Cu electrodeposition from methanesulfonate electrolytes for ULSI and MEMS applications

Article abstract of DOI:10.1149/1.3332729

Methanesulfonic acid (MSA) is an alternative to sulfuric acid electrolyte for metal deposition. The electrochemical nucleation and growth of Cu on a glassy carbon electrode in methanesulfonate was compared with sulfate baths. The overpotential for Cu deposition was much smaller in the MSA bath compared to the traditional sulfuric acid bath, and Cu nucleation occurred at a higher rate in the MSA bath. The measured diffusion coefficient value for Cu deposition from the MSA bath was 6.82× 10^-6 cm² /s. UV-visible spectroscopic results confirmed that the coordination of Cu species was the same in both electrolytes. Cu electrodeposition on Ni sputtered Si substrate from the high efficiency MSA bath was photoresist-compatible with no void formation. One-dimensional Cu nanorods were also deposited through an anodized aluminum oxide template on a Ni evaporated seed layer substrate, showing potential applications as electrical interconnects in ultralarge scale integration (ULSI) and microelectromechanical systems (MEMS).

Full text of DOI:10.1149/1.3332729

D278
Journal of The Electrochemical Society, 157 ͑5͒ D278-D282 ͑2010͒
0
013-4651/2010/157͑5͒/D278/5/$28.00 © The Electrochemical Society
Cu Electrodeposition from Methanesulfonate Electrolytes
for ULSI and MEMS Applications
,
z
Maksudul Hasan* and James F. Rohan**
Tyndall National Institute, University College Cork, Lee Maltings, Cork, Ireland
Methanesulfonic acid ͑MSA͒ is an alternative to sulfuric acid electrolyte for metal deposition. The electrochemical nucleation and
growth of Cu on a glassy carbon electrode in methanesulfonate was compared with sulfate baths. The overpotential for Cu
deposition was much smaller in the MSA bath compared to the traditional sulfuric acid bath, and Cu nucleation occurred at a
higher rate in the MSA bath. The measured diffusion coefficient value for Cu deposition from the MSA bath was 6.82
ϫ 10−6 cm /s. UV-visible spectroscopic results confirmed that the coordination of Cu species was the same in both electrolytes.
Cu electrodeposition on Ni sputtered Si substrate from the high efficiency MSA bath was photoresist-compatible with no void
formation. One-dimensional Cu nanorods were also deposited through an anodized aluminum oxide template on a Ni evaporated
seed layer substrate, showing potential applications as electrical interconnects in ultralarge scale integration ͑ULSI͒ and micro-
electromechanical systems ͑MEMS͒.
2
©
2010 The Electrochemical Society. ͓DOI: 10.1149/1.3332729͔ All rights reserved.
Manuscript submitted December 10, 2009; revised manuscript received January 26, 2010. Published April 7, 2010.
Cu is the present and future interconnect material in high end
microprocessors and memory devices because of its lower electrical
resistivity and higher electromigration resistance than aluminum.
Dual damascene Cu electroplating is now commonly used in semi-
conductor devices usually employing a mixture of CuSO /H SO .
ated direct electrodeposition technique from the MSA plating bath
for interconnects in ultralarge scale integration ͑ULSI͒ and micro-
electromechanical systems ͑MEMS͒ applications.
Experimental
4
2
4
To achieve a so-called bottom-up or superconformal deposit, various
Cu methanosulfonate was prepared as follows. Cu͑OH͒2 was
first precipitated from the sulfate solution using NaOH ͑excess͒ ac-
cording to the following equation
1,2
types of organic additives are also added. Moreover, additives
such as bis-͑3-sodiumsulfopropyl disulfide͒ that are used as accel-
3
erators enhance the bottom-up fill capability of Cu electroplating.
