Nanotechnology

PAPER Related content

- Size-selected growth of ultrathin SiO2
Solid-state dewetting of ultra-thin Au films on SiO2 nanowires on surface and their decoration
by Au nanoparticles
and HfO2 F Ruffino, M Censabella, V Torrisi et al.

- The influence of Au film thickness and

annealing conditions on the VLS-assisted
growth of ZnO nanostructures
To cite this article: G Seguini et al 2014 Nanotechnology 25 495603
K Govatsi, A Chrissanthopoulos, V
Dracopoulos et al.

- Fabrication of periodic arrays of metallic

nanoparticles by block copolymer
View the article online for updates and enhancements. templates on HfO2 substrates
Jacopo Frascaroli, Gabriele Seguini,
Sabina Spiga et al.

Recent citations
- Investigation of Broadband Surface
Plasmon Resonance of Dewetted Au
Structures on TiO2 by Aperture-Probe
SNOM and FDTD Simulations
J. Abed et al

- I.S. Zelenina et al

- Solid state dewetting of thin plasmonic

films under focused cw-laser irradiation
William M. Abbott et al

This content was downloaded from IP address 14.139.62.245 on 05/03/2019 at 17:54

 Nanotechnology

Nanotechnology 25 (2014) 495603 (10pp) doi:10.1088/0957-4484/25/49/495603

Solid-state dewetting of ultra-thin Au films

on SiO2 and HfO2
G Seguini, J Llamoja Curi, S Spiga, G Tallarida, C Wiemer and M Perego
Laboratorio MDM, IMM-CNR, Via C Olivetti 2, I-20864 Agrate Brianza (MB), Italy

E-mail: michele.perego@mdm.imm.cnr.it

Received 5 September 2014, revised 14 October 2014

Accepted for publication 3 November 2014
Published 20 November 2014

Abstract
Ultra-thin Au films with thickness (h) ranging from 0.5 to 6.0 nm were deposited at room
temperature (RT) by means of e-beam evaporation on SiO2 and HfO2. Due to the natural solid-
state dewetting (SSD) of the as-deposited films, Au nanoparticles (NPs) were formed on the
substrates. By properly adjusting the h value, the size and the density of the Au NPs can be finely
tuned. For h = 0.5 nm, spherical-like Au NPs with diameter below 5 nm and density in the order
of 1012 Au NPs cm−2 were obtained without any additional thermal treatment independently
from the substrate. The dependence of the Au NPs characteristics on the substrate starts to be
effective for h ⩾ 1.0 nm where the Au NPs diameter is in the 5–10 nm range and the density is
around 1011 Au NPs cm−2. The effect of a subsequent high temperature (400–800 °C) annealing
in N2 atmosphere on the Au NPs was investigated as well. For h ⩽ 1.0 nm, the Au NPs
characteristics evidenced an excellent thermal stability. Whereas the thermal treatment affects the
cristallinity of the Au NPs. For the thicker films (2.0 ⩽ h ⩽ 6.0 nm), the thermal treatment
becomes effective to induce the SSD. The proposed methodology can be exploited for the
synthesis of Au NPs with diameter below 10 nm on different substrates at RT.

Keywords: Au, solid state dewetting, nanoparticles

(Some figures may appear in colour only in the online journal)

