Sensors and Actuators B 226 (2016) 589–597

Contents lists available at ScienceDirect

Sensors and Actuators B: Chemical

journal homepage: www.elsevier.com/locate/snb

Au nanoparticles decoration of silica nanowires for improved optical

bio-sensing
A. Colombelli a , M.G. Manera a,∗ , A. Taurino a , M. Catalano a , A. Convertino b , R. Rella a
a
IMM-CNR Lecce, Institute for Microelectronics and Microsystems, Strada prov.le Lecce-Monteroni, 73100 Lecce, Italy
b
IMM-CNR Roma, Institute for Microelectronics and Microsystems, Via del Fosso del Cavaliere 100, 00133 Roma, Italy

a r t i c l e i n f o a b s t r a c t

Article history: In this work we study the potential of silica nanowires (NWs) decorated with Au nanoparticles (NPs) as
Received 27 August 2015 novel optical platform for plasmonic biosensing. These composite materials combine the high white-light
Received in revised form scattering of silica NWs with the selective absorption of the localized surface plasmon resonances (LSPRs)
13 November 2015
of the metal NPs, generating efficient light trapping just at the LSPR wavelengths. In addition, the NWs
Accepted 17 November 2015
offer a macroporous framework for a three-dimensional (3D) distribution of the Au NPs, easily accessible
Available online 2 December 2015
by biomolecules. These optical and morphological pecularities, make the decoration of silica NWs with
metal NPs a promising strategy for improved plasmonic biosensing. We fabricated silica NWs by thermal
Keywords:
Surface plasmon
oxidation of silicon (Si) NWs, grown by plasma enhanced chemical vapor deposition (PECVD). The silica
Optical sensing NWs were decorated with Au NPs by evaporating and then de-wetting via thermal annealing Au thin films.
Transducer The biosensing ability of these materials was studied and compared with the performances of similar
Silica Au NPs deposited on glass substrate, forming a planar (two-dimensional, 2D) distribution. A careful
Nanowires morphological analysis about the NWs and their Au NP coverage, including their vertical distribution
across the NW film, was performed by high resolution scanning electron microscopy (HR-SEM) and
energy dispersive X-ray spectroscopy (EDS) experiments. The morphological information was used for
building a theoretical model through 2D finite element modeling (FEM) able to describe the optical
behavior of the new transducers and their functional properties, as well as to predict the evolution of the
sensing performance of the transducer when passing from a 2D to a 3D distribution of plasmonic NPs.
The Au NPs decorated NWs were finally modified with a self-assembled monolayer of BSA (bovin serum
albumin) and the specific binding with the related antibody monitored during the time.
© 2015 Elsevier B.V. All rights reserved.

1. Introduction When compared to other optical transducing platform based on

The need for rapid, specific, sensitive, inexpensive, in-field
and real-time detection of biomolecules analytes has increased
biosensors importance especially in fields such as environmental
monitoring, pollutant detection, medical diagnostics, and biological
warfare defense.
The phenomenon of Localized Surface Plasmon Resonance
(LSPR), which results from the plasmonic response of noble metal-
lic nanoparticles to incident electromagnetic waves [1], can be
exploited in the label-free sensing of analytes. Binding of analytes
on metallic surfaces causes a change in the local refractive indices
(RIs) of the surrounding dielectric, resulting in either a shift in the
spectral extinction peak or a change in the peak intensity [2–4].
∗ Corresponding author. Tel.: +39 0832422523; fax: +39 0832422552.
E-mail address: mariagrazia.manera@cnr.it (M.G. Manera).
plasmon phenomena such as Propagating Surface Plasmon Reso-
nance (SPR) [5,6] or Surface Enhanced Raman Spectroscopy (SERS)
sensors [7,8], LSPR systems have some advantages including the
use of a simple analytical set-ups in addition to being portable
and practical. Transmission spectrometry, which is relatively sim-
ple and inexpensive, can be used for LSPR sensing of 10–100 nm
particles in the UV–vis region [9].
Much work has been devoted to develop sensors from sil-
ver and gold NPs suspended in solution [10,11]. However, the
strong demand for rapid, cost-effective, and high-throughput mea-
surements has led to the development of a chip-based assay
consisting of metal NPs immobilized on a substrate, a so-called
two-dimensional (2D) chip [12–14]. To this purpose, chemical
immobilization can be achieved via compounds such as alka-
nethiols [15,16], but also physical immobilization techniques as
well gave successful results [17–19]. Well-ordered 2D arrays of
monodisperse particles have the advantage of presenting narrow

