Diameter-dependent optical constants of gold mesoparticles electrodeposited on aluminum films containing copper

Article abstract of DOI:10.1021/jp0511707

Electrodeposition of gold mesoparticles on anodized and chemically etched aluminum/copper films deposited on silicon wafers proceeds by instantaneous nucleation and with no diffusion limitations. Both of these phenomena favor the formation of relatively monodispersed gold particles. Under the reported electrodeposition conditions, the relative standard deviation of the particle diameter is 25%. The particle coverage is 7 ?? 108 particles cm ^-2. The mean particle diameter varies as a function of electrodeposition time in the range of 40-80 nm. Optical constants of gold mesoparticles are resolved by spectroscopic ellipsometry. A two-layer optical model is constructed to determine both extinction coefficients and refractive indexes of gold mesoparticles as a function of the mean particle diameter. The absorption peak, associated with surface plasmons, is modeled with two Lorentz oscillators. Absorption peak maximums shift from 610 to 675 nm as the mean particle diameter increases from 42 to 74 nm. Electrodeposition of gold particles on technologically relevant substrates, such as aluminum/copper films, is expected to increase the utility of gold particles and facilitate their incorporation in nanostructured materials and a variety of electronic and optical devices. ? 2005 American Chemical Society.

Full text of DOI:10.1021/jp0511707

J. Phys. Chem. B 2005, 109, 14529-14535
14529
Diameter-Dependent Optical Constants of Gold Mesoparticles Electrodeposited on
Aluminum Films Containing Copper
,†
‡
Dmitri A. Brevnov* and Corey Bungay
Department of Chemical and Nuclear Engineering, Center for Micro-Engineered Materials, The UniVersity of
New Mexico, Albuquerque, New Mexico 87131, and J.A. Woollam Co. Inc., Lincoln, Nebraska 68508
ReceiVed: March 7, 2005; In Final Form: May 23, 2005
Electrodeposition of gold mesoparticles on anodized and chemically etched aluminum/copper films deposited
on silicon wafers proceeds by instantaneous nucleation and with no diffusion limitations. Both of these
phenomena favor the formation of relatively monodispersed gold particles. Under the reported electrodeposition
8
conditions, the relative standard deviation of the particle diameter is 25%. The particle coverage is 7 × 10
-
2
particles cm . The mean particle diameter varies as a function of electrodeposition time in the range of
40-80 nm. Optical constants of gold mesoparticles are resolved by spectroscopic ellipsometry. A two-layer
optical model is constructed to determine both extinction coefficients and refractive indexes of gold
mesoparticles as a function of the mean particle diameter. The absorption peak, associated with surface
plasmons, is modeled with two Lorentz oscillators. Absorption peak maximums shift from 610 to 675 nm as
the mean particle diameter increases from 42 to 74 nm. Electrodeposition of gold particles on technologically
relevant substrates, such as aluminum/copper films, is expected to increase the utility of gold particles and
facilitate their incorporation in nanostructured materials and a variety of electronic and optical devices.
and versatile technique.1
0-22
Electrodeposition of metals on
Introduction
conductive substrates covered with a thin dielectric layer is
suitable for fabrication of particulate films because electrodepo-
sition usually occurs by the three-dimensional island growth
Solution-based methods to synthesize metallic, nano- (1-10
nm) and meso- (10 nm to 1 µm) particles, especially of gold
and silver, have been known for decades.1
-3
These particles
1
0-12
(
Volmer-Weber) mechanism.
This mechanism has been
are considered as building units for nanostructured materials,
1
-5
reported for electrodeposition of metallic particles, including
Au, Cu on Si;
smooth graphite basal plane surfaces; Ag on Si; Ag and Au
on Si; and Pb and Cu on Si. Electrodeposition of metallic
particles has been shown to follow models for either instanta-
due to their optical, electrical, and catalytic properties.
For
13
14,15
16,17
18
19
Cu on TaN;
Ag and Pt on atomically
deposition of metallic particles on a surface, the majority of
reported procedures advocate the employment of a wet-
chemistry method involving a reducing agent and a high
2
0
2
1
22
6
oxidation state metal precursor. Subsequently, free-standing
1
0-12
neous or progressive nucleation.
For both models, the
metallic particles prepared in solution are collected and trans-
ferred to a desired substrate. By using this strategy, particles
can be synthesized in large quantities and with a narrow
distribution of particle diameters, with the latter property being
the primary advantage. Regardless of wide utilization of this
approach in research laboratories, two factors limit its practical
usefulness. First, transferring may result in poor adhesion and,
consequently, no electrical connection of metallic particles to
the underlying support. Second, controlling the particle coverage
and aggregation may be problematic. Therefore, to increase the
utility of metallic particles, it is important to develop deposition
methods, which produce strongly adhesive and particulate films
with a narrow distribution of particle diameters. The fabrication
of these films on technologically relevant substrates would allow
us to fully exploit optical and electrical properties of metallic
particles and facilitate their incorporation in functional materials.
