Novel electrodeposition behavior of Ni on porous anodic alumina templates without a conductive interlayer

Article abstract of DOI:10.1021/jp044936w

During template-assisted electrodeposition, single-crystalline metallic nanowires could be obtained only when the overpotential is low. However, an unusual electrodeposition behavior on the PAA/Si substrate without a conductive interlayer between the template and Si is described in the present study. Through the electrical breakdown of the template, Ni nanodots, nanowires and nanotubes could be obtained by only changing the electrodeposition voltage on the same substrate. The mechanisms leading to the formation of various nanostructures are described in detail and compared with those for the conventional template-assisted electrodeposition process. The electrodeposition first occurred on the pore wall instead of from the underlying substrate, leading to the formation of some Ni nanotubes at a more negative voltage. Besides, single-crystalline Ni nanowires could also be formed even when the electrodeposition voltage was as negative as -40 V, indicating that the formation of single-crystalline metallic nanowires under a large overpotential is possible. ? 2005 American Chemical Society.

Full text of DOI:10.1021/jp044936w

J. Phys. Chem. B 2005, 109, 9575-9580
9575
Novel Electrodeposition Behavior of Ni on Porous Anodic Alumina Templates without a
Conductive Interlayer
M. T. Wu, I. C. Leu,* J. H. Yen, and M. H. Hon
Department of Materials Science and Engineering, National Cheng Kung UniVersity, 701 Tainan, Taiwan
ReceiVed: NoVember 5, 2004; In Final Form: February 6, 2005
During template-assisted electrodeposition, single-crystalline metallic nanowires could be obtained only when
the overpotential is low. However, an unusual electrodeposition behavior on the PAA/Si substrate without a
conductive interlayer between the template and Si is described in the present study. Through the electrical
breakdown of the template, Ni nanodots, nanowires and nanotubes could be obtained by only changing the
electrodeposition voltage on the same substrate. The mechanisms leading to the formation of various
nanostructures are described in detail and compared with those for the conventional template-assisted
electrodeposition process. The electrodeposition first occurred on the pore wall instead of from the underlying
substrate, leading to the formation of some Ni nanotubes at a more negative voltage. Besides, single-crystalline
Ni nanowires could also be formed even when the electrodeposition voltage was as negative as -40 V,
indicating that the formation of single-crystalline metallic nanowires under a large overpotential is possible.
Introduction
Recently, template-mediated growth has become one of the
popular methods to fabricate nanostructures over a large area.1-3
Well-aligned and dimension-controlled nanostructures can be
obtained through the deposition of materials into the nanochan-
nels of a membrane. Among the membrane materials, porous
anodic alumina (PAA) is usually used for the template-mediated
growth due to its ease of production. By varying the anodization
conditions, the characteristics of PAA can be easily controlled.4,5
Because of its thermal stability and chemical inertness, the
fabrication of nanostructures using PAA as a template can be
conducted under many different conditions.
For the template-mediated growth, many processes are
adopted to fabricate nanostructures, such as the vapor-, solution-,
and electrochemistry-based methods, etc.6-8 For the vapor-based
methods, the deposition rate is usually too high so that the pore
mouths will be closed before the materials can be fully deposited
into the channels. For the solution-based methods, the low filling
efficiency is the main problem because the capillarity is the
only driving force to guide the transport of materials in the
channels. For the electrochemistry-based methods (especially
for the electrodeposition), the growth of the nanostructures is
field-assisted and usually selectively occurs at the conductive
area of the substrate, which makes it possible to avoid the
closure of the pore mouth and to increase the filling efficiency.
Besides, the cost of the sources and the equipments for
electrodeposition is relatively low, reducing the capital invest-
ment of the whole process. Therefore, the electrodeposition is
now considered as a powerful tool for template-mediated growth
of nanostructures.
For template-mediated electrodeposition, conventionally, a
conductive layer must be coated on one side of the membrane
as the electrode for electrodeposition. Without a conductive
layer, the filling efficiency of materials into the channels is quite
low. Moreover, for the fabrication of one-dimensional nano-
Figure 1. Dependence of the current on electrodeposition time at
different voltages on 300 nm thick PAA/Si substrates.
structures, single-crystalline nanowires of Cu, Ag, Au and some
low melting point metals (such as Pb and Bi) can be obtained
* To whom correspondence should be addressed. E-mail: icleu@
mail.mse.ncku.edu.tw.