CuSO ·5H O + 2NaOH → Cu͑OH͒ + Na SO + 5H O ͓1͔
4
2
2
2
4
2
The inhibiting effects of poly͑ethylene glycol͒ ͑PEG͒ during Cu
4
electroplating has also been studied. PEG–Cl electrolyte that is
Cu sulfate pentahydrate ͑31.21 g, 0.125 mol; Fisher Scientific, ana-
lytical reagent grade͒ was dissolved in water and slowly poured in
an aqueous solution of sodium hydroxide ͑10 g, 0.25 mol; Sigma-
Aldrich, reagent grade, 97%͒. The bluish slurry Cu hydroxide was
washed with deionized ͑DI͒ water. The filtrate was dried in vacuo ͑1
mm Hg͒ at 50°C and used in the synthesis of Cu methanesulfonate.
The reaction proceeds according to the following equation
commonly used as a suppressor in semiconductor Cu electroplating
bath decomposes to PEG of smaller molecular weight at the cath-
ode. The smaller molecular weight PEG has less adsorption ability
on the electrode surface, and thus its inhibiting effect on Cu reduc-
tion decreases gradually.
In this study, we report that methanesulfonic acid ͑MSA͒ is an
alternative electrolyte system that could replace sulfuric acid in
practical applications. MSA is a strong electrolyte and its conduc-
tivity in water is similar to other strong acids such as sulfuric or
Cu͑OH͒ + 2CH SO H → Cu͑O SCH ͒ ·4H O + 2H O ͓2͔
2
2
2
2
2 2
2
2
The powdered Cu hydroxide ͑12.0 g, approximately 0.125 mol͒ was
dissolved in MSA ͑24.025 g, 0.250 mol; Sigma-Aldrich, 99% anhy-
drous͒ at 80°C for 1 h, and enough DI water was then added to
produce a homogeneous solution. The formation of Cu methane-
sulfonate salt was induced by isopropyl alcohol ͑IPA͒. IPA ͑100 mL͒
was poured into the above solution at 50°C and left overnight for
crystallization, the crystalline precipitate was collected by vacuum
filtration and washed with IPA and ether. The solid product was
brought to constant weight in vacuum ͑1 mm Hg͒. The compound
5
hydrochloric acid and higher than that of other organic acids. Nev-
ertheless, MSA has a “green” character in two different ways. First,
it is odorless and does not generate toxic gas fumes, which make it
5
very safe to handle. Second, it is readily biodegradable and
6
recyclable. Recycling of MSA is readily achieved because of its
excellent solubility in water so that it can be extracted from an
organic phase with small amounts of water. These result in easily
treatable effluents that are less hazardous compared to all other com-
mercial baths such as sulfate, chloride, fluoride, etc. MSA also at-
tracts great attention because of its nonoxidant characteristics com-
pared to other traditional electrolytes, i.e., sulfate, nitrates, etc.
Additionally, MSA solutions are easy to handle because they remain
in liquid form down to −60°C ͑H SO :3°C͒, and they have greater
Cu͑O SCH ͒ ·4H O was characterized by ethylenediaminetetraace-
3
3 2
2
tic acid complexometric titration, CHN analysis ͑Interscience Ce
Instrument EA1110͒.
Cu deposition/dissolution experiments were carried out in a
three-electrode electrochemical cell. The working electrode glassy
carbon ͑GC͒ was constructed from a 3 mm diameter vitreous carbon
2
4
thermal stability than other organic acids. The baths produce a re-
duced amount of waste and sludge, and thus the early investment of
resources can be minimized. Nevertheless, information on the fun-
damental properties of such electrolytes in the “open” scientific lit-
erature is very scarce. This paper reports the basic electrochemistry
of Cu metal ion deposition in aqueous MSA solution by comparison
with the traditional sulfate bath. Here, we present electrodeposition
of Cu on Ni sputtered Si substrate patterned using optical photoli-
thography. We also report the fabrication of one-dimensional ͑1D͒
Cu nanorods based on the anodized aluminum oxide ͑AAO͒ medi-
2
rod with an active surface area of 0.0803 cm ͑geometrical area:
2
0
.07065 cm ͒. The active surface area was determined using the
well-characterized ferricyanide system on the GC electrode, and the
active area was determined using the Cottrell equation. The elec-
trode surface was successively polished with 0.3, 0.1, and 0.05 ␮m
alumina powder ͑Struers͒ to a mirror finish and was ultrasonically
rinsed before the experiment. Two equal molar solutions for these
experiments consisting of 0.05 M CuSO ·5H O ͑Fisher Scientific,
4
2
analytical reagent grade͒ in 1 M H SO ͑Air Products, 96%͒ and
2
4
0
.05 M Cu͑O SCH ͒ ·4H O ͑in house͒ in 1 M MSA ͑Sigma-
3 3 2 2
Aldrich, 99% anhydrous͒ were used. All electrochemical experi-
ments were carried out with a CHI 660C potentiostat ͑CH Instru-
ments, Inc.͒ coupled to a personal computer, a Ti/Pt mesh counter
*
Electrochemical Society Student Member.
