Fabrication of SERS substrate using nanoporous anodic alumina template decorated by silver nanoparticles

Article abstract of DOI:10.1016/j.cplett.2010.07.096

SERS substrate is fabricated using nanoporous anodic alumina membrane decorated by silver nanoparticles, which are self assembled on the surface as well as inside the pores to form a network structure of aligned cylindrical vertical wells creating artificial roughness on the surface of the membrane. When electromagnetic radiation falls on the rough surface, a strong electric field is created due to plasmon excitation on sharp edges of the network structure or at interparticle spacing known as 'hot spots'. This enhanced electric field produces remarkable SERS effect from Raman active R6G molecules attached to silver nanoparticles when excited by a laser source.

Full text of DOI:10.1016/j.cplett.2010.07.096

Chemical Physics Letters 497 (2010) 89–93
Contents lists available at ScienceDirect
Chemical Physics Letters
journal homepage: www.elsevier.com/locate/cplett
Fabrication of SERS substrate using nanoporous anodic alumina template
decorated by silver nanoparticles
B. Mondal, S.K. Saha *
Department of Materials Science, Indian Association for the Cultivation of Science, Jadavpur, Kolkata 700 032, India
a r t i c l e i n f o
a b s t r a c t
Article history:
SERS substrate is fabricated using nanoporous anodic alumina membrane decorated by silver nanoparti-
cles, which are self assembled on the surface as well as inside the pores to form a network structure of
aligned cylindrical vertical wells creating artificial roughness on the surface of the membrane. When elec-
tromagnetic radiation falls on the rough surface, a strong electric field is created due to plasmon excita-
tion on sharp edges of the network structure or at interparticle spacing known as ‘hot spots’. This
enhanced electric field produces remarkable SERS effect from Raman active R6G molecules attached to
silver nanoparticles when excited by a laser source.
Received 11 June 2010
In final form 29 July 2010
Available online 3 August 2010
Ó 2010 Elsevier B.V. All rights reserved.
1
. Introduction
Surface Enhanced Raman Scattering (SERS) has been studied
Although, SERS effect is known for nearly three decades, search
for the suitable substrate is still under active area of research.
Different materials and the various structures are investigated to
use as SERS substrate. In most cases, silver, gold, or copper nano-
particles are used to fabricate the SERS substrate because of its tun-
able SPR peak [9]. Several efforts have been given to prepare SERS
substrate by depositing thermally evaporated silver nanoparticles
on different substrates like glass [10], Anodic Alumina Oxide
(AAO) [11,12], porous silicon [13], aluminum foil [14]. Besides
these, other methods such as magnetron sputtering [15], Electron
Beam Lithography (EBL) [16], and Focused Ion Beam (FIB) [17] or
electro-deposition [18] are also being used to fabricate the suitable
SERS substrate. In most cases, the main disadvantage is the poor
control of the size of nanoparticles, interparticle separation and
the aggregation of nanoparticles, which are the key factors in con-
trolling the plasmon resonance responsible for the electric field
enhancement. Therefore, although, considerable research articles
on the fabrication of SERS substrates have already been published,
the precise control of the surface roughness with the size distribu-
tion of nanoparticles and interparticles spacing through chemical
route is relatively unexplored [19].
In this Letter, we report a simple technique to fabricate the SERS
substrate in which surface roughness has been precisely controlled
with the nanoparticles aggregation. Here, nanoporous AAO
membrane has been decorated by silver nanoparticles with differ-
ent degrees of aggregation just tailoring the amount of silver
nitrate (SN) only. The SPR peak was tuned to cover across a wide
range of 350–460 nm to overlap with a 457.9 nm excitation line
emitted from the laser source. The as-prepared substrates are
highly stable, reproducible and cost effective with high SERS
enhancement efficiency (EF). We have used Rhodamine 6G (R6G)
both theoretically [1] and experimentally [2] for nearly last three
decades as it is a well known tool for the detection and character-
ization of very low concentration of molecules in the chemical and
biological samples [3]. As Raman signal from the molecules is very
weak due to its low scattering cross section and high background
fluorescence [4], various techniques are used to get a detectable
signal. Among various factors to increase the intensity of the signal,
2
chemical enhancement (ꢀ0 to 10 ) and electromagnetic (EM)
4
12
enhancement (ꢀ10 to 10 ) are very important [5]. The chemical
enhancement arises due to charge transfer from metal to the mol-
ecule under consideration but the EM enhancement arises from the
optical excitation of Surface Plasmon Resonance (SPR) of the metal
nanoparticles. When an external EM field is applied on the nano-
particles, the charges get separated (the nanoparticles are polar-
ized) which generate an enormous electric field around the
nanoparticles [6]. The generated electric field is highly influenced
by the aggregation of nanoparticles; interparticle spacing and the
environment [7]. This electric field also enhances the Raman
signals from the molecules adsorbed at the surface. As the SERS ef-
fect is enhanced remarkably on the rough surfaces [8], it is worth
to investigate the fabrication of SERS substrates using ultra fine
metal nanoparticles where the roughness can be controlled by
the size of nanoparticles, interparticle separation and the nature
of aggregation.