10.1021/jp044936w CCC: $30.25 © 2005 American Chemical Society
Published on Web 04/21/2005

9576 J. Phys. Chem. B, Vol. 109, No. 19, 2005
Wu et al.
Figure 2. Surface morphologies of the as-electrodeposited substrates [(a) and (b) for the blank Si substrates, (c) and (d) for the 300 nm thick
PAA/Si substrates, (e) and (f) for the 600 nm thick PAA/Si substrates). The applied voltage was -2 V for that in images (a), (c) and (e), and was
-3 V for that in images (b), (d) and (f).
if a lower overpotential is used during electrodeposition.
However, for the high melting point metals such as Ni, Co and
Pt, the growth of single-crystalline nanowires via electrodepo-
sition is very difficult and has not been reported till now.9,10 In
this study, an unusual electrodeposition behavior of Ni is
presented, using the PAA on Si without a conductive layer as
a template. Ni nanostructures with different aspect ratio could
be obtained through this process. The filling efficiency still
exceeded 90% even though a conductive layer was not
introduced between the PAA and the substrate. Besides, some
single-crystalline Ni nanowires were obtained when the elec-
todeposition was conducted under a large overpotential. For this
reason, this process can be considered as a potential method to
fabricate single-crystalline nanowires of some materials that are
difficult to be obtained in the conventional template-mediated
electrodeposition.
to the removal of the barrier layer because the barrier layer is
partially destroyed. The barrier layer can thus be completely
removed after chemical etching at 30 °C for 20 minutes. A
potentiostat (EG&G, model 263A) was also used to confirm
the complete removal of the barrier layer by monitoring the
current transient. The current density for the barrier layer-
removed alumina/Si substrate is about 10 times that of the as-
anodized one when the applied voltage is -3 V. The cathodic
dc electrodeposition of Ni was performed in an aqueous solution
composed of NiSO4 (330 g/L), NiCl2 (45 g/L) and H3BO3 (35
g/L). The working electrode, counter electrode and reference
were the PAA/Si substrate, a platinum sheet and silver/silver
chloride, respectively. The solution was not stirred during the
whole electrodeposition process. After electrodeposition, a Ni
layer containing nanostructures could be obtained by removing
the PAA template in a 5 wt % NaOH solution. For the sake of
comparison, the electrochemical deposition of Ni was conducted
on PAA with a conductive Au interlayer inserted between the
PAA and the Si substrate. Some blank Si substrates were also
used for electrodeposition to further understand the electrodepo-
sition behaviors of Ni on the PAA/Si substrates. The morphol-
ogies of the electrodeposited specimens were observed by a
scanning electron microscope (Hitachi S4100) and field emission
transmission electron microscope (Hitachi FE-2000).
Experimental Section
The Al film of high purity was deposited on the n-type silicon
substrate (Silicon Inc., (100), 1-10 Ω cm) by e-beam evapora-
tion under the pressure of 10-6 Torr. A two-step anodization
was conducted to improve the ordering of the channels in the
PAA template.11 To improve the filling efficiency, the barrier
layer at the PAA/substrate interface was removed by immersing
the anodized specimen into a 5 wt % H3PO4 solution at 30 °C.