Subsequently, these materials can be utilized for fabrication of
waveguides, solar cells, information storage devices, chemical/
biochemical sensors, and a number of other optical and
electronic devices.7
growth of particles can proceed under kinetic, diffusion, or
mixed control. As in solution-based methods, a narrow distri-
bution of particle diameters can be achieved only if nucleation
is instantaneous and, as a result, separated in time from the
growth of particles. Moreover, it is mandatory that, during
electrodeposition, diffusion zones among growing particles do
12
not overlap. Under these two conditions, the particle coverage
is mainly controlled by the applied potential and the thickness
23
of the dielectric layer. The mean particle diameter is controlled
by electrodeposition time, concentration of a reduced species,
10-12
and temperature of deposition solution.
In addition to being
used for deposition of metallic particles, electrodeposition in
combination with chemical methods has been employed for
fabrication of particles composed of metal oxides and semi-
2
1
conductor materials.
Since metallic particles are frequently prepared in solution,
their extinction coefficients are readily determined by UV-vis
spectrophotometry. The optical properties of metallic particles
-9
25-27
vary slightly as a function of particle diameter.
According
Among a variety of methods employed for deposition of
metallic particles, electrodeposition stands out as an inexpensive
to Kreibig and Vollmer, the dielectric function of metallic
particles with a diameter larger than about 10-20 nm has values
2
5
of bulk metal. Accordingly the size dependence of gold
mesoparticles optical response is “an extrinsic cluster-size effect
governed only by the dimension of the particle with respect to
*
Address correspondence to this author. E-mail: dbrevnov@unm.edu.
The University of New Mexico.
J.A. Woollam Co. Inc.
†
‡
1
0.1021/jp0511707 CCC: $30.25 © 2005 American Chemical Society
Published on Web 07/08/2005

1
4530 J. Phys. Chem. B, Vol. 109, No. 30, 2005
Brevnov and Bungay
the wavelength of the light”.25 The spectra of gold mesoparticles
SE to investigate the position of a surface plasmon band as a
function of the mean particle diameter. While the former was
obtained from the analysis of SE spectra, the latter was
determined from the examination of scanning electron micro-
graphs.
are characterized by a strong absorption band due to surface
plasmons.1
,5-6,9
For gold particles, the position and bandwidth
of the absorption band depend on the mean particle diameter,
the particle shape, the distribution of particle diameters, and
9,25-28
the dielectric constant of the surrounding medium.
For
gold particles in an aqueous solution with mean diameters of
Experimental Section
9
, 15, 22, 48, and 99 nm, maximums of surface plasmon bands
Silicon wafers with an intermediate 0.6-µm layer of SiO2 and
29
were observed at 517, 520, 521, 533, and 575 nm, respectively.
3
-µm metallic layer (99.5 wt % aluminum and 0.5 wt % copper)
For silver particles fabricated on transparent substrates, extinc-
tion maximums were tuned over a wide range of wavelengths
by controlling the particle size, height, and shape and by
40
were fabricated and processed as described previously. Gold
electrodeposition on anodized and etched aluminum/copper films
was carried out at room temperature (22 °C), in 1.0 M Na2SO3
3
0
encapsulating these particles in a dielectric medium.
41
(
pH 8), with 20 mM Na3Au(SO3)2 (Technic Inc. Oromerse
In contrast to straightforward determination of extinction
coefficients of metallic particles in solution, the optical constants
of particulate films on reflective substrates can be resolved only
by either UV-visible reflectance spectroscopy or spectroscopic
Part B gold plating solution). Electrodeposition was performed
with an IM6-e impedance measurement unit (Zahner) by
applying a potential step from an open-circuit potential (OCP)
to -3.95 V vs a Ag/AgCl electrode and under anaerobic
conditions. Current transients were normalized to the electrode
3
1-36
ellipsometry (SE).
The latter technique is inherently more
sensitive than the former. Better sensitivity results from the fact
that SE variables (ψ and ∆), which are defined by the ratio of
Fresnel reflection coefficients for s- and p- polarized light, are
measured as the ratio of two signals. However, SE has been
rarely applied for determination of optical constants of metallic
particles mainly due to difficulties with the deposition of
monodispersed particles.6 In addition, interpretation of ellip-
sometric spectra may be problematic because of uncertainties
with modeling of the background spectrum of a bare substrate.