http://dx.doi.org/10.1016/j.snb.2015.11.075
0925-4005/© 2015 Elsevier B.V. All rights reserved.
590 A. Colombelli et al. / Sensors and Actuators B 226 (2016) 589–597
well defined extinction bands so that an easy theoretical predic-
tion of optical response of such systems using the discrete dipole
approximation (DDA) method is possible [20,21].
However, such plasmonic devices that rely solely on changes in
the dielectric permittivity of the environment often suffer from low
sensitivity (typically <10 nm spectral shift for spherical nanopar-
ticles) [22,23] that originates from fairly small variations in the
refractive index because of the low density of nanoparticles in the
defined surface area.
Recently we have showed as silica nanowires (NWs) deco-
rated with Au and Ag NPs can act as efficient RI sensors [24],
due to their morphological and optical features, which are also
very attractive for biosensing applications. First, they form a dense
three-dimensional (3D) ensemble of metal NPs, easily accessible to
liquids, vapors, or biomolecules due to the macroporous structure
of the NW forest and the large interparticle distance. Second, they
offer absorption enhancement at the LSPR wavelengths because of
light trapping effects in the NW forest. To the best of our knowledge,
only a few attempts have been made to create three-dimensional
(3D) LSPR structures to increase the sensing performance of metal
NPs transducers.
Jiang and co-workers [25] fabricated multilayer structures
composed of Au nanoparticles (NPs) and polyelectrolyte by
spin-assembly or spin-assisted layer-by-layer (SA-LbL) deposition
techniques and discussed their optical properties by correlating
them to the density of the metal NPs into the multilayer struc-
ture. Plasmon resonance peaks from isolated NPs and interparticle
interactions were revealed in their UV–vis extinction spectra. In
Refs. [26,27] metal-coated macroporous films were realized using
silica spheres and used as transducer platform to measure a typ-
ical antigen-antibody binding event. Ye et al. [28] demonstrated
the improvement of bulk refractive index sensitivity of 3D self-
assembled Au NPs multilayer structures obtained by the alternate
deposition of a bifunctional cross-linker and Au NPs. The num-
ber of deposition cycles, in this case, was crucial in obtaining the
desired sensor performance. Analogously, Guo and coworkers [29]
realized a 3D Au nanoarchitecture by layer-by-layer (LbL) deposi-
tion of Au NPs and multiwalled carbon nanotubes (MWCNTs). After
demonstrating the improvement of the bulk refractive index (RI)
sensitivity of these multilayered Au NPs with respect to a mono-
layer of AuNPs on a glass chip, they experimented the beneficial role
of the 3D transducing platform also for the study of a biomolecular
binding experiment.
In this work, we study the biosensing feasibility of silicon
NW forests decorated with Au NPs. In particular, silica NW
were obtained by thermal oxidation of Si NWs grown by plasma
enhanced chemical vapor deposition (PECVD). The NWs were
decorated with metal Au NPs by de-wetting thin metallic films
evaporated on the NWs [24]. The proposed preparation techniques
offer the stability and reproducibility typical of a physical depo-
sition technique while keeping a degree of flexibility required for
producing a functional transducer platform. In order to compare
the sensing performance of the 3D plasmonic structure respect to
a traditional 2D configuration, Au NPs on glass substrate were also
prepared in the same conditions used for the decoration of the NWs.
Structural and morphological information were obtained by
Scanning Electron Microscopy in normal and cross section con-
figuration. They were used to build a theoretical model through
a 2D finite element model (FEM) able to describe the optical
behavior of the transducers and their functional properties. In the
numerical model the scattering in the NWs framework and the
plasmonic absorption of metallic NPs were taken into account.
Numerical results found a confirmation in experimental optical
characterization.
Finally, a functional characterization of the realized transducer
platform was carried out thus demonstrating an increase of the
bulk RI sensitivity in these 3D structure with respect to the classi-
cal 2D chip. The improved performances of the 3D structure with
respect the 2D were also predicted by the theoretical model. After
this calibration step, immobilization of BSA (bovine serum albu-
min) molecules was monitored, and the consequent binding with
anti-BSA antibodies was detected. The label free detection of the
interactions is easily achieved thus confirming the promising out-
comes of the proposed optical sensor.
2. Experimental details
2.1. Silica NWs decorated with Au NPs
Forests of silica NWs were produced by thermal annealing of
Si NWs grown by PECVD. First, gold-catalyzed Si NWs were grown
by PECVD Si(1 0 0) substrates. To induce the NWs growth, a 1 nm
thick Au film was evaporated onto the substrates prior to growth.
The growth was performed by using pure SiH4 as precursor at
a total pressure of 1 Torr and substrate temperature of 360 ◦ C. A
13.56 MHz radiofrequency with power density of 50 mW cm−2 was
used to create the plasma. In these growth conditions 15–20 ␮m
long Si NWs with average diameter of 150–200 nm at the bottom
and tapered shape can be obtained [30]. After the growth, the Si
NWs were oxidized to form SiO2 (silica) NWs by means of a ther-
mal treatment in a convection oven (controlled steam atmosphere)
at 980 ◦ C for 8 h. The decoration of the silica NWs with Au NPs was
obtained by thermal evaporation of a thin film of Au on the silica
NWs. The nominal thickness was fixed at 12 nm and estimated by
considering the deposition rate obtained in the case of evapora-
tion on a flat substrate. Au NPs were obtained as a result of metal
de-wetting at high temperatures (∼800 ◦ C) for 6 min.