In our previous study, the pore bases became arched at the
alumina/Si interface when the Al film was consumed com-
pletely.12 The formation of the arched pore base is beneficial
Results and Discussion
Figure 1 shows the current variations recorded for the
electrodeposition of Ni for different substrate arrangement. For

Electrodeposition Behavior of Ni on Alumina Templates
J. Phys. Chem. B, Vol. 109, No. 19, 2005 9577
the electrodeposition on the PAA/Si substrate with a conductive
Au interlayer under the PAA template, the current variation
revealed the typical three stages of the growth of nanowires,13
as shown in Figure 1a. Stage I indicates the deposition of Ni
into the channels. Stage II shows the formation of caps at the
end of nanowires outside the channel and stage III shows the
formation of planar Ni layer on the PAA surface. Figure 1b
shows the current variation for the electrodeposition of Ni on
the 300 nm thick PAA/Si substrate without a conductive
interlayer. The current decreased in the beginning and reached
a steady state in the subsequent electrodeposition process for
all the applied voltages, quite different from that shown in Figure
1a. This indicates that the deposition mechanisms of Ni were
probably different for the different substrates. Besides, no
obvious fluctuation in current was observed for the electrodepo-
sition on the PAA/Si substrate without a conductive layer, even
when the applied voltage was as negative as -40 V. Only a
few bubbles were found on the electrode for all the electrodepo-
sition voltages, implying that the hydrolysis of water on the
PAA/Si substrate was quite low. It is not clear why the
hydrolysis reaction cannot readily occur on the PAA/Si substrate
without a conductive layer; however, the low degree of the
hydrolysis reaction reduces the possibility of obstructing the
electrodeposition at more negative voltages by gas bubbles.
To further investigate the electrodeposition behavior of Ni
on the PAA/Si substrate without a conductive interlayer, the
electrodeposition was also conducted on the blank Si substrate.
Figure 2a shows the surface morphologies of a Si substrate after
the electrodeposition of Ni at -2 V for 10 min. Some Ni
particles were formed, but with a very low density of about
1.45 × 104 cm-2. Even for a more negative applied voltage
(-3 V), the density of Ni particles only slight increases (3.15
× 104 cm-2), accompanying the formation of Ni particles of
larger size (Figure 2b). The dominant carriers of n-type Si are
electrons. When a negative voltage was applied on the n-type
Si (forward bias),14 the electrons should easily be driven to the
Si/solution interface and reduce the metal ions to form the
deposit. Nevertheless, the results of Figure 2a,b indicate that
the deposit does not prefer to nucleate on the Si substrate. The
low nucleation density of metal on the semiconductor can be
attributed to the low interaction energy between the absorbed
metal atom and the substrate, as well as the required larger
critical nucleus size compared with that on the metallic
substrate.15
Figure 3. High-magnification image of 300 nm thick PAA surface in
the initial period of electrodeposition at -2.0 V for (a) 30 s, (b) 90 s,
and (c) 150 s.
pure oxide.16 Owing to the limited thickness of the template,
electron avalanches due to the electrical breakdown of the anion-
contaminated oxide might occur under the applied electrodepo-
sition voltage in the present study, even though the calculated
field strength was only about 105 V cm-1 for the electrodepo-
sition voltage of -3 V. The surface of the template could be
considered as the cathode for electrodeposition. The electric field
across PAA would decrease as the thickness of PAA was
increased, leading to a reduced nucleation density of Ni because
the nucleation density is dependent on the magnitude of
overpotential.17
Another interesting phenomenon is that the density of Ni
particles is higher on the PAA/Si substrate than that on the blank
Si substrate under the same electrodeposition conditions. In
addition to the factor of interaction energy mentioned above,
other factors may also affect the electrodeposition of metal on
the different substrates used. Because the pore wall of PAA is
composed of pure oxide at the inner part and anion-contaminated
oxide at the outer part,16 the existence of defects on the anion-
contaminated oxide will provide more nucleation sites during
electrodeposition. It may be one possible reason the density of
Ni particle is higher on the PAA/Si substrate than that on the
blank Si substrate.
Part c and d of Figure 2 show the surface morphologies of
300 nm thick PAA/Si substrates after electrodeposition. The
densities of Ni particles were about 2.5 × 106 and 4 × 106
cm-2 when the electrodeposition was conducted at -2 and -3
V, respectively. When a 600 nm thick PAA/Si substrate was
used for electrodeposition, the densities of Ni particles were
7.7 × 105 and 1.18 × 106 cm-2 for voltage at -2 and -3 V,
respectively (Figure 2e,f). These results indicate that the density
of Ni particles decreased with the increase of the PAA thickness.