The great majority of SE spectra have been obtained with
metallic particles deposited on bare or modified silicon wafers,
2
geometric area of 1.4 cm . The surface morphology of deposited
gold particles was evaluated with a Hitachi (S-5200) scanning
electron microscope (SEM) operated at 5 kV. Histograms of
particle diameters were obtained with Spirit software. Single
angle (70.0°) SE experiments and data analysis were performed
with a M-44 ellipsometer (J. A. Woollam Co., Inc.) and
WVASE32 software. Forty-four pairs of ψ and ∆ were collected
over a wavelength range from 418 to 753 nm. The geometric
,32
2
area probed in SE measurements was approximately 0.5 cm .
The background spectrum (ψ and ∆) for the optically thick
anodized and etched aluminum/copper film (without gold
particles) was fit to a model consisting of a substrate and a 50-
nm thick Bruggeman EMA layer. The EMA layer, which
approximates scallops formed on the surface as a result of
anodization and chemical etching, was modeled as a 50%/50%
mixture of substrate optical constants and void (n ) 1, k ) 0).
The thickness of 50 nm was chosen because the depth of
scallops was estimated to be about 50 nm. Optical constants of
the anodized and etched aluminum/copper film (refractive index,
n, and extinction coefficient, k) were obtained by performing a
wavelength-by-wavelength fit of ψ and ∆ to the one layer
model. Thus, the optical constants were allowed to vary
independently at each wavelength. The substrate optical con-
stants had to be measured before depositing the gold particles
because they were different from optical constants of both
aluminum (incorporated in the software database) and pristine
aluminum/copper substrates (not subject to anodization and
etching).
3
4
a substrate which has been thoroughly investigated by SE.
The absorption band due to surface plasmons of gold particles
is typically modeled by using a Lorentz oscillator, with the peak
position around 550 nm.35,36 For gold particles embedded in
dielectric media (e.g. SiO2), effective medium approximation
(
EMA) theories have been applied to extract information about
37-39
the nanocomposite films.
Due to the scarcity of SE data
for particulate films of noble metals electrodeposited on
aluminum substrates, there is an incentive to apply SE for
determination of optical constants of gold particles.
Technologically relevant aluminum/copper films can be made
amenable for subsequent electrodeposition of gold particles by
40
anodization followed by chemical etching of aluminum oxide.
Electrodeposition on anodized and etched aluminum/copper
films covered with a thin layer of aluminum oxide proceeds by
the Volmer-Weber mechanism because the surface free energy
2
of gold (∼1.3 J/m ) is higher than that of aluminum oxide
2
2
(
∼0.64 J/m ). Under galvanostatic conditions (∼1 mA/cm )
Results and Discussion
electrodeposition of approximately 1-µm gold particles occurs
4
0
by instantaneous nucleation and under kinetic control. Both
of these phenomena favored the formation of monodispersed
particles. Gold particles were electrodeposited with a coverage
Figure 1 shows the current transient acquired upon application
of a potential step for 0.05 s from OCP (0.01 s) to -3.95 V
(0.06 s). The electrochemical cell time constant was estimated
5
7
-2
of 10 -10 particles cm . In this case, the gold particle
coverage was too low and the mean particle diameter was too
large for characterization of gold particles by SE.
to be 0.32 ms as the product of the uncompensated resistance
(
40 Ω) and capacitance of the barrier aluminum oxide layer
40
(
∼8.0 µF). Therefore, the charging current decayed to zero
The objectives of this project were 2-fold. The first was to
implement a fast and facile procedure for activation of aluminum/
copper films and electrodeposition of gold mesoparticles with
mean particle diameters less than 100 nm and coverage higher
in about 1.6 ms (5 times the cell time constant), as indicated
by two arrows (Figure 1). Figure 1 demonstrates that the
cathodic current, attributed to the irreversible reduction of
3
-
2
Au(SO3)2 , reached a value of -43 mA/cm and then stayed
constant as a function of time. This observation indicates that
electrodeposition of gold is not under diffusion control. The
increasing absolute magnitude of the cathodic current is due to
the increasing surface area of gold particles during the course
40
than that reported previously. Electrodeposition of gold
mesoparticles with a relatively narrow distribution of diameters
allowed for employment of SE to probe the optical constants
of gold mesoparticles. Under controlled electrodeposition condi-
tions, the particle coverage could be held constant and the mean
particle diameter varied. The second objective was to employ
1
0-12
of electrodeposition.