In order to compare the sensing performance of the 3D plas-
monic structure with respect to a traditional 2D configuration, Au
NPs on glass substrate were also prepared in the same conditions
used for the decoration of the NWs, i.e. thermal evaporation of a
12 nm thick layer of Au onto glass substrates and successive thermal
treatment at 800 ◦ C for 6 min.
2.2. Morphology
The morphology of the silica NWs decorated with metal NPs was
investigated by SEM both in plan-view and cross-sectional geom-
etry. EDS line scan profiles were acquired in cross-section in order
to study the compositional distribution of gold, silicon and oxygen
along the thickness of the NWs film. A Zeiss NVISION 40 dual-beam
Focused Ion Beam machine, equipped with a high resolution SEM
Gemini column and an Oxford 350 x-act EDS spectrometer, was
used for SEM and EDS experiments. SEM observation were also
performed onto 2D samples containing gold NPs deposited on glass
substrates.
2.3. Modeling and simulation
The optical properties of the system were investigated through
finite element simulations. A numerical model for the diffuse reflec-
tivity of disordered NWs decorated with plasmonic nanostructures
was developed with the RF Module of COMSOL Multiphysics
platform. The model has taken into account the morphological
characteristics of the system, the scattering events of the light in
the NWs mats and the plasmonic absorption of metallic NPs. The
model well described the reflectivity measurements and the ability
of the system to perform real-time monitoring of molecular adsorp-
tion. We explored several key parameters in order to know how
the optical properties of the local environment influence the LSPR
conditions.
 A. Colombelli et al. / Sensors and Actuators B 226 (2016) 589–597
Fig. 1. (a) Representation of the used experimental set-up. (b) Schematic diagram of the sensor system.
2.4. Optical characterization
The optical behavior of all the fabricated samples was studied in
the spectral range between 200 and 1100 nm by angle-integrated
total reflection measurements which allowed the collection of
specular light beams and scattered light by using an integrating
sphere (Perkin-Elmer Lambda 950 spectrometer).
2.5. Sensing functional characterization
The bulk RI and antigen–antibody interaction studies were per-
formed by measuring the reflectivity spectra in a flow cell in which
different solutions (water/ethanol solutions at different concentra-
tion volumes) were fluxed.
The optical system used to characterize the biosensing
layer consists of a tungsten halogen light source (LS-1, wave-
length range 360–2000 nm), a spectrophotometer (USB2000
UV–vis, wavelength ranging between 250 and 1100 nm), and
an optical bifurcated fiber probe (R-400-7 UV–vis, fiber core
diameter = 400 ␮m, wavelength range = 250–850 nm). White light
emerging from the optical fiber was perpendicularly shed onto
the Au NPs layer chip. The reflected light was coupled into the
detection fiber probe in the same bundle and analyzed by using
the UV–vis spectrophotometer. All spectra were taken from 400 to
800 nm at room temperature. For minimization of the stray light
in experiments using fluids, the fiber tip was entirely immersed.
To assess the contribution from light scattering, a spectrum was
also obtained with the sample placed directly within an inte-
grating sphere of an UV–vis spectrophotometer Cary 500. The
results indicate comparable contributions from absorption and
scattering to the total extinction. The experimental set-up (shown
in Fig. 1a) allows acquiring the dynamic responses at different
fixed spectral range where the sensing response shows a maxi-
mum. In our case 400–800 nm spectral range. The effect of the
small variations in the RI of the liquid surrounding the NPs or
591
in the antigen–antibody interaction were measured in a dynamic
microfluidic system implemented in our laboratory (Fig. 1a and b).
A peristaltic pump was used to control the flux of the liquid in the
test chamber during the experiment. Controlled room temperature
conditions during all the sensing experiments were guaranteed.
The immobilization of BSA molecules onto Au NPs decorating
the silica forest and the monitoring of the consequent binding with
Anti-BSA antibody was recorded by LSPR spectra changes. The pro-
cedure of the transducer preparation included several sequential
steps as explained in the following sections.
3. Results and discussion
3.1. Morphology of silica NWs decorated with metal NPs
The NWs and NPs morphology was first investigated by plan-
view SEM analyses. Fig. 2a reports an overview of the Au decorated
NWs forest, where a highly dense and disordered NWs distribution
can be observed. The NWs exhibit a tapered shape and an average
diameter of some hundreds of nanometers on the larger sections.
As evidenced by the high magnification image in Fig. 2b, the wire
surface is covered with Au NPs, whose diameters vary from few to
tens of nanometers. The largest particles show an evident faceting
and appear to be isolated, which is a direct consequence of the
dewetting coalescence phenomenon, as reported in literature [31].
These particles are generally observed at the centre of the width
of those wires whose axis is nearly parallel to the substrate; this is
likely to be due to the higher sticking coefficient achieved on these
surfaces in comparison to the surfaces not perpendicularly exposed
to the evaporated gold flux. The size of the particles decreases at the
edges of the wires. The inset of Fig. 2b evidences the capability of
the NPs to grow on the extreme tip of the wires, where the tapering
phenomenon leads to a wire diameter of few nanometers.
EDS cross-sectional analyses were carried out on a cleaved edge
of the sample. Line scan profiles were acquired by scanning the
592 A. Colombelli et al. / Sensors and Actuators B 226 (2016) 589–597