1. Introduction
Many fields of solid-state technology currently use metal
nanoparticles (NPs). The functional properties noticeably
depend on the choice of metal (e.g. Au, Ag, Pt, Co…) [1–6],
the type of the supporting substrate, and the morphology of
the NPs, i.e., size, density, and shape. Small variations of the
metal NPs configuration can lead to dramatic changes in the
electrical and optical properties [7–12].
The deposition conditions of the thin films on a selected
surface along with the post-deposition treatments influence
the morphology of the metal NPs. Indeed, after deposition
with different techniques, such as sputtering and e-beam or
resistive heating evaporation, a thin metallic film on a surface
can develop holes and eventually disintegrate into separated
objects [7, 13–18]. Dewetting of metallic films on a substrate,
which is referred as solid-state dewetting (SSD), is driven by
the reduction of the surface energy of the thin film and of the
interface energy between the film and the substrate. SSD can
be induced or accelerated by thermal annealing, pulsed laser
heating, and ion or electron irradiation [10, 11, 13, 19]. Upon
thermal treatment, SSD proceeds by surface diffusion even in
the solid-state well below the melting temperature of the film
[2, 3, 13]. This behavior is different from the one of poly-
meric films that are in the liquid state even close to room
temperature (RT) [1, 20, 21]. The driving force for SSD
increases and thus the rate of dewetting is accelerated with
decreasing film thickness. Therefore, the temperature at which
SSD occurs decreases with decreasing film thickness. More-
over, the manipulation of the spatial distribution and of the
density of the metal NPs can be achieved depositing a
sacrificial layer of Sb between the metal layer and the sub-
strate oxide [6]. Usually, a further problem is how to control
the SSD process in order to enhance the long-range ordering.
This can be achieved either by controlling the local surface
tension of the substrate by means of patterned surfaces or by a
local perturbation of thin films with a probe or a stamp
[14, 22, 23].
The various technological applications for metal NPs
cover electronic, magnetic, and optical devices, as well as