Due to its high absolute magnitude
(∼60 mA), the cathodic current is likely to be controlled by

Diameter-Dependent Optical Constants of Au Mesoparticles
J. Phys. Chem. B, Vol. 109, No. 30, 2005 14531
Figure 1. A current transient acquired upon application of a potential
step from OCP to -3.95 V vs Ag/AgCl. The deposition solution was
2
3 3 2 2 3
0 mM Na Au(SO ) and 1.0 M Na SO .
the ohmic resistance of the electrochemical cell (40 Ω). Proton
and residual oxygen reduction might also contribute to the
cathodic current. Regardless of some uncertainty with the origin
of the rate-limiting process, no diffusion limitation (no decrease
in the absolute magnitude of the cathodic current even at very
negative potentials) was observed during potential steps applied
for less than 0.05 s. Therefore, diffusion zones formed around
individual gold particles were unlikely to overlap during
electrodeposition. Both the instantaneous nucleation and the
absence of diffusion limitations favored the formation of gold
particles with a relatively narrow distribution of particle
diameters.
To fabricate gold particles with variable diameters, electro-
deposition experiments were performed with the same applied
potential of -3.95 V and variable deposition times. Panels a-c
in Figure 2 show SEM micrographs collected after electrodepo-
sition of gold particles for 0.03, 0.04, and 0.05 s, respectively.
The black pseudohexagonal cells with white edges represent
shallow scallops formed on the surface as a result of anodization
and subsequent chemical etching. Whereas anodization generates
two distinct layers of barrier and porous aluminum oxides,
etching completely dissolves the porous layer and partially
dissolves the layer of barrier aluminum oxide.40 Upon comple-
tion of chemical etching, the thickness of the barrier aluminum
oxide is approximately 1.5 nm.42 As previously suggested, a
combination of anodization and etching results in copper
enrichment in and/or underneath the layer of barrier aluminum
oxide.4
0,43-45
As a consequence, electrodeposition on anodized
and etched aluminum/copper films becomes possible. The
analysis of Figure 2 suggests that electrodeposition proceeds
via the formation of gold particles randomly located on the
surface (no long term order) and with a particle coverage of
8
-2
Figure 2. SEM micrographs obtained after electrodeposition of gold
at -3.95 V vs Ag/AgCl for (A) 0.03, (B) 0.04, and (C) 0.05 s.
approximately 7 × 10 particles cm . By employing a more
negative potential, a 2 orders of magnitude higher coverage than
40
previously reported was achieved in this project. For kinetically
controlled electrodeposition, both the cathodic current and
nucleation density increase exponentially with the applied
sition is 100%. For a one-electron reaction, the mean particle
diameter, D (m), is related to the faradaic charge density, Q
2
1
0
(mC/cm ), by eq 1.
overpotential. The particle coverage varied from sample to
sample by approximately 15% despite the same applied potential
of -3.95 V. These variations were attributed to slight variations
in the capacitance (thickness) of barrier aluminum oxide.
Nucleation density is known to exponentially decrease with the
3
6QM
D )
(1)
x
πFFN
23
thickness of a dielectric layer covering a conductive substrate.
In eq 1, M is the atomic mass of gold (197 g/mol), F is the
6
3
The particle diameters can be estimated by two alternative
ways. The first method assumes that electrodeposited particles
are spherical, their gravimetric density is equal to the bulk
density of gold, and the faradaic efficiency during electrodepo-
density of gold (19 × 10 g/m ), F is the Faraday constant, and
8
-2
N is the particle coverage (7 × 10 particles cm ). The faradaic
charge densities, as determined by integrating the faradaic
current transients, and calculated particle diameters are shown

14532 J. Phys. Chem. B, Vol. 109, No. 30, 2005
Brevnov and Bungay
TABLE 1: Mean Particle Diameters and Positions of Absorption Maximum as a Function of Electrodeposition Time
Q, faradaic
D, particle
diameter as
f(Q)/nm
D, mean particle
diameter by
SEM/nm
electrodeposition
time/s
charge density
aspect ratio
by SEM
absorption
max/nm
-
2
mC cm
0.03
0.04
0.05
1.1
1.5
2.0
68
76
84
42 ( 12
49 ( 12
74 ( 17
1.36 ( 0.36
1.31 ( 0.25
1.19 ( 0.13
610
620
675
in Table 1 (columns 2 and 3). The first method is likely to
overestimate the particle diameter because the faradaic efficiency
is less than 100%. The more accurate second method is based
upon the quantitative analysis of SEM micrographs (Table 1,
column 4). As indicated in Table 1, the mean particle diameter
increases with the electrodeposition time. Three histograms of
particle diameters (Figure 3) corresponding to three micrographs
Examination of Figure 2 reveals that there are two preferential
sites for electrodeposition of gold particles. Some of the gold
particles are located in the middle of scallops. This observation
is not surprising since in the middle of scallops the layer of
barrier aluminum oxide is likely to be the thinnest. Thus, the
electrodeposition current, which exponentially decreases with
the thickness of a dielectric layer, becomes “directed” to the
middle of scallops. In a similar way, the electroless deposition
of silver particles was observed to proceed from the middle of
scallops. Unexpected is an observation that gold particles also
tend to develop at the top edges of scallops, particularly at those
points where three or more individual scallops are adjacent to
each other. The physical origin of the second nucleation site
needs further investigation. No particles are observed on the
slopes of scallops.
