ARTICLE

pubs.acs.org/Langmuir

Au Nanoparticle Monolayers Covered with Sol
Gel Oxide Thin Films:

Optical and Morphological Study
Enrico Della Gaspera, Matthias Karg, Julia Baldauf, Jacek Jasieniak, Gianluigi Maggioni,|| and
Alessandro Martucci*,

Dipartimento di Ingegneria Meccanica Settore Materiali, Universita di Padova, Via Marzolo, 9, 35131 Padova, Italy
School of Chemistry & Bio21 Institute, University of Melbourne, Parkville, VIC, 3010, Australia

CSIRO Materials Science and Engineering, Ian Wark Laboratory, Bayview Avenue, Clayton 3168, Australia
Dipartimento di Fisica, Universita di Padova c/o INFN Legnaro National Laboratories, Viale dell'Universita, 2 35020 Legnaro (Pd) Italy

ABSTRACT: In this work, we provide a detailed study of the inuence of thermal annealing on
submonolayer Au nanoparticle deposited on functionalized surfaces as standalone lms and
those that are coated with sol
gel NiO and TiO2 thin lms. The systems are characterized
through the use of UV
vis absorption, X-ray diraction (XRD), atomic force microscopy
(AFM), scanning electron microscopy (SEM), and spectroscopic ellipsometry. The surface
plasmon resonance peak of the Au nanoparticles was found to red-shift and increase in intensity
with increasing surface coverage, an observation that is directly correlated to the complex
refractive index properties of Au nanoparticle layers. The standalone Au nanoparticles sinter at
200
C, and a relationship between the optical properties and the annealing temperature is
presented. When overcoated with sol
gel metal oxide lms (NiO, TiO2), the optical properties
of the Au nanoparticles are strongly aected by the metal oxide, resulting in an intense red shift
and broadening of the plasmon band; moreover, the temperature-driven sintering is strongly limited by the metal oxide layer.
Optical sensing tests for ethanol vapor are presented as one possible application, showing reversible sensing dynamics and
conrming the eect of Au nanoparticles in increasing the sensitivity and in providing a wavelength dependent response, thus
conrming the potential use of such materials as optical probes.

INTRODUCTION
Noble metal nanoparticles (NPs) have recently been dispersed inside numerous active metal oxide matrices extensively
studied as high-performance materials for sensing,1
3 catalysis,4,5
and within optoelectronic devices.6,7 Conventionally, such nanocomposites are prepared as thin lm conguration through the
use of techniques such as sputtering, physical vapor deposition
(PVD), and chemical vapor deposition (CVD). However, despite each of these techniques requiring expensive deposition
equipment, they produce lms with poor control of NP size, size
distribution, and spatial distribution within the lm.
A simpler methodology relies on a two-step process whereby
NPs are rst chemically synthesized and dispersed in a host
matrix.8
10 The embedding of monodisperse metallic NPs inside
sol
gel matrixes is an example of such a methodology, which
potentially obviates the pitfalls of the above-mentioned techniques, while presenting a cheap and straightforward way to create
nanocomposite materials with tunable optical and electronic
properties. Nevertheless, the practical deposition of homogeneous composite thin lms is not trivial, because there are many
dierent parameters involved in achieving a stable colloidal NP
dispersion, such as pH, solvent type, ligand chemistry, and
complexing agents.11,12
For these reasons, in this work we have decided to develop
composite thin lms through a dierent approach. We have
r 2011 American Chemical Society

employed a multilayer process involving an initial deposition of

Au NP monolayers and subsequent sol
gel lm deposition. By
rst anchoring noble metal NPs to a suitable substrate, a wide
combination of metal oxides and noble metal NPs could be
investigated while circumventing all the problems related with
colloidal stability of the NPs. Moreover, if the metal NPs are
optically active in the visible range due to the surface plasmon
resonance (SPR), the mutual proximity in a close-packed NP
layer may induce coupling of the plasmon frequencies, resulting in potentially novel and interesting optical and electronic
properties.13
Moreover, control of the resonance conditions, like tailoring
NP organization, dielectric environments, and their stacking, is
required in many technological applications like optical sensors
and biosensors,14
17 surface-enhanced Raman spectroscopy,18
deep-colored coatings,19 and catalysis.20,21
In this paper, we present a detailed characterization of Au NP
monolayers that are deposited with dierent surface coverages,
and are subsequently overcoated with an active metal oxide
(TiO2 and NiO). The eect of the Au NP layer surface coverage,
annealing temperature, and type of metal oxide coating has been
Received: June 1, 2011
Revised:
September 28, 2011
Published: October 04, 2011
13739

dx.doi.org/10.1021/la2032829 | Langmuir 2011, 27, 1373913747

Langmuir

ARTICLE

Table 1. Sample Formulations Based on Au NPs Surface

Coverage and Top Layer Composition
sample name

Au NPs amount

top layer

AuL

Low

AuM
AuH

Medium
High

AuLN

Low

NiO

AuMN

Medium

NiO

AuHN

High

NiO

AuLT

Low

TiO2

AuMT

Medium

TiO2

AuHT

High

TiO2

assessed in detail, as well as the inuence of the top layer in

limiting the temperature-driven Au NP sintering and growth.
Optical sensing tests for ethanol vapor detection were also
carried out to exemplify one possible application of these
nanoscale architectures that can be employed in several other
optochemical and optoelectronic devices. The versatility of the
described approach enables it to be easily extended to other types
of metal NPs (Ag, for example) or semiconducting NPs (like
quantum dots as CdSe@CdS or PbSe), and to a variety of
solution-based coatings, from oxides to polymers.

