Electrochemical tailoring of catalyst nanoparticles for CNT spatial-dimension control

Article abstract of DOI:10.1149/1.3280245

Electrodeposition is an excellent method for a site-selective and direct fabrication of catalyst nanoparticles for subsequent carbon nanotube (CNT) growth. Nickel nanoparticles with densities as high as 10¹¹ cm ^-2 and diameters down to 15±4 nm are deposited on blanket and patterned TiN wafers. Ni nanoparticles are selectively placed at the bottom of submicrometer vias by through-mask electrodeposition. The spherical nanoparticles all produce aligned multiwalled CNTs. The CNT diameter is directly determined by the diameter of the as-deposited nanoparticles, which is tailored by the electrodeposition conditions. Such control of CNT size, density, and location represents an important step toward the implementation of CNTs in device fabrication.

Full text of DOI:10.1149/1.3280245

Journal of The Electrochemical Society, 157 ͑3͒ K47-K51 ͑2010͒
K47
0013-4651/2010/157͑3͒/K47/5/$28.00 © The Electrochemical Society
Electrochemical Tailoring of Catalyst Nanoparticles for CNT
Spatial-Dimension Control
d
Ainhoa Romo-Negreira,a,b Daire J. Cott,a Stefan De Gendt,a,c, Karen Maex,
*
Marc M. Heyns,a,b and Philippe M. Vereeckena,
,z
*
aIMEC, Leuven B-3001, Belgium
bMetallurgy and Materials Engineering Department, cChemistry Department, and dDepartment of
Electrical Engineering, Katholieke Universiteit Leuven, Leuven B-3001, Belgium
Electrodeposition is an excellent method for a site-selective and direct fabrication of catalyst nanoparticles for subsequent carbon
nanotube ͑CNT͒ growth. Nickel nanoparticles with densities as high as 1011 cm−2 and diameters down to 15 Ϯ 4 nm are
deposited on blanket and patterned TiN wafers. Ni nanoparticles are selectively placed at the bottom of submicrometer vias by
through-mask electrodeposition. The spherical nanoparticles all produce aligned multiwalled CNTs. The CNT diameter is directly
determined by the diameter of the as-deposited nanoparticles, which is tailored by the electrodeposition conditions. Such control
of CNT size, density, and location represents an important step toward the implementation of CNTs in device fabrication.
© 2010 The Electrochemical Society. ͓DOI: 10.1149/1.3280245͔ All rights reserved.
Manuscript submitted August 31, 2009; revised manuscript received November 27, 2009. Published January 19, 2010.
1
Since their discovery in 1991, carbon nanotubes ͑CNTs͒ have
attracted a great deal of attention in the fields of nanoscience and
nanotechnology. The physical and chemical properties of both
single-walled carbon nanotubes ͑SWCNTs͒ and multiwalled carbon
nanotubes ͑MWCNTs͒ have potential for applications in optoelec-
tronics, nano-electromechanical systems, flat-panel displays, sen-
sors, energy storage, high strength composites, and drug delivery.2-8
One of the major obstacles in achieving high end technology
CNT products is the lack of sufficient process control. For example,
at this point, the direct growth of one type of SWCNT with chirality
control is still impossible, hindering its introduction as field-effect
transistors. Next to an excellent control of CNT diameter and den-
sity, the integration of large numbers of CNT-based devices requires
site-selective surface-bound CNT growth with excellent yield. To
scale up CNT growth, catalytic chemical vapor deposition ͑CVD͒ is
the best-known technique for growing organized networks or arrays
of nanotubes on large substrates. As CVD growth is a catalytic pro-
cess, catalyst placement directly on the substrate and catalyst acti-
vation take an important role in the CNT integration. Moreover, the
substrate, with all its components and materials already present, may
set the temperature budget for the CNT fabrication. For example,
applications targeting CNT on top of complementary metal oxide
semiconductor layers are limited to processing temperatures of
400–450°C to ensure the integrity of the underlying Si devices. The
temperature restrictions apply to both the catalyst formation and
activation steps, as well as to the CNT growth. Even so, most effort
on lowering the process temperature has been on CNT synthesis,9-11
not as much on the formation and activation of the catalyst.
and growth process whereby the deposition potential ͑current͒ and
charge ͑time͒ provide control over particle density and size,
respectively.20-23 Furthermore, ECD is a low temperature process
typically around room temperature that offers the possibility of pre-
paring pure metals or alloys with a controlled composition.