*
* Electrochemical Society Active Member.
z
E-mail: james.rohan@tyndall.ie
electrode, and a H reference electrode. The temperature of the cell
2
Downloaded on 2014-10-30 to IP 155.33.16.124 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 ͑5͒ D278-D282 ͑2010͒
D279
Figure 2. ͑Color online͒ CVs of 1 M CH SO H ͑solid line͒ and 1 M H SO
4
3
3
2
͑
dotted line͒ on a GC disk electrode ͑conditioning/initial potential: 0.4 V, 20
Figure 1. ͑Color online͒ Schematic illustration of the synthesis process of
s; scan rate: 0.01 V/s͒.
the high aspect ratio Cu nanorod arrays.
Results and Discussion
was thermostatically controlled at 20 Ϯ 1°C. Before a series of ex-
periments, the electrolytic solution was deaerated for 2 h with high
purity nitrogen, which was presaturated with water. The ultraviolet-
visible ͑UV-vis͒ spectroscopy experiments of two equimolar solu-
The electrochemical window of the background electrolyte of 1
M aqueous MSA and sulfuric acid solutions ͑Fig. 2͒ in the cathodic
direction was measured by cyclic voltammetry ͑CV͒. The small cur-
rent, which remains constant over the potential interval from 0 to
Ϫ0.3 V, is not due to a faradaic process but is merely a capacitive
current associated with the action of a continuous potential change
during the CV experiment. The sharp current rise at Ϫ0.4 V is due to
hydrogen evolution. Thermodynamically, hydrogen evolution on an
electrode surface can occur starting from 0 VSHE ͑where SHE de-
notes standard hydrogen electrode͒. In this case, an overpotential of
0.3 V for hydrogen evolution reflects the activation barrier that has
to be overcome for the formation of hydrogen gas on the GC elec-
trode in the MSA bath. No electrochemical reaction on the GC elec-
trode takes place at potentials more positive than Ϫ0.3 V.
tions of 0.05 M Cu͑O SCH ͒·4H O and CuSO ·5H O were carried
3
3
2
4
2
out with a CARY 5000 spectrophotometer to compare the coordina-
2
+
tion of Cu ions into solutions with the reference solutions of 1 M
CH SO H and 1 M H SO , respectively.