EXPERIMENTAL SECTION
All chemicals used in the sample preparation have been purchased
from Sigma-Aldrich and used without any further purication.
Au NPs of about 14 nm mean diameter were prepared with the
Turkevich method.22 Briey, 12 mL of 1% trisodium citrate (>99%)
aqueous solution was added to a 200 mL boiling solution of 0.5 mM
HAuCl4 trihydrate (99.9%) in Milli-Q water. After the solution turned a
red-wine color, it was stirred at boiling point for an additional 15 min and
then was cooled down to room temperature. Separately, 11-mercaptoundecanoic acid (MUA, 95%) was dissolved in 10 mL of water and
0.25 mL of ammonium hydroxide solution (33%) yielding a 2 mM
concentrated solution and then added as a complexing agent. The
resulting colloidal suspension was then puried and concentrated
through a precipitation/redispersion process that has been previously
described.11 The glass substrate was functionalized with (3-aminopropyl)trimethoxysilane (APTMS, 97%) using the method reported in ref
23. Briey, the substrates were dipped in a 1% APTMS solution in
toluene at 60
C for 5 min, and subsequently washed with fresh toluene
and dried in a nitrogen stream. Then, Au NP monolayers were formed by
spin-coating the liquid suspensions of gold NPs directly onto the
APTMS monolayers. In this study, we prepared Au NP monolayers
with 3 dierent extents of surface coverage, hereafter indicated as low
(L), medium (M), and high (H). The as-deposited Au NP monolayer
samples were thermally treated at 100
C for 1 h in air. Following this
stabilizing treatment, the samples were used as substrates for sol
gel
thin lm deposition.
NiO sol
gel solutions were prepared as follows: 300 mg of nickel
acetate tetrahydrate (98%) was dissolved in 2 mL of methanol (99.8%),
and then 0.18 mL of diethanolamine (99%) was added. The amine acts
as a complexing agent, as conrmed from the change in color of the
solution (from bright green to dark green
blue, due to the formation of
the complex between Ni2+ ions and nitrogen atoms of the amine24). The
solution was stirred for 30 min prior to deposition.
TiO2 sol
gel solutions were prepared as follows: 0.55 mL of titanium
butoxide (97%) was added to 0.47 mL of ethanol (99.8%) under stirring;
0.27 mL acetylacetone (99%) was subsequently added and the solution

was stirred for 10 min, and then 0.12 mL of water was slowly added
under vigorous stirring. After 20 min, 2.2 mL of ethanol was added, and
the solution was directly used for the deposition process.
All sol
gel samples were deposited by spin-coating at 2500 rpm for
30 s on either SiO2 (HSQ300, Heraeus) or Si (100 oriented, p-type
boron-doped, Silicon Materials) substrates, with and without the Au NP
monolayers, and annealed in a mue furnace at 500
C for 1 h in air,
obtaining crystalline inorganic oxide lms of about 50
60 nm thickness.
A complete list of the samples prepared in this study which utilized Au
NPs underlayers is provided in Table 1.
The lms deposited on SiO2 substrates were characterized by XRD
using a Philips diractometer equipped with glancing-incidence X-ray
optics. The analysis was performed at 0.5
incidence, using Cu K Ni
ltered radiation at 30 kV and 40 mA. The average crystallite size was
calculated from the Scherrer equation after tting the experimental
proles with Lorentzian curves: the diraction peaks used for the
analyses are {111} at 38.2
and {200} at 44.4
for Au (JCPDS no.
040714), {111} at 37.2
and {200} at 43.3
for NiO (JCPDS no.
471049), {101} at 25.3
and {200} at 48.1
for TiO2 (JCPDS no.
841285). The surface structure of the nanocomposite lms deposited on
Si substrates was investigated with an xT Nova NanoLab scanning
electron microscopy (SEM). AFM height proles of samples deposited
on Si substrates were recorded with a Veeco Multimode AFM operating
in tapping mode. Transmission electron microscopy (TEM) measurements of the metal NPs deposited on a carbon-coated copper grid were
taken with a Philips CM10 TEM; the size distribution of the NPs has
been evaluated with Fiji-Image JA 1.44b image analyzer software
measuring a minimum of 150 particles.
Oblique angle attenuated total reectance (ATR) FTIR was performed on Thermo Scientic Nicolet 6700 FT-IR spectrometer with a
Harrick VariGATR attachment. Samples were prepared on polished
silicon wafer, with bare silicon being used as the reference. Measurements were performed at an incidence angle of 62
from normal. Optical
absorption spectra of samples deposited on SiO2 substrates were measured
in the 300
2000 nm range using a Jasco V-570 spectrophotometer.
Transmittance at normal incidence and ellipsometry quantities and
of samples deposited on SiO2 substrates were measured using a J.A.
Woollam V-VASE spectroscopic ellipsometer in vertical conguration,
at two dierent angles of incidence (60
, 70
) in the wavelength range
300
1700 nm. Optical constants n and k were evaluated from , , and
transmittance data using WVASE32 ellipsometry data analysis software,
tting the experimental data with Gaussian and Cauchy oscillators for
absorbing and nonabsorbing spectral regimes, respectively.
Optical sensing tests for ethanol detection were performed in
reection mode on samples deposited on SiO2 substrates using a
custom-built stainless steel cell provided with a heater that enabled
gas sensing tests up to 150
C. The reection spectra were collected at an
incident angle of 90
with a reection probe composed of a tight bundle
of seven optical bers (six illumination bers around one read ber),
connected to a OceanOptics USB2000 spectrophotometer. For ethanol
sensing, a nitrogen stream owed inside through saturated ethanol, and
if needed, the nal stream was diluted with pure nitrogen, to lower the
ethanol concentration. The nal ow rate was set constant at 0.5 L/min.

