted 25 April 2007
Available online 10 May 2007
Abstract
This paper describes the initial stages of electrochemical growth of Cu and Ni nanowires in polycarbonate (PC) track-etched membranes as the
template. The diameters of the wires ranged from 50 to 200 nm. A thin Pt–Pd layer (∼30 nm thick) was sputter-deposited onto one side of the
membrane and used as the cathode. The layer was not thick enough to seal the pore mouths. Cathodic current in the early stages of both Cu and
Ni electrodeposition abruptly decreased after a period proportional to the pore diameter. Growing Cu grains plugged the pore openings causing
the current to decrease, while the Ni deposition initially yielded a hollow tube in each pore resulting in a nanostructure transition of the tube to the
wire at the growth front and a decrease in the current.
© 2007 Elsevier Ltd. All rights reserved.
Keywords: Cu; Ni; Electrodeposition; Nanowire; Nanotube; Nanoporous template
Electrochemical processing with using nanoporous templates
allows us to prepare metallic nanowire arrays with various con-
trolled dimensions [1–5]. The ability to control the composition
along the length of the nanowire is an important feature of the
electrochemical method [6,7]. Understanding of the coupling
between ionic mass transport phenomena and morphological
variation within nanoscopic pores is indispensable to allow a
more quantitative control of nanoscale materials formed by the
electrochemical method. Especially, the control of nucleation
and growth is expected to be widely applicable to the advanced
designs and thus necessary [8–12]. Solution phase synthesis of
nanoparticles is one example pointing out the importance of
the above respect [8–10]. Growth model for electrochemical
nucleation and growth were initially developed in the 1980s.
Scharifker and Hills proposed the models for a diffusion-limited
process, based on analysis of the current transient characteristic
in the early stages of electrodeposition [13]. However, the liter-
ature reported the results significantly deviating the predictions
of the models [14,15], hence the experimental understanding is
still important. Furthermore, the models have been developed to
understand the nucleation and growth in a semi-infinite space,
but it is not applicable to complex structures with small length
scales, such as nanoscopic channels [16–19]. A more detailed
understanding of nucleation and growth during electrodeposi-
tion into nanoscopic pores is critical for designing deposition
process for obtaining desired nanostructures.
In this paper we report on the initial stages of electrochemical
growth of Cu and Ni nanowires. The electrochemical deposi-
tion begins at a sputter-deposited Pt–Pd layer (∼30 nm thick) as
the cathode on a polycarbonate (PC) membrane template. The
presentsputteringmethodentailsopeningtheporemouthsdueto
muchsmallersizeofthesputteredPt–Pdparticlesthanthemouth
diameter [20]. These pore openings can allow us to electroplate
a hollow tube in each pore. Indeed, however, the deposition
process of metal only along the pore walls is not sufficiently con-
trolled [21–24] relative to an electroless plating method [19,25].
This study has been conducted with the aim of understanding the
electrochemical growth mechanism within nanoscopic pores.
Here we show the difference of the initially formed structures
betweenCuandNielectrodepositionintothePCmembranetem-
plate. ThestabilityoftheCunanotubegrowthislostimmediately
after the beginning of the electrodeposition, while the Ni nan-
otube growth lasts for longer deposition times. The importance
of the deposited grain size will be discussed.
∗
Corresponding author. Tel.: +81 75 753 5415; fax: +81 75 753 4719.
E-mail address: fukunaka@energy.kyoto-u.ac.jp (Y. Fukunaka).
0013-4686/$ – see front matter © 2007 Elsevier Ltd. All rights reserved.
doi:10.1016/j.electacta.2007.04.122

206
M. Motoyama et al. / Electrochimica Acta 53 (2007) 205–212
Fig. 2. Cathodic current transients during Cu and Ni electrodeposition into the
PC membrane template with the pores of 200 nm in diameter. The Stages I–IV
in this figure are expressed for the Cu deposition.