In our previous study, it was found that no deposit was observed
on the pore wall and the underlying Si when the electrodepo-
sition voltage was more positive than -2 V.11 As the applied
voltage was more negative than -2 V, Ni began to deposit at
the outer surface of PAA instead of the underlying Si. It was
suggested that the unusual electrodeposition behavior was caused
by the electrical breakdown of PAA that would lead to the
formation of negatively charged pore wall. It is well-known that
the pore wall is composed of the pure oxide at the inner part
and the anion-contaminated oxide at the outer part. The
conductivity of the anion-contaminated oxide is higher than the
Figure 3 shows a high-magnification image of a 300 nm thick
PAA surface in the initial period of electrodeposition at -2.0
V. In the beginning, the deposition occurred in local area (the
white area) on the PAA surface, like those shown in Figure 3a.
The subsequent electrodeposition of Ni prefers to occur at the

9578 J. Phys. Chem. B, Vol. 109, No. 19, 2005
Wu et al.
Figure 4. SEM images of the Ni layer containing nanostructures obtained at different voltages: (a) -2.0 V, (b) -8.0 V, (c) -20 V, (d) -40 V and
(e) -40 V. The thickness of PAA used for images of (a)-(d) is 300 nm, and the thickness of PAA used for (e) is 600 nm.
deposited regions, leading to the growth of Ni particles, as
shown in Figure 3b,c. Two important issues were revealed from
the results in Figures 2 and 3. First, the electrical breakdown
occurs randomly at local regions of PAA. The density of the
breakdown regions increases with the applied voltage. Nucle-
ation will first occur at these breakdown sites, unlike the
homogeneous nucleation of deposit on a conductive substrate.
Second, in the prolonged electrodeposition, a continuous planar
Ni layer will be observed on the PAA surface, implying that
the formation of the Ni layer is contributed by the continuous
nucleation and lateral growth of the as-grown Ni particles.
After the electrodeposition, the PAA template was removed
by immersing the specimen in a 0.5 wt % NaOH solution and
the Ni layer containing nanostructures was observed from the
side facing the PAA/Si interface, as shown in the schematic
diagram of Figure 4. Figure 4a-d shows the Ni nanostructures
obtained under various voltages on the 300 nm thick PAA/Si
substrates. At a less negative voltage (-2.0 V), only Ni nanodots
could be observed on the backside of the Ni deposit, implying
that the filling of ions into pores was terminated soon after the
beginning of the electrodeposition. When the applied voltage
became more negative (-8 V), some Ni nanowires were formed
in addition to nanodots, like those indicated by arrows in Figure
4b. The fraction of nanowires in the deposits increased when
the applied voltage became more negative, as shown in Figure
4c,d. Similar deposition behaviors were also observed if a 600
nm thick PAA/Si substrate replaced the 300 nm thick one. Figure
4e shows the Ni nanowires obtained on the 600 nm thick PAA/
Si substrate (the electrodeposition voltage was -40 V). The
diameter and the length of the nanowires are 60 and 600 nm,
respectively, coinciding with the characteristics of the PAA
template. These results indicate that most of the channels of
PAA can be completely filled as the electrodeposition voltage
is negative enough, though there is no conductive layer
introduced between PAA and Si substrate.
As mentioned above, the electrodeposition of Ni on the PAA
substrate without a conductive interlayer is contributed by the
electrical breakdown of the PAA template. The electron
avalanches caused by the electrical breakdown transformed the
PAA substrates to be negatively charged, which in turn could
be the cathode for the electrodeposition of Ni. Owing to the
geometry factor, the electric field at the pore mouth would be
higher than that at the pore walls, implying that the deposition
rate of Ni would be higher at the pore mouth. Moreover, at a
less negative voltage, the migration of ions into channels driven
by the electric field is low. Therefore, the deposition of Ni into
channels would be self-terminated at the electrical breakdown
regions and keep on growing laterally till the metal spread over
the whole electrode area, leading to the formation of Ni nanodots
by the partial filling of the pores near the top surfaces.