RESULTS AND DISCUSSION

Au Nanoparticle Monolayer. In Figure 1, we show SEM and
AFM images of Au NPs on silicon that correspond to the low
(a,d), medium (b,e), and high (c,f) surface coverage samples
prepared in this study, as well as a TEM micrograph of the asprepared Au colloids (i). It can be clearly seen that at low surface
coverages the Au NPs are homogeneously dispersed on a
micrometer scale, while at higher surface coverage, the formation
of NP islands are observed. High-resolution topographic (g) and
13740

dx.doi.org/10.1021/la2032829 |Langmuir 2011, 27, 1373913747

Langmuir

ARTICLE

Figure 1. AFM (a,b,c) and SEM (d
h) images of Au NPs layers with dierent surface coverage: Low = 0.06 (a,d), Medium = 0.34 (b,e), High = 0.62
(c,f). Image (g) is a higher magnication micrograph image (f); image (h) is a cross-sectional micrograph showing that Au NPs are on one single layer.
Image (i) is a TEM micrograph of the Au colloids used for the nanocomposites preparation.

cross-sectional (h) SEM images show that the Au NPs predominantly exist within a monolayer, and only in the high surface
coverage samples, few multilayered nanoparticles are detected.
To estimate the surface coverage of submonolayer coatings,
the ratio between the projected surface area of the Au NPs and
the total area of the analyzed surface is considered. The mean
particle diameter evaluated from TEM images is D = 14 ( 1 nm
(see Figure 1i, and it was also conrmed by the SEM images
(Figure 1a
h). AFM overestimates the actual particle size
(D = 41 ( 4 nm) due to convolution of the measurement with
the nite angle of the tip.25 For this reason, the Au NP surface
coverage could only be accurately evaluated from the SEM
characterization. The values obtained from the surface coverage
analysis were 62%, 34%, and 6% for the high, medium, and low
coverage samples, respectively. These values highlight that the
substrate functionalization and the Au NPs deposition processes
provide a simple and reproducible way to deposit metal NPs with
submonolayer covering.
During the deposition process, the MUA functionalized Au
NPs must anchor to the APTMS derivatized surface. This may
arise due to (i) a simple electrostatic interaction between the

surface of Au NPs and the amino functionalities on the substrate

and/or (ii) the formation of an amide bond between the amino
and carboxylate groups which originate from the APTMS and
MUA, respectively. It is known that such a reaction is strongly
favored in the presence of activating agents, such as a mixture of
pentauorophenol and 1-(3-dimethylaminopropyl)-3-ethylcarbodiimide hydrochloride (EDC),26 but the formation of the
amide bond in the absence of an activating agent is dicult at
room temperature.27
To investigate the anchoring mechanism, we have employed
ATR FT-IR measurements to study APTMS-functionalized
substrates before and after Au NP monolayer deposition. As
can be seen in Figure 2, the APTMS-functionalized substrate
without Au NPs shows a broad absorption peak at 1650 cm
1,
two very weak peaks at 1465 cm
1 and 1379 cm
1, and a further
two weak peaks at 2851 cm
1 and 2922 cm
1, while Au NPs
layer exhibit also a series of peaks in the 1300
1900 cm
1 range.
The band at 1650 cm
1 can be ascribed to strong in-plane NH2
scissoring absorptions,28 and it is consistent with the presence of
APTMS. The vibrational peak at 1379 cm
1 can be attributed to
symmetric rocking of H
C
H bonds, arising from the APTMS
13741

dx.doi.org/10.1021/la2032829 |Langmuir 2011, 27, 1373913747

Langmuir

ARTICLE

Figure 2. FT-IR spectra of APTMS-functionalized silicon substrates

uncovered (a) and covered (b) with Au NPs layer.

organic chain.29 For the Au NP layer sample, this vibrational peak

should also be observed due to the H
C
H groups arising from
both APTMS and the additional MUA molecules adsorbed to the
surfaces of the Au NPs. However, this vibration cannot be resolved
because it is convoluted with an intense peak at 1411 cm
1. The
peak at 1465 cm
1 is distinctive of the asymmetric C
H bending
originating from the organic carbon chain of both APTMS and
MUA molecules, and it is recognizable in both spectra as well.
The two peaks at 2922 and 2851 cm
1 are ascribable to C
C
and C
H bond vibrations in CH2 groups as reported in several
publications.29
32 Some authors claim that a peak at about
2950 cm
1 can be due to amino groups,33,34 but since this peak
is usually very weak and here it is overlapped with stronger C
H
vibrations, inferring its presence from these spectra will be rather
speculative. These peaks are again due to the presence of organic
chains of both APTMS and MUA molecules, and it is consistent that
they are more intense in the MUA-containing sample, since more
material is probed by the IR beam.
The peak at 1712 cm
1 is assigned to the CdO stretching
band arising from the carboxyl group of the MUA molecule,35 a
further conrmation of Au NPs being present.
The peak at 1560 cm
1 appears only in the sample containing
Au NPs, and it can be ascribed to N
H bending modes of the
amide bond26 or to N
H scissoring modes of adsorbed amine
on metals.36 The CdO stretching of the amide group has been
found at 1660 cm
1 by Whitesides and co-workers,26 but in our
samples, this peak overlaps with primary amine vibrations.
The last peak at 1411 cm
1 is dicult to assign. In the past, it
has been ascribed to C
H vibrations, but the data available are
controversial. Bertilsson and Liedberg37 have observed two peaks
at 1418 cm
1 and 1470 cm
1 which they have related to C
H
vibrations of the organic chain of thiols self-assembled on a gold
surface. These frequencies are relatively close to those in this
study, thus providing a potential origin of the vibrational peak
observed at 1411 cm
1.
The bonds that form during the deposition process enable
submonolayer coatings of Au NPs to be deposited. As the
temperature of such Au NP deposited substrates increases, these
bonds may be insucient to prevent particles from diusing over
the surface and ultimately coalescing. An understanding of such
eects is vital for the application of Au NP based lms. With this
in mind, we have investigated the thermal stability of bare Au NP

Figure 3. SEM images of Au NP layers annealed at dierent temperatures: (a) 100
C; (b) 200
C; (c) 300
C; (d) 400
C.