Fig. 1. SEM images of the sputter-deposited Pt–Pd layer on the PC membrane
template with the pores of 200 nm in diameter. Scale bar: 1.0 ␮m. (Top) The
sputtered side. (Bottom) The Pt–Pd nanotubes liberated from the PC membrane.
[20,26]. Fig. 2 shows the cathodic current transient during Cu
electrodeposition into the PC template with the pores of 200 nm
in diameter at E = −0.4 V versus a Cu quasi-reference electrode.
The cathodic current drops from 1.2 to 0.25 mA before t = 20 s.
We define this period as Stage I. The subsequent stage, Stage
II, then extends to about t = 260 s with progress of the Cu elec-
trodeposition process along the length of the pores. A profile of
the I–t curve in Stage II is responsible for a pore-size distribu-
tion along the length. A cigar-shaped profile characterized by
a heavy ion irradiation technique was already described in the
previous work [26]. Then, the cathodic current increases to 1 mA
at t = 380 s because the effective electrode surface area enlarges
with progress of the electrodeposited Cu nanowires emerging
from the pores in the lateral direction on the PC membrane sur-
face (Stage III). Finally, an electrodeposited Cu film outside the
top surface of the membrane template is further developed in a
free electrolyte volume in Stage IV.
1. Experimental
Two kinds of electrolyte compositions for Cu and Ni
electrodeposition were adjusted as 0.60 mol L−1 CuSO4,
5 × 10−3 mol L−1 H2SO4 (pH 1.7) and 1.0 mol L−1 NiSO4,
0.19 mol L−1 NiCl2, and 0.62 mol L−1 H3BO3 (pH 3.4), respec-
tively. The pore diameters of the PC membrane templates were
50, 80 (Nomura Micro Science Co. Ltd.), 100, and 200 nm (Toyo
Roshi Kaisha Ltd.). Ion sputter coating of Pt–Pd (Hitachi, E-
1030) was conducted onto one side of the template to prepare
the cathode substrate. The deposited Pt–Pd cathode thickness
was roughly 30 nm according to the company’s operational data.
Fig. 1(top) shows the Pt–Pd surface on the template membrane
with the pores of 200 nm in diameter. It is clearly seen that the
pores are not sealed and the original diameter is maintained.
This sputtering method entails coating the pore walls near the
openings with a thin Pt–Pd layer. The membrane can be dis-
solved away with a dichloromethane solution, and the liberated
Pt–Pd tubes as the cathode are imaged via electron microscopy
[see Fig. 1(bottom)]. The height of the nanotubes is roughly
300–500 nm.
Cathodic current transient for electrodeposition of Ni
nanowires can be also characterized by the four stages [26].
Fig. 2 shows the current transient for the Ni deposition at
E = −1.0 V versus a Ni quasi-reference electrode.
2.1. Cu electrodeposition
Fig. 3(left) shows the cathodic current transients for elec-
trodeposition of the Cu nanowires with different diameters. Each
deposition process was interrupted at t = 1 to 5 s in order to pre-
parethespecimensforSEMobservation. Inallcasesthecathodic
current rapidly increases to 1.3–1.5 mA immediately after t = 0 s,
and thereafter decreases abruptly. The interruption times were
determined by appearance of the above abrupt decrease in the
current. The cathodic current starts to decrease earlier as the pore
diameter decreases. Fig. 3 apparently resembles a typical cur-
rent transient behavior associated with nucleation and growth
process for potentiostatic electrodeposition [11–15,27–31].
Cathodic current transient in the early stages of potentiostatic
electrodeposition is often normalized by the current maximum,
Im, and the deposition time at which the instant current maxi-
mum, tm, is recorded. According to Scharifker and Hills [13],
three-dimensional (3D) diffusion zones numerously formed sur-
A conventional three-electrode type cell was employed as
used in the previous papers [20,26]. Potentiostatic electrodepo-
sition into the PC templates was carried out in a circular area of
2.0 mm in diameter restricted by silicon rubber insulator. Cu and
Ni wires were used as the conventional reference electrodes for
Cu and Ni electrodeposition, respectively. Electrode potential,
E, is measured against each reference electrode. The template
membrane was placed on a platinum sheet as the current collec-
tor such that the membrane face with the Pt–Pd layer was down
to the collector, and they were pressed to keep a tight contact.