3
3
2
4
Cu electrodeposition on Si substrates was carried out using a
mask with arrays of ͑100 ϫ 100 ␮m͒ features and a pitch of
4
00 ␮m. These features were patterned in a Ni sputtered Si sub-
strate by a photolithographic technique. A 100 mm diameter Si wa-
fer was used as a substrate, with thin layers of Ti ͑200 Å͒ and Ni
͑
1000 Å͒ physically deposited by the sputtering process. Here, the Ti
layer improves the adhesion of Ni to the Si substrate, which acts as
the conducting seed layer. A positive photoresist AZ 9260 was spun
on the wafer to a height of 12 ␮m, UV exposed, and developed in
an AZ 400k developer. Cu was deposited in the developed photore-
sist mold from 0.1 M Cu͑O SCH ͒ ·4H O/1 M CH SO H at
Figure 3a shows two cyclic voltammograms of the Cu deposition
obtained from methanesulfonate and sulfate baths during the for-
ward sweep to lower potentials. In methanesulfonate, a sharp rise at
around 0.125 V is observed by comparison with 0.05 V in the sulfate
bath, which passes through a maximum at around 0 V, correspond-
ing to Ϫ0.160 V for sulfate bath, and then decreases due to the
increase in the diffusion layer thickness. In the reversed sweep, the
current passes through zero at around 0.275 V before oxidation com-
mences. This crossover potential is a good approximation of the
equilibrium potential. The cathodic peak maximum is shifted by
0.160 V to a more positive potential value in methanesulfonate bath
3
3 2
2
3
3
2
0 Ϯ 1°C, with the Ti/Pt mesh as a counter electrode. After electro-
plating, the resist was stripped in acetone; the wafer was then rinsed
and dried with nitrogen. 1D nanoscale electrodeposition for noble
interconnects in the three-dimensional ͑3D͒ integration of ULSI and
MEMS applications was performed on commercial AAO Anodisc
membranes ͑Whatman, 109 pores/cm2͒. The complete fabrication
͑
full line͒ compared to the sulfate bath ͑dotted line͒, thus the critical
overpotential for Cu deposition on the GC electrode is much lower
in the methanesulfonate. In addition, the cathodic peak current ͑I
process of Cu nanorods is illustrated in Fig. 1. A 500 nm thin layer
of Ni conducting layer was deposited onto one side of the AAO
template by the E-beam evaporation technique ͑Temescal FC-2000͒.
p
2
The electrodeposition was carried out in 0.24
M
= 0.019 A/cm ͒ in the methanesulfonate bath is also higher than
that obtained in the sulfate bath ͑I = 0.014 A/cm ͒, implying that
2
Cu͑O SCH ͒ ·4H O/1.8 M CH SO H at 20 Ϯ 1°C with a slow
3
3 2
2
3
3
p
convection of the electrolyte at a constant current of 40 mA. Based
on the available surface area on the exposed back side of the AAO
the amount of Cu deposit at the same potential interval and sweep
rate is particularly higher in the methanesulfonate bath. The nucle-
ation loops shown in Fig. 3b are an indication of the nucleation and
growth process of Cu on the GC substrate. The initial adsorption and
͑
21 mm in diameter͒ and in the pores of the template, 40 mA may be
2
equated with 6.6 mA/cm . In a two-electrode setup, a Ti/Pt mesh
was the anode and a Ni seed layer on the AAO substrate was the
working electrode ͑cathode͒. A Cu wire was connected to the Ni
conducting side of the template by a silver conductive paint ͑Radi-
onics Ltd., Ireland͒ and left to dry overnight before use. The mor-
phology of the Cu deposits was analyzed by a scanning electron
microscope ͑SEM, FEI Nova 630 Nano-SEM and Hitachi S-4000 at
2
+
nucleation rate of Cu ions at the cathode is faster compared to that
in the sulfate bath. Further, on the reverse sweep, when E Ͼ E , the
eq
anodic peak corresponding to the dissolution of Cu is observed.
CV is not always an accurate quantitative technique because the
electrode potential changes continuously, and so does the kinetics
for electrochemical reactions, during the time scales of the
7
1
5 kV͒ and characterized by an X-ray powder diffractometer ͑Phil-
lips PW3710-MPD with Cu K␣ radiation, l = 1.54056 Å, at 40 kV
5 mA͒, and data were analyzed using a Philips X’pert X-ray dif-
fraction ͑XRD͒ software.
experiment. Therefore, the kinetics of Cu deposition is also inves-
tigated by chronoamperometric experiments. Two chronoampero-
metric current transients from methanesulfonate and sulfate baths
are shown in Fig. 4, in which the electrode potential was instanta-
3
Downloaded on 2014-10-30 to IP 155.33.16.124 address. Redistribution subject to ECS terms of use (see ecsdl.org/site/terms_use) unless CC License in place (see abstract).