layers with medium surface coverage (about 34%) following

annealing at 100, 200, 300, and 400
C for one hour.
In Figure 3, we show SEM images that map out the evolution
of Au NPs layers with increasing temperature. A clear change in
the morphology of the layer can be seen at 200
C and above,
while minor modications in the average particle size can be seen
after the 300
C and 400
C annealing temperatures. From
analysis of the average particle size, it can be noted that the
originally monodisperse Au nanoparticles grow from 14 ( 1 nm
to polydispersed nanoparticle ensembles of 32 ( 13 nm in size
following annealing at 200
C; a progressive increase in the mean
diameter can be seen after the 300
C (37 ( 15 nm) and 400
C
(43 ( 18 nm) annealing steps. This behavior can be attributed to
the close packing of the as-synthesized Au NPs within the
monolayer, which provides adequate interparticle surface contact
(Figure 3a) to enable temperature-driven sintering.38 Following
coalescence, the interparticle distance increases (Figure 3b) and
therefore, further growth is hindered, or even prevented.
It is known that low molecular weight thiols desorb from gold
surfaces at temperatures below 200
C.39,40 This factor may
suggest that the driving force for the observed Au NP sintering at
high temperatures is related to the desorption and/or decomposition of the organic molecules on their surface, i.e., MUA and
APTMS. It has recently been reported41 that Au polymer
core
shell structures when deposited on glass substrates and
annealed at 700
C, produce Au NPs layers that adhere to the
substrate without signicant eect on the metal core size or
spatial distribution. In this study, the surface coverage was substantially higher, ensuring that the greatly reduced interparticle
spacing may have additionally enabled more ecient sintering of
the nanoparticles following the removal of the capping agent. It is
also worth noting that both MUA and APTMS decompose below
500
C;42,43 hence, the presence of residual organic is unlikely in the
sample annealed at 500
C.
Optical spectroscopy together with spectroscopic ellipsometry
are useful for understanding the eects that the surface coverage,
the thermal annealing, and the oxide coating presence have on
the Au NPs. Figure 4a shows UV
vis-NIR spectra for asdeposited Au NP layers with dierent surface coverage that have
13742

dx.doi.org/10.1021/la2032829 |Langmuir 2011, 27, 1373913747

Langmuir

ARTICLE

Figure 5. Optical absorption spectra of (a) Au NP layers covered with

NiO and annealed at 500
C. (b) Au NP layers covered with TiO2 and
annealed at 500
C.

Figure 4. (a) Optical absorption spectra of Au NP layers annealed at

100
C. (b) Optical absorption spectra of AuM sample annealed
between 100 and 400
C; the vertical dashed lines highlight the Au
SPR peak recorded at 520 nm in water. The inset shows a picture of the
samples annealed at 100
C (left) and at 400
C (right); the scale bar is
in cm. (c) Refractive index and (d) extinction coecient dispersion
curves for Au NP layers annealed at 100
C.

been stabilized at 100
C. In accordance with previous reports,

increasing surface coverage results in both an increased intensity
and a red shift of the Au NPs SPR peak.17,23 The increase in
absorbance with higher surface coverage is simply related to the
higher concentration of NPs interacting with the incoming beam,
leading to an increase in absorption as described by the wellknown Lambert
Beer equation.44,45 Meanwhile, the red shift of
the plasmon peak with increased surface coverage can be ascribed
to the decrease in the mutual distance between Au NPs; this
results in a stronger coupling of the plasmon resonances and a
concordant red-shift of the plasmon resonance.46
In Figure 4b, we show absorption spectra of a Au NP layer with
medium surface coverage (34%, the same coverage used for the
SEM measurements shown in Figure 3) annealed between 100
and 400
C. The Au SPR peak clearly depends on the annealing
temperature. Following the 200
C annealing step, the SPR peak
reduces slightly in intensity, red shifts, and broadens. These
observations are consistent with the initial coalescence step,
which also causes necking between particles and the consequent
formation of some elongated particles. However, with increasing
annealing temperature, a progressive blue-shift and a more
pronounced reduction in intensity is observed. This is due to
the subsequent spheroidization of the as-coalesced particles

and their progressive increase in size. In turn, this leads to an

increase of the average interparticle distance, causing a decrease
in SPR coupling. Moreover, the number of particles decreases as
a consequence of the sintering, and so does their optical cross
section, causing a reduction in the absorbance of the sample. The
chromatic eects of the thermal annealing can be seen in the
picture shown in the inset of Figure 4b: the Au NP layer annealed
at 100
C is blue-colored, as a consequence of the plasmon
coupling (the wavelength corresponding to the SPR peak
registered in aqueous solutions is at 520 nm, corresponding to
a red-colored solution), while the Au NPs layer annealed at
400
C is pink, as a consequence of the plasmon decoupling after
the particle growth and the consequent increase in the mutual
distance.
The absorption measurements, combined with the previously
discussed SEM measurements, demonstrate that an annealing
temperature of 200
C is sucient to promote sintering and
coarsening of close-packed Au NPs, leading to a respective change of
their optical properties. For this reason, unless specied, all the
characterizations of uncoated Au NP layers presented in the
following refer to the samples annealed at 100
C.
To understand further the observed absorption properties of
the Au monolayers, we have combined spectroscopic ellipsometry and optical transmission measurements to determine their
optical constants. Figure 4c,d shows the dispersion curves of the
refractive index and the absorption coecient of the Au NPs
layers, respectively. The layers have been modeled as an eective
medium thin lm composed of Au NPs in air. As such, the
resulting curves do not represent the dielectric function for Au
itself, but they are the dielectric function for the layer itself. The
refractive index dispersion curves (Figure 4c) show that increasing Au NP coverages results in a higher eective refractive index
and in accordance with the Kramers
Kronig relation exhibit a
concordant red shifting of the second-order inection point that
is characteristic of the Au NPs SPR absorption. The spectral
changes associated with the refractive indexes translate to a clear red
shift and increased magnitude of the SPR peak extinction coecients (Figure 4d) with increasing surface coverage, thus conrming
the experimental results of the optical absorption measurements.
Au Nanoparticle Monolayer Coated with Sol
Gel Film.
The effect of the sol
gel layer deposited on top of the Au NPs
13743

dx.doi.org/10.1021/la2032829 |Langmuir 2011, 27, 1373913747

Langmuir

ARTICLE

Table 2. Mean Crystallite Diameters Calculated According to

the Scherrer Equation for the Three Crystalline Phases
Detected for the Samples Reported in Figure 6a
diameter (nm)
sample