At a more negative voltage (-8 V), the pore walls of some
channels would suffer severe electron avalanches and be highly
negatively charged, enhancing the migration of ions into these
channels. Some channels were completely filled during elec-
trodeposition and formed nanowires, like those shown in Figure
4b. As the voltage became much more negative (20-40 V),
the fraction of pores with highly negatively charged pore wall
would increase and more pores could be completely filled during

Electrodeposition Behavior of Ni on Alumina Templates
J. Phys. Chem. B, Vol. 109, No. 19, 2005 9579
deserves to be mentioned here. Figure 6 shows the TEM images
of Ni nanowires formed at -40 V. Parts a and b of Figure 6
show the low-magnification image and high-resolution lattice
image of polycrystalline Ni nanowires, respectively. It is clear
that a single nanowire was composed of many grains. Several
rings can be observed in the electron diffraction pattern (the
inset of Figure 6b), indicating that the nanowire is polycrys-
talline. Expect for the polycrystalline ones, some single-
crystalline Ni nanowires were also formed at -40 V. Parts c
and d of Figure 6 show the low-magnification image and high-
resolution lattice image of single-crystalline Ni nanowires,
respectively. The diffraction patterns obtained at the different
points of a single nanowire are all the same, implying that the
nanowire is single-crystalline. For the preparation of metal
nanowires, the growth of single-crystalline nanowires is related
to the 2D nucleation.18 During electrodeposition, generally, new
grains will grow when the size of nucleus exceeds the critical
dimension Nc.19 A larger Nc is favorable to the growth of a single
crystal. For the 2D-nucleation, the critical dimension Nc can be
expressed as Nc ) bsꢀ2/(Zeη)2 , where b, s, ꢀ, Z, and η are a
constant, the area occupied by one atom on the surface of the
nucleus, the edge energy, the effective electron number and the
overpotential,19 respectively. During electrodeposition, the only
parameter one can control directly is the overpotential η. A low
η is favorable for the growth of the single-crystal because Nc is
large. However, in our study, single-crystalline nanowires could
also be formed even under a very large overpotential. Therefore,
there must be some other factors that affect the formation of
the single-crystalline nanowires on the PAA/Si substrate. Except
for the nucleation conditions, the growth of the nanowire is also
time-dependent. A very large overpotential (for example, -40
V in the present study) may prefer the growth of new grains
during electrodeposition. However, for the electrodeposition on
the PAA/Si without a conductive layer, Ni ions would be
strongly driven into some channels with highly negatively
charged pore walls. These pores were completely filled rapidly
before the growth of a second grain, leading to the formation
Figure 5. Morphologies of Ni nanowires and nanotubes obtained at
-40 V on the 600 nm thick PAA/Si substrate.
electrodeposition. In fact, the density of Ni nanowires also
increased with the increase of the applied voltage, as shown in
Figure 4c,d.
Except the nanowires, some nanotubes were also formed as
the electrodeposition was conducted at -40 V, as shown in
Figure 5. The inset in Figure 5 shows the TEM image of the
nanotube, indicating that the one-dimensional structure is
hollow. Because the electrodeposition was contributed by the
electrical breakdown of the template, the pore walls were highly
negatively charged at a more negative voltage. The electrodepo-
sition could first occur on the whole surface of the pore walls,
followed by the growth of Ni toward the center of the channel.
Some pore mouths may be closed before the channels were
completely filled, leading to the formation of the nanotubes.
As the pore mouths were closed after the channels were
completely filled, the nanowires were formed. Figure 5 can
further confirm that the deposition began on the pore wall
instead of from the bottom of the channel if the electrodeposition
was conducted at a more negative voltage. The results in Figures
4 and 5 revealed that the electrodeposition behaviors on the
PAA/Si substrate without a conductive interlayer are mainly
voltage-dependent, but not time-dependent.
Finally, another unusual phenomenon different from those
obtained in the conventional template-mediated deposition
Figure 6. TEM images of Ni nanowires obtained under the electrodeposition at -40 V: (a) bright field and dark field images of the polycrystalline
Ni nanowire, (b) high-resolution lattice image of the polycrystalline Ni nanowire, (c) bright field and dark field images of the single-crystalline Ni
nanowire, and (d) high-resolution lattice image of the single-crystalline Ni nanowire. The insets in images (b) and (d) show the diffraction patterns
of polycrystalline and single-crystalline Ni nanowires, respectively.

9580 J. Phys. Chem. B, Vol. 109, No. 19, 2005
Wu et al.
(2) Keating, C. D.; Natan, M. J. AdV. Mater. 2003, 15, 451-454.
of the single-crystalline nanowires. Therefore, it is believed that
this process is beneficial to the preparation of single-crystalline
nanowires that are difficult to obtain by conventional template-
mediated techniques.