Electrodeposition has been extremely suited for the fabrication of
nanostructures.24-31 Single and multicomponent nanowires have
been made inside high aspect ratio nanoporous templates such as
anodized aluminum oxide, and freestanding nanowires have also
been demonstrated.24-27 Electrochemical nucleation and growth for
thin films has been the topic of fundamental studies for many
years.28,29 The nucleation phenomenon can be successfully used for
controlled nanoparticle formation as well; for example, Iacopi et al.
showed the use of electrodeposited indium nanoparticles as the cata-
lyst for silicon nanowire growth,30 and recently, we have shown the
use of electrodeposited nickel nanoparticles on silicon for CNT
growth.31
Herein, we report on the electrodeposition of nickel nanopar-
ticles. Densities as high as 1011 cm−2 and diameters down to
15 Ϯ 4 nm were obtained on blanket and patterned TiN substrates,
showing room-temperature control over particle density and size.
The as-deposited spherical nanoparticles all produced aligned MW-
CNT with high density. The CNT outer diameter equals that of the
catalyst particle, which is tailored by the electrodeposition condi-
tions. Further, selective CNT growth at specific locations using a
patterned via-hole structure is presented.
The main focus of this work is on room-temperature catalyst
formation for control of the CNT size, density, and location; discus-
sions of low temperature CNT growth are beyond the scope of the
present paper. TiN was used as the bottom contact for CNT bundles
as it is a good diffusion barrier for interconnects,32 which also pro-
vides a good electrical contact for electrodeposition.
Up until now, physical vapor deposited ͑PVD͒ metal films, metal
particle impregnated supports, and spin-coated or spray-coated col-
loidal nanoparticles or metal salts have been commonly used as a
catalyst source for substrate-attached CNT growth.12-17 Even though
these approaches can yield sufficient dimensional uniformity, many
need a high temperature treatment step. Moreover, they all lack se-
lectivity in isolating the catalysts at specific locations. For example,
Sato et al. fabricated CNT bundles inside vias for chip interconnects.
The metal catalyst particles were generated ex situ by laser ablation
at a high temperature ͑1000°C͒, size-separated by a “differential
mobility analyzer,” and then blanket deposited on the wafer.18,19
Therefore, this method allows extremely good control of particle
size without exact control of location. Hence, a low temperature
catalyst deposition method is needed that provides site-selective
catalyst placement with control over particle size and density.
Electrochemical deposition ͑ECD͒ proceeds through a nucleation
Experimental
Catalyst deposition and subsequent CNT growth experiments
were carried out using blanket and patterned silicon wafers with a
70 nm PVD TiN layer on top of 100 nm thermal oxide. Patterns of
via-hole arrays with pore diameters ranging between 150 and 300
nm were etched onto a 300 nm oxide ͑phosphorous silicate glass͒
+30 nm SiC bilayer deposited over the TiN layers. The pattern den-
sities were 6–10%. The via holes were grouped in dense ͑pitch to
diameter ratio of 2͒, semidense ͑pitch to diameter ratio of 3͒, and
isolated arrays ͑pitch to diameter ratio of 8͒, where the pitch is
defined as the distance between two neighboring vias ͑center to
center͒. Unless otherwise stated, 2 ϫ 2 cm samples were used in all
depositions and growth experiments.
Nickel was deposited from an aqueous solution of 0.01 M
*
Electrochemical Society Active Member.
z E-mail: Philippe.Vereecken@imec.be
Ni͑NO3͒2, 1 M NaCl, and 1 M NH4Cl.33 The pH of the solution was
Downloaded on 2014-03-30 to IP 134.208.103.160 address. Redistribution subject to ECS terms of use (see ecsdl.org/site/terms_use) unless CC License in place (see abstract).

K48
Journal of The Electrochemical Society, 157 ͑3͒ K47-K51 ͑2010͒
adjusted to 8.3 with NH4OH, during which the solution color
changed from green to blue as a result of the formation of the
34
2+
6
nickel–ammonia complex Ni͑NH3͒
.