Fig. 4. SEM micrograph of the NPs distribution deposited onto glass substrate.

Fig. 2. SEM plan-view secondary electron images of the NWs forest at low (a) and
high (b) magnification. The inset of figure (b) shows the Au coverage of the extreme
tapered tip of a wire.
electron beam from the Si/SiO2 interface to the top of the NWs
film. Fig. 3 reports the SEM image of the cross section together with
the EDS profiles of Si (in red), O and Au signals acquired during
the electron beam scanning along the 18 ␮m long yellow line. A
silicon dioxide layer of 1.8 ␮m between the Si substrate and the
NWs film can be observed. The analyses of these profiles allowed to
extrapolate quantitative information to be used for the NWs model.
An estimation of the NWs density was obtained from the Si pro-
file. From the value of the Si counts in the substrate, corresponding
to a 100% Si density, the Si percentage in the film was derived from
the profile: a value of (32 ± 2)% was obtained at the bottom of the
NWs film and a value of (10 ± 1)% was obtained at the top of the film.
Similarly, the oxygen concentration gradually decreases; consider-
ing the Si/O counts ratio along the scan profile, it can be noted that
this ratio remains almost constant, being 2.92 ± 0.76 in the NWs
film and 2.77 ± 0.78 in the SiO2 , demonstrating that the NWs have
a uniform composition along the film depth.
Some information about the Au coverage can be also derived
from the EDS profile: the Au signal increases almost linearly when
going from the bottom of the NWs to their top; this trend is opposite
to the one observed for the Si signal, giving a clear evidence that
the Au NPs are mostly distributed on the outer part of the NWs
layer, that mainly exposed to the evaporated gold flux, whereas
the innermost areas remain uncovered.
By using the fabrication process described above to decorate
the Si NWs, a planar distribution of Au NPs on a glass substrate was
realized. The morphology of the 2D samples was also investigated
by plan-view SEM analysis. The 2D architecture is shown in Fig. 4. As
one can see the dimension of Au NPs are similar to those distributed
onto Si NWs ranging between 20–30 nm. In the inset the optical
absorption spectrum of the Au NPs distribution is shown. As one can
see the absorption peak due to the localized plasmons is centered
at about m = 530 nm.