*
Corresponding author. Fax: +91 33 24732805.
E-mail address: cnssks@iacs.res.in (S.K. Saha).
0
009-2614/$ - see front matter Ó 2010 Elsevier B.V. All rights reserved.
doi:10.1016/j.cplett.2010.07.096

9
0
B. Mondal, S.K. Saha / Chemical Physics Letters 497 (2010) 89–93
as the probing molecule to study the SERS performance of as fabri-
cated substrates.
surfaces of the AAO membrane. The membrane containing silver
nanoparticles is washed several times with distilled water and
dried in a vacuum oven at temperature 333 K. Amount of SN and
ammonium hydroxide to form the silver complex for four sub-
strates and amount of formic acid added to the silver complex to
reduce to silver are summarized in Table 1.
2
. Experimental
NaOH and AAO membrane (diameter 13 mm) of pore diameters
00 nm are purchased from s. d. fine – Chem Limited (Mumbai, In-
Optical, morphological and microstructural characterizations of
the as-prepared substrates have been carried out. The silver
nanoparticles decorated AAO membranes are used to take the
absorption spectra with a Cary 5000 (Varian, Mumbai, India)
spectrophotometer in the transmission mode with a virgin AAO
membrane as reference. R6G is taken in the methanol with a
1
dia) and Whatman International Ltd. (Maidstone, England) respec-
tively. SN, ammonium solution (28%) and formic acid were
purchased from Merck India Limited (Mumbai, India). Methanol
(
spectroscopic grade) and R6G were purchased from Sigma–Al-
drich Corporation (Bangalore, India). To prepare the SERS sub-
strate, aqueous solution of SN is added to the excess amount of
NaOH. The gray color precipitate is washed several times with dis-
tilled water to remove the un-reacted NaOH or SN if any and fil-
tered. The filtrate is taken in a beaker and 5 ml of the distilled
water is added to it and stirred. A grayish colloidal solution is ob-
tained. Ammonium solution is added to the gray solution drop
ꢁ4
concentration of 10 M. Five microliters of the solution is spread
on one surface of a substrate covering the whole area with a micro-
pipette. After the evaporation of methanol, Micro-Raman measure-
ment was performed on a Jobin Yvon HR 800 micro-Raman
spectrometer with 457.9 nm excitation from a 20 mW air-cooled
argon ion laser. The laser power at the sample position was about
8 lW. The data acquisition time and the number of accumulations
wise while stirring until a transparent solution due to formation
+
were 1s and 10, respectively. Silver nanoparticles decorated AAO
membrane was used to study the morphology and the structures
by a Scanning Electron Microscope (JEOL, JSM6700F, Tokyo, Japan).
The substrate was ground in a mortar, mixed with acetone using
ultrasonication and one drop was taken on a TEM grid to study
the microstructures of the silver nanoparticles using a Transmis-
sion Electron Microscopy (TEM, JEM-2010, Tokyo, Japan) with an
Electron Diffraction facility at 200KV.
of silver ammonium complex ([Ag(NH
3
)
2
] OH) is obtained. One
pre-heated (ꢀ390 K) porous AAO membrane is dipped into the
solution and desired amount of formic acid is then added to the
beaker containing the membrane. Silver nanoparticles are
produced upon reduction of the complex by formic acid on both
Table 1
Amount of SN taken in 5 ml of water, corresponding amount of ammonium hydroxide
used to form silver complex for four substrates and amount of formic acid added to
the silver complex to reduce.