The complexation of the
Ni2+ ions with NH3 shifts the standard reduction potential from the
U0
= −0.25 V ͓standard hydrogen electrode ͑SHE͔͒ to the
= −0.49 V ͑SHE͒.35
2+
͑Ni /Ni͒
U0
2+
͓Ni͑NH ͒ /Ni͔
3
6
Before electrodeposition, the blanket samples were dipped in 2
wt % HF for 3 min to remove the native oxide. The patterned
samples were cleaned in 2.9 wt % NH4OH for 2 min to eliminate
any polymer residues. To eliminate the native oxide from the pat-
terned TiN surface, a dilute 0.25 wt % HF solution ͑2 min͒ was used
to minimize the loss of the oxide pattern.
A three-electrode electrochemical cell with separated cathode
and anode compartments was used for the nickel deposition. In a
typical experiment, a sample was mounted onto a rotating sample
holder, immersed into the solution of the cell’s main compartment,
and rotated at 500 rpm ͑Metrohm 16280010 rotator and controller͒.
A platinum rod separated from the main compartment by a porous
glass frit was used as the counter electrode. The reference electrode
was a Ag/AgCl/3 M NaCl ͑BAS, 0.215 V vs SHE͒ connected to the
main cell compartment by a Luggin capillary. The current and po-
tential were controlled by
a digital potentiostat ͑Autolab
Figure 1. ͑a͒ Top view SEM micrograph of Ni nanoparticles deposited on
TiN at −1.5 V vs Ag/AgCl for 0.6 s from a solution of 0.01 M Ni͑NO3͒2, 1
M NaCl, 1 M NH4Cl, and NH4OH ͑pH 8.3͒; high magnification insert with
tilted view at 45° showing the spherical geometry of the nanoparticles; ͑b͒
nickel particle density Np vs deposition potential UD at Ϫ1.2 V for 1 s, Ϫ1.3
V for 0.6 s, Ϫ1.4 V for 0.6 s, Ϫ1.5 V for 0.6 s, Ϫ1.6 V for 0.5 s, and Ϫ1.7
PGSTAT30͒. The exposed substrate area was 1.54 cm2, and all ex-
periments were carried out at room temperature.
An Oxford Instruments ͑OIPT Nanofab͒ plasma-enhanced
chemical vapor deposition ͑PECVD͒ reactor was used for CNT
growth at a sample temperature of ϳ650°C. First, the catalyst was
exposed to H2 plasma for 5 min using 200 sccm H2 at 1 Torr ͑200 W
plasma power generated from a 13.56 MHz radio-frequency source͒.
The growth consisted of 96 sccm C2H2 at 1 Torr for 10 min. Note
that the growth temperature for this system could be as low as
500°C.36
V for 0.5 s; ͑c͒ dependence of particle size with deposition time for UD
=
−1.5 V is shown by ͑b͒ and theoretical diameter for growth under steady-
state diffusion ͑Eq. 3͒ is shown by solid line. ͑d͒ SEM image of CNT grown
from ECD Ni nanoparticles ͑Ϫ1.5 V for 0.6 s͒. CNT growth: 96 sccm C2H2
at 1 Torr for 10 min, 14 ␮m long tubes.
Scanning electron microscopy ͑SEM, NOVA 200 NanoSEM͒
was used to characterize the nickel nanoparticles and CNT. Particle
densities and projected particle area were analyzed from the SEM
images using digital imaging software ͑image J͒.37 From the pro-
jected particle area ͑disk͒, the diameters and volumes of the particles
could be determined. The geometry of the particles was spherical, as
determined from the tilted SEM images. For each sample, an aver-
age of several images was made. For ͑high resolution͒ transmission
electron microscopy ͑TEM͒ analysis, the CNT was dispersed in eth-
anol by sonication for 10 min and was drop-casted on a carbon lacy
copper grid. The analysis was performed in a Tecnai F30 operating
at 200 kV to avoid damage of CNT during imaging.
1/3
qNi
d =
͓1͔
ͩ ͪ
KNp
with the charge density qNi ͑C cm−2͒ corresponding to nickel depo-
sition and K ͑C cm−3͒, a material constant that depends on the ge-
ometry of the particles, being 1.53 ϫ 104 C cm−3 for spherical
nickel particles.