Fig. 3. SEM cross-sectional image showing the Si substrate, the SiO2 layer and the NWs forest. The EDS profiles of Si (counts reduced by a factor 5 to compare all the profiles
in the same graph), O and Au are reported in the graph. (For interpretation of the references to color in this figure legend, the reader is referred to the web version of this
article.)
 A. Colombelli et al. / Sensors and Actuators B 226 (2016) 589–597
3.2. Theoretical modeling of highly disordered NWs system
Disordered NWs mats have significant diffuse reflectivity arising
from multiple scattering events that the light undergoes by pass-
ing across NW forests [30,32–34]. Systems of Silica NWs assure,
for example, large light scattering at all wavelengths in the vis-
ible range, appearing as a natural bright white mat. Recently it
has been also demonstrated how the outstanding optical reflec-
tion, especially in the UV range, makes these materials very suitable
in UV lightening devices [32]. The presence of Au particles on the
NWs modifies their reflectivity and their appearance because of
the selective optical absorption of plasmonic nanostructures [24].
How the geometrical properties (shape, dimension, distribution) of
both NWs and metal nanostructures may modify the reflection of
these materials is not yet clarified, and this is a critical issue for the
development of efficient NW/NP based devices. Here we developed
a numerical model with the RF Module of COMSOL Multiphysics,
which calculated the reflectivity of these combined systems. Opti-
cal and electric properties were calculated across a wide spectral
range including substrate and environmental effects. Due to the
highly disordered nature of the system, it is useless to develop a
model with a precise 3D geometry, as it can be done with ordered
vertical NWs. Instead, we developed a finite element simulation
characterized by a 2D multi-layered structure, showing its appli-
cation to silica NWs decorated with Au nano-spheres. As built, the
model can predict easily the optical properties of other disordered
NWs systems.
Fig. 5. (a) Schematic of the layered simulation domain. The geometry is not to scale. (b) Periodic Au NPs distribution characterized by an increasing density in the NW layer.
(c) SiO2 and Air concentrations calculated for the NW layer (left); refractive index distribution calculated for the NW layer (right).
593
3.2.1. Building of the model
The simulation domain was composed of several layers with
optimized thicknesses. By using the results of SEM morphological
analysis and EDS profile, we built a model composed by the fol-
lowing layers visible in Fig. 5a: air environment (10 ␮m), SiO2 NWs
layers (15 ␮m), a suitable 3D distribution of Au NPs Fig. 5b, SiO2
layer (1.8 ␮m) and the Si substrate (10 ␮m). The SiO2 NWs were
modeled with a series of 15 layers of composite material SiO2 /Air,
characterized by an inwards increase of the refractive index over
the NW layer. This variation was due to the decreasing filling fac-
tor of SiO2 along the wires growth direction, which arises from
both tapering of the wires and the presence of a distribution of
the NW length that especially affects the outer part of the film.
The presence of Au nanostructures was simulated by adding, in
the simulation domain, a dense array of nano-spheres with diam-
eter of 20 nm. Both the Au particles distribution and the RI of the
wire layers were optimized by exploiting the results of the EDS
profiles reported in the previous section. The effective distribution
of Si, O and Au from the substrate interface to the surface of the
sample, represented a key factor in the model construction. The
gradual increase of Au signal from the base to the top of the NW
film suggested that Au NPs are mostly distributed on the top of
the wires. In the model, we reproduced a similar particles distribu-
tion, by adding a periodic array of nano-spheres characterized by
an outward increase of the density. The silicon and the oxygen EDS
profiles confirm the decreasing filling factor of SiO2 in the outer
part of the film allowing the estimation of the concentrations of
594 A. Colombelli et al. / Sensors and Actuators B 226 (2016) 589–597
Fig. 6. Simulated and experimental UV–vis spectra comparison of the NWs system
(red lines) and the NWs/AuNPs hybrid structure (black lines).
these elements in the NWs mats. The effective refractive index of
the NWs layers was calculated by considering a SiO2 /Air compos-
ite material with a decreasing SiO2 concentration that varies from
33%, near the substrate interface, to 13% near the sample surface
(Fig. 5c). Finally, complex wavelength dependent RIs were set for
both the SiO2 layer and the silicon substrate.
3.2.2. Optical investigation of the model
In Fig. 6 we plot the inverse of the experimental reflectivity spec-
tra obtained from pristine silica NWs and Au decorated silica NWs
(continous black line) in the wavelength range 500–600 nm. The
spectrum of the pristine silica NWs is characterized by low values
(i.e. high reflectivity values) whereas the Au decorated NWs show
a peak at around 540 nm. Fig. 6 shows also the simulated spectra
(dashed red lines) obtained with 2D finite element model previ-
ously described for both the samples. At the investigated wave-
lengths, Au was modeled with a complex RI. The real and imaginary
components of the Au RI were described by an interpolated
Fig. 7. Normalized integral absorption response curves of the sensing layer exposed to different ethanol concentration (dynamic sensing curves).
version of the often-used experimental data from Johnson and
Christy [35]. By using a parametric sweep in the frequency domain,
we calculated the total electromagnetic (E.M.) field distribution
(given by the sum of the incident and scattered E.M. fields) and the
inverse reflectivity of the system, as function of the light frequency.
We simulated the reflectivity of the system by taking into account
the diffusion optical properties of the nanowires and the plasmonic
absorption of metallic nanoparticles. The experimental spectra are
well reproduced by the model, showing a clear peak for the Au dec-
orated silica NWs at around  = 540 nm in the spectral region where
LSPR is activated. The analysis of the electric field distribution near
the particle surface suggested, indeed, that this resonance mode
corresponds just to an oscillating dipole mode. The higher value of
the electric field near the particle surface was found, as expected, in
the same spectral region where the inverse of reflectivity showed
its peak.
Theoretical and experimental data confirmed an efficient trapp-
ing of light in the SiO2 NWs mat decorated with Au NPs. The
multiple light scattering action of the SiO2 wires enhanced the
light absorption of plasmonic nanostructures, explaining the very
low reflectivity observed in the UV–vis range. The proposed model,
although relatively simplified, demonstrated to describe quite well
the experimental results, clearly indicating that the SiO2 NW layer
is efficiently described by a variation of the refractive index over its
thickness.
3.2.3. Refractive index sensitivity simulation
The modeling of the two investigated transducing nanostruc-
tures allowed us also building a simulated calibration curve relative
to bulk Refractive Index changes in the environment surrounding
the metal NPs, both in a planar and three-dimensional configura-
tion. The results are reported in the following by comparison with
experimental ones.
3.3. Functional characterization of the investigated transducer
platforms
3.3.1. Refractive index sensitivity tests
The sensitivity of the silica NWs/AuNPs and 2D Au array was
initially investigated by recording LSPR spectra after their immer-
sion in solutions with different refractive indices: air (n = 1), pure
ethanol (n = 1.36), water (n = 1.33) and in a water/ethanol solutions
at increasing concentrations. In Fig. 7 we plot the dynamic response,
evaluated as the normalized integrated area (NIA) of LSPR peak, for
both the transducers exposed to different ethanol/water concen-
tration, i.e. by varying the RI of the environment. Changes of the
sensors signal were due to changes in the intensity and wideness of
 A. Colombelli et al. / Sensors and Actuators B 226 (2016) 589–597 595

Fig. 8. (a) Experimental calibration curves of the 2D and 3D systems immersed in different concentrations of ethanol; (b) calibration curves calculated for 2D and 3D periodic
distribution of gold NPs immersed in different concentrations of ethanol.

the LSPR curves as shown in the following Fig. 9. This method of sen-
sitivity measurement can be often found in literature as it ensures a
higher signal-to-noise ratio compared to the peak wavelength shift
measurement [36,37]. The Au decorated NWs are characterized by
very short rise time, in the order of seconds, and can appreciate
RI change as low as 0.2%, values much lower than those detected
(0.67%) from the 2D Au array.
In Fig. 8a we show the experimental sensitivity, i.e. the NIA,
for both the 3D and 2D transducer platforms as function of the RI
variation. Improved performance are exhibited by the decorated

Fig. 9. (a) Illustration of the Au-NPs functionalization for the detection of BSA–AntiBSA biological interactions. (b) Absorption spectra of the silica NWs/Au-NPs system and
(c) dynamic curve (Integrated absorption vs. time) registered during the biosensing measurements.
596 A. Colombelli et al. / Sensors and Actuators B 226 (2016) 589–597

NWs with respect the Au 2D distribution. Indeed experimental

values of 4.6 NIA/RIU and 8.5NIA/ RIU were obtained for 2D and
3D transducer platforms, respectively. The improved performances
exhibited by the decorated NWs in both dynamic and calibration
curves were well predicted by numerical simulation reported in
Fig. 8b where a comparison between planar and 3D sensing archi-
tecture for the detection of very low refractive index changes is
reported. The FEM analysis was clearly performed by taking into
account the changes of environmental RI and the predicted sensi-
tivity gives us for 3D configuration values twice higher than those of
2D array. This demonstrates the robustness and efficacy of the built
theoretical model not only in describing the optical and functional
properties of the investigated plasmonic transducers, but also in
predicting their sensing evolutions upon changes in morphological
and optical properties of the nanostructures.
We expect that these values can be strongly improved by opti-
mizing density and size of the gold NPs and silica NWs, which
depend on the amount of metal evaporated and on the de-wetting
procedure, as well as on density and length of the silica NWs.