(Figure 2) show that the relative standard deviation of the
particle diameter is about 25%. We note that electrodeposition
carried out with the same potential and for longer than 0.06 s
resulted in a bimodal distribution of particle diameters, with a
large population of small particles and a small population of
large particles. Similar observations were reported for electro-
deposition of gold particles on highly oriented pyrolytic
42
4
6
graphite.
Figure 4a,b demonstrates three SE spectra (ψ and ∆)
corresponding to three samples represented in Figure 2 (elec-
trodeposition times are 0.03, 0.04, and 0.05 s, respectively).
Shown also is a background SE spectrum collected for an
anodized and etched aluminum/copper film without gold
particles. The experimental data are shown as symbols, and the
model-generated data are drawn as solid lines. The SE back-
ground spectrum is featureless, as typically observed for
Figure 4. Three SE spectra: (a) ψ and (b) ∆ obtained for three samples
after electrodeposition for 0.03, 0.04, and 0.05 s. Shown also is a
background SE spectrum for an anodized and etched aluminum/copper
film (without gold particles). Experimental SE data are symbols and
model SE data are solid lines.
Figure 3. Histograms of particle diameters (counts vs particle diameter)
corresponding to micrographs shown in Figure 2.

Diameter-Dependent Optical Constants of Au Mesoparticles
J. Phys. Chem. B, Vol. 109, No. 30, 2005 14533
TABLE 2: Results of Modeling of SE Data
electrodeposition time/s
0.04
parameter
0.03
0.05
A
1
B
1
E
1
A
2
B
2
E
2
0.13 ( 0.02
0.62 ( 0.02
2.085 ( 0.016
0.047 ( 0.022
0.63 ( 0.10
1.81 ( 0.06
0.084 ( 0.013
0.65 ( 0.03
0.091 ( 0.04
0.58 ( 0.06
2.002 ( 0.028
0.12 ( 0.04
0.67 ( 0.09
1.73 ( 0.05
/eV
/eV
2.054 ( 0.007
0.031 ( 0.015
0.97 ( 0.14
1.70 ( 0.10
MSE
0.99
0.62
1.74
Both closely positioned Lorentz oscillators (∼2.05 and ∼1.75
eV) are likely to represent the light absorption of gold particles
associated with surface plasmons. In contrast to a single Lorentz
Figure 5. Optical model describing gold particles on top of rough
and optically thick aluminum/copper substrate.
3
5,36
optically thick metallic films. Comparison of the SE background
spectrum and spectra with gold particles indicates that all three
ψ sets collected for samples with gold particles demonstrate a
depression region around 580-620 nm (Figure 4a). ψ is known
to be sensitive to extinction coefficients of deposited films, as
opposed to ∆, which is primarily sensitive to the film thickness.
As it turns out, the ψ depression region in all three SE spectra
corresponds to an absorption peak associated with surface
plasmons of gold particles. Moreover, the position of the
depression region systematically shifts to higher wavelengths
as the mean particle diameter increases from 42 to 74 nm. While
the position of the ψ depression regions correlates with the mean
particle diameter, the absolute magnitude of the depression peak
is likely to be proportional to the coverage of gold particles.
Comparison of three ψ depression regions indicates that gold
particle coverage for three samples is slightly different. Specif-
ically, the sample with gold particles electrodeposited for 0.04
s shows the smallest decrease in ψ as compared to two other
samples. This observation agrees with data obtained from Figure
oscillator used elsewhere,
our model requires two Lorentz
oscillators to describe surface plasmon absorption. The need
for two oscillators may be due to the shape of deposited gold
particles, which are not perfectly spherical. As indicated in Table
1, the aspect ratios are approximately 1.3. Metallic particles with
the aspect ratio of approximately 3 have been previously shown
to have two distinct surface plasmon bands separated by about
250 nm (0.77 eV). In addition, slight polydispersity in particle
diameters for a given sample may also require using the second
Lorentz oscillator.
The optical constants of gold particles (n and k) are calculated
(Figure 6) by using six fitting parameters of two Lorentz
oscillators for each of three samples. As indicated above, three
samples had slightly different coverage of gold particles.
However, in the modeling of SE data all three samples are
assumed to have the same coverage and, correspondingly, the
same thickness of the layer with two Lorentz oscillators. As a
result, magnitudes of calculated optical constants are coupled
4
2
b, which indicate this sample had the lowest particle count
(Figure 2).