Together with nickel deposition, water is decomposed to hydro-
gen gas at these negative potentials.38 Consequently, the integrated
charge accounts for both reactions and therefore cannot be used
Results and Discussion
directly with Eq. 1. However, the nickel deposition reaction at Ϫ1.5
V becomes limited by convective diffusion of the Ni͑NH3͒ ions
2+
6
Nickel nanoparticles were fabricated on blanket TiN substrates
by applying a short potential step between 0.5 and 1 s, depending on
the applied potential. A typical example of electrodeposited nickel
nanoparticles on the blanket TiN substrate is shown in the micro-
graph of Fig. 1a. In tilted view, the spherical shape of the nanopar-
ticles is seen. The density of the nanoparticles can be tailored by the
applied potential as the nucleation kinetics is potential dependent.33
Figure 1b shows the particle densities for deposition potentials be-
tween Ϫ1.2 and Ϫ1.7 V. The particle density rapidly increases from
close to 1010 cm−2 at low overpotentials ͑Ϫ1.2 V͒ to about
1011 cm−2 at Ϫ1.4 V and only increases slightly thereafter. The
reproducibility for multiple deposition runs at the same potential
was above 90%. The particle size can be tailored by the deposition
time: Fig. 1c shows the dependence of the particle size with depo-
sition time for the nickel ECD at Ϫ1.5 V. For the shortest deposition
time of 50 ms, particles with an average diameter of 8.5 Ϯ 3.5 nm
were obtained. After 0.6 s, the average particle diameter had in-
creased to 17 Ϯ 6 nm, and particles of 30–40 nm were obtained
after 2–4 s. Faraday’s law relates the charge to the total mass of
deposited metal and thus to the average volume of an individual
metal particle when the particle density Np is known.31 The particle
diameter, d, is then given by
shortly after the potential step is applied. This means that the depo-
sition current becomes constant and the charge simply increases
proportionally with time
qNi = ilim
t
͓2͔
The limiting current ilim equals 6.5 mA cm−2 for a 0.01 M ion
concentration and a sample rotation rate of 500 rpm.39 Substitution
of Eq. 2 in Eq. 1 gives the relationship for particle diameter with
time
1/3
ilim
t
d =
͓3͔
ͩ ͪ
KNp
The theoretical upper limit for the particle diameter in Eq. 3 is
plotted in Fig. 1c together with the experimental data. As expected,
the discrepancy between the theory and the experiment is largest at
short times where the growth current has not yet reached a steady-
state condition.
Next, dense mats of aligned CNT were grown from the ECD
nickel catalyst nanoparticles, as shown in Fig. 1d. The growth con-
sisted of a catalyst pretreatment step in hydrogen plasma followed
by thermal CNT growth using acetylene gas, both at a sample tem-
Downloaded on 2014-03-30 to IP 134.208.103.160 address. Redistribution subject to ECS terms of use (see ecsdl.org/site/terms_use) unless CC License in place (see abstract).

Journal of The Electrochemical Society, 157 ͑3͒ K47-K51 ͑2010͒
K49
Figure 2. HRTEM of the obtained MW tubes for ͑a͒ 10 nm diameter and 5
walls and ͑b͒ 20 nm diameter and 15 walls. ͑c͒ Cumulative frequency of
ECD Ni nanoparticles diameters ͑ᮀ͒ and CNT diameters ͑᭺͒ extracted from
SEM and HRTEM images, respectively. Ni deposition at Ϫ1.5 V for 0.6 s;
CNT growth recipe: 96 sccm C2H2 at 1 Torr for 10 min.
Figure 3. ͑a͒ SEM images of the patterned structure; ͑top͒ large-area view
showing the border of the dense ͑pitch to diameter ratio of 2͒ and the semi-
dense ͑pitch to diameter ratio of 3͒ arrays separated by a trench; ͑bottom͒
small-area view of vias in a dense array from top down ͑left͒ and in cross
section ͑right͒. ͑b͒ Large-area view showing long CNT bundles selectively
grown from the dense and semidense arrays as well as the trench separating
both with tubes of approximately 6 ␮m long ͑5 min plasma pretreatment at
650°C in 200 sccm H2 at 1 Torr with 200 W plasma 13.56 MHz, followed by
10 min thermal growth in 96 sccm C2H2 at 1 Torr͒. ͑c͒ CNT grown in ͓͑c.1͒
and ͑c.3͔͒ 150 nm and ͓͑c.2͒ and ͑c.4͔͒ 300 nm dense via arrays after ͓͑c.1͒
and ͑c.2͔͒ 5 min and ͓͑c.3͒ and ͑c.4͔͒ 10 min plasma pretreatment followed
by 30 s growth ͓͑same conditions as ͑b͔͒. The CNT grew only from the TiN
bottom substrate where the nanoparticles were deposited.
perature of ϳ650°C and at 1 Torr. The pretreatment and growth
steps were 5 and 10 min, respectively. A tube length of 14 ␮m was
obtained after 10 min of growth corresponding to a growth rate of
about 23 nm s−1. The density of CNT estimated from the SEM
images is around 1011 CNT per cm2 or is similar to the as-deposited
nanoparticle density.