Au

TiO2

NiO

Au Nanoparticle Monolayer Uncoated

AuL

7.4 ( 0.4

AuM
AuH

7.9 ( 2.6
7.5 ( 2.8

/
/

/
/

14.4 ( 2.0

AuM 400
C

Au Nanoparticle Monolayer Coated with NiO

/

16.7 ( 0.4

AuLN

9.3 ( 2.1

15.2 ( 1.3

AuMN

11.1 ( 3.8

14.2 ( 0.2

AuHN

11.6 ( 3.2

13.7 ( 0.4

NiO

Figure 6. XRD patterns of (a) Au NPs layers stabilized at 100
C;

for comparison purposes the sample annealed at 400
C is also reported;
(b) Au NP layers covered with NiO and annealed at 500
C; (c) Au NP
layers covered with TiO2 and annealed at 500
C. Theoretical diraction
lines for Au (black lines) NiO (dashed lines in gure (b)) and TiO2
(dashed lines in gure (c)) are reported at the bottom.

monolayer on the optical properties of the nanocomposite is a

broadening and red-shift of the SPR peak (see Figure 5a,b). This
effect is found to be much more pronounced for TiO2 compared
to NiO, due to the higher refractive index of the former compared
to the latter (2.51 at 590 nm for TiO2,47 2.33 at 620 nm for
NiO48), and also to the interaction between anatase crystals and
the surface of Au NPs, causing spreading and scattering of
conduction electrons as described previously.49,50
Moreover, looking particularly at the Low surface coverage
samples coated with the two dierent oxide layers, distinct high
and low wavelength SPR peaks emerge. The low wavelength
peaks, at about 590 and 630 nm for NiO and TiO2, respectively,
can be considered as arising from noninteracting randomly
dispersed Au NPs inside those matrices. The high wavelength
peaks appear in the Low surface coverage samples around 750
and 1000 nm for NiO and TiO2, respectively, and are found to
progressively red shift and become more prevalent with increasing Au NPs surface coverage. These peaks are consistent
with the surface plasmon coupling between Au NPs.38 We note
that the apparent broad absorption band observed for pure
TiO2 lm is due to optical interference; hence, the Au SPR
peaks of the Au containing samples overlap with this interference fringe.
XRD measurements of the dierent samples studied are
reported in Figure 6. Crystallization was evident in all samples,
as testied by clearly identied diraction peaks. The intensity of
the Au peaks (JCPDS no. 040714) was in good agreement with
the dierent surface coverage, and so with the amount of Au NPs
present inside the lms. For anatase TiO2 (JCPDS no. 841285),
the peaks have similar intensities for each of the samples. This
suggests that the Au monolayers used as substrates do not
signicantly alter the process of matrix formation for this metal
oxide. In contrast, for NiO (JCPDS no. 471049) we observe
broadened NiO diraction peaks within increasing Au NP
coverage. This factor suggests that Au NPs directly inuence
the crystallization of NiO, a point that will be further elucidated
on shortly.

Au Nanoparticle Monolayer Coated with TiO2

20.1 ( 0.2

AuLT
AuMT

7.5 ( 0.4
10.8 ( 0.2

19.2 ( 1.5
21.3 ( 5

/
/

AuHT

11.3 ( 0.4

22.6 ( 4.2

TiO2

The coated samples are annealed at 500
C.

Crystallite sizes evaluated through the Scherrer formula are

listed in Table 2. As can be noticed, the crystallite size of Au NPs
is slightly higher when the monolayer is covered with the sol
gel
coating. This eect is related to the annealing process (500
C for 1 h)
that must induce some coarsening or sintering of the particles.
For comparison, the XRD spectrum of the uncovered, medium
surface coverage, sample annealed at 400
C is reported as well in
Figure 6a; a clear sharpening of the Au diraction peaks is
experienced as a consequence of the previously discussed NP
sintering and growth processes. The crystallite size is about 8 nm
for the uncovered Au layers annealed at 100
C, in the 8
11 nm
range for the sol
gel-coated Au NP layer annealed at 500
C,
and >14 nm for the uncovered Au NP layer with medium surface
coverage annealed at 400
C (the same sample used for SEM and
UV
vis measurements reported in Figure 3d and Figure 4b,
respectively). This comparison shows that the presence of the
sol
gel lms reduces or even prevents the growth of Au NPs.
Notably, the dierence in the mean crystallite size of Au NPs
estimated from XRD peak broadening (7
8 nm) and the mean
particle size evaluated from SEM (14 nm) indicates that the Au
NPs are not monocrystalline.
NiO and TiO2 crystallites sizes are in the 14
17 nm and
19
23 nm ranges, respectively. Interestingly, for NiO we nd that
the crystal size is slightly smaller for samples where the sol
gel
solution is deposited on the high surface coverage Au monolayers,
and slightly higher when deposited on the lower surface coverages,
or on bare substrates. This behavior was not observed for TiO2
crystals, where the data are randomly distributed and their dierence
is within the error bars. The trend observed in the NiO crystal size
can be attributed to the small lattice mismatch between NiO and Au
crystals.51 This factor can permit Au NPs to act as heterogeneous
nucleating sites for NiO crystals. Therefore, as the number of nuclei
is related to the number of Au NPs, i.e., to the surface coverage,
higher Au NP concentrations result in smaller metal oxide crystal
sizes for a given volume of deposited material.
13744

dx.doi.org/10.1021/la2032829 |Langmuir 2011, 27, 1373913747

Langmuir

ARTICLE

Figure 8. ORC plots for the high surface coverage Au NP layers bare
and covered with NiO and TiO2 lms, and for pure NiO and TiO2 lms
when exposed to 180 ppm ethanol at 150
C OT. Zero value of response
is highlighted with a dotted line.
Figure 7. SEM images of Au NP monolayer with medium surface
coverage: (a) covered with NiO annealed at 500
C; (b
d) covered
with TiO2 annealed at 500
C. Au NPs as brighter spots are indicated by
the arrows in (c); in (d), a fragment of the lm ipped over, exposing the
Au NPs on the upper lm surface.