0957-4484/14/495603+10$33.00 1 © 2014 IOP Publishing Ltd Printed in the UK

Nanotechnology 25 (2014) 495603
electromechanical systems, plasmon-enhanced solar cells,
photo electrochemical water splitting, and coatings
[12, 24, 25]. Diagnosis, antimicrobial activity, and advanced
sensing, such as localized surface plasmon resonance sensing,
surface-enhanced Raman scattering, and second harmonic
generation, should be mentioned too [26–28]. In particular,
very small metal NPs maximize the number of surface atoms,
which are fundamental in catalytic systems [29–32]. Beyond
the applications, also the physical properties of thin films in
the presence of an interface earn a fundamental interest
[30, 33–35]. The thickness (h) of the as-deposited films is a
robust technological parameter to manage the final config-
uration of the metal NPs. The superior chemical stability, the
strong plasmonic activity, the catalytic role, and the high
electrical conductivity make Au one of the most interesting
metals [3, 32].
This work systematically investigates the Au NPs char-
acteristics resulting from the SSD of ultra-thin,
0.5 nm ⩽ h ⩽ 6.0 nm, Au films deposited at RT by means of e-
beam evaporation on SiO2 and HfO2 substrates. The proposed
route allows tailoring ultra-small (diameter well below 10 nm)
Au NPs on different substrates at RT. The effects of a sub-
sequent high temperature (T) thermal treatment in N2 atmo-
sphere on the Au NPs population are elucidated as well.
2. Experimental
Either SiO2 and HfO2 oxides (10 nm thick) were prepared on
a H-terminated Si(100) substrates (the native oxide was
removed by HF cleaning). SiO2 films were obtained by
thermal oxidation of the Si surface in O2 ambient at 900 °C
whereas HfO2 was deposited by e-beam evaporation at RT in
high vacuum (base pressure 5 × 10−7 mbar) from HfO2 pellets
[36]. Atomic force microscopy (AFM) measurements on the
surface of both SiO2 and HfO2 revealed a limited root mean
square roughness (≈0.1 nm), thus fixing flat interfaces for the
subsequent Au deposition. The Au deposition was performed
at RT in an e-beam evaporator system equipped with a four-
crucible e-beam gun working in high vacuum. The h of the
evaporated Au films was in situ monitored by a quartz
microbalance. The thermal treatments were performed in a
Rapid Thermal Process machine in N2 flux in a T range from
400 to 800 °C for 60–300 s. The h and the structural prop-
erties of the Au films were measured by means of x-ray
reflectivity (XRR) and x-ray diffraction (XRD) using a con-
ventional laboratory Cu Kα x-ray source equipped with a four
circle goniometer, a parabolic multilayer monochromator, a
scintillator, and a position-sensitive detector. The Au NPs
were characterized by scanning electron microscopy (SEM)
with an accelerating voltage of 15 kV. The characteristic
dimensions (average diameter and density) of the Au NPs
were determined by statistical analysis of the SEM images
using the software Scanning Probe Image Processor (SPIP,
V5.0, Image Metrology A/S).
2
G Seguini et al
3. Results and discussion
3.1. As-deposited Au films
Figure 1 reports the SEM images for the Au films deposited at
different h (0.5, 1.0, 2.0, 4.0, and 6.0 nm) on SiO2 (left col-
umn) and HfO2 (right column). The analysis of the SEM
images qualitatively allows to study the evolution of the
surface morphology and to highlight the different growth
regimes as a function of h. For 0.5 ⩽ h ⩽ 1.0 nm, the SEM
images clearly evidence a nucleation regime of the Au NPs.
With increasing h to the 1.0–2.0 nm range, the system enters a
growth regime with nuclei that start to grow forming bigger
and bigger Au NPs. The reported SEM images for the case
h = 2.0 nm indicate that this transition is slightly anticipated
for the SiO2 substrate. Further increasing h from 2.0 to
4.0 nm, these structures start to coalesce with the formation of
islands. Finally, for h = 6.0 nm, the islands merge forming a
connected structure without yet reaching a complete coverage
of the surface. Apart from the slight difference at h = 2.0 nm,
the SEM images do not evidence any other significant dif-
ference between the as-deposited Au films on SiO2 and on
HfO2 substrates.
The effective h of the Au films and the corresponding
electronic densities were derived by XRR analysis; data are
reported in table 1, whereas the comparison between the
effective h and the nominal h is reported in figure 2(a) for
both SiO2 and HfO2. The effective h values measured by
XRR are proportional to the nominal h measured by the
quartz microbalance during the evaporation. The proportion-
ality factor is 1.30 for the deposition on the HfO2 substrate
while it is 1.34 for the deposition on the SiO2 substrate. This
linear dependence allows a well-defined nanometric control
on the h of the deposited Au films. On the other hand, the
electronic densities measured by XRR (table 1), which are
lower than the reference bulk value (4.42 e− Ǻ−3) of Au,
corroborate the SEM observation that the Au films are not
continuous on the substrate surface. Moreover, the significant
difference between the density values on the two oxides
suggests that the effective amount of Au deposited on SiO2
and HfO2 is considerably different. In order to highlight this
point the equivalent h of the deposited films, which is defined
as the virtual h of a continuous Au film formed by the amount
of Au effectively deposited on the surface, was calculated.
Based on the h and the electronic density measured by XRR,
the percent of Au in the volume of the deposited film was
extrapolated (table 1). Figure 2(b) reports the equivalent h
versus the nominal h. The trend of the equivalent h evidences
a peculiar behavior of the two substrates: the amount of Au
effectively deposited on SiO2 and HfO2 is different. In par-
ticular, the proportionality factor for the deposition on SiO2
doubles that for the deposition on HfO2. Because the eva-
poration on the two substrates were simultaneously per-
formed, these differences should be attributed to the peculiar
interaction of the impinging Au atoms with the host surface.
Namely, the sticking coefficient, i.e. the ratio of the incoming
atoms over the desorbed atoms, of the SiO2 substrate doubles
that of the HfO2.
Nanotechnology 25 (2014) 495603 G Seguini et al

Figure 1. (On the left column) SEM images of the Au thin film (h = 0.5, 1.0, 2.0, 4.0, and 6.0 nm) as-deposited at RT by means of e-beam
evaporation on SiO2. (On the right column) SEM images of the Au thin film (h = 0.5, 1.0, 2.0, 4.0, and 6.0 nm) as-deposited at RT by means
of e-beam evaporation on HfO2.

3
Nanotechnology 25 (2014) 495603 G Seguini et al

Figure 2. (On the left) XRR measured h versus the nominal h of the as-deposited at RT Au films. (On the right) the equivalent h of the Au
films is calculated considering the electronic density measured by XRR and plotted versus the nominal h.