AgNO
g)
3
taken
Amount of NH
l)
4
OH
Formic acid added
l)
3. Results and discussion
(
(
l
(l
0
0
0
0
.1
.2
.3
.4
200
400
600
800
400
–
–
The AAO membranes were decorated by silver nanoparticles
following the method as described earlier. The amount of ammo-
nium solution added is very important as addition of less amount
of ammonium solution may keep some amount of silver oxide
–
Fig. 1. (a)–(d) are surface morphology of the AAO membranes decorated with 0.1, 0.2, 0.3 and 0.4 g of SN, respectively. It is seen from the images that particles size as well as
the roughness increases with the increase in SN due to the aggregation of the silver nanoparticles. However, for SN amount 0.4 g, the effective roughness decreases due to
filling of AAO membrane pores.

B. Mondal, S.K. Saha / Chemical Physics Letters 497 (2010) 89–93
91
intact while addition of excess amount may increase the basicity of
the solution, which may lead to the dissociation of the AAO mem-
brane. So, an optimum amount of ammonium solution is added.
However, the amount of formic acid only controls the reduction
process. Formic acid being a weak acid, cannot dissociate the mem-
brane in the time span of the experiment. Hence, we kept the
amount of formic acid constant in our as-prepared substrates to
study the effect of degree of aggregation of silver nanoparticles
on the AAO membrane varying only the amount of SN. For the solu-
tion containing 0.1 g of SN, silver started to deposit in the form of
nanoparticles on the surface of the AAO membrane after reduction,
keeping the pores open as shown in Fig. 1a. For higher SN content
are non-radiative in nature are excited. As only the dipole transi-
tion leads to Raman scattering, the higher-order transitions will
cause a decrease in the overall efficiency of the enhancement. So,
proper tuning of the SPR peak with suitable selection of the nano-
particles size and their aggregation is the primary criterion for
nanoparticles based SERS substrates. Njoki et al. reported a size
dependant SERS enchantment and revealed a critical size [22].
The electric field generated around the nanoparticles becomes
more intense at the sharp edges or between two nanoparticles sep-
arated by nanometric distances as investigated by Koh et al. by
Electron Energy Loss Spectroscopy (EELS) [23], typically known
as ‘hot spots’. When two nanoparticles in the aggregate are very
close to each other, the generated electric field around each nano-
particle is coupled and their plasmon peaks become broad due to
hybridization of the plasmons. The basic physics behind the plas-
mon hybridization is that the interaction between the primitive
plasmons of individual components give two energy levels, one
anti-bonding (w+) orbital and the other bonding (wꢁ) orbital.
The separation between the bonding and anti-bonding energy lev-
els depends both on the size as well as the distance between the
nanoparticles. The splitting of the bonding and anti-bonding en-
ergy levels causes the peak broadening due to the hybridization.
Fig. 3 shows the UV–vis absorption spectra for four substrates
from which it is seen that the plasmon band of silver nanoparticles
splits into several peaks. For 0.1 g of SN substrate, the splitting is
not so prominent but the peak is broad ranging from approxi-
mately 350 to 422 nm, however, for 0.3 g of SN substrate, splitting
of the plasmon band into sharp peaks is observed and the SPR peak
is red shifted up to nearly 460 nm. As mentioned above, silver
nanoparticles are self assembled to form a network structure of
vertically aligned cylindrical wells creating sharp edges, which
produce an artificial roughness on the surface of the AAO mem-
brane. The splitting of the plasmon band for 0.1 g of SN substrate
is due to hybridization of the primitive plasmons of self assembled
silver nanoparticles. The effect increases as amount of SN increases
and for 0.3 g of SN substrate, the band splits into several intense
peaks. This is due to the fact that with increasing SN amount,
roughness on the surface increases and for 0.3 g of SN substrate
the splitting is remarkable with formation of intense peaks due
to strong enhancement of the electric field produced on the sharp
edges where silver nanoparticles are self assembled. The broaden-
ing of SPR peak or the increase in the absorption cross section in
the order of 0.1 g < 0.2 g < 0.3 g of SN substrates gradually overlaps
(
0.2 g) silver nanoparticles are self assembled to form a network
structure as shown in Fig. 1b, which creates an artificial roughness
on the surface of the membrane. The effect is remarkable for SN
content 0.3 g, when a large number of silver nanoparticles are
formed on the surface to form an assembly of vertical wells creat-
ing a large surface roughness as shown in Fig. 1c. If the SN content
is further increased to 0.4 g, the pores of the AAO membrane starts
to fill up thereby reducing the effective surface roughness (Fig. 1d).