2. Results and discussion
Current–time, I–t, transient for electrodeposition of the
nanowire arrays in the PC template is divided into four stages

M. Motoyama et al. / Electrochimica Acta 53 (2007) 205–212
207
Fig. 3. (Left) Cathodic current transients during the Cu electrodeposition into the PC membrane templates with the pores of a: 50 nm, b: 80 nm, c: 100 nm, and d:
200 nm in diameters. (Right) SEM images of the electrodeposited Cu grains on the sputtered Pt–Pd layer. Scale bar: 200 nm.
rounding the nuclei start to interact to form a planar (2D)
diffusion zone parallel to the substrate, and the Im is recorded in
the early stages corresponding to a transition from the 3D to the
2D diffusion transport.
as the several 3D diffusion zones in the internal space of the
pores start to interact to turn to the 2D field. A longer deposition
time should be consumed for the formation of the 2D diffusion
zone in each pore with a larger diameter.
Cu clusters are nucleated uniformly on the Pt–Pd surface
including the edge of the pores (see Fig. 4). The 3D diffusion
zones formed surrounding the Cu nuclei prevail into the pores.
The transient diffusion field will develop along the pore length
The above speculation apparently looks sound. The deposi-
tion times at which the instant current maximums were recorded,
tm, in Fig. 3 ranged from 1 to 3 s. However, these measured
periods are too long to quantitatively support the above spec-
ulation for analyzing the relationship between the Im and the
tm recorded in the present experiments. That is because the
√
Nernst diffusion layer thickness, πDt, reaches to 50–200 nm
within deposition times of 2–30 ␮s for diffusivity of Cu2+ ions,
D = 5 × 10−6 cm−2 s−1 [32]. Another reason must explain the
current transient behavior in Stage I.
Fig. 3(right) shows SEM images of the Pt–Pd side observed
after the Cu deposition with the current transients shown in
Fig. 3(left). The present electrolytic cell design forms a narrow
gap between the membrane face with the Pt–Pd layer and the Pt
sheet. The electrolyte solution permeate the template membrane
resulting in leakage of the electrolyte to the gap, and thereby
electric field lines penetrate through the pore openings at the bot-
tom to reach the Pt–Pd surface. Thus, the Cu deposition occurs
on the Pt–Pd layer and even on the Pt sheet.
With the pores of 50 nm in diameter, numerous Cu particles
smaller than 100 nm are deposited (Fig. 3(right) (a)). A particle
density of 4 × 109 cm−2 is measured from the SEM image. The
Cu particles agglomerate or coalesce along with the progress of
the electrodeposition, but more than a half of the Pt–Pd surface
still remains uncovered with Cu grains.
Fig. 4. Schematic illustration of diffusion zones surrounding Cu nuclei prevail-
ing into the pores. The upper direction is the growth direction of the nanowires.

208
M. Motoyama et al. / Electrochimica Acta 53 (2007) 205–212
Table 1
The larger pore size requires a longer deposition time to cause
Comparison between the calculated amount of electricity, Qc, and the measured
amount of electricity, Qm
the current to start to decrease abruptly. The Cu grains reach the
sizes of 50 to 200 nm. The bare Pt–Pd surface is partially seen,
and the pores are not covered perfectly. There is no observable
morphological difference between (b): 80 nm and (c): 100 nm.
With the pores of (d): 200 nm, the electrodeposited Cu gains
cover all over the Pt–Pd surface. Almost all the grains are larger
than 100 nm. Several dark valleys or deep caldera left among the
Cu particles are created due to the pore openings, although they
are completely covered with the Cu grains.