Tube characteristics were studied in detail with high resolution
transmission electron microscopy ͑HRTEM͒ revealing MWCNT
with a typical bamboo structure, as shown in Fig. 2a and b. CNT
diameters ranged between 10 and 20 nm with 5–15 walls. A statis-
tics of CNT diameter was made based on the HRTEM images of 25
different tubes taken from the same sample. In Fig. 2c, the CNT
diameter distribution is compared with that of the ECD nanopar-
ticles obtained from the SEM image analysis. An excellent agree-
ment between the diameters of the as-deposited nanoparticles and
the subsequently grown CNT was found with average diameters of
17 Ϯ 6 and 18 Ϯ 4 nm for the nanoparticles and CNT, respec-
tively. It was not possible to repeat such detailed statistics for all
samples; however, from the SEM analysis, a good relationship be-
tween the particle size and the CNT diameter was found for particle
diameters between 15 and 25 nm. For larger particles ͑30–40 nm͒,
the CNT diameter remained at around 20 nm as the particles re-
shaped during the growth step ͑not shown͒.40 Our observations con-
firm earlier reports that the CNT diameter is closely linked to the
diameter of the catalyst nanoparticles.41-43 More importantly how-
ever, in our case, the CNT diameter can be directly tailored by the
ECD parameters because they determine particle size, size distribu-
tion, and density. Note that the CNT curve in Fig. 2 shows a nar-
rower size distribution than that for the nanoparticles. Nevertheless,
this might simply be due to a difference in precision obtained from
the SEM imaging vs HRTEM measurements.
sition. Even though the same deposition potential of Ϫ1.5 V was
used, much shorter deposition times were necessary for the pat-
terned surfaces compared to the blanket substrates due to the reduc-
tion in the active surface area and the consequent increase in the flux
of Ni͑NH3͒2+ ions toward the TiN microelectrodes. The Ni nanopar-
6
ticle diameters were 17 Ϯ 6 and 15 Ϯ 5 nm, respectively, after 0.6
and 0.1 s of deposition on blanket and patterned substrates.
Figure 4b shows a top view image of nanoparticles deposited in
150 and 300 nm vias located in dense and isolated via arrays. Simi-
lar particle densities ͑2 ϫ 1011 cm−2
͒
and average sizes
͑15 Ϯ 5 nm͒ were found in both 150 and 300 nm dense and iso-
lated vias. As the effect of neighboring vias on the flux of
2+
6
Ni͑NH3͒ ions toward the bottom of a single via is still limited at
short times, no significant pattern density effects were observed dur-
ing the nanoparticle deposition stage.44,45 Note that the obtained
densities are comparable for the patterned and blanket depositions
͑ϳ1011 cm−2͒. Hence, information obtained from blanket substrates
can be directly transferred to patterned substrates, which consider-
ably simplifies process optimization.
As ECD nickel nanoparticles are formed at room temperature, no
additional high temperature shaping treatments are needed. To con-
firm that ECD Ni nanoparticles do not coalesce at the growth tem-
perature, particles have been recounted after plasma pretreatment at
650°C, and the same size and density were found for the as-
deposited and thermal-plasma-treated nanoparticles.
Next, CNT growth from selectively electrodeposited Ni catalysts
in patterned structures was pursued. Patterned structures of 150–300
nm via-hole arrays on top of TiN were employed. An SEM image of
two neighboring via-hole arrays is shown in Fig. 3a. First, nickel
was selectively electrodeposited inside the via holes, as shown in the
cross-section SEM images of Fig. 4a. As the SiO2 dielectric masks
the TiN substrate, except in those areas where the hole arrays were
etched, nickel nanoparticles are formed directly and selectively at
the bottom of the via holes ͑cf. through-mask plating͒, thus allowing
site-selective growth of CNT later on.31,39
Finally, CNTs were selectively grown from the ECD nickel
nanoparticles placed inside the patterned via holes. As expected, the
CNTs grew only in the patterned domains and not in the surrounding
field, as shown in Fig. 3b. At high magnification, the selective
growth of CNT bundles from the via’s bottom can be seen ͑Fig. 3c͒.