D280
Journal of The Electrochemical Society, 157 ͑5͒ D278-D282 ͑2010͒
Figure 4. ͑Color online͒ Chronoamperograms of Cu deposition from meth-
anesulfonate ͑solid line͒ and sulfate ͑square line͒ baths at Ϫ0.1 V on a GC
disk electrode ͑conditioning/initial potential: 0.4 V, 10 s, deposition time: 5
s͒. The inset is the enlarged initial time region.
sion coefficient value of Cu2+ ions is 6.82 ϫ 10−6 cm /s ͓from the
2
ͱ
initial part of the Cottrell plot ͑I/A vs 1/ t͒ with a correlation coef-
ficient value of 0.99994͔. There is no prior literature available for
2
+
the diffusion coefficient value of Cu ions in methanesulfonate so-
lutions. The diffusion coefficient measured for the MSA electrolyte
−
6
2
is comparable to the mean literature value of 7.54 ϫ 10 cm /s
8
2+
cited by Quickenden and Xu for Cu in sulfate systems.
Figure 5 shows the UV-vis spectrometry of two solutions, where
the maximum of the absorption band in the CuSO ·5H O solution is
4
2
at 810 nm and no shift of the maximum absorption is apparent in the
Cu͑O SCH ͒ ·4H O solution. This indicates that the Cu species in
3
3 2
2
both electrolytes have the same coordination with respect to sur-
rounding anions and water molecules ͓Cu͑H O͒ ͔. The high effi-
2
6
ciency of the methanesulfonate bath can be interpreted, at least par-
tially, in terms of the methanesulfonate anion acting as an
accelerator when added as the Cu salt itself.
Figure 3. ͑Color online͒ ͑a͒ CVs of Cu deposition from methanesulfonate
͑
solid line͒ and sulfate ͑dotted line͒ baths on a GC disk electrode and ͑b͒
Figure 6a shows a Cu deposit array in a 12 ␮m thick AZ 9260
photoresist mold after photoresist removal. The experiment demon-
strated the chemical stability of the photoresist in the methane-
sulfonate bath and the Cu electrodeposition efficiency without addi-
close-up view of the nucleation loops of copper deposition ͑conditioning/
initial potential: 0.4 V, 20 s; scan rate: 0.05 V/s͒.
neously changed from 0.4 V ͑a value where no electrochemical ac-
tivity occurs͒ to Ϫ0.1 V in the overpotential deposition region. Both
the transients increase through a maximum and reach the same cur-
rent value at longer times. This behavior is expected for reduced
species that are insoluble in the electrolyte such as in metal nucle-
ation and phase growth. However, Fig. 4 shows that at the same
potential steps, the current transient in the methanesulfonate bath is
significantly steeper than that in the sulfate bath, which represents a
faster nucleation rate in this bath at similar potentials.
2
+
The determination of the diffusion coefficient of Cu ions in the
7
methanesulfonate bath is estimated using the Cottrell equation
D
I͑t͒ = nFAC
ͩ
ͱ
ͪ
͓3͔
␲
t
in which n is the number of electrons transferred per molecule, F is
Faraday’s constant ͑C/mol͒, A is the electrode area ͑cm ͒, D is the
diffusion coefficient ͑cm /s͒, and C is the concentration of the spe-
cies in the solution ͑mol/cm ͒. A potential step was applied to the
2
2
3
GC electrode from an initial value of 0.4 V ͑where no electrochemi-
cal reaction occurs͒ to a final value of Ϫ0.125 V and held at this
potential for 5 s. Afterward, the copper deposit was removed by
applying an anodic potential of 0.7 V for 30 s. The measured diffu-
Figure 5. ͑Color online͒ The absorbance as a function of wavelength ͑␭͒ in
methanesulfonate ͑solid line͒ and sulfate ͑filled square line͒ baths.
Downloaded on 2014-10-30 to IP 155.33.16.124 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 ͑5͒ D278-D282 ͑2010͒
D281
Figure 7. ͑Color online͒ Sheet resistance curve for 1 ϫ 1 cm square size
and 1.02 ␮m thick Cu film.