For quantitative analysis of SE spectra, it is necessary to
postulate a physically adequate model. The pseudodielectric
function of electrodeposited gold mesoparticles is described by
using a 50-nm layer consisting of two Lorentz oscillators (eq
2).
Ak
hc 2
ꢀ
(λ) ) ꢀ +
(2)
∞
∑
k
hc
2
E -
- iBk
k
( )
( )
λ
λ
In eq 2, ꢀ∞ is an offset, h is Planck’s constant, c is the light
speed, and λ is the wavelength. For the kth oscillator, Ak is the
amplitude (oscillator strength), Ek is the center energy, and Bk
is the broadening. The model assumed that the particulate film
of gold had a constant thickness of 50 nm, because of a strong
correlation between film thickness and oscillator strengths. The
complete two-layer optical model (Figure 5) describing gold
particles on a rough aluminum/copper substrate includes the
layer with two Lorentz oscillators, the 50-nm thick Bruggeman
EMA layer (as described in the Experimental Section), and the
optically thick aluminum/copper substrate. The model does not
include a layer of barrier aluminum oxide because it is optically
thin (1.5 nm). Table 2 summarizes the six model parameters
obtained by fitting the acquired SE data to the postulated model
for each of three samples (electrodeposition times of 0.3, 0.4,
and 0.5 s). Shown also is the mean-squared error (MSE), which
is defined as a sum of the squares of the differences between
measured and calculated data, with each difference weighted
by the standard deviation of that measured data point.
Figure 6. Three sets of calculated optical constants of gold particles:
(a) extinction coefficient, k, and (b) refractive index, n. Gold particles
were obtained after electrodeposition for 0.03, 0.04, and 0.05 s.

14534 J. Phys. Chem. B, Vol. 109, No. 30, 2005
Brevnov and Bungay
to the actual particle coverage. This point is illustrated by noting
that the largest deviation of ψ (electrodeposition time of 0.05
s) from the background signal (Figure 4a) corresponds to the
largest magnitude of k (Figure 6a). Regardless of the coupling
between the particle coverage and magnitude of optical con-
stants, the particle coverage is not expected to influence the
position of absorption bands. This assumption can be justified
by examining Figure 2, which indicates that on average gold
particles are located sufficiently far away from each other.
Therefore, their optical constants are not affected by the presence
of the neighboring particles.
be extended in order to deposit other metallic particles (e.g.
Cu, Cd), which subsequently can be transformed to semicon-
ductor salts by chemical oxidation and displacement reactions.49
Thus, this method represents an opportunity to fabricate
semiconductor particles on technologically relevant substrates.
Conclusions
Electrodeposition of gold mesoparticles on anodized and
chemically etched aluminum/copper films provides a facile and
inexpensive method to achieve the 2D arrangement of metallic
particles on a technologically important substrate. Under the
reported experimental conditions, electrodeposition of gold
particles occurs by instantaneous nucleation and with no
diffusion limitations. Taking advantage of these two character-
istics of the electrodeposition process, gold particles are
Analysis of Figure 6a results in three conclusions. First, the
position of absorption maximum, as indicated by labels (Figure
6a), shifts to longer wavelengths as the mean particle diameter
increases (Table 1). This trend is consistent with previously
reported results. In the extrinsic size region (the diameter of
gold particles is more than 25 nm), absorption maximums are
8
electrodeposited with the particle coverage of 7 × 10 particles
-2
cm , the mean particle diameter in the range of 40-80 nm,
4,5,25,26
red shifted with increasing particle diameter.
Second, in
and the relative standard deviation of the particle diameter of
comparison with gold particles of the same size (50-70 nm)
in solution, electrodeposited gold particles have absorption bands
shifted about 80-100 nm to longer wavelengths. This shift is
most likely due to the fact the gold particles are deposited on
an aluminum/copper substrate. The dielectric constant of a
substrate is known to affect the positions of absorption bands.
For example, the extinction maximum of localized surface
plasmons of silver nanoparticles was shown to shift to longer
2
5%. The optical constants of gold mesoparticles electro-
deposited on optically thick aluminum/copper films are deter-
mined by SE. The absorption peak, associated with surface
plasmons, shifts from 610 to 675 nm as the mean particle
diameter increases from 42 to 74 nm. Refractive indexes of
particulate gold films are very close to that of air (n ∼ 1.02 at
6
33 nm). Electrodeposition of gold particles with a narrow
distribution of particle diameters on aluminum/copper films is
expected to increase the utility of gold particles and facilitate
their incorporation in nanostructured materials and devices.
4
7
wavelengths with increasing refractive index of the substrate.