The reactivity of the nanoparticles for CNT growth and the degree
of CNT alignment was strongly dependent on the catalyst activation.
Maintaining the 5 min plasma pretreatment step ͑200 sccm H2 at 1
Torr, 200 W͒ used for the catalyst activation on blankets, vertically
aligned CNT bundles were grown from the 300 nm dense and iso-
lated vias; however, only sparse growth was observed in the 150 nm
vias, as can be seen by comparing Fig. 3c, 1c, and 2. Hence, particle
activation was not sufficient after 5 min in the smaller openings.
When a longer plasma pretreatment of 10 min instead of 5 min was
used, excellent particle reactivity was achieved for both the 300 and
150 dense arrays. Hence, longer plasma pretreatment times are nec-
essary to activate the catalyst nanoparticles inside the smaller cavi-
Figure 4 shows nickel nanoparticles obtained after 0.1 s of depo-
Downloaded on 2014-03-30 to IP 134.208.103.160 address. Redistribution subject to ECS terms of use (see ecsdl.org/site/terms_use) unless CC License in place (see abstract).

K50
Journal of The Electrochemical Society, 157 ͑3͒ K47-K51 ͑2010͒
direct control of the CNT diameter and density through the catalyst
deposition step as well as a selective placement of the catalyst. Both
are important steps toward the device integration and the demonstra-
tion of potential CNT applications.
Acknowledgments
This work was partly supported by the EU projects CARBon-
CHIP ͑NMP4-CT-2006-016475͒, SEA-NET Pro-Nano ͑IST-
027982͒, and METACEL project ͑IWT-SBO project 060031͒,
funded by IWT-Vlaanderen. The authors thank Hefin Griffiths of
Oxford Instruments Plasma Technologies U.K. for the experiments
in their PECVD reactor, and Hugo Bender and Olivier Richard of
IMEC for the HRTEM.
IMEC assisted in meeting the publication costs of this article.
References
1. S. Iijima, Nature (London), 354, 56 ͑1991͒.
2. V. N. Popov, Mater. Sci. Eng. R., 43, 61 ͑2004͒.
3. R. H. Baughman, A. A. Zakhidov, and W. A. de Heer, Science, 297, 787 ͑2002͒.
4. P. G. Collins and P. Avouris, Sci. Am., 283, 62 ͑2000͒.
5. F. Kreupl, A. P. Graham, G. S. Duesberg, W. Steinhogl, M. Liebau, E. Unger, and
W. Honlein, Microelectron. Eng., 64, 399 ͑2002͒.
6. W. Hoenlein, Jpn. J. Appl. Phys., Part 1, 41, 4370 ͑2002͒.
7. N. Srivastava, R. V. Joshi, and K. Banerjee, in IEEE/IEDM, IEEE, p. 257 ͑2005͒.
8. M. Nihei, D. Kondo, A. Kawabata, S. Sato, H. Shioya, M. Sakaue, T. Iwai, M.
Ohfuti, and Y. Awano, in Proceedings of IEEE/IITC, IEEE, p. 234 ͑2005͒.
9. S. Maruyama, R. Kojima, Y. Miyauchi, S. Chiashi, and M. Kohno, Chem. Phys.
Lett., 360, 229 ͑2002͒.
10. S. Hofmann, B. Kleinsorge, C. Ducati, A. C. Ferrari, and J. Robertson, Diamond
Relat. Mater., 13, 1171 ͑2004͒.
11. M. Cantoro, S. Hofmann, S. Pisana, V. Scardaci, A. Parvez, C. Ducati, A. C.
Ferrari, A. M. Blackburn, K. Y. Wang, and J. Robertson, Nano Lett., 6, 1107
͑2006͒.
12. G. Zhang, D. Mann, L. Zhang, A. Javey, Y. Li, E. Yenilmez, Q. Wang, J. P.
McVittie, Y. Nishi, J. Gibbons, et al., Proc. Natl. Acad. Sci. U.S.A., 102, 16141
͑2005͒.
13. M. Hu, Y. Murakami, M. Ogura, S. Maruyama, and T. Okubo, J. Catal., 225, 230
͑2004͒.