Transmission Electron Microscopy (TEM) was performed to inves-
tigate size distribution of the nanoparticles and their aggregation.
Fig. 2 shows the TEM image of the substrate containing 0.3 g of
SN from which it is seen that most of the nanoparticles are 5–
1
0 nm in diameter, however, some bigger size nanoparticles which
are nothing but the aggregation of smaller size nanoparticles are
also observed (indicated by an arrow in the Fig. 2). The Selected
Area Electron Diffraction (SAED) pattern of the silver nanoparticles
is shown in the inset of Fig. 2. The circular rings obtained from the
silver nanocrystals corresponding to (1 1 1), (2 0 0), (2 2 0), (3 1 1)
planes of silver, clearly indicates the all possible orientation of the
silver nanoparticles.
The shape, size and the aggregation of nanoparticles strongly af-
fect the efficiency of the SERS enhancement because these factors
influence SPR absorption spectra. The definition of the plasmon
does not hold for very small nanoparticles as well as very large
nanoparticles because of the fact that in case of very small nano-
particles number of electrons are not sufficient for collective exci-
tation [20] and for very large nanoparticles multipoles [21] which
Fig. 2. (a) Transmission Electron Micrograph (TEM) image of the substrate with
0
.3 g of SN to investigate the size distribution of the nanoparticles. It shows average
Fig. 3. The absorption spectra recorded from four SERS substrates. The variation in
the absorption spectra with SN is prominent. The broadening as well as intensity of
the SPR peaks in the absorption spectra increases with increase in SN up to 0.3 g.
However, for 0.4 g of SN the substrate becomes opaque.
size of the nanoparticles as 5–10 nm with some larger particles (15–20 nm). Inset in
the image is the Selected Area Electron Diffraction (SAED) pattern of the
nanoparticles.

9
2
B. Mondal, S.K. Saha / Chemical Physics Letters 497 (2010) 89–93
Fig. 4. SERS spectrum of the substrates. (a) The Raman signals recorded from the substrates with its background emission. The intensity of the Raman signals increases with
increase in SN. For the substrate with SN amount of 0.4 g the intensity decreases. (b) The Raman signals after subtraction of its background emission.
with the excitation wavelength of laser at 457.9 nm thereby pro-
ducing ‘hot spots’. However, all the pores of the AAO membrane
are covered by silver nanoparticles if the SN amount is increased
further to 0.4 g, reducing the effective roughness and making the
substrates nearly opaque for transmission of light. Opacity in
transmitted light in case of 0.4 g SN substrate results a broad hump
in absorption spectra shown in Fig. 3. This phenomenon diminishes
the intensity of the Raman signal from the surface adsorbed
molecules.
molecules spread uniformly under the laser spot on the fabricated
3
SERS substrate. Assuming the scattering volume as 1 lm and molar
ꢁ5
3
volume of the R6G to be 36.84 ꢃ 10 m , NRef is calculated to be
2
0
ꢁ4
approximately 16.34 ꢃ 10 . If five microliters of a 10 M of R6G
in methanol solution is spread on the membrane having a total sur-
ꢁ
6
2
face area of about 132.75 ꢃ 10 m , NSERS comes to be nearly
1
5
2.27 ꢃ 10 . The value of G calculated from the SERS peak at
ꢁ1
5
5
612 cm for three substrates are (i) 7.72 ꢃ 10 (ii) 8.3 ꢃ 10 (iii)
5
5
9.1 ꢃ 10 and (iv) 8.54 ꢃ 10 .