To evaluate the morphology of the metal oxide coated Au

monolayer following annealing at 500
C, SEM analyses have
been carried out. As can be seen from Figure 7, NiO coated
samples are rougher and with more irregularities than those
coated with TiO2 lms. To investigate the Au NP morphology
below the TiO2 sol
gel lms, we gently scratch the surface with a
scalpel. This enabled us to observe the underlying Au NPs
(Figure 7c), as well as overturn the 50
60 nm lm to expose
the underlayer containing the Au NP monolayer (Figure 7d). In
both cases, the Au NPs could be easily detected due to their
higher contrast (highlighted by arrows in Figure 7c).
From the SEM, the size of the Au NPs was estimated as being
16.1 ( 1.8 nm; a value that is only slightly higher compared to the
as-synthesized colloids. Thus, it can be concluded that the metal
oxide lm is not damaging the Au NPs layer during the sol
gel
solution spinning process, and is also strongly limiting the growth
of the metal particles by providing a physical diusion barrier
between neighboring particles.
Having characterized the Au NP layers covered with NiO and
TiO2 sol
gel lms, we will now evaluate the practical use of such
systems as a gas sensor to detect ethanol vapor. To ensure the
stability of the Au monolayer without metal oxide layers, the
sensor was operated at a temperature (OT) of 150
C and its
optical detection mode was reection (see Experimental section
for details). All tests were performed on the Au NP layers with
high surface coverage. The uncovered sample was annealed at
150
C for 6 h prior to gas sensing measurements. Following
this annealing time, no changes to the optical absorption
spectrum could be detected, conrming its thermal stability at
this temperature.
The eect of ethanol vapors is shown in Figure 8. The
dierence in reection intensity between the spectrum collected
during ethanol exposure and during nitrogen exposure (optical
reection change, ORC = ReEtOH
ReN2) is plotted as a
function of wavelength, and it is reported for uncovered Au NPs

layer, Au NPs layers covered with NiO and TiO2, and pure NiO
and TiO2 lms. As can be seen, outside the Au SPR peak
wavelength range, the ORC is similar to the response of the
metal-oxide matrix with no Au NPs monolayer. Interestingly, in
these ranges the ORC parameter is negative for NiO lm,
positive for TiO2 lms, and null for the uncovered layer. The
SPR wavelength range shows a wavelength-dependent response,
this being true for the uncovered Au NPs layer as well. This arises
as a consequence of the role played by the noble metal particles as
optical probes for the target analyte detection. Moreover, compared to the pure metal oxide lms, the covered Au NPs layers
exhibit a higher response in a limited wavelength range. This
provides conrmation of the SPR enhancement to the sensing
behavior obtained by combining the close-packed Au NPs layer
and the oxide active lms. In fact, especially for TiO2, the ORC
maximum of the covered Au NPs layers is higher than the sum of
the Au NPs layer and the oxide lm alone, so a synergetic eect
between the two components is likely to occur. Moreover, the
eect of the dierent metal oxide can be also seen: the ethanol
eect on NiO is to reduce the reection intensity, while its
interaction with TiO2 causes an increase in reection. This fact
can be explained by considering the dierent electric nature of
the two semiconducting oxides, i.e., NiO being a p-type and TiO2
an n-type semiconductor.
As reported in the literature, volatile organic compounds
(VOCs) can be oxidized on the surface of semiconducting
materials; for example, in the case of ethanol (C2H5OH), the
main reaction mechanisms can be described as the following:52,53
2C2 H5 OH O2 f 2CH3 CHO 2H2 O
C2 H5 OH f C2 H4 H2 O
In the rst reaction, ethanol is oxidized to acetaldehyde
(CH3CHO) by dehydrogenation of the ethanol molecule and
a subsequent reaction with oxygen leads to water formation. This
reaction can proceed further, with successive oxidation of acetaldehyde to acetic acid. The second reaction is a direct dehydration
of ethanol to ethylene (C2H4) with water formation. There are
other possible ethanol oxidation reactions leading, for example,
13745

dx.doi.org/10.1021/la2032829 |Langmuir 2011, 27, 1373913747

Langmuir

ARTICLE

interfering vapors or gases: these nanocomposites are currently

under investigation in order to analyze their gas sensing response
when exposed to dierent gases and vapors, according to their
composition and the operative temperature, and the results are
intended to be published in a separate paper. Nevertheless, these
preliminary results are encouraging, considering also that the
synthetic approach described in this work can be easily extended
to a great variety of active layers, simply changing the type of NPs
and the material for the top coating, tailoring the properties of the
nanocomposite by an appropriate choice of the active materials
and a proper optimization of their organization and spatial
distribution.