(3) Dick, B.; Sit, J. C.; Brett, M. J.;Votte, I. M. N.; Bastiaansen, C.
W. M. Nano Lett. 2001, 1, 71-73.
(4) Jessensky, O.; Mu¨ller, F.; Go¨sele, U. J. Electrochem. Soc. 1998,
145, 3735-3740.
Conclusion
(5) Li, A. P.; Mu¨ller, F.; Birner, A.; Nielsch, K.; Go¨sele, U. J. Appl.
Phys. 1998, 84, 6023-6026.
In the present study, an unusual electrodeposition behavior
on the PAA/Si without a conductive layer between the template
and Si was described. The current decreased continuously during
the whole electrodeposition process, unlike that obtained on the
PAA/Si substrate with a conductive layer between the template
and Si. No obvious hydrolysis of water on the substrate surface
was observed even when the applied voltage was as negative
as -40 V. The deposition occurred at the local regions of PAA
in the beginning, and a planar Ni layer was finally formed due
to the continuous nucleation and lateral growth of Ni particles.
At a less negative voltage (-2 V), a Ni layer containing only
nanodots was formed. As the electrodeposition voltage became
more negative, some nanowires would also be formed and the
density of nanowires increased with the increase of the applied
voltages. Besides the polycrystalline nanowires, some single-
crystalline nanowires were also formed when the electrodepo-
sition voltage was -40 V. The results in the present study show
that the electrodeposition behaviors on the PAA/Si substrate
without a conductive layer between the template and Si are
mainly voltage-dependent.
(6) Hobbs, K. L.; Larson, P. R.; Lian, G. D.; Keay, J. C.; Johnson, M.
B. Nano Lett. 2004, 4, 167-171.
(7) Lakshmi, B. B.; Dorhout, P. K.; Martin, C. R. Chem. Mater. 1997,
9, 857-862.
(8) Wang, Y. C.; Leu, I. C.; Hon, M. H. J. Mater. Chem., 2002, 12,
2439-2444.
(9) Yi, G.; Schwarzacher, W. Appl. Phys. Lett. 1999, 74, 1746-1748.
(10) Zhang, J.; Wang, X.; Peng, X.; Zhang, L. Appl. Phys. A 2002, 75,
485-488.
(11) Wu, M. T.; Leu, I. C.; Yen, J. H.; Hon, M. H. Electrochem. Solid
State Lett. 2004, 7, C61-C63.
(12) Wu, M. T.; Leu, I. C.; Hon, M. H. J. Mater. Res. 2004, 19, 888-
895.
(13) Ohgai, T.; Hoffer, X.; Fabian, A.; Gravier, L.; Ansermet, J.-P. J.
Mater. Chem. 2003, 13, 2530-2534.
(14) Plummer, J. D.; Deal, M. D.; Griffin, P. B. Silicon VLSI Technology;
Prentice Hall Inc.: Upper Saddle River, NJ, 2000; Chapter 1.
(15) Oskam, G.; Long, J. G.; Natarajan, A.; Searson, P. C. J. Phys. D:
Appl. Phys. 1998, 31, 1927-1949.
(16) Thompson, G. E.; Wood, G. C. Treatise on Materials Science and
Technology, Corrosion: Aqueous Processes and PassiVe Films; Academic
Press Inc. Ltd.: London, 1983; Vol. 23, Chapter 5.
Acknowledgment. We acknowledge the financial support
from the National Science Council of the Taiwan, Republic of
China under Contract No. NSC 92-2120-M-006-003.
(17) Budevski, E.; Staikov, G.; Lorenz, W. J. Electrochemical Phase
Formation and Growth; Weinheim-VCH: Weinheim, 1996; Chapter 3.
(18) Tian, M.; Wang, J.; Kurtz, J.; Mallouk, T. E.; Chan, M. H. W.
Nano Lett. 2003, 3, 919-923.
References and Notes
(1) Cao, H.; Qiu, X.; Luo, B.; Liang, Y.; Zhang, Y.; Tan, R.; Zhao,
M.; Zhu Q. AdV. Funct. Mater. 2004, 14, 243-246.
(19) Paunovic, M.; Schlesinger, M. Fundamental of Electrochemical
Deposition; Weily: New York, 1998; p 108.