SERS effect of the substrate is investigated using R6G as the Ra-
man active compound (analyte) with a micro-Raman set up. Fig. 4a
gives the SERS spectra for the virgin AAO membrane as well as the
relative enhancement of the Raman signal for silver nanoparticles
decorated AAO membranes creating different degree of roughness.
After subtraction of the background emission, the spectra look very
sharp and the intense characteristic peaks of R6G are displayed in
It is to be mentioned that the enhancement of Raman signal is
the maximum when wavelength of the laser source is comparable
to the plasmon excitations. In our case we have used excitation
wavelength of 457.9 nm, which is the higher limit of the SPR peaks.
Remarkable effect of enhancement in the Raman signal would be
obtained if we could use the excitation line around 400 nm by
proper choice of the laser source and the optical arrangement. In
addition to laser wavelength, the intensity of the Raman signal
may be increased using potential strategies such as core–shell
nanostructures and tuning the shape (cubic, triangular, rods) of
the nanoparticles. Recently, Kumar et al. has reported enhance-
ment of electric fields using core–shell nanostructures with Ag core
and Au shell and observed large enchantment of Raman signal for
Au:Ag = 0.4 [27]. However, formation of core–shell nanostructures
on the AAO membrane by coreducing the mixture of gold and sil-
ver complexes and investigation of its SERS activity is underway.
ꢁ1
Fig. 4b. The characteristic peaks of R6G at 612, 773 and 1182 cm
are associated with C–C–C ring in-plane, C–H out of plane bend
mode and C–C stretching vibrations, respectively. The bands at
ꢁ1
1
360, 1496 and 1642 cm are usually assigned to aromatic C–C
stretching vibrations of the R6G molecules but none of the Raman
ꢁ1
lines from methanol at 1035 and 1465 cm can be seen in the
spectrum. We have noticed that there is a small shift (ꢀ4 to
ꢁ1
1
7 cm ) in the Raman peak positions relative to its literature val-
ues [11]. Niu and his co-workers fabricated a ‘silver nano-particles/
sample molecules/silver film system and showed that the spectral
shift in the Raman spectra arise due to the interaction of the nano-
particles with the analyte molecules [24]. In the present work, the
spectral shift arises due to the interaction between the silver atoms
and the carboxyl group of R6G. From Fig. 4a it is seen that the R6G
molecule attached to the virgin AAO membrane does not show any
Raman signal. The response starts when the silver nanoparticles
(
0.1 g) are formed on the surface. Intense Raman signals are ob-
served for 0.3 g of SN substrate when an artificial roughness is cre-
ated by silver nanoparticles self assembled to form a network
structure of aligned vertical wells. The Raman Enhancement Effi-
ciency (EF) per molecule is expressed as [25,26],
I
I
SERS
N
Ref
G ¼
ꢂ
N
Ref
SERS
where ISERS is the enhanced intensity due to SERS effect of the sur-
face adsorbed R6G molecules, IRef is the spontaneous scattering
intensity from the bulk R6G molecules under the laser spot on the
virgin AAO membrane, NRef is the number of molecules excited by
the laser on the virgin AAO membrane and NSERS is the number of
Fig. 5. The graph showing variation in intensity of the Raman signal with amount of
SN. The intensity of the SERS signal starts to decrease for SN amount of 0.4 g.

B. Mondal, S.K. Saha / Chemical Physics Letters 497 (2010) 89–93
93
The variation in intensity of the Raman signal of as-prepared sub-
References
strates with amount of SN used is demonstrated in Fig. 5. It is seen
that the intensity of the Raman signal of the surface adsorbed R6G
molecules is enhanced with increase in amount of SN. As amount
of SN increases, more and more pores are decorated with silver
nanoparticles as well as aggregation of nanoparticles takes place.
This aggregation of nanoparticles produces large number of ‘hot
spots’ which enhances Raman signals from R6G molecules up to
[
1] E.C. Le Ru, E. Blackie, M. Meyer, P.G. Etchegoin, J. Phys. Chem. C 111 (2007)
3794.
2] W. Li, Y. Guo, P. Zhang, J. Phys. Chem. C 114 (2010) 6413.
1
[
[3] P. Nielsen, H. Hassing, O. Albrektsen, S. Foghmoes, P. Morgen, J. Phys. Chem. C
13 (2009) 14165.