Figure 9. Dynamic tests for AuHT sample at 594 nm and 150
C OT

under repeated cycles of nitrogen
180 ppm ethanol.

to the formation of diethyl ether or diethylacetal, but they are less

common and require specic reaction conditions.53 The preferred
path is generally related to the type of the metal oxide (basic
oxides usually promote the rst reaction, acid oxides the second
one), to the presence of adsorbed oxygen species on the surface
of the oxide, or to the presence of oxygen in the gas phase (as can
be seen, oxygen is necessary for the rst reaction). In any case, the
oxidation of the ethanol molecules will lead to electron injection
into the metal oxide. In the case of TiO2, the ethanol oxidation
will lead to an increase in conductivity, because electrons are
injected into the conduction band of anatase. For NiO, the
opposite is true, as oxidation will lead to a decrease in conductivity, due to electron
hole recombination. As a consequence, it is reasonable to suppose that a dierence in the
electronic properties of the metal oxides produces a dierence
in the reection intensity, and that this dierence is of opposite
sign according to the type of semiconducting material.
Within our experimental setup, we employ a low-resolution
spectrophotometer and collect light under reection mode.
These factors contribute to low signal-to-noise ratios. Despite
this, the observed ORC trends clearly exemplify the synergetic
eect of coupling Au NP layers and metal oxide lms to enhance
the sensing properties.
While the above experiments were obtained under static
conditions, we also performed dynamic tests at a xed wavelength of 594 nm (corresponding to the maximum of the
response, as obtained from Figure 8) on the AuHT sample, the
most sensitive among the tested ones. The results, which are
depicted in Figure 9, show a reversible signal during repeated
cycles of nitrogen
ethanol exposure to the sensor. Although the
sensing dynamics are not ideal, because both response and
recovery times are occurring in a time scale of few minutes, the
results are promising, considering the low thickness of the
samples and the low resolution of the setup used. In fact, an
easily detectable variation in the reectivity during ethanol
exposure has been observed, with good reproducibility after
repeated nitrogen/ethanol cycles (see Figure 9). These results
show promise for applications of such materials in transmission
mode or in devices where the reection is enhanced, like SPR
congurations or on the surface of unclad optical bers. One of
the main issues with gas sensors is the cross sensitivity between

CONCLUSIONS
We have demonstrated that Au NP layers covered with
sol
gel oxide lms constitute an eective design for materials
to be used in optoelectronic applications. Nearly monodisperse
Au NPs were deposited on properly functionalized substrates
with good control of the surface coverage. Detailed optical and
morphological studies have been presented, showing a relationship between the Au NP surface coverage, annealing temperature
and optical properties of the uncovered monolayers; moreover,
the bond formation between Au NPs and the APTMS functionalized substrate has been deduced from infrared spectroscopy
measurements. The presence of a sol
gel oxide lm deposited
on top of the Au NP layers aects the optical properties of the
nanocomposite and also provides a physical barrier between
neighboring Au NPs, strongly limiting the extent of their
temperature-driven sintering. Preliminary gas sensing measurements on these systems show that ethanol vapor induces a
reversible and reproducible response, conrming the role of
Au NPs in increasing the sensitivity of the oxide lm itself and
providing a wavelength-dependent response.
AUTHOR INFORMATION
Corresponding Author

*E-mail: alex.martucci@unpd.it.

ACKNOWLEDGMENT
This work has been supported through Progetto Strategico
PLATFORMS of Padova University. E.D.G. thanks Fondazione
CARIPARO for nancial support. A.M. thanks the Universities
of Melbourne and Padova for their support through the University academic exchange program. J.J. acknowledges the Australian Research Council for support through the APD grant
DP110105341. M.K. acknowledges the Alexander von Humboldt foundation for a Feodor Lynen research fellowship.
REFERENCES
(1) Della Gaspera, E.; Antonello, A.; Guglielmi, M.; Post, M. L.;
Bello, V.; Mattei, G.; Romanato, F.; Martucci, A. J. Mater. Chem. 2011,
21, 4293.
(2) Joy, N. A.; Settens, C. M.; Matyi, R. J.; Carpenter, M. A. J. Phys.
Chem. C 2011, 115, 6283.
(3) Kolmakov, A.; Klenov, D. O.; Lilach, Y.; Stemmer, S.; Moskovits,
M. Nano Lett. 2005, 5, 667.
(4) Larsson, E. M.; Langhammer, C.; Zoric, I.; Kasemo, B. Science
2009, 326, 1091.
(5) Formo, E.; Lee, E.; Campbell, D. Nano Lett. 2008, 8, 668.
13746