Table 1. Nominal thickness, measured thickness, electron density, and percentage of Au considering an electron density of Au of 4.42 e− Ǻ−3
for the Au films deposited on SiO2 and on HfO2 substrates.
SiO2 HfO2
Nominal thick- Measured thick- Electronic density Measured thick- Electronic density
ness (nm) ness (nm) (e/Ǻ3) % Au ness (nm) (e/Ǻ3) % Au
1 3.35 1.50 33.9 4.30 1.44 32.5
2 5.40 2.42 54.7 5.54 1.89 42.7
4 7.92 3.19 72.1 8.20 1.90 42.4
6 10.30 3.79 85.6 — — —

In order to characterize the resulting Au NPs population,
a detailed statistical analysis of the SEM images of the
samples with h = 0.5 nm and h = 1.0 nm was carried out to
derive the average diameter and the density of the Au NPs.
For these ultra-thin films deposited at RT, the SSD occurs
without any additional thermal treatment in agreement with
the fact that the T at which SSD occurs decreases with h.
Figure 3 reports the statistical distribution of the Au NPs
diameters. The size distributions were fitted using a Gaussian
function. The quantitative analysis of the SEM images evi-
dences that for h = 0.5 nm the mean diameter of the Au NPs is
the same on the two substrates and that the density is above
1012 Au NPs cm−2 in both cases. On the contrary, in the case
of h = 1.0 nm, the two distributions start to diverge in terms of
the average value and width. Indeed, for h = 0.5 nm the dia-
meter is 2.8/2.9 ± 0.9 nm for both the substrates. Whereas, for
h = 1.0 nm, although the equivalent h is the same for both
substrates, the Au NPs distribution reveals that the growth
evolution is affected by the specific substrate. Precisely, the
diameter is larger (7.3 ± 1.6 versus 4.6 ± 1.3 nm) and the
density is smaller (6.2 × 1011 versus 10.6 × 1011 Au NPs cm−2)
for SiO2 than for HfO2 [6]. For h ⩾ 1.0 nm, the interaction
with the substrate is clearly different, which results in a higher
equivalent h (figure 2(b)) along with a larger Au NPs diameter
and smaller density for the SiO2 than for the HfO2 substrate
(figure 3). It could be speculated that for h = 0.5 nm the
interactions between the Au atoms are stronger than those
between the atoms and the substrate. This corresponds to an
incubation stage of the SSD during which the contact area
with the substrate is too small to affect the aggregation pro-
cess of the Au film.
Comprehensively, the average diameter and the density
of the Au NPs evidence that the capability to produce dense
distribution of Au NPs well below 10 nm diameter by means
of e-beam evaporation at RT of Au films with h ⩽ 1.0 nm is
already effective upon natural SSD without any additional
thermal treatments. Noticeably, for h = 0.5 nm the SSD is
independent on the substrate.
3.2. Annealed Au films
The stability of the Au NPs population against thermal
treatment at different T is crucial for the applicative exploi-
tation of this manufacturing method and for the investigation
of SSD process in thin films as well. Figure 4 reports the SEM

4
Nanotechnology 25 (2014) 495603 G Seguini et al

Figure 3. Statistical evaluation of the mean diameter, width, and density of the Au NPs on SiO2 (on the left) and HfO2 (on the right) for as-
deposited at RT Au films with h = 0.5 and 1.0 nm.

Figure 4. SEM images of the Au thin film (h = 0.5, 1.0. 2.0, 4.0, and 6.0 nm) deposited by means of e-beam evaporation on SiO2 upon a
thermal treatment at T = 500 °C for 60 s. The size distribution reported in the histograms allowed the evaluation of the mean diameter and
width of the resulting Au NPs.