2r (nm)
Qc (C cm−2
)
Qm (C cm−2
)
50
80
100
200
3.4 × 10−3
1.4 × 10−2
2.6 × 10−2
1.5 × 10−1
(a) 1.3 × 10−2
(b) 5.5 × 10−2
(c) 5.6 × 10−2
(d) 1.6 × 10−1
Hemispherical nuclei with a diameter of 2r and the nucleus density of
N = 4 × 109 cm−2 are assumed.
Averaged sizes of the Cu grains are approximately 50, 80, 80,
and 150 nm corresponding to the SEM images of (a), (b), (c), and
(d), respectively. ThecomparisonbetweenFig. 3(left)and(right)
suggests that the abrupt decrease in the cathodic current is asso-
ciated with the plugging process of the pore openings with the
Cu grains. That is, the cathodic current starts to decrease once the
size of the Cu grains reaches to the pore mouth diameter. Thus,
the observed tm is not explained by the diffusion-limited growth
as described above, but corresponds to a moment at which the
pore opening is plugged with the growing grains occluding cur-
rent path to the Pt–Pd layer as well as to the Pt sheet (see Fig. 5).
The Cu electrodeposition takes place on the Pt–Pd surface and
the Pt sheet, but it does only inside the pores once the pore
mouths are covered up with the Cu grains. That induces the
abrupt decrease in the cathodic current in Stage I.
of the nuclei may be introduced around the edge of the pores, but
such contribution is neglected here. The calculated amount of
electricity, Qc, and 2r are listed in Table 1. The right end column
shows the amount of electricity, Qm, recorded by the time when
the cathodic current starts to abruptly decrease in Fig. 3(left).
Such deposition times of (a) to (d) correspond to 0.3, 1.3, 1.3,
and 3.0 s, respectively. The Qc agrees with the Qm with the pores
of 200 nm. However, the smaller the pore diameter, the larger the
deviation of the Qc from the Qm. An extra amount of Cu, which
is not taken into account for the calculation, was deposited on
the Pt sheet, and the Qm is suspected to become less constant
with the smaller pores. These situations support that Table 1
reasonably provides the presumption that the abrupt decrease in
the current is induced as the growing grains start to cover up the
pores at the bottom.
An amount of electricity consumed by the above grain growth
process is estimated by the following Avrami equation [33].
ꢀ
ꢁ
ꢀ
ꢁ
2.2. Ni electrodeposition
Vex
V0
2
f = 1 − exp
−
= 1 − exp − πNr2
(1)
3
Fig. 6(left) shows the cathodic current transients for the elec-
trodeposition of the Ni nanowires with different diameters. After
the cathodic current quickly increases after starting the elec-
trodeposition, it shows about 1.5 mA with a slower decrease
over a certain period proportional to the pore diameter. Then, a
large decrease in the current is measured abruptly. The period
of time to keep the slower decrease in the current is longer as
the pore diameter increases. Those current transient behaviors
are similar to the case of the Cu electrodeposition in the several
respects.