dx.doi.org/10.1021/la2032829 |Langmuir 2011, 27, 1373913747

Langmuir
(6) Cavanna, E.; Segaud, J. P.; Livage, J. Mater. Res. Bull. 1999,
34, 167.
(7) Dhas, V.; Muduli, S.; Lee, W.; Han, S.-H.; Ogale, S. Appl. Phys.
Lett. 2008, 93, 243108.
(8) Selvan, S. T.; Bullen, C.; Ashokkumar, M.; Mulvaney, P. Adv.
Mater. 2001, 13, 985.
(9) Sundar, V. C.; Eisler, H. J.; Bawendi, M. G. Adv. Mater. 2002,
14, 739.
(10) Manera, M. G.; Spadavecchia, J.; Buso, D.; de Julian Fernandez,
C.; Mattei, G.; Martucci, A.; Mulvaney, P.; Perez-Juste, J.; Rella, R.;
Vasanelli, L.; Mazzoldi, P. Sens. Actuators, B 2008, 132, 107.
(11) Buso, D.; Pacico, J.; Martucci, A.; Mulvaney, P. Adv. Funct.
Mater. 2007, 17, 347.
(12) Hong, C. S.; Park, H. H.; Moon, J.; Park, H. H. Thin Solid Films
2006, 515, 957960.
(13) Freeman, R. G.; Grabar, K. C.; Allison, K. J.; Bright, R. M.;
Davis, J. A.; Guthrie, A. P.; Hommer, M. B.; Jackson, M. A.; Smith, P. C.;
Walter, D. G.; Natan, M. J. Science 1995, 267, 1629.
(14) Szunerits, S.; Das, M. R.; Boukherroub, R. J. Phys. Chem. C
2008, 112, 82398243.
(15) Galopin, E.; Niedziolka-Jonsson, J.; Akjouj, A.; Pennec, Y.;
Djafari-Rouhani, B.; Noual, A.; Boukherroub, R.; Szunerits, S. J. Phys.
Chem. C 2010, 114, 1176911775.
(16) Haes, A. J.; Van Duyne, R. P. J. Am. Chem. Soc. 2002, 124, 10596.
(17) Gao, S.; Koshizaki, N.; Tokuhisa, H.; Koyama, E.; Sasaki, T.;
Kim, J. K.; Ryu, J.; Kim, D. S.; Shimizu, Y. Adv. Funct. Mater. 2010, 20, 78.
(18) Shuming, N.; Emo, S. R. Science 1997, 275, 1102.
(19) Beyene, H. T.; Tichelaar, F. D.; Verheijen, M. A.; van de
Sanden, M. C. M.; Creatore, M. Plasmonics 2011, 6, 255.
(20) Zhang, Z.; Fu, Q.; Li, X.; Huang, X.; Xu, J.; Shen, J.; Liu, J. J. Biol.
Inorg. Chem. 2009, 14, 653.
(21) Huang, T.; Meng, F.; Qi, L. J. Phys. Chem. C 2009, 113,
1363613642.
(22) Enustun, B. V.; Turkevich, J. J. Am. Chem. Soc. 1963, 85, 3317.
(23) Buso, D.; Palmer, L.; Bello, V.; Mattei, G.; Post, M.; Mulvaney,
P.; Martucci, A. J. Mater. Chem. 2009, 19, 2051.
(24) Cotton, F.A.; Wilkinson, G. In Advanced Inorganic Chemistry,
3rd ed.; Interscience Publishers, John Wiley and Sons: New York, 1972;
pp893
894.
(25) Mulvaney, P.; Giersig, M. J. Chem. Soc., Faraday Trans. 1996,
92, 3137.
(26) Lahiri, J.; Ostuni, E.; Whitesides, G. M. Langmuir 1999,
15, 20552060.
(27) Vogel, A. I. Practical Organic Chemistry; Longman: United
Kingdom, 1974; p 424.
(28) Maria Chong, A. S.; Zhao, X. S. J. Phys. Chem. B 2003, 107,
1265012657.
(29) Innocenzi, P.; Falcaro, P.; Grosso, D.; Babonneau, F. J. Phys.
Chem. B 2003, 107, 4711.
(30) Kung, K. H. S.; Hayes, K. F. Langmuir 1993, 9, 263.
(31) Vracken, K. C.; van der Voort, P.; Gillis-DHamers, J.; Vansant,
E. F. J. Chem. Soc., Faraday Trans. 1992, 88, 3197.
(32) Li, D.; Seddon, Ou, A. B. J. Non-Cryst. Solids 1997, 210, 187203.
(33) Piers, A.; Rochester, C. H. J. Chem. Soc., Faraday Trans. 1995,
91, 359.
(34) Klonkowsky, A. M.; Widernik, T.; Grobelna, B.; Mozgawa, W.;
Jankowska-Frydel, A. Microporous Mesoporous Mater. 1999, 31, 175186.
(35) Martinet, C.; Paillard, V.; Gagnaire, A.; Joseph, J. J. Non-Cryst.
Solids 1997, 216, 77.
(36) Jasieniak, J.; Califano, M.; Watkins, S. E. ACS Nano 2011,
5, 5888.
(37) Bertilsson, L.; Liedberg, B. Langmuir 1993, 9, 141.
(38) Toderas, F.; Baia, M.; Baia, L.; Astilean, S. Nanotechnology 2007,
18, 255702.
(39) Delamarche, E.; Michel, B.; Kang, H.; Gerber, C. Langmuir
1994, 10, 4103.
(40) Xiao, X. D.; Wang, B.; Zhang, X.; Yang, Z.; Loy, M. M. T. Surf.
Sci. 2001, 472, 41.

ARTICLE

(41) Jaber, S.; Karg, M.; Morfa, A.; Mulvaney, P. Phys. Chem. Chem.
Phys. 2011, 13, 5576.
(42) Shi, W.; Sahoo, Y.; Swihart, M. T. Colloids Surf., A 2004, 246, 109.
(43) Mitrikas, G.; Trapalis, C. C.; Kordas, G. J. Non-Cryst. Solids
2001, 286, 41.
(44) Yu, W. W.; Qu, L.; Guo, W.; Peng, X. Chem. Mater. 2003,
15, 2854.
(45) Grabar, K. C.; Smith, P. C.; Musick, M. D.; Davis, J. A.; Walter,
D. G.; Jackson, M. A.; Guthrie, A. P.; Natan, M. J. J. Am. Chem. Soc. 1996,
118, 1148.
(46) Freeman, R. G.; Grabar, K. C.; Allison, K. J.; Bright, R. M.;
Davis, J. A.; Guthrie, A. P.; Hommer, M. B.; Jackson, M. A.; Smith, P. C.;
Walter, D. G.; Natan, M. J. Science 1995, 267, 1629.
(47) Cozzoli, P. D.; Kornowski, A.; Weller, H. J. Am. Chem. Soc.
2003, 125, 14539.
(48) Powell, R. J.; Spicer, W. E. Phys. Rev. B 1970, 2, 21822193.
(49) Alemany, P.; Borse, R. S.; Burlitch, J. M.; Homann, R. J. Phys.
Chem. 1993, 97, 8464.
(50) Lee, M.; Chae, L.; Lee, K. C. Nanostruct. Mater. 1999, 11, 195.
(51) Buso, D.; Guglielmi, M.; Martucci, A.; Mattei, G.; Mazzoldi, P.;
Sada, C.; Post, M. L. Cryst. Growth Des. 2008, 8, 744749.
(52) Rao, B. B. Mater. Chem. Phys. 2000, 64, 6265.
(53) Gucbilmez, Y.; Dogu, T; Balci, S. Ind. Eng. Chem. Res. 2006,
45, 34963502.

13747

dx.doi.org/10.1021/la2032829 |Langmuir 2011, 27, 1373913747