14. J. Li, Q. Ye, A. Cassell, H. T. Ng, R. Stevens, J. Han, and M. Meyyappan, Appl.
Phys. Lett., 82, 2491 ͑2003͒.
15. S. Sato, M. Nihei, A. Mimura, A. Kawabata, D. Kondo, H. Shioya, T. Iwai, M.
Mishima, M. Ohfuti, and Y. Awano, in Proceedings of IEEE/IITC, IEEE, p. 204
͑2007͒.
Figure 4. ͑Color online͒ ECD nickel nanoparticles selectively deposited
inside via patterns at Ϫ1.5 V vs Ag/AgCl for 0.1 s: ͑a͒ SEM cross section of
150 and 300 nm vias from dense arrays; ͑b͒ top view SEM images showing
similar particle density and size in 150 and 300 nm isolated ͑pitch/diameter
of 8͒ and dense ͑pitch/diameter of 2͒ arrays; ͑b.1͒ 2.3 ϫ 1011 cm−2
15 Ϯ 7 nm; ͑b.2͒ 2.0 ϫ 1011 cm−2, 17 Ϯ 5 nm; ͑b.3͒ 2.4 ϫ 1011 cm−2
15 Ϯ 3 nm; and ͑b.4͒ 2.3 ϫ 1011 cm−2, 14 Ϯ 3.5 nm.
,
,
16. T. Iwasaki, R. Morikane, T. Edura, M. Tokuda, K. Tsutsui, Y. Wada, and H.
Kawarada, Carbon, 45, 2351 ͑2007͒.
17. S. Hermann, R. Ecke, S. Schulz, and T. Gessner, Microelectron. Eng., 85, 1979
͑2008͒.
18. S. Sato, A. Kawabata, M. Nihei, and Y. Awano, Chem. Phys. Lett., 382, 361
͑2003͒.
19. S. Sato, A. Kawabata, D. Kondo, M. Nihei, and Y. Awano, Chem. Phys. Lett., 402,
149 ͑2005͒.
ties of the patterned structures compared to larger features and blan-
ket substrates. The need for a longer pretreatment step for efficient
catalyst activation is most probably due to the profile of the pat-
terned substrates as it may hinder the plasma activation, especially
in the smaller cavities.
The growth rate was independent of catalyst activation time pro-
vided that the particles were sufficiently activated. About 1 ␮m
long CNT bundles were grown after a 30 s exposure to 100 sccm
acetylene at 1 Torr and at 650°C in the 300 nm dense and isolated
vias for both the 5 and 10 min pretreatment conditions. For the 5
min pretreatment condition, increasing the growth time to 1 min
resulted in twice as long tubes with similar densities ͑i.e., linear
growth rate of 33 nm s−1 for the first minute͒. After 10 min of
growth ͑as shown in Fig. 3b͒, 6 ␮m long aligned tubes had grown
out of the vias, which was about half of the length obtained on the
blankets. Moreover, unlike the CNT bundles grown out of the vias,
the CNT part that was still inside the vias now showed a disordered
and defective morphology. As a tip growth process was observed
and the quality of the CNT inside the vias was fine at short growth
times, the damage after such long growth times must have been
introduced after the growth. Indeed, a significant increase in the TiN
surface roughness was observed at these high temperatures after
annealing the substrate. At this point, it is not completely clear what
causes this postgrowth damage; however, changes in the substrate
morphology could be one explanation.
20. G. Oskam, J. G. Long, A. Natarajan, and P. C. Searson, J. Phys. D: Appl. Phys.,
31, 1927 ͑1998͒.
21. Y. Tu, Z. P. Huang, D. Z. Wang, J. G. Wen, and Z. F. Ren, Appl. Phys. Lett., 80,
4018 ͑2002͒.
22. D. Liu, J. Chen, W. Deng, H. Zhou, and Y. Kuang, Mater. Lett., 58, 2764 ͑2004͒.
23. M. Tamura, Y. Kemmochi, Y. Murakami, N. Chino, M. Ogura, S. P. Naik, M.
Takai, Y. Tsuji, S. Maruyama, and T. Okubo, Appl. Phys. A: Mater. Sci. Process.,
84, 247 ͑2006͒.