where f is fractional of transformation, Vex is extended volume of
nuclei, V0 [=1 (cm2) × r] is an entire volume to be transformed,
N is the nucleus density, and r is the nucleus radius. The hemi-
spherical nuclei are assumed in Eq. (1). It is questionable if the
aboveequationcanbeappliedtothePt–Pdsurfacewithanumber
of the pores. However, these pores occupy a fractional area less
than 0.09. Such a small fractional number may make possible a
first-order approximation that the Pt–Pd surface is assumed as
the uniform surface for the above calculation. A limited number
SEM images of the Pt–Pd side are shown in Fig. 6(right). The
images of (a), (b), (c), and (d) correspond to morphologies of
the Ni grains formed through the electrodeposition process of
(a)–(d) in Fig. 6(left), respectively. The morphologies observed
in these images are clearly different from those of the Cu depo-
sition. The case of (a): 50 nm shows the abrupt decrease in the
current at t = 3 s. The Ni deposition on the Pt–Pd surface uni-
formly and densely forms the 30–50-nm-sized grains to cover
almost all the Pt–Pd area with a Ni thin film. In the cases of the
larger diameters: (b)–(d), the period of time to keep the slower
decrease in the current is longer. The Ni film growth on the Pt–Pd
side further progresses as observed in the images of (b)–(d). The
individual grains are recognized at (a) t = 3 s, but the observable
boundaries disappear over 10 s. Some cracks are introduced in
the film by enlarging pits near the pores. The cracks are devel-
oped more deeply to connect with each other as the pore size
increases. It is probably due to the residual stress generated in
the Ni film covering the vicinity of the pores. The dense Ni mor-
Fig. 5. Schematic illustration of the pore opening plugged with Cu grains.

M. Motoyama et al. / Electrochimica Acta 53 (2007) 205–212
209
Fig. 6. (Left) Cathodic current transients during the Ni electrodeposition into the PC membrane templates with the pores of a: 50 nm, b: 80 nm, c: 100 nm, and d:
200 nm in diameters. (Right) SEM images of the electrodeposited Ni grains on the sputtered Pt–Pd layer. Scale bar: 200 nm.
phology composed of the much finer grains contrasts with the
coarser grains observed in the case of the Cu deposition.
The cathodic current transient of (d): 200 nm in Fig. 6(left) is
again shown in Fig. 7(left). It was interrupted at the end of Stage
I at t = 32 s. The Ni films electrodeposited over three different
deposition times, which are indicated by dotted line segments
of a to c, were prepared for SEM observation. The SEM images
of the Pt–Pd side are shown in Fig. 7(right). The image (b) in
Fig. 7 is exactly the same as Fig. 6(d). The pore mouth diameter
decreases to about 100 nm at (a) t = 7 s. The progress of the Ni
film growth makes the pores narrower after the deposition times
of (b) 15 s and (c) 32 s. There is no essential difference between
both morphologies except for the development of the cracks in
(c). These images suggest that the deposition process to decrease
the mouth diameter is limited and the residual stress generated
near the pores is saturated with the development of the cracks.
Fig. 7. (Left) Cathodic current transient during the Ni electrodeposition into the PC membrane template with the pores of 200 nm in diameter. Dotted line segments
of a to c correspond to t = 7, 15, and 32 s, respectively. (Right) SEM images of the electrodeposited Ni grains on the sputtered Pt–Pd layer. Scale bar: 200 nm.

210
M. Motoyama et al. / Electrochimica Acta 53 (2007) 205–212
Fig. 9. Schematic illustration of the pore plugging at the growth front of a Ni
hollow tube.
Fig. 8. SEM images of Ni nanostructures formed at the deposition times of
(a)–(c) in Fig. 7. Scale bar: 200 nm.
While the Ni film grows on the Pt–Pd side, progress of the
deposition also occurs in each pore. The resulting structure
was observed after the template membrane was dissolved in
dichloromethane (see Fig. 8). The images with the same alpha-
bet: (a) to (c) in Figs. 7 and 8 were from the same specimen.
Hollow tubes with wall thicknesses of 20–40 nm are clearly
observed at (a) t = 7 s. This image indicates the proof of the
deposition of nickel only along the pore walls from the edge of
the Pt–Pd tubes as the cathode.
The Ni tube walls become thicker gradually. The wall thick-
ness shows roughly 50 nm at (b) t = 15 s. Finally, the top surface
no longer shows a hollow cross-section at (c) t = 32 s. The abrupt
decrease in the cathodic current is associated with the nanostruc-
ture transition from the tube to the wire occurring on its top. The
current path through each pore to the Pt–Pd layer and the Pt
sheet under the template is blocked at the top of the growing
nanostructure (see Fig. 9), while the same situation occurs at
the pore openings at the bottom in the case of the Cu deposi-
tion. The hollow cross-section formed in Stage I is not filled
with nickel even after the Ni deposition yields the wires in
Stage II.