[
[
1
4] Y.S. Li, Y. Wang, Appl. Spectrosc. 46 (1992) 142.
5] H. Guo, F. Ruan, L. Lu, J. Hu, J. Pan, H. Yang, B. Ren, J. Phys. Chem. C 113 (2009)
10459.
0
.3 g of SN. However, for SN amount of 0.4 g, intensity of the Ra-
[6] P. Zhang, S. Smith, G. Rumbles, M.E. Himmel, Langmuir 21 (2005) 520.
[
7] Z. Cai, C. Tian, L. Wang, W. Zhou, B. Wang, H. Fu, J. Raman Spectrosc. (2010), in
press, doi:10.1002/jrs.2640.
man signals decreases due to filling of nanopores of AAO mem-
brane creating a smooth surface.
[
[
8] C. Noguez, J. Phys. Chem. C 111 (2007) 3806.
9] S. Kruszewski, Surf. Interface Anal. 21 (1994) 830.
[
10] C.L. Haynes, R.P. Van Duyne, J. Phys. Chem. B 107 (2003) 7426.
[11] D. Pristinski, S. Tan, M. Erol, H. Du, S. Sukhishvili, J. Raman Spectrosc. 37 (2006)
62.
4
. Conclusion
7
[
[
12] R.J. Walsh, G. Chumanov, Appl. Spectrosc. 55 (2001) 1695.
13] C. Viets, W. Hill, Sens. Actuators B 51 (1998) 92.
SERS substrate is fabricated using AAO membrane decorated by
silver nanoparticles via a simple straightforward chemical route.
Silver is deposited in the form of nanoparticles in such a way that
a network structure is formed on the surface of AAO membrane to
produce an artificial roughness which creates strong enhancement
of the electric field. Intense SERS effect is observed due to electric
field enhancement at the sharp edges on the surface of the AAO
membrane. We believe this technique will be very useful to fabri-
cate the SERS substrates in a controllable way where the electric
field enhancement can be controlled in a precise way by tailoring
the particle size distribution etc.
[14] G.H. Gu, J.S. Suh, J. Phys. Chem. C 114 (2010) 7258.
[
[
[
[
15] N. Nuntawong, M. Horprathum, P. Eiamchai, K. Wong-ek, V. Patthanasettakul,
P. Chindaudom, Vacuum xxx (2010) 1.
16] T. Bhuvana, G.V. Pavan Kumar, G.U. Kulkarni, C. Narayana, J. Phys. Chem. C 111
(2007) 6700.
17] A. Dhawan, Y. Du, F. Yan, M.D. Gerhold, V. Misra, T. Vo-Dinh, IEEE Sens. J. 10
(
2010) 608.
18] Y. Du, L. Shi, T. He, X. Sun, Y. Mo, Appl. Surf. Sci. 255 (2008) 1901.
[19] P. Nielsen, O. Albrektsen, S. Hassing, P. Morgen, J. Phys. Chem. C 114 (2010)
459.
3
[
20] M. Moskovits, Surface-enhanced raman spectroscopy: a brief perspective, in:
Surface-Enhanced Raman Scattering – Physics and Applications, Springer,
Berlin, 2006, pp. 1–18.
[
[
21] K.S. Thorne, Rev. Mod. Phys. 52 (1980) 299.
22] P.N. Njoki et al., J. Phys. Chem. C111 (2007) 14664.
Acknowledgements
[23] A.L. Koh et al., ACS Nano 3 (2009) 3015.
[
[
[
24] Z. Niu, Y. Fang, Spectrochim. Acta Part A 66 (2007) 712.
25] G.H. Gu, J. Kim, L. Kim, J.S. Suh, J. Phys. Chem. C 111 (2007) 7906.
26] G.N. Xiao, S.Q. Man, Chem. Phys. Lett. 447 (2007) 305.
B.M. acknowledges CSIR, New Delhi, Govt. of India, for his
fellowship and SKS acknowledges DST Unit on Nanoscience and
Center for Nanotechnology for infrastructural facilities.
[27] G.V. Pavan, S. Shruthi, B. Vibha, B.A. Ashok Reddy, T.K. Kundu, C. Narayana, J.
Phys. Chem. C 111 (2007) 4388.