␲
d a
,
G = t ·
· C
ͩ
ͪ
͓5͔
ln͑2͒
s d
where C is the additional correction factor due to the finite dimen-
sion of the film, a and d are the length and breadth, respectively, and
s is the probe spacing. Figure 7 shows the sheet resistance curve for
which current and voltage are linearly proportional. Using the value
of G suggested by Smits, the estimated sheet resistivity is
.12 ␮⍀ cm ͑G = 4.2209 for a = d, 10 mm and s = 10 m; here
11
−
3
2
−
6
t = 1.02 ϫ 10 m, slope = 0.00494 ⍀ m͒; this value is compa-
1
2
rable to available literature data.
Figure 8a shows the SEM image of the well-ordered and homo-
geneous growth of Cu nanorod arrays from a bath of 0.24 M
Cu͑O SCH ͒ ·4H O in 1.8 M CH SO H. The average diameter of
3
3 2
2
3
3
the Cu nanorods is measured to be 290 nm, which correlates with
the value of the pore diameter of the AAO template. The deposition
rate is 0.55 ␮m/min at 40 mA with an additive-free solution by
1
3
comparison with the slower deposition rate of 0.32 ␮m/min for a
Cu sulfate bath operated under the same conditions. When the typi-
−
cal additives such as PEG and Cl were added to the MSA electro-
lyte, the templated deposits were solid wires rather than tubes as
achieved in the sulfate system described in Ref. 13. Thus, it would
appear that the suppression effects of that combination of additives
do not govern the deposition type in the MSA bath, and high rate
solid structures are achieved. Indeed, it may be that the MSA anion
acts as an accelerator for Cu deposition and leads to the enhanced
kinetics for deposition observed in voltammetry and chronoamper-
ometry. The XRD pattern reveals that Cu nanorods are polycrystal-
line, having a preferred orientation of ͑111͒; the peaks from the Ni
seed layer are also apparent ͑see Fig. 8b͒. The preferred crystalline
orientation of deposits depend on the conditions of electrodeposition
such as current density, bath composition, electrolyte convection,
Figure 6. ͑Color online͒ ͑a͒ SEM of Cu micropad array in a 12 ␮m thick
AZ 9260 photoresist mold after resist removal. ͑b͒ Magnified single pad
showing linear square features with void-free deposit.
tive influence. The Cu was deposited to fill the open structures at
2
1
0 mA/cm for 1 h. The uniformity of the deposited Cu from the
MSA electrolyte is shown in Fig. 6b. It can also be seen that the
dimension of the developed features is a linear square size, i.e., no
sidewall undercutting or voids are observed. Therefore, the
photolithography-assisted Cu interconnect fabrication is viable in
the MSA electrolyte.
1
4
temperature, pH, etc. It has been observed that the increasing ca-
thodic current from 40 to 100 mA or bath temperature from 20 to
4
0°C resulted in the development of the ͑200͒ and ͑220͒ orientations
at the same pH ͑0.45͒.
The properties of metallic films differ remarkably from the bulk
materials when the thickness of the metallic films decreases to the
Conclusion
9,10
submicrometer or nanoscale length.
Here, the electrical resistivity
MSA has been shown to be a potential electrolyte for Cu elec-
trodeposition. In this paper, we presented a preliminary study of Cu
electrodeposition on the GC electrode from the MSA and sulfuric
acid electrolytic solutions. The CV results indicate that the overpo-
tential for the Cu deposition is much lower in the methanesulfonate
bath. Consequently, from the current transients, it is evident that for
a fixed value of the overpotential, the kinetics of Cu nucleation is
larger in the MSA bath compared to that in the sulfuric acid bath.
of Cu thin film is estimated using a four-point probe technique ac-
cording to the following equation
G͑s,t͒V
␳
=
͓4͔
I
G is the geometric factor, has a unit of length, and is given by the
11
following equation
Downloaded on 2014-10-30 to IP 155.33.16.124 address. Redistribution subject to ECS terms of use (see ecsdl.org/site/terms_use) unless CC License in place (see abstract).