Fig. 10(top) shows the side views of the Ni nanostructures
observed in Fig. 8(a)–(c). The height is plotted versus the depo-
sition time in Fig. 10 (bottom). When a linear regression line
is applied to the height variation with time, the slope and the
intercept at t = 0 s are estimated as 0.030 ␮m s−1 and 0.46 ␮m,
respectively. The slope is slightly smaller than the averaged
growth rate of 0.043 ␮m s−1, estimated from Fig. 2, over the
Fig. 10. (top) SEM images of the side views of the electrodeposited Ni nanos-
tructures. (bottom) Variations of the height and the wall thickness of a nanotube
as functions of time. Solid circle and blank circle indicate the height and the
wall thickness, respectively.

M. Motoyama et al. / Electrochimica Acta 53 (2007) 205–212
211
whole 10 ␮m-length. The intercept is close to the height of the
Pt–Pd nanotubes presented in Fig. 1(bottom). This good agree-
ment supported that the electrochemical deposition process of
the nanotubes begins at the top of the Pt–Pd nanotubes. The
observedwallthicknessisalsoplottedversustimein Fig. 10(bot-
tom). A linear time-dependence of the tube wall thickness is
also obtained. The soundness of this linearity is supported by
Fig. 6(left) where the induction time of the abrupt decrease in
the cathodic current is proportional to the pore diameter. The
tube wall thickness at t = 32 s is approximated as 100 nm. The
slope and the intercept of the regression line of the tube wall
thickness versus time are roughly 0.003 ␮m s−1 and 0.01 ␮m,
respectively. The rate of the growth along the length is about 10
times faster than that of the tube wall thickening. As a result, a
hollow Ni tube is obtained in each pore before the current starts
to decrease abruptly. The maximum length of the Ni nanotubes
with a 200 nm outer diameter was about 1 ␮m at pH 3.4 and
E = −1.0 V.
A control of surface diffusion of adatoms or clusters
at an interface of three phase of electrolyte/gas/metal or
PC/electrolyte/metal is one of the ultimate aims in materials pro-
cessing for nanostructured devices. Tailored interface promises
a better performance of the devices and new applications.
3. Conclusions
Cu and Ni nanowires were electrodeposited into the PC
membrane templates containing the cylindrical pores with diam-
eters of 50 to 200 nm. The sputter-deposited Pt–Pd layer as
the cathode was so thin that the pores at the bottom were
opened to pathways for the electrolyte. The abrupt decrease
in the cathodic current in Stage I during the Cu deposition
was correlated with the progress of the Cu grain growth with
the careful observation. It was found that the cathodic current
abruptly decreased when the pores on the Pt–Pd side are plugged
with the coarse Cu grains, and the nanowires were then grown.
On the other hand, a similar abrupt decrease in the Ni depo-
sition current was associated with a nanostructure transition
from the tube to the wire occurring at the growth front. The
growth rate of the tubes along the length was one order faster
than that for the wall thickening. The important combination
of the nanoporous template with a size distribution of metal
grains must be emphasized when the nanostructured interface is
designed.
The electrochemical growth process of the Ni nanowires has
been examined carefully. It is considerably different from that
of the Cu nanowires. The key feature of the electrochemical
deposition process of the Ni nanotubes is that the fineness of
the Ni grains can form the nanotube walls in the early stages.
Therefore, it is deduced that metallic nanotubes cannot be fabri-
cated by electrodeposition if the growing grains readily become
comparable to an employed pore size. Davis and Podlaha suc-
cessfully produced Cu nanotubes with also the PC template
having the pores partially covered with the sputter-deposited
cathode layer [24]. The outer diameters and the tube wall thick-
ness were reported as 800 and 200–300 nm, respectively. Their
electrolyte composition is slightly different from ours, but it
is expected that many coarser grains of 200 nm in size were
deposited in the pores with a much larger diameter than ours.