5
Nanotechnology 25 (2014) 495603
Figure 5. SEM images of the Au thin film (h = 1.0 nm) deposited by means of e-beam evaporation on SiO2 upon a thermal treatment at
different T: 400, 500, 600, 700, and 800 °C for 60 s.
images for the Au films with different h on SiO2 after a
thermal treatment in N2 at 500 °C for 60 s. The samples with
h = 0.5, 1.0, and 2.0 nm evidence spherical-like Au NPs. It is
interesting to note that the as-deposited samples reported in
figure 1, exhibiting spherical-like nanostructures (h = 0.5 and
1.0 nm), preserved their configuration with a limited variation
of the average diameter of the nanostructures. Conversely, the
h = 2.0 nm film underwent a significant morphological evo-
lution from irregular and quite dense islands (figure 1) to
spherical-like structures (figure 4) suggesting that the addi-
tional thermal treatment drives the system toward the
achievement of a thermodynamic equilibrium configuration.
Similar evolution occurred for h = 4.0 and 6.0 nm where the
films started with an almost continuous Au layer (figure 1)
and rearranged upon annealing forming well-defined and
separated nanostructures (figure 4). This reconfiguration of
the Au NPs is associated with a significant increase of their
mean heights as indicated by the contrast of the SEM images
and confirmed by AFM measurements (data not shown).
Summarizing, for h < 2.0 nm, the Au films naturally (as-
deposited at RT) and upon thermal treatment form Au
6
G Seguini et al
spherical-like nanostructures with diameter below 10 nm. On
the other hand, for h ⩾ 2.0 nm, the thermal treatment drives
the as-deposited coalesced nanostructures towards isolated
nanostructures with irregular shape and average size above
10 nm. Due to the rate of SSD being accelerated with
decreasing h, these data fix the transition from natural to
thermally induced SSD at h ≈ 2.0 nm. Only above this
threshold, an additional thermal treatment is required to
induce the SSD process. Similar behavior (not shown) was
detected for the HfO2 substrate. The quantitative analysis
allowed to obtain the size distributions reported in figure 4.
Both the average diameter and the width of the size dis-
tribution increase with increasing h. The variation of the
width upon annealing, for h values ranging from 0.5 nm to
1.0 nm, is negligible suggesting that a thermodynamic equi-
librium condition has been already achieved during the RT
deposition. Differently, the width distribution for h = 2.0 nm
after thermal treatment is very large, corroborating the
observation of an early stage of the thermal induced SSD.
The Au NPs resulting from h = 1.0 nm already evidence
the effect of the different substrates, but they are quite
Nanotechnology 25 (2014) 495603
0.5 nm
Diameter (nm) 15
10
5
20
Density (x 1011 ncs/cm3)
15
10
0
0 400
Figure 6. Statistical comparison of the mean diameter and density for the Au NPs deposited on SiO2 and HfO2 for different thermal
treatments and for h = 0.5, 1.0, and 2.0 nm.
unaffected by the thermal treatment. For this reason this h
value was selected as the representative one in order to evi-
dence the thermal stability upon different annealing T.
Figure 5 reports the SEM images for h = 1.0 nm after
annealing in N2 for 60 s at different T: 400, 500, 600, 700,
and 800 °C. These images qualitatively evidence that for this
starting Au thickness, where the natural SSD process was
already effective in the as-deposited state, the variations upon
thermal treatment are very limited. These images therefore
indicate a very good thermal stability against the high T
thermal treatments. Because the natural SSD is already
effective to drive the system toward a thermodynamic equi-
librium, the subsequent thermal treatment scarcely affects the
system [3]. Software analysis of these images allowed
quantitatively to study the evolution of the Au NPs diameter
G Seguini et al
1 nm 2 nm
SiO2
HfO2
SiO2
HfO2
800 0 400 800 0 400 800
Annealing temperature (°C)
and density as a function of the annealing T. This study was