D282
Journal of The Electrochemical Society, 157 ͑5͒ D278-D282 ͑2010͒
background electrolyte ions and Cu2+ on the GC interface surface,
2
+
where the coordination strength of Cu ions with the surrounding
anions or water molecules influences the formation of a critical
nucleus that precedes further growth. The enhanced kinetics and
lower overpotential for the Cu electrodeposition using MSA may
result from accelerator effects of the MSA by comparison with the
sulfuric acid electrolyte.
Void-free Cu deposits have also been fabricated by a through-
mold electroplating process on Ni sputtered Si substrates. Cu square
pads ͑100 ϫ 100 ␮m͒ with a pitch of 400 ␮m have been demon-
strated. The electrical resistivity of the as-deposited Cu thin film is
measured to be 2.12 ␮⍀ cm, which correlates with literature val-
ues for Cu from sulfate baths. Cu nanorods were also electrodepos-
ited through an AAO template on Ni evaporated substrate to inves-
tigate submicrometer deposition from the MSA bath. The Cu
methanesulfonate bath is suitable for high rate deposition of void-
free Cu micrometer-scale pads and nanorod interconnects for
additive-free 3D ULSI and MEMS applications.
Acknowledgment
The authors acknowledge Enterprise Ireland Technology Devel-
opment Project CFTD/05/IT/317. M.H. is grateful for the help of
Professor L. Heerman.
University College Cork assisted in meeting the publication costs of this
article.
References
1
2
. T. P. Moffat, D. Wheeler, W. H. Huber, and D. Josell, Electrochem. Solid-State
Lett., 4, C26 ͑2001͒.
. L. T. Koh, G. Z. You, S. Y. Lim, C. Y. Lim, and P. D. Foo, Microelectron. J., 32,
9
73 ͑2001͒.
3
4
5
6
. R. Akolkar and U. Landau, J. Electrochem. Soc., 151, C702 ͑2004͒.
. J. J. Kelly and A. C. West, J. Electrochem. Soc., 145, 3472 ͑1998͒.
. M. D. Gernon, M. Wu, T. Buszta, and P. Janney, Green Chem., 1, 127 ͑1999͒.
. J. Ma, Y. Kale, and M. Turnbull, Biocatalysis/Biodegradation Database, University
of Minnesota ͑1997͒.
7
. A. J. Bard and L. R. Faulkner, Electrochemical Methods. Fundamental and Appli-
cations, p. 218, John Wiley & Sons, New York ͑1980͒.
8
9
. T. I. Quickenden and Q. Xu, J. Electrochem. Soc., 143, 1248 ͑1996͒.
. M. D. Uchic, D. M. Dimiduk, J. N. Florando, and W. D. Nix, Science, 305, 986
͑
2004͒.
0. Z. P. Bažant, Z. Y. Guo, H. D. Espinosa, Y. Zhu, and B. Peng, J. Appl. Phys., 97,
73506 ͑2005͒.
1. F. M. Smits, Bell Syst. Tech. J., 37, 711 1958.
Figure 8. ͑Color online͒ ͑a͒ SEM image of Cu nanorod arrays ͑inset is top
view͒ and ͑b͒ XRD patterns of Cu deposit.
1
0
1
1
2. P. Dixit, C. W. Tan, L. Xu, N. Lin, J. Miao, J. H. L. Pang, P. Backus, and R.
Preisser, J. Micromech. Microeng., 17, 1078 ͑2007͒.
3. T. Chowdhury, D. P. Casey, and J. F. Rohan, Electrochem. Commun., 11, 1203
The diffusion coefficient for the Cu deposition from MSA is esti-
mated to be 6.82 ϫ 10−6 cm /s. The difference in nucleation behav-
ior is most likely due to the difference in the adsorption of the
1
2
͑2009͒.
14. M. Hasan, M.S. Thesis, Katholieke Universiteit Leuven, Belgium ͑2005͒.
Downloaded on 2014-10-30 to IP 155.33.16.124 address. Redistribution subject to ECS terms of use (see ecsdl.org/site/terms_use) unless CC License in place (see abstract).