Thus, the above mechanism does not interfere with their results.
The grain size relative to the pore size is therefore essential to
designing the nanostructure formed in each pore.
Acknowledgments
Part of this work was supported by financial aides from the
21st Century COE Program and the Ministry of Education, Sci-
ence and Culture (Giant-in-Aid for Exploration Research No.
15360402), for which the authors are granted.
References
According to classification of metals by Winand [34,35],
Cu and Ni are categorized into different groups i.e., interme-
diate metals with moderate melting points (Au, Cu, Ag) and
inert metals with high melting points (Fe, Ni, Co, Pt, Cr, Mn),
respectively. As the melting point is lower, smoothness of the
film is more difficult to obtain due to high surface diffusion
coefficient for adatoms. In general, the grain size is propor-
tional directly to the linear growth rate of nuclei and inversely to
the nucleation rate [36–38]. If the surface diffusion coefficient
for adatoms is large enough, they diffuse to increase the island
rather than form a new nucleus. Cluster coalescence phenomena
such as Ostwald ripening and cluster migration also increase the
grain size. An understanding of diffusion of adatoms or clusters
on the cathode surface adjacent to the membrane walls, which
may play a role similar to surfactant additives, as well as the
growth kinetics of metal islands [39] can lead to control of the
grain size. Interestingly, H2 evolution can sustain deposition of
metal only on the pore walls [24,25]. Simulations are required to
optimize a grain size distribution for desired structures, includ-
ing contribution from ionic mass transfer accompanied with H2
evolution.
[1] C.R. Martin, Science 266 (1994) 1961.
[2] C. Ji, P.C. Searson, Electrochem. Solid-State Lett. 81 (2002) 4437.
[3] Y. Wu, G. Cheng, K. Katsov, S.W. Sides, J. Wang, J. Tang, G.H. Freder-
ickson, M. Moskovits, G.D. Stucky, Nat. Mater. 3 (2004) 816.
[4] L. Sun, Y. Hao, C.-L. Chien, P.C. Searson, IBM J. Res. Dev. 49 (2005) 79.
[5] F. Li, J. He, W.L. Zhou, J.B. Wiley, J. Am. Chem. Soc. 125 (2003) 16166.
[6] L. Piraux, J.M. George, J.F. Despres, C. Leroy, E. Ferain, R. Legras, K.
Ounadjela, A. Fert, Appl. Phys. Lett. 65 (1994) 7.
[7] C. Ji, G. Oskam, Y. Ding, J.D. Erlebacher, A.J. Wagner, P.C. Searson, J.
Electrochem. Soc. 150 (2003) C520.
[8] C.B. Murray, D.J. Norris, M.G. Bawendi, J. Am. Chem. Soc. 115 (1993)
8706.
[9] X. Peng, M.C. Schlamp, A.V. Kadavanich, A.P. Alivisatos, J. Am. Chem.
Soc. 119 (1997) 7019.
[10] L. Manna, E.C. Scher, A.P. Alivisatos, J. Am. Chem. Soc. 122 (2000)
12700.
[11] R.M. Penner, J. Phys. Chem. B 106 (2002) 3339.
[12] M.J. Williamson, R.M. Tromp, P.M. Vereecken, R. Hull, F.M. Ross, Nat.
Mater. 2 (2003) 532.
[13] B. Scharifker, G. Hills, Electrochim. Acta 28 (1983) 879.
[14] D. Grujicic, B.W. Sheldon, E. Chason, A.F. Bower, Appl. Phys. Lett. 81
(2002) 1204.
[15] O.E. Kongstein, U. Bertocci, G.R. Stafford, J. Electrochem. Soc. 152(2005)
C116.