systematically performed on the h = 0.5 nm (where the Au
NPs characteristics are substrate independent) and h = 2.0 nm
(where the thermal induced SSD is effective) as well. The
overall results are reported in figure 6. These data allow
several considerations. First, increasing h the mean diameter
increases, while the density decreases. Moreover, considering
the cases h = 0.5 nm and h = 1.0 nm the mean diameter slightly
increases with increasing the annealing T. Then, for the as-
deposited sample with h = 2.0 nm, the absence of regular
structures without thermal induced SSD inhibits an accurate
statistical analysis of the mean diameter. Finally, for
h ⩾ 1.0 nm the Au NPs formed on HfO2 substrate have
smaller diameter and larger density, than those formed, under
the same experimental conditions, i.e. h and T, on SiO2. In
7
Nanotechnology 25 (2014) 495603
Figure 7. XRD spectra of the Au NPs formed on SiO2 and HfO2 substrates as function of h (1.0 or 4.0 nm) of the deposited Au film and of the
annealing conditions (as-deposited and annealed at T = 500 °C for 60 s).
addition, for h = 2.0 nm (where the thermal induced SSD is
effective) a further test was performed to verify the stability of
the Au nanoparticles upon prolonged thermal treatments. The
samples were annealed at T = 600 °C for different annealing
times ranging from 60 s to 300 s. The values of the mean
diameter and of the density of the Au NPs are unaffected by
the prolonged annealing indicating that the final equilibrium
condition was achieved by the system already in 60 s.
Settled h as the reference fabrication parameter, the
overall results on the thermal stability of the RT deposited
films evidence the dependence of the Au NPs characteristics
on the type of substrate. The proper selection of the process
conditions (annealing T) and the type of substrate reveal a
technological strategy to reach the desired Au NPs properties.
3.3. Structural characteristics
In order to deeply evidence the Au NPs peculiarities, the
structural properties of the Au NPs were studied as well.
Figure 7 reports the XRD spectra for the Au films deposited
on SiO2 (left column) and HfO2 (right column) for two
representative Au film thickness, h = 1.0 nm, for the thinner
regime, and h = 4.0 nm, for the thicker regime. The investi-
gation considered both the as-deposited and annealed samples
upon a thermal treatment in N2 at 500 °C for 60 s. XRD
analysis evidences the preferential orientation along the 111-
direction perpendicular to substrate for the Au NPs on the
SiO2 substrate. This feature is present for h = 1.0 nm after the
thermal treatment. Similarly, for HfO2 substrates, after
annealing a signature of a preferential orientation with 111-
direction perpendicular to the substrate emerges. The
8
G Seguini et al
signature of the preferential orientation increases by increas-
ing h and it is higher in the films grown on HfO2 than SiO2. It
is worth to note that no evidence of crystal structure was
observed neither in the SiO2 nor in the HfO2 as-deposited
substrates as opposed to the thermal treated HfO2 substrates
that undergo crystallization with a mixture of monoclinic and
cubic (or orthorhombic, or tetragonal) structures. For
h = 1.0 nm the SSD dewetting is already effective in the as-
deposited samples and the effects of the thermal treatment on
the Au NPs characteristics are limited. However, the thermal
treatment improves the crystallization of the Au NPs. On the
contrary, for thicker samples, i.e. h = 4.0 nm, where the SSD
was effective only upon the thermal treatment, the preferential
111-direction perpendicular to the substrate is already evident
in the as-deposited state. Considering h as the reference
parameter to monitor the Au NPs characteristics, these facts
indicate that the SSD and the crystallization processes in the
Au NPs follow quite independent paths.
4. Conclusions
The peculiarities of the SSD of e-beam deposited ultra-thin
Au films with 0.5 ⩽ h ⩽ 6.0 nm have been investigated. In