3.3.2. BIO sensing performance of metal NPs decorated silica NWs

transductors
Fig. 9 shows the related LSPR spectra of the Au decorated silica
NWs corresponding to each step involved in the biosensing exper-
iment. As can be noticed, surface modification of the transducers
resulted in a redshift of the recorded curves toward higher wave-
length. Thiol functionalization as well as BSA immobilization and
consequent hybridization with the corresponding antibody pro-
duced a change, namely an increase of RI at the interface Au/water
which is responsible of the change in LSPR curves (Fig. 9b).
The Au functionalization step is sketched in Fig. 9a. In brief, pre-
liminary cycles of a solution of 10% ethanol in water and pure water
were sent into the test chamber to clean the sample surface and to
get a stable baseline of the signal for the LSPR sensing. Then, a mix-
ture of NHS/EDC was introduced to activate the thiol surface and
finally the sample was cleaned with the solute of 10% ethanol in
water and pure water (this sequence is not shown here). Finally, an
aqueous solution of 50 ppm of BSA molecule was sent into the test
chamber and left in dynamic condition for about 20 min prior wash-
ing in pure water. After washing, the sensor signal remains almost
unchanged, thus ensuring that BSA molecules cannot be removed
and therefore it has been correctly immobilized onto the trans-
ducer and ready for the following detection of the correspondent
antibody.
A blocking step with ethanolamine (1 M) was necessary to
avoid non-specific interactions of the target biomolecule. Finally,
when the anti-BSA (10 ppm) antibody solution was injected into
the sensor flow-cell, fast increase in the sensing signals occurred.
This increase was associated with the specific binding BSA a-BSA
molecules at the Au NPs surface. After switching back to the PBS
buffer solution, the signal dropped slightly and levelled off at a
lower level. The difference between this new level and initial base-
line was associated with the amount of bound a-BSA.
The dynamic response reported in Fig. 9c demonstrates also that
LSPR curve is able to record regeneration of the biosensor surface
for its eventual re-use. Regeneration in HCl allowed for releasing
the binding between BSA and the corresponding antibody, thus
restoring the surface with BSA molecules still present onto the
surface.
For high a-BSA concentrations, the initial fast binding was
followed with a slower binding, thus demonstrating a kind of sat-
uration of the sensor surface toward further a-BSA binding. Fig. 10
reports a typical calibration curves realized by monitoring the
biosensing response as a function of different ppm concentration
of a-BSA. Here the sensor response is defined as the ratio of the
signal intensity at the steady state at each analyte concentration
Fig. 10. Calibration curve of the silicon NWs/AuNPs functionalized bioassay
immersed in different concentrations of Anti-BSA antibodies.
and the baseline signal recorded in the buffer solution before the
first analyte injection. As can be noticed, the normalized LSPR sig-
nal gives an interesting sensor response in the investigated analyte
concentration ranges.
Label-free detection of the interaction was easily achieved, thus
confirming the promising outcomes of the proposed NW/NP struc-
ture. Increasing in sensitivity of these new transducers materials
by optimizing density and morphology of both the metal NPs and
silica NWs is to be expected. In particular we expect that the
performances of this hybrid system can be strongly improved by
optimizing NPs density and size, as well as changing the density
and length of the silica NWs.
4. Conclusions
In conclusion, we reported a new class of plasmonic materials
consisting of silica NWs decorated with metal NPs characterized by
morphological and optical features very attractive for biosensing
applications. The microporous structure of the NWs forest allowed
the allocation of 3D ensemble of metal NPs, easily accessible to
liquids, vapors, or biomolecules. In addition these materials com-
bined the large light scattering by silica NWs with the selective
absorption resonances provided by metal NPs, showing a strong
absorption enhancement just at the correspondent LSPR frequen-
cies. We developed a numerical model for the diffuse reflectivity of
thick disordered nanowire mats decorated with plasmonic nano-
structures. In spite of its handiness, the model demonstrated it
efficacy in describing quite well the optical and functional prop-
erties of the investigated transducers, predicting the improvement
in sensing performance when passing from a 2D to a 3D distribu-
tion. The improved performance of the decorated NWs with respect