Our SE results at 620 nm show that the optical constants of
anodized and etched aluminum/copper substrate are n ) 1.1
and k ) 6.1. Thus, the substrate may be responsible for the
observed red shift of absorption maximums. Third, our SE
measurements do not allow us to establish whether there is any
increase in the bandwidth of the surface plasmon absorption
band with an increase in the mean particle diameter as observed
Acknowledgment. This work was supported in part by the
Center for Micro-Engineered Materials (UNM). The SEM
laboratory was supported by the New Mexico EPCoR NSF grant
and the NNIN grant. We thank Todd M. Bauer (Sandia National
Laboratories, Albuquerque, NM) for fabrication of silicon wafers
with an aluminum/copper layer, Tim S. Olson (UNM) for
quantitative analysis of micrographs, and Prof. Gabriel P. Lopez
(UNM) for critical reading of this paper.
4
in the extrinsic size region. This fact may be explained by the
coupling of the calculated optical constants to the coverage of
gold particles.
In addition to the analysis of extinction coefficients, it would
be interesting to examine refractive indexes of deposited
particulate films (Figure 6b). Under our electrodeposition
conditions, gold particles occupy only a small volume fraction
on the substrate surface (Figure 2). Our optical model describes
the gold particles as a continuous planar layer. This layer
consists of widely distributed gold particle inclusions imbedded
in a sea of void. Thus, refractive indexes of particulate gold
films are very close to that of air (n ∼ 1.02 at 633 nm). The
previously reported refractive indexes of nanocomposite gold
References and Notes
(
1) Bradley, J. S.; Schmid, G. In Nanoparticles: From Theory to
Application; Schmid, G., Ed.; Wiley-VCH, Weinheim, Germany, 2004; p
186.
(2) Goia, D. V. J. Mater. Chem. 2004, 14, 451.
(
3) Daniel, M. C.; Astruc, D. Chem. ReV. 2004, 104, 293.
(4) Link, S.; El-Sayed, M. A. J. Phys. Chem. B 1999, 103, 8410.
(5) El-Sayed, M. A. Acc. Chem. Res. 2001, 34, 257.
(6) Shipway, A. N.; Katz, E.; Willner, I. ChemPhysChem 2000, 1, 18.
(
7) Corti, C. W.; Holliday, R. J.; Thompson, D. T. Gold Bull. 2002,
5, 111.
8) Hutter, E.; Fendler, J. H. AdV. Mater. (Weinheim, Ger.) 2004, 16,
1685.
3
35,36
films were significantly higher (∼1.6)
due to the employ-
(
ment of organic molecules to immobilize gold particles on a
substrate.
(9) Xia, Y.; Halas, N. J. MRS Bull. 2005, 30, 338.
(10) Oskam, G.; Long, J. G.; Natarajan, A.; Searson, P. C. J. Phys. D:
Combined analysis of optical constants, k and n, allows us
to conclude that electrodeposition on aluminum/copper films
results in the formation of gold mesoparticles, which show an
optical response measurable by SE. Moreover, the absorption
maximum shifts to longer wavelengths as the mean particle
diameter increases. Therefore, electrodeposition is demonstrated
to be a facile method to fabricate particulate films with distinct
optical properties. To our knowledge, electrodeposition has not
been previously shown to produce gold particles on a substrate
under conditions which enable their absorption bands due to
surface plasmons to be resolved. The optical properties of
electrodeposited gold particles can be exploited in a variety of
electronic, optical devices and biosensors.48 Moreover, the
procedure for electrodeposition on aluminum/copper films can
Appl. Phys. 1998, 31, 1927.
(11) Budevski, E.; Staikov, G.; Lorenz, W. J. Electrochim. Acta 2000,
5, 2559.
4
(
(
(
12) Penner, R. M. J. Phys. Chem. B 2002, 106, 3339.
13) Oskam, G.; Searson, P. C. J. Electrochem. Soc. 2000, 147, 2199.
14) Krumm, R.; Guel, B.; Schmitz, C.; Staikov, G. Electrochim. Acta
2000, 45, 3255.
(15) Ji, C. X.; Oskam, G.; Searson, P. C. J. Electrochem. Soc. 2001,
48, C746.
1
(16) Radisic, A.; Cao, Y.; Taephaisitphongse, P.; West, A. C.; Searson,
P. C. J. Electrochem. Soc. 2003, 150, C362.
(17) Radisic, A.; Oskam, G.; Searson, P. C. J. Electrochem. Soc. 2004,
1
51, C369.
18) Zoval, J. V.; Stiger, R. M.; Biernacki, P. R.; Penner, R. M. J. Phys.
Chem. 1996, 100, 837.
19) Zoval, J. V.; Lee, J.; Gorer, S.; Penner, R. M. J. Phys. Chem. B
1998, 102, 1166.