particular, the attention was focused on ultra-thin thicknesses
(h ⩽ 2.0 nm). The SSD was able to produce Au NPs with
diameter well below 10 nm and density around 1012 Au
NPs cm−2. For h = 0.5 nm Au films, the SSD results in iden-
tical Au NPs characteristics on both SiO2 and HfO2 sub-
strates. At this thickness, the interaction with the substrate is
not yet affecting the film organization at RT. The resulting Au
Nanotechnology 25 (2014) 495603
NPs are stable against thermal treatments up to T = 800 °C for
60 s. The dependence of the Au NPs characteristics on the
substrate begins for h ⩾ 1.0 nm: the Au NPs formed on HfO2
have smaller diameter and larger density than those formed on
SiO2. Furthermore, the thermal treatment starts to induce the
SSD at h = 2.0 nm. The preferential orientation of the Au NPs
along the 111-direction perpendicular to the substrate
increases upon the thermal treatment and increasing h. This
signature is higher in the films grown on HfO2 than SiO2. The
fine control on the Au NPs properties as a function of h and
their excellent thermal stability evidence that this methodol-
ogy allows tailoring ultra-small Au NPs at RT. Furthermore,
the oxide dependent interaction with the surface offers an
additional option in order to tune the diameter and the density
of the Au NPs. The dependence of the Au NPS on the
characteristic of the surface where the SSD takes place was
already investigated depositing a sacrificial layer of Sb
between the oxide and the Au [6]. The concurrent sublimation
of the Sb atoms influences the dewetting characteristics of the
Au films. Differently, in this work the two oxide surfaces
(SiO2 and HfO2) affect the SSD evolution without any con-
current detectable evolution of their surface.
The EU via project NanoPV (FP7-NMP3-SL-2011-
246331) partially supported this work. Mario Alia (MDM,
IMM-CNR) and Giovanni Onida (University of Milan, Italy)
are gratefully acknowledged for technical assistance and for
fruitful discussions, respectively.
References
[1] Krishna H, Sachan R, Strader J, Favazza C, Khenner M and
Kalyanaraman R 2010 Thickness-dependent spontaneous
dewetting morphology of ultrathin Ag films Nanotechnology
21 155601
[2] Favazza C, Kalyanaraman R and Sureshkumar R 2006 Robust
nanopatterning by laser-induced dewetting of metal
nanofilms Nanotechnology 17 4229–34
[3] Govatsi K, Chrissanthopoulos A, Dracopoulos V and
Yannopoulos S N 2014 The influence of Au film thickness
and annealing conditions on the VLS-assisted growth of
ZnO nanostructures Nanotechnology 25 215601
[4] Galinski H, Ryll T, Elser P, Rupp J L M,
Bieberle-Hütter A and Gauckler L J 2010 Agglomeration of
Pt thin films on dielectric substrates Phys. Rev. B 82 235415
[5] Strobel S, Kirkendall C, Chang J-B and Berggren K K 2010
Sub-10 nm structures on silicon by thermal dewetting of
platinum Nanotechnology 21 505301
[6] Farzinpour P, Sundar A, Gilroy K D, Eskin Z E,
Hughes R A and Neretina S 2012 Altering the dewetting
characteristics of ultrathin gold and silver films using a
sacrificial antimony layer Nanotechnology 23 495604
[7] Gupta G, Tanaka D, Ito Y, Shibata D, Shimojo M, Furuya K,
Mitsui K and Kajikawa K 2009 Absorption spectroscopy of
gold nanoisland films: optical and structural characterization
Nanotechnology 20 025703
[8] Beyene H T, Weber J W, Verheijen M A, Sanden M C M and
Creatore M 2012 Real time in situ spectroscopic
ellipsometry of the growth and plasmonic properties of au
nanoparticles on SiO2 Nano Res. 5 513–20
9
G Seguini et al
[9] Trollmann J and Pucci A 2014 Infrared dielectric function of
gold films in relation to their morphology J. Phys. Chem. C
118 15011–8
[10] Sun H, Yu M, Wang G, Sun X and Lian J 2012 Temperature-
dependent morphology evolution and surface plasmon
absorption of ultrathin gold island films J. Phys. Chem. C
116 9000–8
[11] Tesler A B, Maoz B M, Feldman Y, Vaskevich A and
Rubinstein I 2013 Solid-state thermal dewetting of just-
percolated gold films evaporated on glass: development of