the traditional 2D Au NPs distribution to detect even small changes
in the RI of the environment was shown as well as their potential
as plasmonic biosensors by monitoring the immobilization of BSA
and detecting the consequent binding at very low concentration of
Anti-BSA antibodies. The robustness of the model, suggested us the
possibility to follow an analogue route to investigate the effect of
 A. Colombelli et al. / Sensors and Actuators B 226 (2016) 589–597
the order in plasmonic nanostructures on their sensing ability, in
addition to the three dimensionality parameter.
Acknowledgment
The authors want to acknowledge M. Maiani and G. Montagna
for technical support and the financial support from Italian MIUR
by the grant number RBFR10OAI0 (FIRB-Nanoplasmag).
References
[1] S. Link, M.A. El-Sayed, J. Phys. Chem. B 103 (1999) 8410–8426.
[2] A. Neval, S.B. Cinel, Ö. Ekmel, Opt. Express 20 (2012) 2587–2597.
[3] J. Spadavecchia, P. Prete, N. Lovergine, L. Tapfer, R. Rella, J. Phys. Chem. B 109
(2005) 17347–17349.
[4] P.J. Rivero, M. Hernaez, J. Goicoechea, I.R. Matias, F.J. Arregui, SPIE Proc. 9157
(2014) 9157–9161.
[5] A.P.F. Turner, Science 290 (2000) 1315–1317.
[6] H.J. Lee, D. Nedelkov, R.M. Corn, Anal. Chem. 78 (2006) 6504–6510.
[7] K.E. Shafer-Peltier, C.L. Haynes, M.R. Glucksberg, R.P. Van Duyne, J. Am. Chem.
Soc. 125 (2) (2003) 588–593.
[8] M. Kahraman, M.M. Yazici, F. Sahin, O.F. Bayrak, M. Culha, Appl. Spectrosc. 61
(5) (2007) 479.
[9] J.N. Anker, W.P. Hall, O. Lyandres, N.C. Shah, J. Zhao, R.P. Van Duyne, Nat. Mater.
7 (2008) 442–453.
[10] R. Elghanian, J.J. Storhoff, R.C. Mucic, R.L. Letsinger, C.A. Mirkin, Science 277
(1997) 1078–1081.
[11] C.A. Mirkin, R.L. Letsinger, R.C. Mucic, J.J. Storhoff, Nature 382 (1996) 607–609.
[12] M.D. Malinsky, K.L. Kelly, G.C. Schatz, R.P. Van Duyne, J. Am. Chem. Soc. 123
(2001) 1471–1482.
[13] N. Nath, A. Chilkoti, Anal. Chem. 74 (2002) 504–509.
[14] N. Nath, A. Chilkoti, Anal. Chem. 76 (2004) 5370–5378.
[15] K.C. Grabar, P.C. Smith, M.D. Musick, J.A. Davis, D.G. Walter, M.A. Jackson, A.P.
Guthrie, M.J. Natan, J. Am. Chem. Soc. 118 (1996) 1148–1153.
[16] J. Schmitt, G. Decher, W.J. Dressick, S.L. Brandow, R.E. Geer, R. Shashidhar, J.M.
Calvert, Adv. Mater. 9 (1997) 61–65.
[17] C.L. Haynes, A.D. McFarland, L.L. Zhao, R.P. Van Duyne, G.C. Schatz, L. Gunnars-
son, J. Prikulis, B. Kasemo, M. Kall, J. Phys. Chem. B 107 (2003) 7337–7342.
[18] M.D. Malinsky, K.L. Kelly, G.C. Schatz, R.P. Van Duyne, J. Phys. Chem. B 105
(2001) 2343–2350.
[19] T.R. Jensen, G.C. Schatz, R.P. Van Duyne, J. Phys. Chem. B 103 (1999) 2394–2401.
[20] H. Gehan, C. Mangeney, J. Aubard, G. Levi, A. Hohenau, J.R. Krenn, E. Lacaze, N.
Felidj, J. Phys. Chem. Lett. 2 (2011) 926–931.
[21] W.T.S. Huck, S. Kim, N. Cheng, J.R. Jeong, S.G. Jang, S.M. Yang, Chem. Commun.
31 (2008) 3666–3668.
[22] K.M. Mayer, J.H. Hafner, Chem. Rev. 111 (2011) 3828–3857.
[23] A.C. Templeton, J.J. Pietron, R.W. Murray, P. Mulvaney, J. Phys. Chem. B 104
(2000) 564–570.
[24] A. Convertino, M. Cuscunà, F. Martelli, M.G. Manera, R. Rella, J. Phys. Chem. C
118 (2014) 685–690.
[25] C. Jiang, S. Markutsya, V.V. Tsukruk, Langmuir 20 (2004) 882–890.
[26] H.H. Chen, H. Suzuki, O. Sato, Z.Z. Gu, Appl. Phys. A: Mater. Sci. Process. 81
(2005) 1127–1130.
[27] Z.Z. Gu, R. Horie, S. Kubo, Y. Yamada, A. Fujishima, O. Sato, Angew. Chem. Int.
Ed. 41 (2002) 1153–1156.
[28] J. Ye, K. Bonroy, D. Nelis, F. Frederix, J. D’Haen, G. Maeb, G. Borghs, Colloids Surf.,
A: Physicochem. Eng. Aspects 321 (2008) 313–317.
[29] L. Guo, G. Chen, D.-H. Kim, Anal. Chem. 82 (2010) 5147–5153.
[30] A. Convertino, M. Cuscunà, F. Martelli, Nanotechnology 21 (2010) 355701.
[31] F. Ruffino, M.G. Grimaldi, Phys. Status Solidi A: Appl. Mater. Sci. (2015) 1–23.
[32] T. Shi, L. Zhang, D. Liu, W. Lai, Z. Tang, Thin Solid Films 529 (2013) 115–118.
[33] R.A. Street, P. Qi, R. Lujan, W.S. Wong, Appl. Phys. Lett. 93 (163109) (2008)
1–3.
[34] A. Convertino, M. Cuscunà, S. Rubini, F. Martelli, J. Appl. Phys. 111 (114302)
(2012) 1–9.
[35] P.B. Johnson, R.W. Christy, Phys. Rev. B: Condens. Matter 6 (1972) 4370–4379.
[36] S.-H. Wu, K.-L. Lee, A. Chiou, X. Cheng, P.-K. Wei, Small 9 (2013) 3532–3540.
[37] L. Olofsson, T. Rindzevicius, I. Pfeiffer, M. Kall, F. Hook, Langmuir 19 (2003)
10414–10419.
597
Biographies
Adriano Colombelli obtained his degree in Physics in 2011 at the University of
Salento in Lecce (Italy). In July 2015 he received a Ph.D. degree in Materials and
Structural Engineering at the Innovation Engineering Department of the University
of Salento, discussing a thesis on new strategies for modulation of nano-structured
plasmonic materials. Since 2012 he joined as scholarship holder the Institute for
Microelectronics and Microsystem of the National Research Council (CNR) in Lecce.