24. Y. Xia, P. Yang, Y. Sun, Y. Wu, B. Mayers, B. Gates, Y. Yin, F. Kim, and H. Yan,
Adv. Mater., 5, 353 ͑2005͒
25. A. J. Yin, J. Li, W. J. A. J. Benett, and J. M. Xu, Appl. Phys. Lett., 79, 1039 ͑2001͒.
26. G. Hautier, J. D’Haen, K. Maex, and P. M. Vereecken, Electrochem. Solid-State
Lett., 11, K47 ͑2008͒.
27. N. S. Pesika, A. Radisic, K. J. Stebe, and P. C. Searson, Nano Lett., 6, 1023 ͑2006͒.
28. M. J. Williamson, R. M. Tromp, P. M. Vereecken, R. Hull, and F. M. Ross, Nature
Mater., 2, 532 ͑2003͒.
29. A. Radisic, J. B. Hannon, P. M. Vereecken, P. C. Searson, and F. M. Ross, Nano
Lett., 6, 238 ͑2006͒.
30. F. Iacopi, P. M. Vereecken, M. Schaekers, M. Caymax, N. Moelans, B. Blanpain,
O. Richard, C. Detavernier, and H. Griffiths, Nanotechnology, 18, 505307 ͑2007͒.
31. A. Romo-Negreira, O. Richard, S. De Gendt, K. Maex, M. M. Heyns, and P. M.
Vereecken, Sci. Adv. Mater., 1, 86 ͑2009͒.
32. A. E. Kaloyeros and E. Eisenbraun, Annu. Rev. Mater. Sci., 30, 363 ͑2000͒.
33. A. Romo-Negreira, P. M. Vereecken, C. M. Whelan, G. Groeseneken, and K.
Maex, ECS Trans., 2͑6͒, 409 ͑2007͒.
34. W. Davison and J. A. Harrison, J. Electroanal. Chem. Interfacial Electrochem., 36,
399 ͑1972͒.
Conclusions
35. G. O’Mallory, in Electroless Plating: Fundamentals and Applications, G.
O’Mallory and J. B. Hajdu, Editors, ASCE ͑1990͒.
36. A. Romo-Negreira, D. J. Cott, A. S. Verhulst, S. Esconjauregui, N. Chiodarelli, J.
Ek-Weis, C. M. Whelan, G. Groeseneken, M. M. Heyns, S. De Gendt, et al., Mater.
In summary, we showed that electrodeposition allows the room-
temperature formation of catalyst nanoparticles with a fixed size and
density within specific areas on the wafer. This method provides a
Downloaded on 2014-03-30 to IP 134.208.103.160 address. Redistribution subject to ECS terms of use (see ecsdl.org/site/terms_use) unless CC License in place (see abstract).

Journal of The Electrochemical Society, 157 ͑3͒ K47-K51 ͑2010͒
K51
Res. Soc. Symp. Proc., 1079E, N06-01 ͑2008͒.
42. C. L. Cheung, A. Kurtz, H. Park, and C. M. Lieber, J. Phys. Chem. B, 106, 2429
37. http://rsbweb.nih.gov/ij/, last accessed Jan 2010.
͑2002͒.
38. M. Fleischmann and A. Saraby-Reintjes, Electrochim. Acta, 29, 69 ͑1984͒.
39. A. J. Bard and L. R. Faulkner, Electrochemical Methods, Fundamentals and Ap-
plication, 2nd ed., John Wiley & Sons, New York ͑2000͒.
40. A. Romo-Negreira, Ph.D. Thesis, Katholieke Universiteit Leuven, Leuven, Bel-
gium ͑2010͒.
41. E. F. Kukovitsky S. G. L’vov, N. A. Sainov, V. A. Shustov, and L. A. Chernoza-
tonskii, Chem. Phys. Lett., 355, 497 ͑2002͒.
43. Y. Zhang, Y. Li, W. Kim, D. Wang, and H. Dai, Appl. Phys. A: Mater. Sci. Process.,
74, 325 ͑2002͒.
44. S. Mehdizadeh, J. O. Dukovic, P. C. Andricacos, L. T. Romankiw, and H. Y. Cheh,
J. Electrochem. Soc., 139, 78 ͑1992͒.
45. K. Stulík, C. Amatore, K. Holub, V. Marecek, and W. Kutner, Pure Appl. Chem.,
72, 1483 ͑2000͒.
Downloaded on 2014-03-30 to IP 134.208.103.160 address. Redistribution subject to ECS terms of use (see ecsdl.org/site/terms_use) unless CC License in place (see abstract).