(
(

Diameter-Dependent Optical Constants of Au Mesoparticles
J. Phys. Chem. B, Vol. 109, No. 30, 2005 14535
(
20) Stiger, R. M.; Gorer, S.; Craft, B.; Penner, R. M. Langmuir 1999,
(36) Cant, N. E.; Zhang, H.-L.; Critchley, K.; Mykhalyk, T. A.; Davies,
G. R.; Evans, S. D. J. Phys. Chem. B 2003, 107, 13557.
(37) Joerger, R.; Gampp, R.; Heinzel, A.; Graf, W.; Kohl, M.; Ganten-
bein, P.; Oelhafen, P. Sol. Energy Mater. Sol. Cells 1998, 54, 351.
(38) Dalacu, D.; Martinu, L. J. Appl. Phys. 2000, 87, 228.
(39) Girardeau, T.; Camelio, S.; Babonneau, D.; Toudert, J.; Barrranco,
A. Thin Solid Films 2004, 455-456, 313.
1
5, 790.
(21) Marquez, K.; Staikov, G.; Schultze, J. W. Trans. Inst. Met. Finish.
2
002, 80, 73.
(
22) Ziegler, J. C.; Reitzle, A.; Bunk, O.; Zegenhagen, J.; Kolb, D. M.
Electrochim. Acta 2000, 45, 4599.
23) Radisic, A.; Oskam, G.; Searson, P. C. J. Electrochem. Soc. 2004,
51, C369.
24) Hsiao, G. S.; Anderson, M. G.; Gorer, S.; Harris, D.; Penner, R.
M. J. Am. Chem. Soc. 1997, 119, 1439.
25) Kreibig, U.; Vollmer, M. Optical Properties of Metal Clusters;
Springer-Verlag: Berlin, Germany, 1995; p 13.
(
1
(40) Olson, T. S.; Atanassov, P.; Brevnov, D. A. J. Phys. Chem. B 2005,
109, 1243.
(
(41) Green, T. A.; Liew, M.-J.; Roy, S. J. Electrochem. Soc. 2003, 150,
C104.
(42) Brevnov, D. A.; Olson, T. S.; Lopez, G. P.; Atanassov, P. J. Phys.
Chem. B 2004, 108, 17531.
(43) Strehblow, H. H.; Melliar-Smith, C. M.; Augustyniak, W. M. J.
Electrochem. Soc. 1978, 125, 915.
(44) Habazaki, H.; Zhou, X.; Shimizu, K.; Skeldon, P.; Thompson, G.
E.; Wood, G. G. Electrochim. Acta 1997, 42, 2627.
(45) Paez, M. A.; Bustos, O.; Thompson, G. E.; Skeldon, P.; Shimizu,
K.; Wood, G. C. J. Electrochem. Soc. 2000, 147, 1015.
(46) Boxley, C. J.; White, H. S.; Lister, T. E.; Pinhero, P. J. J. Phys.
Chem. B 2003, 107, 451.
(
(
(
(
(
(
26) Link, S.; El-Sayed, M. A. Int. ReV. Phys. Chem. 2000, 19, 409.
27) Hilger, A.; Tenfelde, M.; Kreibig, U. Appl. Phys. B 2001, 73, 361.
28) Mulvaney, P. Langmuir 1996, 12, 788.
29) Link, S.; El-Sayed, M. A. J. Phys. Chem. B 1999, 103, 4212.
30) Jensen, T. R.; Malinsky, M. D.; Haynes, C. H.; Van Duyne, R. P.
J. Phys. Chem. B 2000, 104, 10549.
(
(
(
31) Greef, R. Thin Solid Films 1993, 233, 32.
32) Suni, I. I. J. Appl. Electrochem. 1997, 27, 1219.
33) Urban, F. K.; Hosseini-Tehrani, A.; Khabari, A.; Griffiths, P.;
Bungay, C.; Petrov, I.; Kim, Y. Thin Solid Films 2002, 408, 211.
34) Kooij, E. S.; Wormeester, H.; Brouwer, E. A. M.; van Vroonhoven,
E.; van Silfhout, A.; Poelsema, B. Langmuir 2002, 18, 4401.
35) Zhang, H.-L.; Evans, S. D.; Henderson, J. R. AdV. Mater. 2003,
5, 531.
(47) Malinsky, M. D.; Kelly, K. L.; Schatz, G. C.; Van Duyne, R. P. J.
Phys. Chem. B 2001, 105, 2343.
(48) Haes, A. J.; Zou, S.; Schatz, G. C.; Van Duyne, R. P. J. Phys.
Chem. B 2004, 108, 6961.
(
(
1
(49) Penner, R. M. Acc. Chem. Res. 2000, 33, 78.