He is involved in the development of chemical and biochemical sensors based on
optical transduction mechanism. His research activities are focused on numeri-
cal simulation and experimental application of surface plasmon resonance based
sensors, modeling and fabrication of micro-fluidics devices for medical applications.
M.G. Manera received her University degree in Physics and Ph.D. degree in “Mate-
rial and new technologies” from the University of Lecce, Italy in 2003 and 2006. Her
research interests lie in the area of optical characterization of organic and inorganic
materials for optochemical sensing and biosensing applications exploiting surface
Plasmon resonance transduction methodologies. Recently she has been involved
also in the study and functional characterization of metal oxide based gas sensors,
particularly when exposed to light illumination. Currently she is involved in the
functional characterization of magneto-plasmonic materials as novel magnetoop-
tical transducing platforms for improving the gas and bio-sensing performances of
standard surface plasmon resonance sensors.
Antonietta Taurino received the Laurea (with honors) and Ph.D. degrees in physics
from the University of Lecce, Lecce, Italy, in 1995 and 1999, respectively. Since 2001,
she has been a Researcher with the Institute for Microelectronics and Microsys-
tems, National Research Council (IMM-CNR), Lecce. Her competences are related
to the analysis of the morphological, structural, and compositional properties of
semiconducting materials by transmission electron microscopy (TEM) techniques.
Her main fields of interest are related to nanostructured thin films for gas sensors
and quantum confined semiconductors for optoelectronic applications. Recently,
her activity has been extended to advanced SEM techniques, like scanning TEM and
electron-beam-induced current, as well as to focused-ion-beamtechniques for the
nanofabrication, nanodeposition, and nanomanipulation of materials and devices.
Massimo Catalano graduated in physics in 1987, with full honours. Since 1998
he has been employed as a permanent researcher by the Institute for Microelec-
tronics and Microsystems of CNR. Presently he is Senior Scientist in the same
Institute, where he coordinates the Microstructural and Microanalytical Character-
ization Laboratory. His competences cover the microstructural and microanalytical
characterization of materials and devices by means of both Scanning and Trans-
mission Electron Microscopy techniques (SEM-TEM). The main fields of interest are
related to nanostructured thin films for gas sensing and quantum confined semi-
conductors for optoelectronic applications. Recently, the activity has been extended
to Focused Ion Beam (FIB) techniques for the nanofabrication, nanodeposition and
nano-manipulation of materials and devices.
Annalisa Convertino received her Laurea degree in Physics at the University of Bari
(Italy) in 1994 and her Ph.D. in Physics at the University of Bari in 1999. She is cur-
rently researcher at the for Microelectronics and Microsystems of CNR in Rome,
Italy. Currently she is involved in the research on organic–inorganic nanocomposite
functional materials for the development of optical gas sensors and in nanostruc-
tured semiconductors (nanorods and quantum dots) for optoelectronic and photonic
applications.
R. Rella, physicist, from 2001 senior researcher of the National Research Council
at the Microelectronics and Microsystems Institute (IMM-CNR) unit of Lecce—Italy.
Adjunct professor of material’s science and technology at the University of Salento.
The research activity is connected with the optimization of prototypes of new
devices for optoelectronic sensing and biosensing applications. Optics and opto-
chemical sensors with new organic and inorganic material systems like metal
oxide semiconductors, organic microcycles, conjugated polymers, dots and rods
of metal oxide semiconductor, gold nanoparticles, organic/inorganic multilayer
systems. Deposition in thin film form by using various preparation techniques,
in particular self-assembly, spinning, thermal evaporation, self-assembly and
Langmuir–Blodgett technique. The emphasis of this work is on the fabrication of
electronic devices incorporating these layers in the realization of chemical (gas,
vapor and liquid) sensors, self-organized structures and nanoscale devices. Analysis
of the sensing performances and optimization of the sensing parameters. Develop-
ment and application of surface-sensitive spectroscopic techniques such as surface
plasmon resonance (SPR) imaging, magneto-plasmonic SPR, localized SPR. Chemical
modification of metal surfaces for adsorption based biosensors (BIOMEMS).