212
M. Motoyama et al. / Electrochimica Acta 53 (2007) 205–212
[16] J.S. Beck, J.C. Vartuli, W.J. Roth, M.E. Leonowicz, C.T. Kresge, K.D.
Schmitt, C.T.-W. Chu, D.H. Olson, E.W. Sheppard, S.B. McCullen, J.B.
Higgins, J.L. Schlenker, J. Am. Chem. Soc. 114 (1992) 10834.
[17] H. Masuda, K. Fukuda, Science 268 (1995) 1466.
[18] M.E. Davis, Nature 417 (2002) 813.
[19] C.C. Harrell, P. Kohli, Z. Siwy, C.R. Martin, J. Am. Chem. Soc. 126 (2004)
15646.
[20] Y. Konishi, M. Motoyama, H. Matsushima, Y. Fukunaka, R. Ishii, Y. Ito,
J. Electroanal. Chem. 559 (2003) 149.
[30] M. Morisue, Y. Fukunaka, E. Kusaka, R. Ishii, K. Kuribayashi, J. Elec-
troanal. Chem. 559 (2003) 155.
[31] Y.L. Kawamura, T. Sakka, Y.H. Ogata, J. Electrochem. Soc. 152 (2005)
C701.
[32] Y. Konishi, Y. Nakamura, Y. Fukunaka, K. Tsukada, K. Hanasaki, Elec-
trochim. Acta 48 (2003) 2615.
[33] M. Avrami, J. Chem. Phys. 7 (1939) 1103;
M. Avrami, J. Chem. Phys. 8 (1940) 212;
M. Avrami, J. Chem. Phys. 9 (1941) 177.
[21] J. Verbeek, O.I. Lebedev, G. Van Tendeloo, L. Cagnon, C. Bougerol, G.
Tourillon, J. Electrochem. Soc. 150 (2003) E468.
[22] W.-C. Yoo, J.-K. Lee, Adv. Mater. 16 (2004) 1097.
[23] D.M. Davis, E.J. Podlaha, Electrochem. Solid-State Lett. 8 (2005) D1.
[24] Y. Fukunaka, M. Motoyama, Y. Konishi, R. Ishii, Electrochem. Solid-State
Lett. 9 (2006) C62.
[34] R. Winand, Electrocrystallization, in: Application of Polarization Mea-
surements in the Control of Metal Deposition, Elsevier, Amsterdam, 1984,
47.
[35] M. Tian, J. Wang, J. Kurtz, T.E. Mallouk, M.H.W. Chan, Nano Lett. 3
(2003) 919.
[36] E. Budevski, G. Staikov, W.J. Lorentz, Electrochemical Phase Formation
and Growth. An Introduction to the Initial Stage of Metal Deposition, VHC,
Weinheim, 1996.
[25] M. Delvaux, S. Demoustier-Champagne, Biosens. Bioelectron. 18 (2003)
943.
[26] M. Motoyama, Y. Fukunaka, T. Sakka, Y.H. Ogata, S. Kikuchi, J. Elec-
troanal. Chem. 584 (2005) 84.
[37] B. Mutaftschiev, The Atomistic Nature of Crystal Growth, Springer-Verlag,
Berlin Heidelberg, 2001.
[27] G.J. Hills, D.J. Schiffrin, J. Thompson, Electrochim. Acta 19 (1974) 657.
[28] J. Amblard, M. Froment, G. Maurin, N. Spyrellis, E. Trevisan-Souteyrand,
Electrochim. Acta 28 (1983) 909.
[29] A. Radisc, J.G. Lomg, P.M. Hoffmann, P.C. Searson, J. Electrochem. Soc.
148 (2001) C41.
[38] M. Ohring, Materials Science of Thin Films, Deposition and Structure, 2nd
ed., Academic Press, San Diego, 2002.
[39] S. Morin, A. Lachenwitzer, O.M. Magnussen, R.J. Behm, Phys. Rev. Lett.
83 (1999) 5066.