the morphology and optical properties J. Phys. Chem. C 117
11337–46
[12] Wang C-C, Wu J-Y, Chiou Y-K, Chang C-H and Wu T-B
2007 Charge storage characteristics of Au nanocrystals
embedded in high-k gate dielectrics on Si Appl. Phys. Lett.
91 202110
[13] Thompson C V 2012 Solid-state dewetting of thin films Annu.
Rev. Mater. Res. 42 399–434
[14] Gentili D, Foschi G, Valle F, Cavallini M and Biscarini F 2012
Applications of dewetting in micro and nanotechnology
Chem. Soc. Rev. 41 4430–43
[15] Schaub A, Slepička P, Kašpárková I, Malinský P,
Macková A and Svorčík V 2013 Gold nanolayer and
nanocluster coatings induced by heat treatment and
evaporation technique Nanoscale Res. Lett. 8 249
[16] Malinský P, Slepička P, Hnatowicz V and Svorčík V 2012
Early stages of growth of gold layers sputter deposited on
glass and silicon substrates Nanoscale Res. Lett. 7 241
[17] He J-Y, Lu J-X, Dai N and Zhu D-M 2011 Surface plasma
resonance spectra of Au nanoparticles formed from dewetted
thin films J. Mater. Sci. 47 668–76
[18] Ruffino F, Canino A, Grimaldi M G, Giannazzo F,
Bongiorno C, Roccaforte F and Raineri V 2007 Self-
organization of gold nanoclusters on hexagonal SiC and
SiO2 surfaces J. Appl. Phys. 101 064306
[19] Ruffino F, Canino A, Grimaldi M G, Giannazzo F,
Roccaforte F and Raineri V 2008 Kinetic mechanism of the
thermal-induced self-organization of Au/Si nanodroplets on
Si(100): size and roughness evolution J. Appl. Phys. 104
024310
[20] Stange T G, Evans D F and Hendrickson W A 1997 Nucleation
and growth of defects leading to dewetting of thin polymer
films Langmuir 13 4459–65
[21] Seemann R, Herminghaus S and Jacobs K 2001 Dewetting
patterns and molecular forces: a reconciliation Phys. Rev.
Lett. 86 5534–7
[22] Wang D, Ji R and Schaaf P 2011 Formation of precise 2D Au
particle arrays via thermally induced dewetting on pre-
patterned substrates Beilstein J. Nano. 2 318–26
[23] Jang J-W, Park B and Nettikadan S 2014 Generation of
plasmonic Au nanostructures in the visible wavelength using
two-dimensional parallel dip-pen nanolithography
Nanoscale 6 7912–6
[24] Atwater H A and Polman A 2010 Plasmonics for improved
photovoltaic devices Nat. Mater. 9 205–13
[25] Rodriguez A W, Capasso F and Johnson S G 2011 The casimir
effect in microstructured geometries Nat. Photonics 5
211–21
[26] Schuller J A, Barnard E S, Cai W, Jun Y C, White J S and
Brongersma M L 2010 Plasmonics for extreme light
concentration and manipulation Nat. Mater. 9 193–204
[27] Linic S, Christopher P and Ingram D B 2011 Plasmonic-metal
nanostructures for efficient conversion of solar to chemical
energy Nat. Mater. 10 911–21
[28] Pillai S, Catchpole K R, Trupke T and Green M A 2007
Surface plasmon enhanced silicon solar cells J. Appl. Phys.
101 093105
[29] Ma Z and Dai S 2011 Design of novel structured gold
nanocatalysts ACS Catalysis 1 805–18
Nanotechnology 25 (2014) 495603
[30] Čechal J, Polčák J and Sikola T 2014 Detachment limited
kinetics of gold diffusion through ultrathin oxide layers
J. Phys. Chem. C 118 17549–55
[31] Zhang Y, Cui X, Shi F and Deng Y 2012 Nano-gold catalysis
in fine chemical synthesis Chem. Rev. 112 2467–505
[32] Du M, Sun D, Yang H, Huang J, Jing X, Odoom-wubah T,
Wang H, Jia L and Li Q 2014 Influence of Au particle size
on Au/TiO2 catalysts for CO oxidation J. Phys. Chem. C 118
19150–7
[33] Müller M C and Spolenak R 2013 Dewetting of Au and AuPt
alloy films: a dewetting zone model J. Appl. Phys. 113
094301
10
G Seguini et al
[34] Carrey J, Maurice J-L, Petroff F and Vaurès A 2001 Growth of
Au clusters on amorphous Al2O3: evidence of cluster
mobility above a critical size Phys. Rev. Lett. 86 4600–3
[35] Trice J, Thomas D, Favazza C, Sureshkumar R and
Kalyanaraman R 2007 Pulsed-laser-induced dewetting in
nanoscopic metal films: theory and experiments Phys. Rev.
B 75 235439
[36] Perego M and Seguini G 2011 Charging phenomena in
dielectric/semiconductor heterostructures during x-ray
photoelectron spectroscopy measurements J. Appl. Phys.
110 053711